Home

USER MANUAL VECTIS-MAX Version 1.1

Contents

1. 279 17 2 GroupBox lor Restart Reading ss oase saate a lc A a 280 17 3 GroupBox for Restart Writing s 4 4 ane oh ae bp ane ed a a 281 17 4 GroupBox for Initial Values usas er A a a 281 17 5 GroupBox for Pressure Monitoring s s s oe sots mac acsee ea a o a ia a a ee a 282 i726 Retere ce Values LIneEditS soso asea do e a la Gp a A 282 17 7 Monitoring Points Panel occiso rea a a dees 283 17 8 Initial Conditions Padel s e s ncc aia a E da do E eA SS 284 17 9 Postprocessing Output Panels cs puso a a OS be a des 285 1710Species Output Panel as s os a A a a AA A ae a a 285 17 11Report Regions Arbitrary Surface Panel oo o ee 287 17 12Postprocessing Output Panel Solid Domain 288 17 18 Sold Material Panel c 4 soneca Gee e e a e ee 288 17 14 Initial Conditions Panel Solid Domain e 288 17 15Live update amp XY Canvas pressure versus iteration number o o 290 17 16Live Update Value GTOUPBOR 0004 ra a A A oa ew 290 17 17 Solver Control panels 4 as 4 rea we we aa a ee e we 291 17 18Solver Control via command line run_control 2 ee es 291 17 19Launcher buttons from left to right vmesh vpre amp vsolve oo 292 17 20Mesher launcher dialog DOR 3 2 mates 2 45 Badko ba ee ad he deeds ee eee 292 17 21 Vpre launcher dialog DOX 6 0306 ol ge A eh a ed a ee a 293 17 22V solve launcher dialog box wit ena e
2. Add Boundary Delete Boundary Show all Hide All Toggle Compress Reduce Paint All Paint Face 1 Auto Paint le Paint Line Motion Refinement Motion Info 12000 Auto Paint Angle 45 118101 ymax 0 071992 zmin 0 078000 zmax 0 080000 ZUTISTIS 38100 NOUES 20 18 13 Model dimensions 87500 xmax 0 2 Figure 19 65 The coolant jacket boundary setup Ricardo VECTIS Phase 1 coolant tri 4 10 xj Fie Edit View Toolbars Operations Help ages E e m 1 Ne Yl Options Stitch Mesh Setup view View Style mjujmjo Mesh Lines EJEJE E 0 Information UK Refinement m uj nj Add Delete Edit Refinement Parameters 0 078000 O 6 ORDERING TRIANGLE NORMALS 6 FINISHED ORDERING TRIANGLE NORMALS 0 SWAPS DONE 20 22 46 WRITING MODEL DATA TO D RPe training_tutorials V4 coolant coolant tri Figure 19 66 The mesh setup toolbar Ricardo Software December 2009 420 19 TUTORIALS 19 3 COOLANT FLOW Ricardo VECTIS Phase 1 coolant tri 10 xi Fie Edit View Toolbars Operations Help B a a 8 sade A N a a BRIL E al pH ra fest feia place a Horizontal Mesh Line 20 22 46 IG TRIANGLE NORMALS 20 22 46 FINISHED ORDERING TRIANGLE S DONE 20 22 46 WRITING MODEL DATA TO D RFe training_tutoriels V4 coolant coolant tri
3. Figure 13 7 Setting up an Outlet Boundary type in R Desk There are two options available for the outlet boundary Given Flow Split and Given Mass flow Rate Figure 13 7 shows the Given Flow Split option where the Split Factor should be entered in the InputBox When Given Mass flow Rate option is selected then Split Factor is disabled and Mass flow Rate is enabled In this case the user should always select the Out option as indicated in Figure 13 7 right 13 7 Symmetry Plane The flow symmetry is enforced at this type of boundary by setting to zero the velocity component normal to the symmetry plane This results in a zero convective flux Similarly a zero diffusion flux is obtained by setting to zero the normal derivatives of all other variables Ricardo Software December 2009 223 13 BOUNDARY amp INTERFACE CONDITION TYPES 13 8 WALL 13 7 1 Setting up asymmetry boundary condition To set up the symmetry plane condition left click on the boundary region in the Solver Setup Tree and the boundary set up panel is displayed to the right Under Boundary Condition Type select Symmetry from the ListBox and this sets up a Symmetry Boundary as in Figure 13 8 Bnd_Reg_4 Symmetry Boundary 5 K Region Name Bnd_Reg_4 Region ID 4 Material ID 1 Y Boundary Report Coupled Link Number 0 Boundary Condition Type Symmetry 4 gt Figure 13 8 Setting up a Symmetry Boundary type i
4. 8 3 2 Ideal gas model An ideal gas is described by the equation of state P R M 2 p RT where R R M 8 28 is the gas constant R 8314J kmol K is the universal gas constant and M denotes molecular weight In case of incompressible and weakly compressible gas Mach number Ma lt 0 3 the reference pressure pref can be used instead of the local pressure p in the above equation Pref RT 8 29 p The following relationships hold for an ideal gas B 1 T 1 p 6 1 xp cp T cy T Cp T G T R k T cp T amp T de cy T dT or e GT and dh cp T dT or h cT The speed of sound can be expressed as a v KRT and c and c are related to K and R as i K cp T Es i amp T R 8 30 For a number of common gases Air H2 CO O2 H20 CO2 Cp increases with temperature while it is nearly constant 2 5R for monatomic gases such as Ar Ne and He The polyno mial form cp R f T involving five coefficients Equation 8 26 is available in literature Moran and Shapiro 1992 Wark 1983 It covers the temperature range from 300 to 1000 K and can be easily integrated to obtain the averaged specific heat R as a function of temperature For air these coefficients are a 3 653 ay 1 337 x 1073 a3 3 294 x 1076 a4 1 913 x 10 as 0 2763 x 1071 giving cp 300K R 3 499 or cp 1004 1 J kgK K 1 400 Ricardo Software December 2009 135 8 MO
5. EME 28 p gt TR E T FIBERS m LINE SORTING AND ADDRESSING COMPLETE 177273 LINES WERE FOUND O LINES NEED 5 Figure 19 63 Gasket Hole Detail ORicardo Software December 2009 418 19 TUTORIALS 19 3 COOLANT FLOW 19 3 3 1 Selecting Boundaries Once the geometry has been loaded in to Phasel the first action is to paint the inlet and outlet boundaries Do this using the Operations gt Boundary Painting tools Y Ricardo VECTIS Phase 1 coolant tri 10 x Fie Edit View Toolbars Operations Help cama E A g E Nada MM Options Stitch Mesh Triangles AAA A iJ Part and Bo A i 6 A Parts Boundaries al A F 7 pH hos Tri tee Refine ae es A iz Marking 4 mer Add Bounda Delete Bounda ER Show All Hide All 1 E Toggle Compress Chopping Reduce Paint Alt e FA Paint Face Auto Paint fy ES A Paint Line Motion Refinement Motion Info gt Lo MN auto Paintangie ss Information ml a 9 ZUTIBTIS SUreT T 20 18 13 Model dimensions xmin 0 187500 xmax 0 212000 ymin 0 118101 ymax 0 071992 zmin 0 078000 zmax 0 080000 Figure 19 64 The Boundary Painting Panel The inlet and outlet boundaries should be painted as universal boundaries 2 and 3 respectively as shown by figure 19 65 using the Paint Face tool Also th
6. Ricardo Software December 2009 105 READING 8 MANIPULATING MESHES 5 1 Introduction Vpre is required to be run after the mesh grid file is generated and prior to the running of the solver vsolve The main argument is the grid file in addition to various other arguments vpre options fluid GRD Vpre automatically performs optimal cell reordering Reverse Cuthill McKee if grid file is un ordered Optional arguments to vpre are as follows inp file Input file boundary definition same input file supplied to vsolve metis meth Mesh partitioning method METIS see Section 5 2 np num Number of partitions for parallel run rest Repartition restart files along with mesh files see Section 5 4 agi num Define Arbitrary Grid Interface boundaries see Section 5 3 export form Export boundary mesh to specified format where form can be one of unv Uni versal dx OpenDX gnu gnuplot or x3D nomat Do not equally divide materials during decomposition of a multi material mesh see Section 5 2 jtype num Mesh joining method see Section 5 3 o file Output file name when joining meshes overrides default COALESCED GRD Controlling environment variables VECPRE_REORDER_MATERIAL Set to OFF to disable material reordering AGI EXTRUDE Specify which agi boundary to extrude during ex trusion process Set to 1 to indicate user interac tion
7. get a list of bnd conditions l_bc ibct ir ibout return Set object name boundary interested in usr_obj_name air usr_reg_name Inlet Tf usr_obj_nam sot usr_reg_ name return ro Get a phase or species index usr_obj_id for the selected usr_obj_name iget iget_phase Ricardo Software December 2009 IR 18 USER PROGRAMMING if var_name spec_mass_frac iget iget_specs if var_name mass_frac_ps iget iget_ps call get_id usr_obj_name usr_ob3J_id iget lGet a region index usr_reg_id for the selected region name usr_reg_name call get_id usr_reg_name usr_reg_id iget_breg if abs idt usr_obj_id and ir usr_reg_id then select case var_name Case temperature call get_field idt iget_bnd var_name fib lupr gt start call get_grid_geom xyz_bndf_c xbnd do jb jbnd1 jbnd2 fib jb xbnd 1 jb xxbnd 2 jb xbnd 3 jb end do xbnd gt null lupr gt end fib gt null Case heat_flux call get_field idt iget_bnd var_name fib lupr gt start do jb jbnd1 jbnd2 end do lupr gt end fib gt null case velocity call get_field idt iget_bnd var_name fib_vec lupr gt start do jb jbnd1 jbnd2 end do lupr gt end fib_vec gt null case pressure The user specified static pressure must be relative to the reference pressure ref_press while the total stagnation pressure is alwa
8. 3 5 1 View Menu The load and save operations available from the View Menu allow a view to be saved to disk or recovered and reset re centres the active model and realigns the axes with the screen Edit View Toolbars Operations Help Load Ctrl L Save Ctrl Reset Ctrl R Transformation Ctrl T Options v 45 Degrees 15 Degrees 5 Degrees 1 Degrees The transformation option pops up the view transformation panel This allows rotation and trans lation of the model and saving of the current view or the loading of a view to and from the trans mat dat file Ricardo Software December 2009 20 3 GEOMETRY 3 6 TRIANGLE PROCESSING ll Transformation x Pan a 0 01 gt X Axis Ej fas oe y pan Whoo AI Yaw o Zon N pes our zas Ol hs O m View Load Save i 3 6 Triangle Processing This section of the program is concerned with producing a triangulated surface model which meets the requirements of the VECTIS mesh generator The mesher requires that a surface model should be completely closed such that any point in space can be unambiguously determined as being inside or outside the model This can only be achieved if every edge in the model lies on exactly two triangles Equivalently every triangle must be connected to exactly one other triangle on each of its three sides Exceptions to these rules happen for example when a VDA is triangulated Since a VDA file usually co
9. NTERNAL_ POINT 0 074374 0 045267 0 003891 Ricardo Software December 2009 50 3 GEOMETRY 3 15 BOUNDARY PROCESSING 3 15 Boundary Processing 3 15 1 Part and Boundary Definition In Phase it is possible to split a model up into parts This facilitates model manipulation and part substitution Y Part and Boundary Panel Parts Boundaries 2 3 Model Parts EY 1 Chamber 4 03 2 Body 4 03 3 Engine E 1 wall WB 2 zero_grad 2 BB 3 inlet 3 BB 4 boundary_4 BB 5 boundary_5 WB 6 boundary_6 29 6 Part_6 H a 7 Part Add Part Delete Part Add Boundary Delete Boundary Show All Hide All Part Panel The part tree displays the nested part structure of the current model Part pop up menu operations Selecting a part with the Right Mouse Button causes the part pop up menu to be displayed Ricardo Software December 2009 af 3 GEOMETRY 3 15 BOUNDARY PROCESSING 5 3 Model Parts EL m 1 E New Part Delete Part E Chop Part Chop Part Children Unchop Part Unchop Part Children Add Boundary DRE fab Gah Properties Part pop up menu The Part Pop up menu allows the user to add and delete parts cut and paste parts into other parts chop and unchop parts and their children add new boundaries and set part properties Add Delete Part The Add Part function adds a part definition without assigning boundaries or trian
10. d To get starting and ending boundary faces for each boundary region then integer iwp pointer il i2 Call get_reg ise_reg_face il 12 subroutine get_reg var_name ival Input Output var_name cha ival integer pointer l_reg_cond boundary region control parameters ival 1 nbcop 1 n_regions l_reg_opts boundary value option ival 1 nte 1 n_regions subroutine get_reg var_name rval Input Output var_name cha rval real pointer specs_value region boundary values for species rval 1 nspreg specs_flux region boundary fluxes for species rval 1 nspreg ps_value region boundary values for passive scalars rval 1 npsreg ps_flux region boundary fluxes for passive scalars rval 1 npsreg subroutine get_reg var_name rval Input Output var_name cha rval real pointer l_reg_value region boundary condition array rval 1 nbcopr 1 n_regions vel_direction region boundary velocity direction rval 1 3 1 n_regions reg_flux region boundary fluxes rval 1 nte 1 n_regions phase_value region boundary values for phases rval 1 nte l nphreg subroutine get_reg var_name il 12 Input Output var_name cha il integer pointer i2 integer pointer ise_reg_face starting boundary face for each ending boundary face for each boundary region 11 1 n_regions boundary region 12 1 n_regions Table 18 14 Subroutine ge
11. real output source term of the discretised equation Table 18 39 Subroutine upr_sources to add source terms to equations 18 4 4 User sources routine This routine is called every outer iteration in the context of SIMPLE algorithm for each phase species transport equation which is solved for Here idt is the domain type index associated with the memory address of equation field For fluid phase equation it is equal to the negative phase index For single phase fluid mixture of fluid phases or solid domain it coincides with the corresponding domain indices In case of species mass fraction equation idt is the idt of the parent phase This routine is used to add source terms either explicitly by adding to src_fi or implicitly by adding to ap the central coefficient For a user selected phase or species object a conditional test is required to match isp or iph with the corresponding user object id via get_id Note it is not possible to modify the source term for the selected phase and for more than one equa tion at the same time In case of a solid material the phase name should be the same as material domain name Depending on the nature of the additional source the user might need to declare a pointer array for the variable whose source need to be modified and also additional pointer arrays if other variables are required for example the density is often required or if the source of turbulence energy is modified the dissip
12. 1 Case energy lupr gt start Get temperature field if required var_name temperature call get_field idt iget_cell var_name fi call get_grid_geom xyz_cell_c xc do ic icell icel2 Define the source term s src tol 0 05 if xc 1 ic lt 2 tol and xc 1 ic gt 2 tol and amp xc 2 ic lt 3 5 tol and xc 2 ic gt 3 5 tol then src per unit volume else cycle end if src 1 e6 fi_old fi ic sign small fi ic src srcxvol_ph ic src_fi 1 ic src_fi 1 ic max 0 src ap_fi ic ap_fi ic min 0 src fi_old end do nullify fi lupr gt end end select 1 end if Next usr_obj_name and select case eq_name copy amp edit the above code 1 end subroutine upr_sources ORicardo Software December 2009 363 18 USER PROGRAMMING 18 7 EXAMPLES This routine sets sources for various equations Equation name is derived for ieq via the call to get_name The phi dependent part of the source is added to the source or to ap according to the source sign 18 7 5 Example of the user generic routine This example is used to write a temperature line out at x 0 5 to a file at the end of each iteration The icp_beg_run case block is used to determine the total size over the parallel partitions of the array to store the values Then the global y position array is determined for the first partition 1 via the use of concat_array The icp_end_iter proceeds similarly storing the tempera ture
13. 1 Time in units of the specified time base 2 Region mass flow rate kg s 3 The boundary region area not used currently 4 Region mean temperature K 5 Region mean absolute pressure Pa 6 Region mean turbulent intensity 7 Region mean turbulent length scale mm 8 Region mean value of a passive scalar mass fraction 9 Entries 9 to 12 Region mean values of wave species mass fractions i e mass fractions for air vapour burnt air and burnt fuel are specified while the value of liquid fuel is omitted assumed zero Ricardo Software December 2009 231 13 BOUNDARY amp INTERFACE CONDITION TYPES 13 10 INTERFACE CONDITIONS In order to perform an unsteady run with time dependent boundary data the global control variable Time Dependent Boundary Conditions from the Timebase panel need to be activated ticked on If it is not activated any time dependent settings for individual boundary regions will be disabled If the real simulation solver time in seconds becomes larger than the ending time specified within a boundary region file the run will switch to time dependent data corresponding to the start time specified in the region file Apart from the time dependent boundary conditions supplied via boundary region files realistic initial velocity pressure temperature turbulence and other variable fields are crucial in obtaining the expected simulation results Thus it is user s responsibility to provide t
14. A panel is popped up in which the user enters the number of triangles to grow the set by and inactive triangles adjacent to the active ones are made active up to the specified depth Triangles on both active and Inactive boundaries are made visible with this operation Swap Chopped and Unchopped Button Es This swaps the visible and invisible portions of the model that lie on active boundaries created via the Chop Area command Show All Button Eni Ricardo Software December 2009 36 3 GEOMETRY 3 12 TRIANGLE INTERROGATION OPERATIONS This makes all the triangles active and visible 3 12 Triangle Interrogation Operations Co ordinate display Button lo ji je While this mode is selected the co ordinates of the nodes that form the vertices of the triangles may be selected and their coordinate locations displayed Pressing the left mouse button near a node will display the number of that node and its 3d co ordinates If another node is selected by single clicking the left mouse button the spatial distances between the nodes are printed for the three co ordinate directions as well as the direct path length distance If another node is selected by double clicking the left mouse button then the cumulative distances and path lengths are printed and this can be repeated for many nodes 0 045000 z 0 013480 0 045000 z 0 013673 000000 z 0 000193 Coordinates of point 423 x 0 063412 y Coordinates of point
15. Browse Fan Model Bnd_Reg_1 Inlet Given Velocity Boundary SSS SSS sss Phase_1 Boundary Phase Extract Interface Regions Sub domains Report Regions Show Mesh Preview Solid Domains mn Grid Extract Dialog x Grid Extract Dialog x Global Domain Fluid Domain Fluid Domain Fluid Domain Solid Domain Solid Material Solid Material Solid Material Solid Material oe cna Co Cere Figure 7 2 R Desk setup Creating domain structure components Respecting this ordering corresponding fluid and solid domains need to be defined by allocating each fluid material from the grid file into a fluid domain and each solid material into a solid domain By right clicking on the Global Domain entry additional fluid and solid domains can be created Additional materials can be added to a solid domain by right clicking on the relevant Solid Domain The number of remaining materials that have not been allocated is indicated in the Materials Remaining box For instance if two materials exist in the grid file but only one fluid domain has been created then there will be one material remaining On the other hand if more solid materials or fluids are defined than exist in the grid file then the number will be negative When all materials have been assigned to the fluid and solid domains the required computational domain components will be created by clicking on the OK button The Solver Setup Tree will be
16. Figure 19 43 The Solution Control Panel Output Output File Outer Iterations Time Steps Here the frequency of data reporting output is chosen Firstly the monitoring and convergence data written to the screen is also written to the out file at the intervals specified If zero values are used then the output is not written Post processing Frequency The frequency of post processing file writing is specified in the Post processing File Frequency box Ricardo Software December 2009 401 19 TUTORIALS 19 2 STEADY STATE PORT FLOW Co Simulation Post processing Frequency Controls the frequency at which post processing data is written when WAVE requests postprocessing output during a coupled simulation Additional output Additional output file can be written for monitoring data boundary data report regions and domain data These can be either in ASCII column format or Ricardo SDF binary format The frequency of the two types of file can be chosen independently by entering the required intervals in to the Report Frequency boxes Again an interval of zero will mean that the data is not written The ASCII files include header rows detailing the data in each of the columns Separate files are written for the different data types Domain Monitoring IO and Wall If reporting is selected for the relevant boundary Arbitrary Surface Additionally the residual data can be chosen to be saved to an ASCII file and the SDF rep
17. In terms of VECTIS MAX multi domain structure a porous media region is associated with the porous fluid sub domain an example is shown in Figure 7 4 The appropriate mesh joining pro cedure should be used to obtain conformal meshes at interfaces between a porous sub domain and its parent fluid domain In some situations the sub domain geometry can be approximated by infinitely thin region represented by cell faces i e by interface between cells rather than a cell region This kind of simplification is described by so called porous jump conditions and it is not yet supported A porous medium is typically characterised by directional pressure drop which is either linear or non linear function of the local velocity For each porous media type the pressure drop is deter mined by the corresponding flow resistance model The following porous media types resistance models are available General porous media isotropic or orthotropic media with the flow resistance defined in terms of viscous and inertial resistance tensors Forchheimer s model O Catalytic converters orthotropic media with Darcy s viscous resistance model i e the pressure drop depends linearly on the local velocity General orthotropic media inertial resistance model where the pressure drop is a quadratic function of the local velocity Radiators orthotropic media with both viscous and inertial resistance models included General heat exchang
18. Pressure correction scheme s kisas aaa a Se 248 14 2 2 Face velocity Tor mass conservation lt i soca scr sace a a ee Ea 248 14 2 3 Pressure COMECUONS sose daimo a A ew Bee 249 14 2 4 Boundary conditions for pressure corrections o o o 251 14 2 5 Setting up pressure correction algorithm o o e 252 14 200 Linear SO VEES kdd ae ae da ATA E a e de 252 14 3 Implementation of boundary conditions o e 253 144 Poor Quality Cell Treatment Visa a a A AS iaa a 254 14 5 ParallelisatiOm ss 4 bos ge a a Be A ake E eas a 256 15 MODELLING RADIATION 257 15 1 Introduction And Overview ooo ee we ee 257 152 Mesh File Setup sy aa a a a Rd ae ae eet Oh 257 13 3 Radiation Setup s s ia Se he Ge Oa a ee OE A Ow arta 258 15 3 1 Running of radprep o cce ve eae a Eh ea eee EES A 261 15 32 RUDILDS OL TAVE e a a a ee a ee Oe oe ae we we ee ee ee 263 154 Radyvim Theory eo 4 24445 ede Saw doe RG eRe RRR ERG SE RO aS GEOR eS 264 Ricardo Software December 2009 viii 15 4 1 Assembling the view factors matrix e ee ee 15 4 2 Calculation of view factor 2 oe a eos enii a eared oh ee ek ad ee 15 4 3 Calculation of Super patch view factors 2 o o 0 ee eee ee 15 4 3 1 Hemisphere Base Projection Methods 4 15 5 Radsolv Theoty s eare aaa hod 2 aS See eae ek a PEP ORR ERDAS SESS 16 MODELLING FANS 16 1 Introduction And OvervieW ee
19. Tij Puju puju 6 33 O Turbulent heat flux e di ph u ph 6 34 Note that specific enthalpy his usually defined in terms of of specific heat Cp and temperature T for both single component and multi component phase i e h re a Therefore the turbulent heat flux can be expressed as dy pepT ul p tpT u 6 35 O Work of body forces i e production or destruction of turbulent kinetic energy by body force see Equation 6 40 o R pr ne pP fu 6 36 Ricardo Software December 2009 120 6 SOLVER FUNDAMENTALS 6 4 REYNOLDS AVERAGED EQUATIONS With regards to the corresponding molecular fluxes J j Tij and q their constitutive relations are given by Fick s Stokes s and Fourier s laws respectively The effects of property fluctuations appearing in the above constitutive laws mass diffusivity 2 viscosity u and thermal conduc tivity A respectively are insignificant cf Barre et al 2002 Huang et a al 1995 Speziale a Further the contributions of averaged fluctuating mass fractions c C C velocities U U and temperatures T T T should be negligible in regions where molecular trans a is important This means that the Favre averaged variables cj and T which govern the constitutive relations can be used The total enthalpy H in Equation 6 31 is defined as Pp a P U U UU Herk ee ee ae 6 37 P P 2 2 where the Favre ave
20. ered laminar for Ra lt 10 and the transition to turbulence happens over the range 108 lt Ra lt 10 Ra GrPr 10 21 The gravity force is conveniently cast in terms of constant reference and buoyancy source terms Fe Pref P Pres E 10 22 The first term is then included in the solver working pressure Equation 10 14 as P Ps Pref P Pref 10 23 In case of variable density flows where density is a function of temperature the buoyancy source p Pre f g does not require any modelling standard approach If the constant density is going to be used for the buoyancy driven flow throughout all equations then the Boussinesq model is used Expressing density according to the Boussinesq approximation Equation 8 27 the buoyancy source becomes P Pref 8 PrefB T Tref 8 10 24 The Boussinesq model should not be used for large density variations i e it is applicable for B T Tref lt lt 1 The buoyancy model is selected in the Body Force sub panel which is a part of the Fluid Domain panel As Figure 10 8 shows by selecting the body force option via three RadioButtons the user can either disable gravity default option or enable the Boussinesq model for constant density flows or enable the standard approach for variable density flows If the gravity is not disabled then the components of the Force per unit mass in x y z directions must be specified Fluid_1 Fluid D
21. gt gt ATP 1 fi nTn f 1 Fi wie Dic w Dic 14 46 j _ y J Try The interface conductivity T is defined as the harmonic mean of the conductivities at P and N z ia A 14 47 r 2 fEp 1 fj Pw Solid cell A Figure 14 8 Control volumes near fluid solid interface The first normal diffusion term of the interface flux is included implicitly in the discretised equa tions for the cells P and N For the second cross diffusion term an explicit treatment is used This term can be limited as discussed earlier The interface temperature T is updated after the solution of temperature and then used to compute cell gradients Ricardo Software December 2009 247 14 NUMERICAL SOLUTION 14 2 SIMPLE BASED SOLUTION PROCEDURE 14 2 SIMPLE based solution procedure The outcome of the discretisation of Equation 14 4 is a set of algebraic equations one for each control volume and for each transport equation and has the following general form nj apop Y aj0p Sp 14 48 j l where nj is the number of internal cell faces ap and aj are the central and neighbour coefficients and S is the source term For this coupled system a segregated SIMPLE like solution algorithm applicable to both incom pressible and high speed compressible flows on arbitrary grids is employed The SIMPLE al gorithm ensures that the calculated flow field once converged simultaneously satisfies bo
22. 1 Point2 Temperature 973 15 Viscosity 1 Figure 8 6 R Desk setup Property calculation options Piecewise linear Figure 8 6 shows the piecewise linear option having the minimum Number of Points 2 Temper ature at Point1 and Point2 corresponds to 7 and T in relation 8 47 respectively Viscosity at Pointi and Point2 corresponds to Q and n 1 in relation 8 47 Piecewise polynomial functions A material property can also be defined with a polynomial function of temperature and may consist of one or more polynomial functions defined in each range or interval Consider 2 piecewise polynomial functions defined with intervals 7 7 1 and Tn 1 7n 2 as in Figure 8 5 In general form a polynomial function in interval T 7 41 is defined as T a aT a3T a4T 8 49 with the following constraint Tp lt T lt Tp 1 Similarly for interval T 1 7 2 the general form of the polynomial function is T by boT b3T b4T 8 50 with the following constraint 7 4 lt T lt Tp 2 Thus the general form to specify a material property using any number of piecewise polynomial Ricardo Software December 2009 142 8 MODELLING CONTINUA 8 6 SELECTING CONTINUUM AND ITS PROPERTIES functions is given in relation 8 51 a aT azT a4T for Tp lt T lt That by boT b3T b4T for Thai lt T lt Trio 9 T Cl eel e3T Lar for Ta42 lt T lt Tn 3 8 51
23. 1 m Refinement Specification Destination Save specification in triangle file Save specification in mesh file ea M Figure 4 8 Boundary refinement depth 0 blending distance O 4 7 Problems With Quality of Input Surface In comparison with VECTIS 3 the solver of VECTIS MAX is more powerful However its de mands for quality of used grids are higher Therefore in VMESH program a lot of effort is spent Ricardo Software December 2009 9 4 MESHING 4 7 PROBLEMS WITH QUALITY OF INPUT SURFACE x Specification for boundary 2 lt gt Delete Refinement depth at boundary RO Refinement blending distance i m Refinement Blending Blend to boundary depth Blend to boundary depth 1 m Refinement Specification Destination Save specification in triangle file Save specification in mesh file Figure 4 9 Boundary refinement depth 2 blending distance 1 RD 1 option set on x Specification for boundary 2 E gt Delete Refinement depth at boundary 2 Refinement blending distance EN Refinement Blending Blend to boundary depth Blend to boundary depth 1 H m Refinement Specification Destination Save specification in triangle file Save specification in mesh file we e dh Figure 4 10 Boundary refinement depth 2 blending distance 2 RD 1 option set on
24. 18 4 2 User initialisation routine subroutine upr_init var_name idt icell icel2 Arguments Description var_name char name of field that can be modified idt integer domain type id icell integer starting cell id of material domain for idt var_name icel2 integer ending cell id of material domain for idt var_name Table 18 37 Subroutine upr_init to modify initial field Here var_name can be one of velocity pressure temperature turb_energy dissipation phase_vol_frac spec_mass_frac mass_frac_ps For fluid phase equation idt is equal to the negative phase index and for single phase fluid mixture of fluid phases or solid domain it coincides with the corresponding domain indices In case of Ricardo Software December 2009 349 18 USER PROGRAMMING 18 4 USER PROGRAMMABLE ROUTINES species mass fraction equation idt is equal to species indices Typically the user would specify the name of a phase or species object e g steel and the name of a variable e g temperature for which the corresponding cell values can be modified The object id for this object name would then be retrieved via get__id A conditional block could then be written to test for the object id and variable name match Inside this block a call to get_ field provides the current cell values via pointer arrays These values can be now modified by looping over material domain cells according to the user spe
25. 6 2 2 Instantaneous equations The global laws of continuum physics the conservation of mass momentum and energy are usually applied to a certain spatial region control volume CV rather than to a given mass of continuum The former is known as the Eulerian or control volume approach while the latter is the Lagrangian approach For the infinitesimally small control volume the conservation equations can be transformed into either a differential coordinate free form or a form specific to a chosen coordinate system The VECTIS MAX solver employs a Cartesian coordinate system x y z Throughout this manual a compact tensor notation is often used Accordingly if a monomial con tains a repeated index a summation over the range of index values is applied Einstein convention A free or un repeated index is the one which appears once in a monomial Two free indices define the second order tensor one free index defines vector while scalar does not have a free index Let us consider an arbitrary control volume V bounded by generally moving surface A whose velocity is U g In this case a moving grid it is convenient to formulate the conservation equations in a moving frame of reference This formulation is usually referred to as Arbitrary Lagrangian Eulerian ALE approach When the CV surface velocity U is zero CV fixed in space the Eulerian form of equations is obtained When however the CV surface moves with the fluid velocity vector U
26. Data J_sets Steps ax Step Iteration 1 1398 SetRange IV Auto apply Apply EDE E x p 019565 3 Y p 130499 3 Zz p 144915 Increment 0 00500002 gt Reset re x fo SF 32 3 Increment 0 1 Reset IV Preview FF Clip mesh v Jj ee al ee commana gt f DA Colour Map Deformations Vectors Connections a2 Stano Figure 19 61 Slice plot showing velocity vectors When using the potential flow solver with a mass flow and pressure boundary condition the pres sure value on the mass flow boundary may need to be modified to reduce the specified pressure drop between the two boundaries When using potential flow initialisation it is important to ensure that the reference values specified are representative as they are used by the potential solver Additional the temperature specified in the initial conditions section is also used by the potential flow solver and therefore must also be a representative value Ricardo Software December 2009 415 19 TUTORIALS 19 3 COOLANT FLOW 19 3 Coolant Flow Get the necessary files for the coolant flow tutorial http www software ricardo com support tutorials vectismax TutorialFiles CoolantFlow 19 3 1 Introduction The purpose of this tutorial is to illustrate the steps needed to set up a simple engine coolant flow analysis However the general process could also be applied to any incompressible flow simulati
27. Density Option Polynomial f T 5 Number of Ranges 2 Tmin 300 o Tmax 400 Number of Coefficients 3 Coeff1 1 Coef2 1 Coef3 1 Tmin 400 Tmax 500 Number of Coefficients 2 Coeff 1 1 Coeff2 1 Figure 8 7 R Desk setup Property calculation options Polynomial Piecewise polynomial Figure 8 7 shows the polynomial option where the Number of Ranges defines how many piecewise polynomial functions would be used The default is 1 For each range the minimum and maximum temperature Tmin and Tmax should be specified continuously Each range can contain any Number of Coefficients that can be inserted in the appropriate edit boxes For example in Figure 8 7 for range 1 Coeff1 Coeff2 and Coeff3 correspond to a1 a2 and az in relation 8 51 respectively whereas for range 2 Coeff1 and Coeff 2 correspond to b and b in relation 8 51 respectively Other temperature functions f T in Table 8 1 are defined in terms of the coefficients cn Such an example is the Sutherland formula As Figure 8 8 shows these coefficients should be specified in the available edit boxes Coeff 1 corresponds to the c in the Table 8 1 etc Multi component phase If the fluid phase has been defined as the multi component phase see Figure 8 3 the Solver Setup Tree will show Species node with two defined species Figure 8 9 left Note that Mixture calculation option should be assigned by defa
28. Inlet Given Velocity Boundary Phase_1 Boundary Phase Bnd_Reg_2 Inlet Given Velocity Boundary Phase_1 Boundary Phase E Bnd_Reg_3 Inlet Given Velocity Boundary Phase_1 Boundary Phase Bnd_Reg_4 Inlet Given Velocity Boundary Phase_1 Boundary Phase Interface Regions Report Regions Solid Domains Figure 19 40 Solver setup input tree after grid import Global_domain_1 Global Domain A Domain Name Global_domain_1 Pa IV Set Filename port GRD Convergence Criterion ie 05 Figure 19 41 The Global Domain Panel Timebase Click on the Timebase entry in the Solver Setup Tree This allows the calculation timebase options to be defined In this port case the Steady State option will be used with a maximum of 1000 iterations The other options available are detailed below Timebase QA Time Mode le Steady C Unsteady Time Scheme Steady y Time Base Steady pe Number Of Iterations 1000 Figure 19 42 The Timebase Panel Time Mode Here the type of calculation time mode is selected 1 Steady A maximum number of iterations is chosen for the calculation If all the residuals for all the equations are less than the specified tolerance the calcu lation will terminate Ricardo Software December 2009 400 19 TUTORIALS 19 2 STEADY STATE PORT FLOW 2 Unsteady Time Scheme For transien
29. ORicardo Software December 2009 User Manual i Contents xiii List f Tables cios a dh O ds a A noe dE a di E xvi Listo RISUTES ici CM AGL Se o did iaa XXIV Preface 1 1 1 Contents of This Manual 0 ein A AA Re RE ee E 1 L2 Other Manuals 4 5 stes tad Mie dod we Bk wees Si a a lA aa 2 1 3 Acknowledgements ooo Bae bb eG a bee eG La 3 INTRODUCTION 4 2 1 Main Features and Capabilities e scra ose neca es 4 21 1 Meshing vmesh xs es wea eek A a Re a Re ee Re as 4 DAD SO VEE VSOIVE a eae ia a ee eS as 5 2 1 3 Pre and post processing R Desk 1 2 ee 8 2 2 Using VECTIS MAX Brief Guide 2 ee ee 9 GEOMETRY 11 Sul IO dUCHO sa si Sow a bd ee a a ee ee ed 11 22o User Interfaces paa ed a ee Ee AE AE oe de EN ee ee 11 3 2 1 The Button Bar commands sa oe ea ew 12 33 Graphics Interactions ss s ooa ae a Ro a ee ee a 14 3A Model MPOr isla a A Ge Be Bal a Bia de 15 3A Model A Sep ee ee ards A AE ge wah ge He rary ee a e 15 34 2 ASCII Triangle Files ocios 6264340440 eb eae i ea ba abe 16 3 43 VDA File Triangulation ses ew A ee EO ee Oe E 16 34A Meane 22s go sed amp Bea AN 17 3 5 General View Options 2 lt sonc g a cda ew a a we EE eo A 17 Bde View Men st ge ee eo eR a ew Be ee AR a ae EN ee ore 20 3 6 Triangle Processing ack aa Mea Me baw eee eee ae he ee ee dk as 21 3 7 Triangle Creation Operations gt o coca wa soomi eke ee a e a 22 3 8 Triangle Slicing and Intersection Oper
30. Ricardo Software December 2009 90 4 MESHING 4 7 PROBLEMS WITH QUALITY OF INPUT SURFACE Blend to boundary depth 1 a boolean information yes no which specifies whether the blending should start from refinement depth 1 Boundary refinement is applied after IJK refinement blocks i e it will override IJK refinement blocks However the forced refinement level and refinement depth are always only increased by boundary refinement specifications never decreased In Phasel boundary refinement specification can be stored to the trifile or meshfile the input ascii file In trifile the information is stored as 2D integer array VEC BOUNDARY_REFINE MENT as three values for each boundary BN boundary number DEEP refinement depth and BLEND blending distance signed as negative value where Blend to boundary depth 1 is switched on In meshfile the specification is BOUNDARY REFINEMENT NB DEEP BLEND There are several examples of combinations of boundary refinement depth and blending dis tances on figures 4 8 4 11 Comparison of figures 4 10 and 4 11 can shed light on meaning of the option refinement depth 1 both cases have refinement depth 2 blending distance 2 and the only difference is refinement depth 1 on off x Specification for boundary 2 E gt Delete Refinement depth at boundary fi Refinement blending distance i m Refinement Blending Blend to boundary depth Blend to boundary depth
31. The liquid phase can either be Water or Coolant 50 water 50 ethyl glycol This selection determines the method used to calculate the saturation temperature Tsat Since the Fluid Domain can have any number of phases the phase ID for the 2 phases liquid and vapour used in boiling have to be selected by entering the ID number of the phase in Liquid Phase ID and Vapour Phase ID respectively Calculation Option for properties such as Latent Heat of Evaporation and Surface Tension can either be Constant Values or Polynomial f Tsat available from the drop down menu In case of Mixture Model Option Boiling Models Cavitation Models Homogeneous Mixture Figure 12 3 R Desk setup for multiphase mixture models Boiling Model RPI Model VECTIS 3 Model Boiling Factors Evaporation 1 Condensation 1 Boiling Properties Based on Saturation Temperature Tsat Latent Heat of Evaporation Heat Flux 1 Option Constant Values Type of Liquid Value 2e 06 Coolant 50 Water 50 ethly glycol Water Surface Tension Liquid and Vapour Phase IDs Option Constant Values 2 Liquid Phase ID 1 Value 0 0715 Vapour Phase ID 2 Figure 12 4 R Desk setup for boiling models and parameters Ricardo Software December 2009 208 12 MODELLING MULTIPHASE FLOWS 12 4 PHASE CHANGE MODELLING polynomial option these two properties are dependent on saturation temperature 1
32. Y vecris tal Example of Wrapping Feature resolution size 0 02m no maximize feature resolution no decimation a tal Example of Wrapping Feature resolution size 0 01m maximize feature resolution threshold angle 5 degrees 3 14 2 Batch Mode The geometry wrapper can be run in batch mode The batch mode wrapping command line arguments are Ricardo Software December 2009 46 3 GEOMETRY 3 14 GEOMETRY WRAPPING phasel w fr feature_resolution ld leak_detection_mode da decimation_angle dd decimation_distance dv deviation_distance od offset_distance st surface_thckness ds distant_node_smoothing aa approximation_assessment i input_parameters_filename o output_filename input_tri_filename A default fr size is applied if no value is specified A default da is 2 degrees By default ds is applied If no parameter is specified in command line it will be read from the parameters file Boundary fr sizes and maximum refinement boundaries are read from the parameters file The no output filename is specified the file extension wrp is used for the output filename Feature resolution size This parameter approximately corresponds to the smallest features of the original geometry that will be automatically recovered If the geometry contains close surfaces fr size should be 1 3 of the smallest distance between them The algorithm is optimised for the fr value
33. bd wk Ga eRe dd REAR SRE We eR SS BS es 412 19 56 Launch Vsolve gt 0600s Aa oa ee eed ee eee Se AA BOR Ee Ee ee 412 L959 lave Update s remenna Yd e SR a a a a E ae Sea Ree 413 19 60Pressure variation at the model surface o s s cesson aos aa ai ee ee ee a 414 19 61Slice plot showing velocity vectors sura iom aae o 415 19 62The coolant jacket geometry when loaded into Phasel oaaae a 418 19 6 3Gasket Hole Detail s ie a smati m ae dd Ok aTa ee So a kaa ai aa Bw a R Sed 418 19 64 The Boundary Painting Panel 20d ia a ao eide e eA EE 419 19 65 The coolant jacket boundary setup ee 420 19 66 The mesh setup toolbar ausa s a a Bree th og Rg a wR a a a be Rw 420 19 67Re positioning the external red control lines o o e 421 19 68Using meshlines to control the cell sizes located in the gasket region 421 19 69Example control mesh number of sub divisions to give approximately 5mm size cells 422 Ricardo Software December 2009 xxiii LIST OF FIGURES LIST OF FIGURES 19 70Using meshlines to control the cell sizes located in the gasket region 423 19 71Chopping the geometry by IJK region o s sosoca sonada a E ee 425 19 72 Growing the chopped geometry lt e oc a sc ba ea ba eb bee eb ee eee a a a 425 19 73 Overlapping geometry region after repair 2 ee ee ee 426 19 74Problematic region in the geometry 2 eee 426 19 73Importing VECTIS 3 solver setup oca oe a a
34. can indicate the incompressible 1_bc ibcomp ir incompres weakly compressible 1_ bc ibcomp ir iwcompres Mach number Ma lt 0 3 subsonic 1_bc ibcomp ir isubsonic 0 3 lt Ma lt 1 and supersonic flow condition 1_bc ibcomp ir isupsonic Ma gt 1 O ibmat2 7 index of neighbouring material domain in case of interface regions Note that 1_bc ibmatl ir lt 1_bc ibmat2 ir During the solution procedure the initial boundary values can be fixed or can be updated depend ing on the boundary type The data retrieved using 1_reg_opts option integer iwp pointer r_opts Call get_reg l_reg_opts r_opts determines what kind of update if any will be done Thus for every equation ieq 1 to nte and region ir 1 to n_regions the following integer parameter can be assigned to r_ opts ieq ir ibmirr 0 perform zero order extrapolation mirror values O ibzerog 1 enforce the zero gradient boundary condition O ibextr 2 perform the 2nd order extrapolation from inside o ibflux 3 calculate the boundary variable from its flux O ibfix 4 keep values fixed and respectively O ibexpr 5 use another expression b To get a list of region boundary values for species then use real wph pointer sp_val call get_reg specs_value sp_val Array sp_val is used in conjunctio
35. copy amp edit the above code end subroutine upr_bnd_cond 18 7 4 Example of the user sources routine subroutine upr_sources mat ieq idt iph isp icell icel2 vol_ph ap_fi src_fi 1 This routine is called every outer iteration in the context of SIMPLE lalgorithm for each phase species transport equation which is solved for Modules used imported type definitions parameters scalars and arrays use upr implicit none integer iwp intent in mat amp index of material domain ieqg amp equation index ridt amp domain type index iph amp fluid solid phase index p isp amp fluid species index icell icel2 Istart end cell indices real wph intent in vol_ph icell icel2 cell volumexvolume_fraction real wph intent inout ap_fi icell icel2 6 amp equation central coeff src_fi 1 3 icell icel2 equation source term s Local Variables Scalar or velocity field corresponding to the equation index ieq real wph pointer o EL Ci gt null cell values of scalar real wph pointer Hi u gt null cell values of velocity Declare other pointers as required real wph pointer den gt null density field character len amp len_var_name usr_obj_name amp user selected name for leither phase or species eq_name amp name of transport equation Var_name Iname of variable field integer iwp be lobj amp phase species object indx usr_obj_
36. dispersed vapour gt From Lahey and Drew 2001 197 12 2 R Desk setup Setting solution of the volume fraction equation 203 12 3 R Desk setup for multiphase mixture models oaoa o e 208 12 4 R Desk setup for boiling models and parameters ooa 208 12 5 R Desk setup for cavitation models and parameters ooo o 211 13 1 Boundary and interface region nodes in the R Desk Solver Setup Tree 215 13 2 Boundary condition setup panel and boundary condition types in R Desk 216 13 3 Setting up an Inlet Boundary type outline of phases species and passive scalars in R Desk 218 13 4 Setting up a Mass Flow Rate Boundary type in R Desk nananana 219 13 5 Setting up a Pressure Boundary type in R Desk o o o o o ooo 221 13 6 Setting up a Stagnation Boundary type in R Desk o o o oo e 222 13 7 Setting up an Outlet Boundary type in R Desk o o o o oo o e 223 13 8 Setting up a Symmetry Boundary type in R Desk o o o o o e 224 13 9 Setting up a Wall Boundary type in R Desk onanan ee ee ee 226 13 10Setting up wall thermal conditions in R Desk drop down list of all conditions top and input variables for the Combined Convection amp Radiation in conjunction with Thin Wall bottom 227 13 11 Using the RTHERM Boundary Setting option to associate the wall boundary with the RTHERM A o AAA Bnd Sok F bake Heke o Bol oe
37. replaced by p The mass balance Equation 14 7 is already enforced by the corrected mass fluxes rit m There fore the second pressure correction equation can be derived from the following balance equation Vp j 2 Vo a 14 62 d i Cpp pp Vp L m 0 14 64 J Ricardo Software December 2009 250 14 NUMERICAL SOLUTION 14 2 SIMPLE BASED SOLUTION PROCEDURE Notably the second pressure correction equation has the same coefficient matrix as the equation for p Its inclusion can be beneficial in terms of improved convergence for highly non orthogonal grids i e for the grids with angles between A and dj less than 45 14 2 4 Boundary conditions for pressure corrections At all boundaries with known velocities the mass flux corrections are zero mp m m 0 Therefore at these boundaries the boundary coefficients a are zero In other cases the boundary velocities and mass fluxes need to be corrected In fully compressible flows Ma gt 0 3 this is the case at open flow boundaries inlets outlets free streams where the pressure is either directly or indirectly prescribed At these boundaries the flow is subsonic for the Mach number at boundary Ma lt 1 and supersonic for Ma gt 1 where the Mach number is defined as 0 e yYR T 14 65 Cc Ma with c being the speed of sound and y is the isentropic exponent y 1 4 for an ideal gas Here we outline the implementation o
38. 0 Boundary Condition Type Wall Prescribed Temperature a ad Given Heat Flux Boundary Condition Op Prescribed Temperature Boundary Setting Uniform Values Roughness Height 0 Heat Flux 0 Roughness Constant 0 5 Wall Velocity Temperature 293 15 xlo vfo z o Figure 13 9 Setting up a Wall Boundary type in R Desk These conditions will be displayed as shown in Figure 13 10 top after selecting the Thermal Condition Op ListBox Each selected wall thermal condition will be accompanied with corresponding LineEdit boxes where the user need to enter required variable values Figure 13 9 right illustrates the inputs for the Prescribed Temperature and Given Heat Flux An adiabatic wall is defined by setting a zero heat flux value In case of external convection and or external radiation the required input variables are shown in Figure 13 10 bottom For external Convection the values of heat transfer coefficients HTC and external Temperature should be specified If external Radiation heat transfer has been enabled the external Emissivity and external radiation Temperature need to be provided 13 8 3 Thin Wall Model The default wall thickness is assumed to be zero However the wall thickness can be taken into account by using the thin wall model which can be used in conjunction with any of the above wall thermal conditions If the thin wall mode
39. 16 2 Subdomain model scs s s a e ea we e a ee ee ha ee a 16 35 LD model vrs a aoe Sine So ea ae Be ea SS dow Bode se eae ds 17 USING SOLVER 17 1 InwodWctOn 2 4 bebe o dd Oe EERE a 17 2 Global Domain nas s ssa aa la AE a Mak ek BRE eee ea ee 173 Restat Control meo esoe a d m a e p Serko ke 2 Peas He de Pee cae ee 174 Fid Domain s deea asiaa a A is A a as a a 17 3 Monitoring Points s asnes oe ei e ae a a a as a ai IE O UA a 176 File QUtp t iosi i 2225 04 da a dd a dd e iea os TT BIO Phase 00000 a s a a A E a we A e a 17S Fluid SPECIES id o e a RA E AE e os IA a AE 17 9 Boundary Region Definition acs asii wa poe ee a A ea a 17 10 Report Resi OMS o sarae ye ee ee ae A ee ee 17 ALSolid Domain lt e a o Me db wee a ee ea oe Oe Gah Xd as 17 12 Y BCTIS Piles oc ci yaw dake Dah bathe eh bee ee bee ee RSG do A 17 13 Monitoring The Somuon sa s a onie eng ee ee ee ao ed a ed ed Bed bee So 17 14 Solver Control sida 24 28 Sag Bawa a ko Ha 2 Bede OS be Sed o eae des 17 15 Voolchain Launcher ui ca ee e ge A a ew 17 16WAVE VECTIS Co simulation 2 2 em ne 18 USER PROGRAMMING 18 1 Introduction And Overview 20000004 44 id es ee ee a AAG BO AN er eke 18 2 Functionality And Calling Sequence s sess sah aos a eo we ia a 18 3 Accessing Solver Variables ccros 2 a ade ak SM eee dk Aad es ORicardo Software December 2009 271 271 273 275 278 278 279 280 281 282 283 284 284 284 287 287 289 289 289 292 293
40. 1986 and Yakhot et al 1992 offered an alternative RNG k model Compared to the basic k model the equation has now an additional production term Sz 2 n 1 2 k ZB P 9 21 1 Bn k k n where no 4 38 and B 0 012 9 3 3 Standard k e model with the realisable time scale bound TSB Imposing the realisability constraint on the eddy viscosity relation for the Reynolds stresses Equa tion 9 4 written in the principal axes of S Durbin 1996 derived an upper bound on the turbu lent time scale 7 a y e A A 3C us where the constant 1 Combining this upper bound and the lower bound pertinent to the Kolmogorov time scale the turbulent time scale Equation 9 14 becomes k 1 u T max min C 9 23 ER m The realisability constraint on the time scale ensures that normal Reynolds stress components T are always positive 9 22 9 3 4 The k model coefficients The model coefficients appearing in the above k equations are assigned the values given in Table 9 1 Table 9 1 Model coefficients for the k models Model Cu Ox Og Ce Ce Ce Ce4 C 0 Mo B Pr SKE R 0 09 1 0 13 144 190 P P 2 2 2 Ce1 3 1 1 0 9 RNG 0 0845 0 72 0 72 1 42 1 68 P P 2 2 2 Ce1 3 4 38 0 012 0 9 9 4 Near Wall Modelling Wall Functions The k e models presented so far are of the
41. 2 Press the Add button 3 Whilst pressing and holding the left mouse button drag the mouse to draw a box around the cells for which the refinement should be applied 4 Change the view 5 Press the Edit button and then again whilst pressing and holding the left mouse button drag the mouse to draw a box around the cells for which the refinement should be applied 4 4 m1 IJK Refinement Parameters DEEP 2 FORCE 1 Lio oa The location of the IJK refinement block can be altered using the Edit button The DEEP and FORCE cell side division values are applied in the same way as the global refine ment IJK Cell Refinement DEEP and FORCE Options Cell refinement is the capability for varying the refinement level across the mesh Each global cell can have associated with it 1 Its maximum allowed refinement level IDEEP 2 A forced refinement level FORCE Ricardo Software December 2009 69 3 GEOMETRY 3 17 MESH SET UP VIEW OPTIONS The forced refinement level specifies a level of refinement to be applied to the cell before any ordinary refinement due to the presence of the boundary takes place For example if a cell has IFORCE 1 and IDEEP 2 it will first be split into a 2x2x2 set of cells and refinement down another level will then be applied if the boundary passes through or close to the cell Thus it is possible to effectively have a finer mesh localised to a particular area e g the valve
42. 7 DP op se Ceon o 1 Gy 5 if P lt P 12 76 B 1 Both Cevap and Ceonq are given values of 1 0 for this model 12 4 2 4 Setting up a Cavitation Model Two phases need to be defined as in the boiling case the liquid and vapour phases It is recom mended that the compressibility for the vapour phase should be set to weakly compressible To set up a cavitation model select the appropriate Fluid Domain from the Solver Setup Tree This opens up the Fluid Domain set up panel Under Multiphase Modelling select the Mixture option As aresult a Multiphase option is added to the Solver Setup Tree Left clicking this option opens the Multiphase panel Under Mixture Model Option select Cavitation Models Three different cavitation models are available Singhal et al Zwart Gerber Melamri and Schnerr Sauer Cavitation Model Schnerr Sauer Singhal et al Zwart Gerber Belamri Cavitation Properties Saturation Pressure Cavitation Factors Option Constant Values Evaporation 0 02 Value 3533 3 Condensation 0 01 Surface Tension Liquid and Vapour Phase IDs Option Constant Values Liquid Phase ID 1 Value 0 0715 Vapour Phase ID 2 Figure 12 5 R Desk setup for cavitation models and parameters Ricardo Software December 2009 211 12 MODELLING MULTIPHASE FLOWS 12 4 PHASE CHANGE MODELLING The default values for evaporation and c
43. Due to thin interface wall thermal resistance the interface tem perature has dual values Thus one has to distinguish between the temperature at the lower interface interface side towards a material with a lower index and between the temperature at the upper interface interface side towards a material with a higher index Note that the thin wall model will be inactive if its thickness is set to zero Other attributes can not be changed they have been extracted from the non partitioned grid file during Domain Structure creation and they are used by the VECTIS MAX solver to check the validity of a grid file and extraction process ORicardo Software December 2009 233 14 NUMERICAL SOLUTION The numerical solution of the modelling mean flow equations is based on the finite volume dis cretisation and on a SIMPLE like segregated procedure for the velocity pressure density cou pling Patankar 1980 Karki and Patankar 1989 Demirdzic et al 1993 Przulj and Basara 2002 The following sections describe O Finite Volume Discretisation considering the discretisation of a generic transport equation O Poor Quality Cell Treatment and O Solver Parallelisation 14 1 Finite Volume Discretisation The solution domain is discretised with an unstructured numerical grid described by edges of convex polyhedra cells All solution variables are located at cell centroids The governing equations can be conveniently expre
44. Equation 8 7 The ideal mixtures are supported in VECTIS MAX For such mixtures the species thermodynamic density p m V is related to the partial volume of the species which is the volume V that would be occupied by mass m at the pressure and temperature of the mixture As Y V V the mixture density p m V follows from the identity Ns Ns N L A 1 Vi _ A mi Pi X Cj 1 t p 8 37 V m p rz ta Pi and it is given as Nsp o a 8 38 Ricardo Software December 2009 136 8 MODELLING CONTINUA 8 5 PROPERTIES OF MULTIPHASE MIXTURE Other properties of the multicomponent phase are calculated from Ns Pp ci i 8 39 i l where j is the property value of the i th species The above expression can be used to evaluate the composition dependent laminar viscosity u specific heat cp thermal conductivity A and gas constant of an ideal gas mixture R Note that the specific volume of the mixture 1 p also conforms to Equation 8 39 8 4 1 Reacting mixture flows In a chemically reacting flow where reactants disappear and products are formed it is necessary to specify specific internal energy and enthalpy with reference to the standard reference state usually defined at Tye 298 15 K and pref 101325 Pa At this enthalpy datum a value of zero is assigned to the stable elements e g H2 O2 N2 and the enthalpy of formation is assigned to compounds e g water H20 carbon dioxide CO2 The enthal
45. Monitoring Monitoring points are specified within the model domain where values of the solution are output at the end of each time step A cell ID number or X Y Z co ordinate can be specified for monitoring velocity values Additional monitoring points can be specified using the Add button Values for the first monitoring point are written to the screen Values for all of the monitoring points are written to files if reporting options are set in the Output Panel Next the fluid phase needs to be set up Fluid Phase The fluid phase panel contains data for each fluid phase Each phase may be made up of a single or multiple species Phase Name Each phase can be given a name to aid reference Mixture of Species Option Each phase can be made up of a number of different compo nents species In this tutorial a single component phase is considered More details of the different options can be found in Setting a multiphase mixture Ricardo Software December 2009 403 19 TUTORIALS 19 2 STEADY STATE PORT FLOW Fluid_1 Fluid Domain Figure 19 45 The Monitoring pint panel Ricardo Software December 2009 404 19 TUTORIALS 19 2 STEADY STATE PORT FLOW Phase_1 Fluid Phase x Phase Name Phase_1 Mixture Of Species Option Single Component Phase zl Phase Property Filename be a wj Phase ji SyS Phase Type le Gas C Liquid Compressibiity Fuly Compressible S
46. Multiphase flow modelling deals with flow situations where more than one fluid phase occupies the fluid domain In comparison to the multicomponent fluid flow the multiphase flow is mixed at the Ricardo Software December 2009 137 8 MODELLING CONTINUA 8 6 SELECTING CONTINUUM AND ITS PROPERTIES macroscopic level The most common is the Euler Euler modelling approach which requires the solution of volume fraction equations for fluid phases For a small control volume V the volume occupied by a fluid phase k Vp is Vi OV where is a volume fraction The volume fraction is a function of space and time As the volume of a phase can not be shared by another phase the sum of all volume fractions is equal to one Y amp 1 The mass fraction of the k th phase ck mg m m Y my is related to the volume fraction through the expression e mg PrVk om Pk 8 44 m Pm V Pm where p m V is the phase thermodynamic density and p is the multiphase mixture density Pm 7 Y m V Y mi04 Ve Y Pros 8 45 k k k Depending on the selected Euler Euler model each fluid phase can have its own flow field the full Eulerian Eulerian model or it can share some common fields VOF model for immiscible fluids and the mixture model The properties that govern the transport equations of VOF and mixture models therefore depend on the properties of constituent fluid phases In case of Np fluid phases the volume fraction mean p
47. NUMERICAL METHOD lt Transforming models into a system of algebraic equations O Spatial meshing amp time discretisation Discretisation of modelling equations Initial and numerical boundary conditions Solution algorithm linearisation amp linear equation solver NUMERICAL RESULTS Post processing In what follows the fundamental mathematical models that underpin the solver design are pre sented O conservation equations Ricardo Software December 2009 113 6 SOLVER FUNDAMENTALS 6 2 CONTINUUM CONSERVATION EQUATIONS O their closure problem O Reynolds and Favre averaging which leads to the O Reynolds averaged equations Other solver related chapters describe modelling of Spatial domains O Continuum material properties Turbulence O Single phase flows and O Multi phase flows The next two chapters are devoted to the implementation and setting of Boundary Conditions and to the Numerical Solution 6 2 Continuum Conservation Equations Before presenting basic conservation equations describing both single phase and multi phase flow several terms relevant to continua and multi phase flows are introduced 6 2 1 Terminology Continuum The term continuum is used to describe the matter which is continuously distributed in space Mathematically properties of continuum such as pressure density temperature and velocity are defined as single valued continuous functions o
48. Note that 12 is optional for some of the variables in Table 18 8 If it is used its content will not be altered Upper bounds for arrays i1 i2 can be obtained using function get number Ricardo Software December 2009 305 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES subroutine get_ domain var_name il 12 Input Output var_name cha i1 integer pointer optional i2 integer pointer optional n_dom_mat list_dom_mat ise_phase ise_bnd_reg ise_specs ise_ps ise_cell ise_bnd_face ise_internal_face ise_high_interface ise_low_interface number of materials for each domain 11 1 n_domains list of domain materials 11 1 2 n_mat_doms starting phase index 11 1 n_domains starting boundary region index 11 1 n_domains starting species index 11 1 n_domains starting passive scalar index 11 1 n_domains starting cell index 11 1 n_domains starting boundary face index 11 1 n_domains stating internal face index 11 1 n_domains starting higher interface index 11 1 n_domains starting lower interface index 11 1 n_domains ending phase index 12 1 n_domains ending boundary region index 12 1 n_domains ending species index 12 1 n_domains ending passive scalar index 12 1 n_domains ending cell index 12 1 n_domains ending boundary face index 12 1 n_domains ending internal face index 12 1 n_domains ending highe
49. TUTORIALS 19 2 STEADY STATE PORT FLOW FORCE refinement These are set to 2 and 1 by default They divide the mesh blocks within them into 2x2x2 or 4x4x4 blocks for values of 1 and 2 respectively Force values divide up all mesh blocks whereas deep values only divide the mesh blocks that also have a boundary in them 19 2 8 Boundary Refinement Mesh refinement can additionally be specified on a per boundary region basis This can useful when the region to be refined is complex or when refinement is used to capture particular geometric features Boundary refinement is specified in the Boundary Painting panel used initially to define the sep arate boundary regions Click on the refinement column for boundary 2 the boundary containing the back of the valves Set the refinement level Refinement Depth at Boundary to be 3 This means that cells in adjacent to the boundary will be sub divided 3 times Next set the Blend ing Distance to 2 this determines that the refinement will be applied to two global cells before decreasing by a level Then select Blend to boundary depth 1 this forces the specified level of refinement to be used only in the subdivided cells that remain adjacent to the boundary The remainder of the global cell will be refined at one level lower than the specified level If Blend to Boundary Depth is used then the entire global cell adjacent to the boundary will be sub divided to the specified level Furthe
50. The canvases can be tiled using the window gt Tile options Drag the slice plot into the second canvas The data plotted on the slice can be changed using the data panel More information on the general post processing capabilities can be found in the R Desk help 19 2 15 Initialisation Tips In general correct initialisation of the calculation will give faster convergence In some cases were the initialisation of the calculation is poor the simulation will diverge very quickly after starting Frequently there are a few things that can be done to improve the initialisation of port flow cases particularly when the flow rate is high When using the potential flow solver with pressure boundary conditions on both the inlet and outlet the potential scale factor may need to be reduced The potential solver neglects any viscous effects and the initial mass flow can therefore be over predicted A value of 0 7 to 0 8 is usually sufficient to give a more stable initialisation although lower values may be necessary in extreme cases ORicardo Software December 2009 414 19 TUTORIALS 19 2 STEADY STATE PORT FLOW Dx File Edit View Options Window Help OBS 3 Pots ax 121342 2 plots Eh da Plot pi Canvas 3d1 al pr zi Plot Properties ax Lines Elmer y Faces 7 Color 1 5 Opacty foie CC H WP eee A FRPP amp Plot Properties
51. amp Transport Properties Properties arising from the constitutive laws are often called transport properties The constitutive laws define the viscous stress tensor 7 heat flux vector q and mass flux vector of a chemical species J in terms of the velocity vector U temperature T and species concentration mass frac tion c respectively 8 1 1 Momentum transport in fluids and molecular viscosity For Newtonian fluids the constitutive relation between the viscous stresses T and the rates of deformation strain tensor 1 OU OU S 1 8 1 J 2 5 F is given by the Stokes law 2 OU where u denotes the molecular dynamic viscosity 130 8 MODELLING CONTINUA 8 1 CONSTITUTIVE RELATIONS The viscosity describes the flow resistance of simple fluids At moderate pressures the viscosity in gases increases with increasing temperature whereas decreases in liquids The Sutherland formula 8 3 based on a kinetic theory is a good approximation for viscosity of dilute gases The viscosity Uo is the viscosity at reference temperature T while S is an effective temperature in K Sutherland constant The following values are valid for air at moderate pressures and temperatures Up 1 716 x 105 kg ms To 273 15 K and S 111K Another expression for dilute gases is the power law formula T n e 8 4 U Uo To gt where iy denotes the viscosity value at reference temperature 7p and n is the
52. defined as AT sub Tsat T 12 35 Enthalpy of the fluid is calculated as follows h Cpfmin Tsat T Usq 12 36 where us describes the work done by the fluid motion and is calculated as squared velocity vector i Usg 7458 12 37 Enthalpy of saturated liquid hsf Ces fT sat US Usq 12 38 Enthalpy of saturated gas hsg hsf hye 12 39 O Boiling at a wall surface The total wall heat flux is expressed as 4 spl Gnuc 12 40 where qspj is the single phase heat flux The nucleate boiling heat flux qnuc is calculated ac cording to Rohsenow 1952 empirical correlation as amp Pf P c AT yy 293 qnuc Csfhfh fg Pr Ps a 12 41 O hare where C f is an empirical coefficient varying with the liquid surface combination g is gravity O is the surface tension ATs is the degree of super heating Other quantities like cpf and Pr represent the specific heat and Prandtl number of the liquid phase respectively The coefficient AT yup is given by ATsup Twatl Tsat 12 42 where Twall Tsat and T are heated wall temperature the local saturation temperature and mean flow temperature respectively O Evaporation rate from the wall Using the total heat flux q from relation 12 40 the evaporation rate from the wall is then calcu lated as 7 107 Fam Es 12 43 hfe Cp fAT sub with m calculated from 1 y 12 44 p x_mass ratio_d 1 x_mass and x_mass following for rela
53. flow and outflow boundaries For incompressible or weakly compressible flows Mach number Ma lt 0 3 the total pressure Prot can be used at the inlets to internal flows instead of the static pressure pp The Stagnation Inlet type is a better choice for the sub sonic flows Ma gt 0 3 The following expression relates the static pressure to the total pressure 12 Pp U Pb Ptot al n for inflow 13 7 Ptot for outflow The implementation of pressure boundary conditions is discussed in section describing Boundary conditions for pressure corrections The treatment of other dependent variables such as temperature and turbulence quantities depends on the flow direction Hence at inflow boundaries the specified inlet values will be used whereas at outflow boundaries all variables will be extrapolated 13 4 1 Setting up a pressure boundary condition A pressure boundary can be set up by left clicking on a boundary region in the Solver Setup Tree and the boundary setup panel is displayed Under Boundary Condition Type select Pressure and the Pressure Boundary panel with relevant settings is displayed as in Figure 13 5 A number of options is available for the pressure boundary type Given Static Pressure Total and Average For each of these options the boundary condition can be setas Inflow or Outflow under Inflow Outflow ListBox as indicated in Figure 13 5 right Setting up boundary values for phases species and passive scalars is do
54. high Reynolds number type in that they are not ap plicable in the near wall region More precisely they can t account for viscous and wall blocking effects In order to account for these effects they ought to be either used with Ricardo Software December 2009 154 9 MODELLING TURBULENCE 9 4 NEAR WALL MODELLING Wall functions formulae that describe the flow variables within the fully turbulent logarithmic layer and bridge the viscous affected layers O or re formulated as Low Reynolds number models that can be integrated up to the wall As a very dense numerical mesh is required to resolve the near wall region the wall functions require significantly less of computing time than the low Re number models The log law velocity profile forms the basis of the the standard wall function approach The purpose of this approach is to link the solution variables at near wall cells placed within the log law layer with those at the wall However the key difficulty is provision of the desirable mesh clustering near the walls especially in case of automatic mesh generation For example the Cartesian cut cell grid generators always produce the near wall grids with very different cell volumes Therefore the first grid points i e the near wall cell centres can lie within either viscous sub layer or buffer layer or logarithmic law region This means that any type of wall functions have to provide formul
55. i e U U the control volume becomes identical to the control mass and the Lagrangian form of equations is recovered Noticing that the instantaneous velocity vector U and pressure p as well as other flow quantities vary both with a position x and time both coordinate free integral and Cartesian differential form of conservation equations are presented 6 2 2 1 Mass Conservation lt S pav 45 0 0 d 0 6 1 22 p oon s In the above equations P represents density If the continuum is a multi component phase a mixture of species the mass conservation for each species i must be satisfied This conservation is usually expressed in terms of the local concentration mass fraction which is the ratio of the mass of i th species m to the total mass m of the mixture gt Mm Ci 6 3 m The corresponding transport equations are d Fa ee gt B o ae f PdV PA 0 0 dA Te dA s dV 6 4 dt Jy A A V Ricardo Software December 2009 115 6 SOLVER FUNDAMENTALS 6 2 CONTINUUM CONSERVATION EQUATIONS das dia ee d x amp Ba pei 0 Ues Kai a 6 5 ot Ox j Ox j where i j is diffusion flux of species and Sc denotes the species source The source can be the net rate of production per unit time and volume due to chemical reactions If there is Nsp species then the definition of mass fractions implies Nsp y o 1 6 6 1 6 2 2 2 Momentum Conservation d ee RATA Ss gt LS PE lt
56. if P lt P 12 69 P Prat 0 5P p 12 70 where Psat is the liquid saturation vapour pressure and Port represents the local values of the turbulent pressure fluctuations and is estimated as Pr p 0 39pk 12 71 where k is the local turbulent kinetic energy To account for laminar flow set the minimum for vk to 1 The turbulence contribution in relation 12 70 is neglected P Pyar 12 4 2 2 Zwart Gerber Melamri model The Zwart et al 2004 model is based on the assumption of a constant bubble size The phase change sources I and I q are calculated as evap con 3 Anuc 1 v 2P P ue Cea Je E iPP 12 72 B l 30 0 2 P P od Re E A ifP lt P 12 73 where Cevap and Ccona are the evaporation and condensation coefficients with values 50 and 0 01 respectively nuc is the nucleation site volume density and is given the value 5x10 Re is the bubble radius and is given the value of 10 m Ricardo Software December 2009 210 12 MODELLING MULTIPHASE FLOWS 12 4 PHASE CHANGE MODELLING 12 4 2 3 Schnerr Sauer model The Schnerr and Suaer 2001 model uses a variable bubble radius diameter g directly in phase change sources and is defined as 1 3 an z aa 12 74 The optimal value for bubble number density n is in the region of 10 The phase change sources I p and rk ond are then defined as 7 2P P Cerap orth 1 E a if P gt P 12 75 B 1
57. if more than one agi boundary to specify 106 5 READING amp MANIPULATING MESHES 5 2 MESH PARTITIONING 5 2 Mesh Partitioning In order to run the solver vsolve in parallel on multiple processors it is first necessary to de compose partition the single mesh file into the corresponding parallel files Mesh partitioning is controlled via the np and optionally the metis argument For example vpre np 4 metis rb test GRD will create 4 new mesh files in the corresponding directories P001 P002 etc using the recur sive bisection method There are 3 possible partitioning methods all METIS based rb Recursive Bisection kway K way vkway VK way In general the default partitioning method used is recommended By default for np lt 8 the recursive bisection method is used otherwise the kway method is selected For further infor mation visit the site http glaros dtc umn edu gkhome views metis If the grid file contains more than one material then METIS will by default also attempt to equally divide the materials amongst the partitions This behaviour can be disabled via the nomat option In addition to the automatic mesh partitioning the user can manually define mesh cut planes via an input file PRECUT DEF This file is used only if it is found in the working directory The typical format within is PLANE ORIGIN PLANE ORIGIN PLANE ORIGIN 0 LOFO 1 15 0 NORMAL 0 1 0 eg He
58. optional domain id iget_dom domain id material id iget_mat material id phase id iget_phase phase id parent name species id iget_specs species id passive scalar id iget_ps passive scalar id boundary region id iget_breg boundary region id interface region id iget_ireg interface region id Table 18 24 Subroutine get_parent to get parent identifier If the domain type does not exist null value is returned eq_idt 0 For example if you want to access momentum equation variables for fluid material domain mat 2 the command idt eq_idt ieq ifmom iobj mat iget iget_mat will return idt which is then used to access members of momentum type eg the velocity field is described by tmom idt Su Note however that not all variables i e members of derived type are used in the given run For example the old velocity vector field within the momentum type does not exist for a steady run Since all type members are Fortran 95 2003 pointers you can check if the variable values are available by using the Fortran 95 2003 logical inquiry function associated pointer Thus associated tmom idt u returns either t rue or false value Furthermore obj is related to iget i e iget_dom is used then obj must be domain id etc function eq_idt ieg iobj iget Arguments Description eq_idt integer index of domain type for the given equation ieq iobj gt integer either domain or material or phase or species or
59. or solid domain sub domain boundary or interface region or sub domain interface is selected it will be highlighted and shown in the preview window In Figure 7 3 the fluid domain has been selected and previewed while Figure 7 4 previews a pair of sub domain interface regions Fluid_1 Fluid Domain Monitoring Points Algorithm Turbulence Model Discretise air Fluid Phase Initial Condition Postprocessing Output Equations amp Solver Fan Model Boundary Regions Bnd_Reg_1 Mass Flow Rate Boundary air Boundary Phase Bnd_Reg_2 Pressure Boundary air Boundary Phase Bnd_Reg_3 Wall Boundary Bnd_Reg_4 Wall Boundary Bnd_Reg_5 Wall Boundary Bnd_Reg_6 Wall Boundary Bnd_Reg_7 Unjoined Boundary Bnd_Reg_8 Unjoined Boundary Bnd_Reg_9 Wall Boundary Interface Regions Port_porous_block Sub domain Sub domain Interface Regions Subdominterface_1 Sub domain Interface Region Subdominterface_2 Sub domain Interface Region Report Regions Solid Domains A Figure 7 4 R Desk setup Example of a single fluid domain tree with a porous sub domain left and its sub domain interface regions right Ricardo Software December 2009 128 7 MODELLING SPATIAL DOMAINS 7 5 CREATING A DOMAIN STRUCTURE Note that in the case of multi domain structure some boundary regions can have the Unjoined Boundary attribute This attribute indicates that a small number of boundary faces of
60. pointer optional i2 integer pointer optional mat_type type fluid or solid mat_compress parent_dom ise_phase ise_bnd_reg ise_interf_reg ise_specs ise_ps ise_cell ise_bnd_face ise_internal_face ise_high_interface n_low_interface isa_low_interface L_low_interface n_low_interf_reg isa_low_interf_reg l_low_interf_reg 11 1 n_mat_doms compressibility flag 11 1 n_mat_doms list of domains 11 1 n_mat_doms starting phase 11 1 n_mat_doms starting boundary region 11 1 n_mat_doms starting interface region 11 1 n_mat_doms starting species 11 1 n_mat_doms starting passive scalars 11 1 n_mat_doms starting cell 11 1 n_mat_doms starting boundary face 11 1 n_mat_doms starting internal face 11 1 n_mat_doms starting high interface 11 1 n_mat_doms number of lower faces at material interfaces 11 1 n_mat_doms initial address for lower interfaces in 1_low_interface 11 1 n_mat_doms list pointer of faces at lower material interfaces 11 1 n_interf_regs number of lower regions 11 1 n_mat_doms initial address for 1_low_interf_reg 11 1 n_mat_doms list of lower interface regions 11 1 n_interf_regs ending phase 12 1 n_mat_doms ending boundary region 12 1 n_mat_doms ending interface region 12 1 n_mat_doms ending species i2 1 n_mat_doms ending passive scalars 12 1 n_mat_doms ending cell 12 1 n_mat_doms ending boundary face 12 1 n_mat_doms end
61. pointer xbnd gt null lupr gt Diamond case real wph pointer ub Cees gt null teb gt null reg_bval gt null zinl 0 50 UZ 0 50 tez 0 50 U2Z V2Z UVZ zfac uzk tezk integer iwp ie k idm ndata ifun real wp pointer real wp parameters scalars and arrays amp domain type index amp boundary region index amp starting bnd region face ending bnd region face name of a variable field pointer array for a scalar field fib_vec gt null pointer array for a vector field pointer array for bnd conditions amp user selected name for either phase or species object luser selected variable name amp flag to get phase species index amp index of phase or species object Icorresponding to upr_obj_name amp index of user selected reg amp boundary face index cell index lbnd face centre coordinates amp bnd face velocity lbnd turbulent kinetic energy region wise var bnd values amp coordinate data for the profile amp velocity profile amp turbulent energy profile amp normal turbulent stresses amp shear uv stress amp interpolation factor amp inerpolated bnd velocity linterpolated turb energy amp loop index amp domain type index for momentum amp number of data file unit for reading bnd cond Do nothing for a symmetry or outlet boundary type call get_reg 1l_reg_cond 1_bc if 1l_bc ibct ir ibsym or
62. puav pu U 0 dh f pl di dd pfav 6 7 dt Jy A A A V O fas O les fa Op a a A 01 zz PO B U F ay Gi 94 6 8 where I is the unit tensor and T 7 j denotes the viscous stress tensor One can notice that the Cauchy stress tensor O pl 7 or Gij Pi Tj 6 9 is decomposed into the pressure and viscous stress tensor 6 is is Kronecker symbol 6 1 if i j and 6 0 otherwise The gravitational body force and other external body forces are represented as fi 6 2 2 3 Energy Conservation The First law of thermodynamics applied to the given control volume can be cast in the form of the total energy E o Ne e E 6 10 Ed 00 das 6f 0 a av A k 1 e PE rae PE 0 U a e yi Fey OG BRO 8 6 11 Ot Ox p J 8J OX qj E kc J 19 PJiVi qv where the total energy is given as the sum of internal kinetic and other forms of energy 22 a U E e cs other forms 6 12 Further 7 is the heat flux vector hy represents the specific enthalpy of k th species and gq is the energy source or sink term The source term can include the energy source due to chemical Ricardo Software December 2009 116 6 SOLVER FUNDAMENTALS 6 3 CLOSURE PROBLEM AND AVERAGING reactions between species The term involving specific enthalpies of species takes into account transport of different enthalpies by individual species The energy conservation expressed in terms of total energy is comm
63. sii 2 Ox E Dx eR The term ME represents the inter phase momentum exchange for the k th phase and requires modelling Energy Conservation k a fut 12 7 a a put x a p n ut Of el oe HEN oT nee ET z fo E Prk Cp dx t of oF Ox U oht th H 12 8 where H is the total enthalpy H Cr Ukuk 2 k and Hk describes the energy inter phase transfer terms Equation of state If a phase is an ideal gas its density is calculated from the equation of state P 5 12 9 with i being the phase gas constant Turbulence Equations The turbulent viscosity u is predicted in conjunction with the modelled equa tions for the turbulent kinetic energy k and its dissipation rate e These equations are similar to the single phase ones Equations 9 19 9 20 They read as d kyk kakrkrrk ak pk pk nakok d uN okt k z a oK 5 PKU a Prot 5 at H E Fx 0210 d kokok lar Cerri Ceap CesPh e TF k k pk ok d Mr del ko mk 0Ce4p E Sn Fo 2 uk R 2 12 11 xj Of x Ay kk 2 pk pic EL 12 12 In the above equations K k and Z k represent inter phase turbulent transfer rates for k and e respectively The unknown inter phase correlations associated with transport of mass I momentum MK energy Hk and with turbulent quantities Kk and ak require closure in terms of dependent variables The complexities of multi fluid mo
64. tangential movements rotate about z whilst radial movements rotate about x and or y as appropriate The centre and edge behaviours blend smoothly into one another Model zooming In conjunction with the CTRL key the middle mouse button performs zooming moving up and to the right zooms in and moving down and to the left zooms out The right mouse button is used to perform zooming and translation If the button is depressed and released in the same place that point is moved to the centre of the window If however the mouse button is held down while the mouse is moved the user can drag out a rectangle with one corner at the point where the mouse was first depressed The rectangle has the same aspect ratio as the graphics window and defines the area that will be visible when the mouse button is released Thus the user can zoom in very rapidly on a chosen portion of the screen Keyboard shortcuts The manipulations described above can also be accomplished from the keyboard Ricardo Software December 2009 14 3 GEOMETRY 3 4 MODEL IMPORT F1 and F2 x rotation F3 and F4 y rotation F5 and F6 z rotation F7 and F8 x translation F9 and F10 y translation Fll and F12 Zoom The near and far clipping planes are moved with the following keyboard keys U key Near plane outwards Ikey Near plane inwards O key Far plane inwards Pkey Far plane outwards Holding down the SHIFT key while moving the clipping planes moves both planes at on
65. var_name cha ival integer pointer turb_ctrl_vars list of turbulence control variables ival 1 nturb_par 1 n_domains subroutine get_ turb var_name rval Input Output var_name cha rval real pointer cmu C_mu rval 1 n_fluid_doms cappa von Karman const rval 1 n_fluid_doms log_law_const log_law constant rval 1 n_fluid_doms yptrans nondimensional wall distance rval 1 n_fluid_doms ypvisc viscous sub layer non dim thickness rval 1 n_fluid_doms subroutine get_turb var_name rval Input Output var_name cha rval real pointer eps_const prandtl_number k eps eps eq const rval 1 n_domains 1 6 turb prandtl numbers rval 1 n_domains I nte Table 18 19 Subroutine get_turb to get turbulence variables defined in access group iacc_turb subroutine get_run_ctrl var_name ival id Input Output var_name cha id integer ival integer current_iter end_iter current_tstep end_tstep post_freq current iteration ival ending iteration ival current time step ival ending time step ival id for the global domain files ival frequency for printing into project subroutine get _run_ctr1 var_name proj_nums character Input Output var_name cha proj_nums project_name proj_run_number name of current project proj_nums c
66. 1 and can be arranged as x mj dj VS ag OPOP ee J eo fj if mj lt 0 fj ae A 14 13 with fi being the flow orientated interpolation factor The blending factor depends on the grid resolution and for sufficiently fine grids the values close to one can be used The alternative is to adopt the high resolution bounded scheme which were discussed in Przulj and Basara 2001 but also covered here 14 1 3 1 Boundedness criteria Let us consider transport of a scalar along the local coordinate which passes through the upstream U central C and downstream D computational nodes Figure 14 3 The actual la Ricardo Software December 2009 237 14 NUMERICAL SOLUTION 14 1 FINITE VOLUME DISCRETISATION belling of these nodes depends on the velocity direction i e on the sign of mass flow rate m1 We require that the cell face value fulfils the following two conditions O itis bounded by the neighbouring cell values fc and p and O enforces monotonicity of the local function amp which passes through the points gy and dc The first condition represents the interpolative boundedness Gaskell and Lau 1988 and can be generalised as E pe pa E ey 14 14 dp Oc 1 c The second monotonicity condition reads Figure 14 3 Definition of upstream central and downstream nodes E j dc c du c In the above equations the normalised variable Leonard 1988 is introduced 0 0 14 1
67. 115e 05 12e05 125e 05Mn 9 41 rr Figure 19 85 Pressure variation at the model surface 19 3 9 1 Data extraction Typically the purpose of performing a coolant flow analysis is to determine two main things ORicardo Software December 2009 434 19 TUTORIALS 19 3 COOLANT FLOW O System pressure drop across the coolant circuit O Mass flow split through the gasket holes The most straightforward way to determine the system pressure drop from inlet to outlet is to view the ascii io files or the binary rep report file This will show the pressures and mass flows at all the boundaries Alternatively if the pressure drop is desired from locations away from the boundaries then slice plots can be created in the locations of interest and the static pressure plotted on the slice Then right click on the slice plot in the plottree and select Calculate gt Mean Weighted gt Area_Mean This will report the average pressure for the slice plot The mass flow split through the gasket holes can be determined in a couple of ways Arbitrary surfaces may be used to output this data as the simulation is running See the Port flow tutorial for instruction on using arbitrary surfaces Additionally the mass flow split can also be extracted during post processing in rdesk a ob MR Fie Edt View Optons Window Hep la x DSa x By gt e va 125 000 Reset iv Animation gt gt F auto apply Apply F Aut
68. 17 9 BOUNDARY REGION DEFINITION 1 n XAVE Xidi Ars Outer iteration number steady state Time unsteady Time step unsteady Total area Average temperature Total mass flow rate Average density Average static pressure Average total pressure Table 17 4 Table of all the variables contained in io files Fcoeff_X Fcoeff_Y Fcoeff_Z VFcoeff_X VFcoeff_Y VFcoeff_Z Ypstar_MIN Ypstar_MAX Heat_Flux Wall_T NearWall_T Htc Similarly wall report files will typically contain the following values Total wall force normalized x component Total wall force normalized y component Total wall force normalized z component Viscous wall force normalized x component Viscous wall force normalized y component Viscous wall force normalized z component y minimum value over boundary y maximum value over boundary Total heat flux Average wall temperature Average near wall temperature over fluid cells Average heat transfer co efficient Table 17 5 Table of all the variables contained in wall files Ricardo Software December 2009 where yp is the dimensionless wall distance turbulent see Section 9 4 1 pCR k 2yp u 17 1 where X is the value and A is the area at a given face i on a boundary comprising of n faces 17 2 286 17 USING SOLVER 17 10 REPORT REGIONS and yp is the perpendicular distance from the neighbouring cell centre to the wall The wall fo
69. 19 The Fluid Phase Output Panel 300 degK The input file should now be complete for the analysis Save the file as tube inp file gt save as 19 1 9 Grid Preparation In a command window run the vpre mesh pre processor This will re order the grid for use with the solver vpre can also be used to re partition the mesh for parallel calculations Type vpre tube GRD in the cmd window 19 1 10 Running the solver Running the solver The simulation is now ready to be run Type vsolve tube GRD in a command window The solver will start ORicardo Software December 2009 383 19 TUTORIALS 19 1 BASIC TUTORIAL Bnd_Reg_1 Wall Boundary 2 a x Region Name End_Reg_1 Region ID fi Boundary Condition Op Presenbed Temperature 3 Inlet Given Velocity Mass Flow Rate Boundary Setting fun form Values Outlet Roughness Height 0 Symmetry Wall Roughness Constant 0 5 Pressure Wall Velocity sfo mfo Bk Ba bon Figure 19 20 The Boundary Region Panel 19 1 11 Live Update Live update is a utility in R Desk that allows a simulation to be monitored whilst it is running Firstly open a xy canvas using the new XY canvas button Then in the Live Update panel browse to the directory where the simulation is running The available data files are presented in the Files window The data files available are determined by the ascii files selected in the reporting section of the 19 1 8 1Glob
70. 29 Refer to Equation 9 26 for mathematical description Consider an example where boundary region with id ir 1 fluid material mat 1 and idt 1 The following example illustrates the use of deln_star function Ricardo Software December 2009 337 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES subroutine get_grad idom idt gfi_out var_name fi_c fi_b fi_ui fili Arguments Description idom integer fluid solid domain id idt integer domain type index efi _out real pointer scalar gradient of var_name var_name character optional variable name _c gt gt real optional scalar variable cell values fi_b real optional scalar boundary values fi_ui real optional scalar upper interface values fi_li real optional scalar lower interface values subroutine get_grad idom idt gx_out gy_out gz_out var_name fi_c fi_b fi_ui fi_li Arguments Description gx_out real pointer grad vector x component gy_out real pointer grad vector y component gz_out real pointer grad vector z component fic real pointer optional vector variable cell values fi_b real pointer optional vector boundary values fi_uil real pointer optional vector upper interface values TEE real pointer optional vector lower interface values Table 18 28 Subroutine get_grad to get gradient of scalar or vector
71. 3 10 Triangle Marking Operations Mark Line Button This operation is used to select a part of the model to which further commands can be applied Ricardo Software December 2009 31 3 GEOMETRY 3 10 TRIANGLE MARKING OPERATIONS The starting point is selected by clicking on the model with the left mouse button A line will be drawn from the location clicked with the other end attached to the mouse Further clicks of the left mouse button will create corners in the line Clicking twice in the same place will end the line Pressing the escape key or selecting another operation will cancel the line marking operation Once the line has been drawn all of the triangles cut by the line will be highlighted in a shaded red colour Once the operation has been completed Phase 1 will return to the Manipulate View state Note that the view may not be manipulated while the line is being drawn Mark Face Button This operation is used to select a part of the model to which further commands will be applied A triangle is selected by clicking with the left mouse button and that triangle becomes the starting triangle for the construction of a face The face contains all of the connected triangles going outwards from the first triangle until a boundary of marked triangles is met Therefore to mark an area of the model the perimeter of that area should be marked with the Mark Line command and then any triangle within that perimeter selected for
72. 4 oc ee ie a wd ek be hha a Ee ee wae ae BR oe es as 62 3 17 Mesh Setup View Options x s os 24 eae ele ee Pe ee ea aa ed Ba Oe ee es 64 Ricardo Software December 2009 li 3 171 UK Refinement Blocks ios Ge A e BG A a Bee Rew 3 17 2 Control Mesh Setup Suggestions s o s u o ee 3 18 Warning and Error Messages c ce q oca ca ba ea ba vd bweeiwe tee ei ed bas SLOP The Veetisichs Files ie s s a lt a Hh eee ew od Hh HE HH Se AE ee Bie Gs 4 MESHING 4 1 IntroducttOii css chek ek bao ARS Ae EERE Oe eR be wh wd bas 4 2 Howto R n VMESH ou teria aora a ba a ke od a 4 3 Setting Up the Input Pile s s s a eek we we Hw a ew ee wed a 44 Command Line OPluONS uo c4 ios Bade ea aa Paka 2a Pe de PS be Peds nae de 4 5 Basic Scheme of VMESH coseno ae Ee ee ERE de ee 4 6 Generation Ot BOXES sse osre eS Ge bow adr a A ae OE Oe ee ae we ee 4 6 1 Box Generation Procedure s ses taone cna eed ee oe Gid A a el a a wd 4 6 2 Parameters controlling generation of boxes o o e 4 7 Problems With Quality of Input Surface 2 ee ee ee 4 8 Meshing for Multidomain Simulations 2 0 000000 02 eee eee 4 9 Warnings and tons ia eR ee A A a A as e 410 Grd Data SUCHTE nt ok ek O E a gene eh sapdb ge deh Be eae ios SE O aces 410 1 Grid components 2 2 2 misa ba ee iweb ela Ca bd a 4 10 2 Calculation of grid geometric properties s sce scaccu a aaae daa eaaa 4 11 Mesh Imports 22 2 24 3 ps de a ata eed ee
73. COOLANT FLOW STATISTICAL DATA OF GENERATED MESH Number of generated cells 63113 boundary 46967 internal 16146 Patching method Number of cells processed by Marching Cubes 36534 73 85 Number of cells processed by Exact Fit 8424 17 03 Number of volumes broken by Cell Splitting 2503 5 06 of boundary cells 618 attempts to split a volume failed 19 80 of all attempts 262 attempts gave incorrect number of volumes 356 attempts gave volumes with too low quality so the undo was applied Cell quality Number of correct boundary cells 47461 95 93 NO PROBLEMS with negative volumes Number of cells with problems of gaps 2 0 00 There were 2281 small volumes they are deleted now 4 61 Number of cells which had to be deactivated 2009 4 06 Total mesh volume 1 75013e 003 u3 1 09839e 003 u3 16146 inner cells 6 51745e 004 u3 47634 boundary cells WRITING THE MESH FILE successfully finished Total time elapsed 52 seconds SUCCESSFULLY DONE Now that the computational mesh has been generated we can move onto the next stage setting up the input file for the solver 19 3 5 Solver Setup Coolant analyses for gasket optimisation and general coolant jacket development are usually steady state since both the geometry and flow properties are constant with time Therefore for this example the analysis will be setup to use the VECTIS MAX steady state solver This is an example analysis wh
74. DNS Results and Modelling Journal of Fluid Mechanics vol 305 pp 185 218 6 4 6 4 Ishii M 1975 Thermo Fluid Dynamic Theory of Two Phase Flow Eyrolles Paris 12 1 12 2 Jayatillaka C V 1969 The Influence of Prandtl Number and Surface Roughness on the Resistance of the Laminar Sublayer to Momentum and Heat Transfer Prog Heat Mass Transfer vol 1 p 193 9 4 2 Jones W and Launder B 1972 The Prediction of Laminarization with a Two Equation Model of Turbulence International Journal of Heat and Mass Transfer vol 15 pp 301 314 9 2 1 Kader B A 1981 Temperature and Concentration profiles in Fully Turbulent Boundary Layers Int J Heat Mass Transfer vol 24 no 9 pp 1541 1544 9 4 2 Kalitzin G and laccarino G 2003 Toward Immersed Boundary Simulation of High Reynolds Number Flows Center for Turbulence Research Annual Research Briefs pp 369 378 Stanford University 14 4 Kalitzin G Medic G laccarino G and Durbin P 2005 Near Wall Behavior of RANS Turbulence Models and Implications for Wall Functions Journal of Computational Physics vol 204 pp 265 291 14 4 Karki K C and Patankar S V 1989 Pressure Based Calculation Procedure for Viscous Flows At All Speeds in Arbitrary Configurations AZAA Paper 86 0207 vol 27 pp 1167 1174 14 Kenning D B R and Victor H T 1981 Fully developed nucleate boiling overlap of areas of influence and interference between bubble sites Int J He
75. December 2009 447 19 TUTORIALS 19 4 IMPORTING THIRD PARTY MESHES Leonard B P 1988 Simple High Accuracy Resolution Program for Convective Modelling of Discontinuities International Journal for Numerical Methods in Fluids vol 8 pp 1291 1318 14 1 3 1 Leschziner M A Batten P and Loyau H 2000 Modelling Shock Affected Near Wall Flows with Anisotropy Resolving Turbulence Closures International Journal of Heat and Fluid Flow vol 21 pp 239 251 9 3 Meijerink J A and Van Der Vorst H A 1981 Guidelines for the Usage of Incomplete Decompositions in Solving Sets of Linear Equations As They Occur in Practical Problems Journal of Computational Physics vol 44 pp 134 155 14 2 6 Menter F 1994 Two Equation Eddy Viscosity Turbulence Models for Engineering Applications AZAA Jour nal vol 32 pp 1598 1605 9 2 1 Moran M J and Shapiro H N 1992 Fundamentals of Engineering Thermodynamics 2nd ed John Wiley amp Sons Inc New York 6 2 1 8 3 2 Moser R Kim J and Mansour N 1999 Direct numerical simulation of turbulent channel flow up to Re 590 Physics of Fluids vol 11 no 4 pp 943 document 9 4 4 9 2 9 3 Murthy J Y and Mathur S R 1998 A Conservative Numerical Scheme for the Energy Equation ASME Journal of Heat Transfer vol 120 pp 1081 1086 14 1 6 14 1 6 Muzaferija S 1994 Adaptive Finite Volume Method for Flow Predictions Using Unstructured Meshes and Multigrid
76. Either the number of sub divisions can be entered or a cell length The cell length will be adjusted when it is found to be approximate to be the length that results in the nearest integer number of divisions To remove a main mesh line click on the Delete Mesh Line button and click on the mesh line to be removed The main mesh line nearest to the point clicked will be deleted The Delete Mesh Line button El deletes a main mesh line 3 17 Mesh Set up View Options Show Mesh Setup T Number cells 3d Subdivide IJK Block Display C Off Solo All Selecting the Show Mesh Set up option adds the outline of the mesh to the view of the model Ricardo Software December 2009 64 3 GEOMETRY 3 17 MESH SET UP VIEW OPTIONS If a mesh has not been defined for the current model then a default mesh is created and displayed This default mesh is the simplest mesh that satisfies the mesher requirement for a halo of cells around the model The Number cells option will add the IJK indices of each of the mesh cells to the view The Show Mesh Set up option will allow the IJK blocks that may be defined during mesh set up to appear in a 3d view Normally IJK blocks may only be seen in the 2D view projections to prevent the view from being cluttered Ricardo Software December 2009 65 3 GEOMETRY 3 17 MESH SET UP VIEW OPTIONS The 3d Subdivide option performs a similar function allowing th
77. In addition other time dependent related parameters can also be specified such as Time Dependent Boundary Conditions Number of Steps Time step Maximum Number of Outer Iterations and Convergence Criterion Ricardo Software December 2009 236 14 NUMERICAL SOLUTION 14 1 FINITE VOLUME DISCRETISATION h x Time base Time Mode Steady e Unsteady Time Scheme 1st Order Euler v Time Base Seconds v Time Mode Time Dependent Boundary Conditions o Steady Unsteady Number Of Steps 1000 i Y Timestep 0 001 Number Of Iterations 3 Maximum Number Of Outer Iterations 10 Convergence Criterion 1e 05 Figure 14 2 R Desk setup for steady and unsteady simulations 14 1 3 Interpolative schemes for cell face values Obviously the cell face values and gradients are important ingredients of the discretisation proce dure They can be calculated using linear interpolation CDS between the values at adjacent cells P and P Figure 14 3 0 fj0r 1 fj op 14 10 where the interpolation factor fj is given as 7p F j 14 11 fj ke j IFP Fjl F Fpl and F denotes the position vector Since CDS is not always appropriate for the convective term a simple and efficient approach is to blend the CDS scheme with a small amount of the UDS scheme Demirdzic and Muzaferija 1995 Ferziger and Peric 1997 The blending is done by introducing a blending factor 0 lt y lt
78. It should be a monotonically decreasing curve i e Pressure as VolumeF low 1 16 2 Subdomain model Selecting this model assumes that a subdomain has already been defined in the corresponding mesh file This model is currently only applicable to axial fans with straight blades It is also assumed that the inlet flow velocity is uniform and has no tangential component For a constant fluid density throughout the fan the axial flow velocity V xial at inlet must equate to that at the outlet as enforced by mass conservation The fan is assumed to impart a tangential force Fg on the fluid along its blades and is directly related to the change in tangential velocity between the fan s inlet outlet Eq 16 1 Fo PranVaxiatA VE Vie 16 1 As the inlet flow is assumed to be purely axial Vj is taken to be zero For Vg if the Ricardo Software December 2009 273 16 MODELLING FANS 16 2 SUBDOMAIN MODEL fluid is assumed to leave along the blade surface in the blade s frame of reference and the axial component is V jq then the relative tangential velocity is V iqitan B see Fig 16 3 Va relative Vaxial V relative blade angle ll wr blade velocity Figure 16 3 Velocity vectors at fan s outlet Thus the absolute outlet tangential velocity is ye or Vaxial tan B 16 2 where is the fan s rotational frequency rad s r is the prependicular distance from the fan s axis B is th
79. MAIN X MESH LINES EXCEEDS MAXIMUM ALLOWED VALUE LINE NOT CREATED The user has tried to insert a new mesh line in the mesh set up section of Phase 1 which would cause an array overflow Corresponding warnings 0415 and 0416 are for the other two directions ERROR 2201 VECTIS PHASE1 CONS consname ON SURFACE surfname DOES NOT START AT THE END OF THE PREVIOUS CONS FACE WILL BE TRIANGULATED AS A SURFACE Ricardo Software December 2009 75 3 GEOMETRY 3 19 THE VECTIS CFG FILE it it and ERROR 2202 VECTIS PHASE1 CONS LOOP NUMBER n ON SURFACE surfname HAS DIFFERENT START AND END POINTS FACE WILL BE TRIANGULATED AS A SURFACE The VDA standard requires that a set of CONS curves form a closed loop in order to define a FACE a trimmed surface The program has detected a non closed loop which can not be closed by joining its end points see warning 0204 The trimming information is therefore discarded and the untrimmed surface is triangulated ERROR 2206 VECTIS PHASE1 ELEMENT elemname IS NOT A CONS ON SURFACE surfname PROGRAM TERMINATING and ERROR 2207 VECTIS PHASE1 CONS consname FOR SURFACE surfname NOT FOUND PROGRAM TERMINATING and ERROR 2208 VECTIS PHASE1 ELEMENT elemname IS NOT A SURFACE PROGRAM TERMINATING These are errors in a VDA file associated with the naming and referencing of elements ERRORS 2501 2504 Various errors in the format of STL files 3 19 The Vectis cfg File The VECT
80. Operations on fully stitched closed volumes in both GUI interactive mode and also batch mode These are Union Subtraction and Intersection operation involving at least two separate geometry entities 3 9 1 GUI interactive mode The volume operation command buttons are So o for Volume Union Volume Subtract and Volume Intersect respectively When these command buttons are selected the user is prompted to select the volumes to operate on A volume is selected by left clicking the mouse on a triangle that belongs to the required volume The selected volume is then highlighted in red Once at least two volumes have been selected the user is asked confirm the operation by firstly left clicking the mouse on the last volume selected and then choosing the OK button from the confirmation pop up window More than two volumes can be selected by continuing to left click on different volumes so that the boolean operation once confirmed will then perform sequential operations using the volumes in the order they were selected The Volume operations preserve boundary and part definitions so that the single resultant volume will be composed of the same parts and boundaries as the original volumes Ricardo Software December 2009 29 3 GEOMETRY 3 9 BOOLEAN VOLUME OPERATIONS Cylinder and Port Volumes red highlight Cutaway Showing Volume Intersection Ricardo Software December 2009 30 3 GEOMETRY 3 10 TRI
81. Option Constant Values Value 0 Value 0 Turbulent Schmidt Number Initial Species 0 Option Constant Values 4 Value 0 9 Figure 8 10 R Desk setup Setting properties of individual species The calculation options for species properties are specified in a similar way as for the single component fluid in accordance with Table 8 1 The Initial Species edit box is provided to specify the initial value of species mass fraction Phase with WAVE properties As Figure 8 3 indicates a phase can have the WAVE attributes which means it is a multi component fluid phase consisting of five species components air fuel vapour burned air burned fuel and liquid fuel When this type of phase is selected the Fluid Phase Setup panel Figure 8 11 left will display the From Wave option for a Density and Specific Heat For all other properties the Mixture option is pre defined Density Phase_1 Fluid Phase Initial Condition Option From Wave a Output Species air Fluid Species Viscosity Output Option Mixture a fuel_vapour Fluid Species Properties Output Conductivity burned_air Fluid Species Properties Output burned_fuel Fluid Species Properties Specific Heat Output liquid_fuel Fluid Species Properties Output Option Mixture Option From Wave Figure 8 11 R Desk setup Setting properties of WAVE components species Ricardo Software December 2
82. Parts F Holes F Sharp Edges _ Number Nodes Part and Boundary Panel Parts Boundaries Show Tri Type Inlet Oi Inlet OL Add Boundary Delete Boundary Show All Hide All Toggle Compress J Reduce Paint All Mesh View Paint Face Auto Paint n a d qo Paint Line Motion 14 08 15 Model dimensions xmin 0 040000 xmax 0 010000 ymin 0 025000 ymax 0 005000 _ zmin 0 000000 zmax 0 100000 Refinement Motion Info Auto Paint Angle as vi Figure 19 3 Geometry Preparation and Boundary Painting In this tutorial case the above is achieved in the following steps 1 Open Phasel either typing phasel in a command prompt or from the start menu 2 Open the file tube tri File gt Open then select tube tri 3 Open the boundary painting panel operations gt boundary painting 4 Add two new boundaries 5 Paint each end of the tube as a separate boundary using paint face Ricardo Software December 2009 369 19 TUTORIALS 19 1 BASIC TUTORIAL 6 Change the two ends to inlet outlet boundaries click on type in boundary panel The prepared geometry should appear as in Fig 19 3 19 1 4 Defining the Global Mesh The purpose of this section of Phase 1 is to set up the global control mesh lines used as the basis for mesh generation The mesh setup tools are activated by selecting
83. Preconditioning options are given in Figure 14 5 A number of ILU Preconditioning Options for Parallel Runs are also provided such as Localised Mixed and Global The Number of Inner Solver Iterations and Number of Preconditioning Iterations can be specified directly including the Solver Tolerance value 14 3 Implementation of boundary conditions Considering boundary faces the convective and diffusion fluxes are calculated in the same manner as for the inner cell faces The convective fluxes are evaluated by using the upwind scheme In case of diffusion flux Equation 14 33 is modified as D T A p bp r on VO A Ay is 14 73 2 SS SS b b bb A b OP obV OP b ae where the subscript b signifies the boundary face The above equation accommodates both Dirichlet conditions specified boundary values and Neumann conditions prescribed boundary flux In the latter case it is used to compute the boundary values In case of momentum and energy the diffusion coefficients at wall boundaries I are defined by Equations 9 40 Ricardo Software December 2009 253 14 NUMERICAL SOLUTION 14 4 POOR QUALITY CELL TREATMENT 14 4 Poor Quality Cell Treatment In VECTIS MAX poor quality cells are identified applying the following criteria O Maximum cell face and boundary face skewness angle see Equation 14 34 O Minimum volume change ratio this is defined as the ratio between a cell volume and
84. Ricardo Software December 2009 407 19 TUTORIALS 19 2 STEADY STATE PORT FLOW Y Ricardo VECTIS Phase 1 port tri E lt lo x File Edt view Toobars Operations Help 6 a E m e a Noa Y Yl Options Stitch mesn Tnangles AA JA a a 2 BA OR DUPLICATE LINES W LINE NUMBERS COMPLETE 70464 LINES WERE FOUND O LINES NEED STITCHING Figure 19 49 Cut the model in the location of the arbitrary surface is required Y cardo VICTIS Phase 1 po aloj x SaaS m Optons Stitch mesn 2 Tnangles Ha j A 2 Delete marked mangles oe ee TS 17 ESTAS u SSSSU Imak Us S7ES7U m TABLE ENTRIES AFTER LINE SORTING 3 BAGE COLLECTION FOR DUPLICATE LINES TABLE ENTRIES 38 LINE SORTING AND ADDRES 70464 LINES WERE FOUND O LINES NEED STITCHING Figure 19 50 Delete the marked area Ricardo Software December 2009 408 19 TUTORIALS 19 2 STEADY STATE PORT FLOW lol x Fle Edt View Toobars Operatons Help aaa m e a INAS ASA Optons Stitch Mesh Dr T Tnangles gt E JO del Se A Aj BIC lIa o IES E de d 5 E FOUND O LINES NEED STITCHING Lol Figure 19 51 Model cut at location of arbitrary surface lol Fie Edt View Toobars Operators Hel
85. Series von Karman Institute for Fluid Dynamics 14 1 4 Benzi M 2002 Preconditioning Techniques for Large Linear Systems A Survey Journal of Computational Physics vol 182 pp 418 477 14 2 6 Bo T 2004 CFD Homogeneous Mixing Flow Modelling to Simulate Subcooled Nucleate Boiling Flow in SAE International 2004 01 1512 12 4 1 1 Brackbill J Kothe D and Zemach C 1992 A Continuum Method for Modelling Surface Tension Journal of Computational Physics vol 100 pp 335 354 12 3 2 Bradshaw P 1994 Turbulence The Chief Outstanding Difficulty of Our Subject Experiments in Fluids vol 16 9 1 9 3 Davydov B I 1961 On Statistical Dynamics of an Incompressible Turbulent Fluid Soviet Physics Doklady Fluid Mechanics vol 6 pp 10 12 9 2 1 de Lemos M J S 2005 Fundamentals of the Double Decomposition Concept for Turbulent Transport in Permeable Media Mat Wiss u Werkstofftech vol 36 no 10 pp 586 593 11 1 2 de Lemos M J S and Pedras M H J 2001 Recent Mathematical Models for Turbulent Flow in Saturated Rigid Porous Media ASME Journal of Fluids Engineering vol 123 pp 935 940 11 1 3 11 1 3 Demirdzic I Lilek Z and Peric M 1993 A Collocated Finite Volume Method for Predicting Flows At All Speeds Int Journal for Numerical Methods in Fluids vol 16 pp 1029 1050 14 14 2 3 14 2 4 445 19 TUTORIALS 19 4 IMPORTING THIRD PARTY MESHES Demirdzic I and Muzaferija S 1995
86. The last term in the above equation is modelled with the help of thermal tortuosity conductivity tensor nin a ATn dA e 11 31 In conjunction with eddy viscosity k e modelling the volume averaged turbulent heat flux d is evaluated according to Equation 9 9 n o ay PeT L a wat 11 32 J where the turbulent thermal conductivity A is given by Equation 9 10 The thermal dispersion term p H U j is usually modelled as a function of the intrinsic temperature gradient Le o xx 0 YT p AU pe TU Pri 11 33 P gt pe PO ager 11 33 by introducing the thermal dispersion conductivity tensor Ags One can define an effective fluid conductivity tensor e A Ar Sj A AGE 11 34 and calculate the sum of various heat fluxes in terms of the effective heat flux xx 1 0 yf EENS t A e G aj G PTO Mi 11 35 Ricardo Software December 2009 187 11 MODELLING POROUS MEDIA 11 1 THEORETICAL BACKGROUND The corresponding volume averaged energy equation for the solid medium can be written in a similar way as for the fluid medium subscript s denotes solid 1 aT 1 9 T7s Y PsHs sito As MS jdA Ps9sv Wj A 11 36 Ot YP s Ox nabs yf ax YsPs4s qs j Y s jk Dx where the effective solid conductivity is defined as AS jk As Oj A 11 37 Noting that the unit interface surface vector ns nj Ys 1 y and assuming con
87. There are additional measurable mass fluxes caused by a pressure gradient pressure diffusion in centrifugal separators and by external body forces force diffusion in plasma technol ogy Normally the ordinary diffusion term is dominant and other types of diffusion including the thermal one are currently neglected in VECTIS MAX In turbulent flows Equation 8 9 is replaced by Je gt a Ve 8 10 Scis where Sc is the turbulent Schmidt number Mr Sci 8 11 a PD aa Lu is the turbulent viscosity and Y denotes the turbulent mass diffusion coefficient see section about eddy viscosity formulation The turbulent Schmidt number is of the order of unity default value 0 9 and the same constant values can be used for all species As the turbulent diffusion usually overwhelms the laminar one the constant values of laminar diffusion coefficients Y are an appropriate choice 8 2 Equation of State amp Thermodynamic Properties The thermodynamic state of a simple compressible system at equilibrium is fixed by any two independent thermodynamic properties In CFD pressure p and absolute temperature T are Ricardo Software December 2009 132 8 MODELLING CONTINUA 8 2 EQUATION OF STATE used as primary solution variables Therefore density p internal energy e thermal enthalpy h e p p and any other thermodynamic property could be determined as functions of p and T p p p T e e p T h h p T and o p T N
88. This is necessary to ensure the same positions of meshlines when generating cells on the complemen tary meshes Also a different settings for deleting of small cells is applied here see the option smallint MAKE BOUNDARIES CONFORMAL operates on finished gridfiles conform GridfileA bA1 bA2 bAn GridfileB bB1 bB2 bBn When this switch is used the two gridfiles GridfileA and GridfileB are loaded and all patches belonging to boundaries bA1 bA2 bAn on the A geometry are tried to be made conformal to boundaries bB1 bB2 bBn from the B geometry The changed gridfiles are written down to files with postfix _conform In order to perform this task successfully it is necessary to use int interface command when generating both gridfiles before using this tool conformrew GridfileA bA1 bA2 bAn GridfileB bB1 bB2 bBn When this option is used the same actions are performed as in the case of conform The only difference is that the given gridfiles GridfileA and GridfileB will be rewritten instead of creating _conform GRD files Ricardo Software December 2009 83 4 MESHING 4 5 BASIC SCHEME OF VMESH cb_tolrnlevs number_of_levels This option controls the tolerance used for linking corresponding vertices of the two complementary geometries The tolerance is calculated as the diagonal from the bounding box of the two geometries divided by number_of_levels The default value of number_of_levels is 1 0 x 107 cb_
89. VECTIS consists of several stages beginning with Geometry Import and Preparation The is performed using the Phase Preprocessor This has a number of main functions 1 To read in geometry from an external source and convert it to a form acceptable for the VECTIS mesh generator vmesh 2 To identify regions of the geometry as different boundaries for the CFD calculation 3 To set up the global mesh and other control parameters for vmesh 4 To start up the mesh generator and view the resulting mesh Mesh Generation Vmesh is the VECTIS mesh generator It is fully automatic and produces a locally refined Carte sian mesh which is suitable for fluid flow analysis using the VECTIS solver vsolve The input to the mesh generator is a completely closed fully connected triangulated surface which contains boundary information as described in the Phasel section Phase 1 is used to supply this surface and a further file containing the meshing control parameters Vmesh writes the computational Ricardo Software December 2009 2 INTRODUCTION 2 2 USING VECTIS MAX BRIEF GUIDE grid GRD file which contains information about the mesh geometry and its associated boundary faces Mesh Preparation The vpre mesh preparation stages allows further manipulations of the mesh generated by the vmesh program 1 Mesh sub division into separate partitions for parallel calculations 2 Conformal joining of grid files 3 Impor
90. VOF MODELS 12 3 2 VOF equations The VOF model is essentially homogeneous multi fluid model Thus the mixture modelling equa tions in the previous section describe the VOF model if the terms associated with the drift slip velocity are omitted Several HRIC schemes to deliver the sharp interfaces have been designed The recent one and quite popular is the scheme of Ubbink 1997 As the VOF method deals with the interfaces between liquid and gas the surface tension force might have a significant contribution to the inter phase momentum transfer depending on the values of the Reynolds Re pLU u Capillary Ca uU 0 and Weber We pLU 0 numbers where is the surface ten sion and U L are the characteristic velocity and length Generally the surface tension force is negligible for Ca gt gt 1 or We gt gt 1 where the Ca number is important for Re lt lt 1 and the We number for Re gt gt 1 At a curved interface the surface tension force is resolved into two components along the unit normal vector ni directed from liquid to gas and along the tangential unit vector f do dt where K is the surface curvature The surface tension force is usually implemented by using the Continuum Surface Force CSF model of Brackbill et al 1992 The CSF model can also take into account the effects of wall adhesion in situations when the interface is in contact with the wall fio Fioni figti OKn 12 24 12 3 3 Setting
91. Y Ricardo VECTIS Phase 1 coolant tri 10 x Fie Edit View Toobers Operations Help Aaga gastos J INA ASANI Options Stitch mesn Triangles VER pra Model IJK Chop Mesh Input File lant mesh Import Mesh Set up File Max 1 42 42 e7 l reset Limit Fe cal K 3 as reset OK Apply Cancel mput file read Figure 19 71 Chopping the geometry by IJK region Y Ricardo VECTIS Phase 1 coolant tri 10 x Fie Edit View Toolbars Operations Help BARA SACI J JING Y M Options Stitch mesh gt Triangles AAA VEND A AD Depth of growth region A as l a Enter depth 3 OK Apply Cancel Figure 19 72 Growing the chopped geometry Ricardo Software December 2009 425 19 TUTORIALS 19 3 COOLANT FLOW Y Ricardo VECTIS Phase 1 coolant tri 10 x Fie Edit View Toobars Operations Help 2 el ul el Mea _ a SI 2 e Y Y Options Stitch Mesh Triangles TE A Al TITZST T TT TIT ut 11 25 58 Select point 2 with the left mouse button 11 25 58 Siet point 3 with the left mouse pose tton ated triangle 11 26 00 Select point 1 with the left mouse butt r X EN Figure 19 74 Problematic region in the geometry Using the delete triangle and cap hole tools this imperfection can be removed 19 74 Ricardo Software December 2009 426 19 TUTORIALS 19 3
92. a first order ordinary differential equation where the term 9 represents the fluxes and sources as nf np ny t a j l j l E sa 4 04 j 1 j Equation 14 8 is integrated over each time step t which advances the solution step by step in time Then this equation is discretised according to the implicit scheme There are 2 implicit schemes available first order accurate Euler scheme and second order accurate three time level scheme Both of these schemes are discussed in further in Transient Term section 14 1 2 1 Setting up a steady or unsteady flow Using R Desk left click on Global Domain node in the Solver Setup Tree and the Global Domain panel is displayed to the right Now parameters for steady and unsteady flow can be set up by selecting the appropriate radio button as in Figure 14 2 In case of steady state simulation Number of Iterations and Convergence Criterion can be specified When the flow is unsteady then the user can select the Time Scheme using the ListBox to either be 1st Order Euler or 2nd Order Implicit three time level scheme Next the user must select the Time Base which can be one of Seconds Non dimensional 2 stroke and 4 stroke For the Non dimensional case the Cycle Time Number of steps per cycle Start Time and End Time must be provided When using the 2 Stroke or 4 Stroke option the Time Step Start Time End Time and Engine Speed are required in the units indicated e g degrees and rev min
93. a pair of inlet outlet boundaries or sub domain interfaces O Using Solver chapter is intended to provide the user with information not contained in the pre vious chapters how to monitor the solution use the restart control interact with the solver do WAVE VECTIS MAX co simulation and understand the VECTIS MAX output and alpha numeric reports O User Programming tells the users how to write and compile their own Fortran 95 2003 source codes Various tasks can be performed such as the definition of boundary and initial conditions material properties or output results O Tutorials are the starting point for those who want to try VECTIS MAX immediately without reading other chapters of this manual This chapter contains the basic tutorial followed by two application examples steady state port and coolant flows and by mesh import tutorial O The manual also contains a bibliography used in various chapters 1 2 Other Manuals Currently there are no other manuals related to VECTIS MAX Ricardo Software December 2009 2 1 PREFACE 1 3 ACKNOWLEDGEMENTS 1 3 Acknowledgements VECTIS MAX uses the following 3rd party software O MPI Message Passing Interface http www mpi forum org O METIS Graph Partitioning from Karypis Lab University of Minnesota http glaros dtc umn edu gkhome views metis O TETGEN Tetrahedral Mesh Generator from Weierstrass Institute http www wias berlin de O CGNS CFD General Notatio
94. a wh ae ee E A a te aa 383 19 19 The Fluid Phase Output Panel 2 2 c ar cado A eee eee a 383 19 20The Boundary Regi n Panel s e s asos sadet 2 e eae ew eRe ee 384 19 21 Comparison of IO mass flow convergence with different flow initialisation 386 19 22Using the chop geometry command o assia ck ee ee ee ee 388 19 23 Inlet outlet and main wall boundaries sos s cerca a a a 389 19 24Back of valve defined as separate wall boundary to allow specification of boundary refinement 390 19 25Example of Global Mesh 2 430 240 4044 ia 444 Hho YR Ee ee he da 390 1926Settine UK refinement o sos dana deet ie a See ee BOR ee Ee eA 391 19 27 Setting Boundaty Refinement 2 44 42244 ako ba ede ad be dea eds be ead 392 19 28 The New Project dialog DOX coi 6S oe A eh ew a ed a ee a 393 19 29 Visualising the geometry file in R Desk 0 0 000000 04 2 eee 393 19 30 The Launch Mesher dialog DOR s o e 2 64 b4 4 b4 64 4d wae ee ew a 394 19 31 Screen output fromthe mesher e a dce esa we we ek eae ee eed ed 394 19 32Removine a plot trom acanvas i s ic eani a ee a Eee e A eA eS 395 19 33Drop down list allowing panels to be opened or closed o o o o o ooo o 396 19 34 Removine a plot froni aCanvasS sa acs s redii ca A AR AR ee a E 396 19 35Open a new VECTIS project ek oocococo so as ee ee 397 19 36Default panel layout for VECTIS Project 2 2 o e 397 19 37 The Solver Setup input tree add oe ee AE
95. bridge region in a coolant flow which does not extend throughout the mesh in the x y and z directions Forced refinement should be used with care level 1 produces 8 cells from each global cell level 2 produces 64 level 3 produces 512 Similarly using too large a level IDEEP refinement should be avoided There can be an enormous increase in number of cells produced when using IDEEP 2 as com pared to IDEEP 1 Depths greater than two should be used only with caution The way the refinement level is varied is done by specifying IDEEP and IFORCE for I J K blocks of cells This information appears in the MESH INP file as follows IJK_BLOCK is ie js je ks ke ideep iforce and can be defined in the mesh setup section of Phase 1 Any number of lines of this type can be put in the file anywhere between the comment line line 2 and the mesh coordinates Where defined blocks overlap the values of IDEEP and IFORCE from later blocks overwrite the values from earlier ones 2PEEP _ 22 4 for surface cells Therefore the global surface cells can be divided into 4 cells in each co ordinate direction 2FORCE 21 2 for internal cells All cells inside the IJK refinement block will be sub divided by 2 in each co ordinate direction A slice of the computational mesh created using the control mesh shown above is shown below The local refinement region can clearly be seen Ricardo Software December 2009 70 3 GEOMETR
96. by using R Therm module This file contains thermal condition data for various engine components wall boundaries which are represented by Ricardo s SFE Standard Finite El ement sets as shown in Table 13 1 Note that the full RTH set names got the prefix FE Within VECTIS MAX solver a thermal condition type for each wall is assigned according to the Solver Name types which are also given in Table 13 1 Ricardo Software December 2009 227 13 BOUNDARY amp INTERFACE CONDITION TYPES 13 8 WALL Boundary Setting RTHERM Roughness Height 0 Roughness Constant 0 5 gt Wall Velocity Xx 0 Y0 zZ 0 Heat Flux 0 Temperature 300 Convection HTC 0 Temperature 0 Radiation Emissivity 0 Temperature lo Thin Wall RTHERM Set Selection Set cylinder CYL_1 1D_headGasFace Figure 13 11 Using the RTHERM Boundary Setting option to associate the wall boundary with the RTHERM Set Selection Engine Component RTH Base Set Name Liner Flame Face Inlet Port Exhaust Port Inlet Valve Contact Exhaust Valve Contact cylinder CYL_x ID_bore cylinder CYL_x ID_headGasFace port CYL_x VAL_ ly port CYL_x VAL_Ey valveSeat CYL_x VAL_ ly valveSeat CYL_x VAL_Ey Solver Name Wall Condition cylinder variable HF cylinder variable HF port uniform EHT port uniform EHT valveSeat uniform HF valveSeat uniform HF Table 13 1 Descriptio
97. can be specified by selecting one of the following Boundary Condition Op Normal Flow Direction Flow Direction Along Grid Lines and Given Flow Direction By Unit Vector For the latter option Velocity Direction must be defined in terms of the unit vector X Y Z components In the next step the user should enter values for Mach Number an estimated value Total Pressure an absolute pressure value and Total Temperature Setting up boundary values for phases species and passive scalars can be done in a similar way as Ricardo Software December 2009 22 13 BOUNDARY amp INTERFACE CONDITION TYPES 13 6 FLOW OUTLET Bnd_Reg_8 Stagnation Boundary 8 K Region Name Bnd_Reg_8 Region ID 8 Material ID 1 y Boundary Report Coupled Link Number 0 Boundary Condition Type Stagnation 2 _ zzz Boundary Condition Op Normal Flow Direction Flow Direction Along Grid Lines AA Boundary Setting Uniform Values 4 Given Flow Direction By Unit Vector T Mach Number 0 Total Pressure 100000 Total Temperature 300 Velocity Direction x lo yfo z o Figure 13 6 Setting up a Stagnation Boundary type in R Desk for to the Velocity inlet boundary condition see also Figure 13 3 top right Considering phase variables only Turbulent Intensity Turbulent Length and Volume Fraction values might be required 13 6 Flow Outlet This type describe
98. case the unv file included some node sets These are converted to boundary def initions by vpre and applied to the GRD file For imported files that do not contain supported boundary definitions all faces will be assign the Boundary 1 The imported sets are shown in Figure 19 91 Ricardo Software December 2009 438 19 TUTORIALS 19 4 IMPORTING THIRD PARTY MESHES R Desk 341 p1 tube GROY ol FE Fie Edt View Optons Window Hep _ lajx 0 oB alje 3 Je 20 ae gt en c gt tube GRD ds Plot pi Canvas 3d1 38 mep pe Es 123 Show Selected wth wreframe new gt est osere F Auto apply Apply ION D RPe training_tutorials V4 mesh_import tube GRD 11 01 42 INFORMATION Closing fie D RPe trairing_tutorials V4 mesh_import tube GRD 3 Figure 19 91 The GRD file when loaded into R Desk 19 4 2 1 Creating Boundary Regions In this case we wish to define each ends of the tube to be separate boundaries with all the remaining faces to be in an additional boundary As such a third boundary face set needs to be created This can be done by right clicking on the face entry in the set tree and selecting Create new face set Figure 19 92 L R Desk 3d1 p1 tube GRD MR File Edt View Options Window Hep eel gt ala Br gt 01 42 DEBUG created 85076 lines 325510 dupicates 11 01 42 INFORMATION Opening fie D RPe training_tutorials V4 mesh_i
99. defined by triangulated surfaces in the Ricardo standard data file SDF format typically called tri or sdf The arbitrary surface reports integrated quantities for the surface defined by the triangle file into the rep sdf file and optionally into text output files Values reported are similar to those for IO files In this case an arbitrary surface will be defined within the port to allow the angular momentum to be monitored during the simulation A straightforward way to define the surface is to chop the geometry in phasel at the location where the arbitrary surface is required Open the port tri file in phasel Rotate the model to view the outlet cylinder Using the horizontal slice button slice the model in the desired position see Figure 19 49 The model will be sliced with one half of the model marked in red Figure 19 50 Use the delete marked area button to delete the marked triangles This will leave just the unmarked area of the model Figure 19 50 Use the cap hole tool to fill the hole this will generate a surface in the location where the geometry was sliced Figure 19 52 Then with the Figure 19 53 chop geometry tool select the triangles that make up the newly created surface Finally save the model as a new file File gt Save As When prompted select save only the active triangles This will save only the triangles currently displayed which should all belong to the arbitrary surface Figure 19 54
100. distance is set to 2 then the cell refinement is reduced by 1 every 2 global cells in each co ordinate direction until the global mesh cell size is reached Surface refinement set to 2 with a blending distance of 2 and blend to boundary depth 1 Now the Blend to boundary depth is set to 2 so the surface refinement is stepped back to the global mesh cell size using two global cells in each co ordinate direction Ricardo Software December 2009 61 3 GEOMETRY 3 16 MESH SETUP x Specification for boundary 2 El El Delete Refinement depth at boundary EX pe Refinement blending distance As Refinement Blending Blend to boundary depth 1 m Refinement Specification Destination Save specification in triangle file Save specification in mesh file Gea Lo If the Refine to boundary depth is used the complete global is refined to the specified refinement level Surface refinement set to 2 with a blending distance of 2 and blend to boundary depth Now the Blend to boundary depth option is used so that the complete global cells next to the boundary are sub divided The blend distance is still set to 2 so again 2 cells in each direction are used to step the mesh cell size back to the global cell size 3 16 Mesh setup A separate toolbar is provided with functions to set up the global mesh lines which the mesher uses as the basis for mesh generation This toolb
101. domain in most cases coincides with the computational domain and it can have any number of fluid solid domain Figure 4 12 right The fluid domain is made of a single fluid ma terial domain whereas the solid domain can be made of one or more materials A fluid domain can be further sub divided into a fluid sub domain to include additional fluid transport equations Boundaries represent the external boundaries of the global domain as in Figure 4 12 right These are sub divided into non overlapping boundary regions according to the boundary condition type Interface regions are implicitly defined at fluid solid and solid solid material interfaces see Fig ure 4 12 right For every pair of adjacent material domains there is one interface region ORicardo Software December 2009 100 4 MESHING 4 10 GRID DATA STRUCTURE Quadrilateral Hexahedron Tetrahedron Triangle Prism Pyramid Figure 4 14 Various 2D and 3D cells CVs 4 10 2 Calculation of grid geometric properties Geometric properties of cells and faces play a vital role in the discretisation procedure For M properly ordered face vertices Figure 14 1 the face surface vector Af is computed by using for mulae for a triangular face A since each polygon can be decomposed into M M 2 triangles Thus the following expression is used 14 A a 22 Fi 1 71 x Fi F1 4 1 Note that the use of the above equation ensures the geometric conservation la
102. dsize gt integer number of values for each partition array integer pointer input local array Also output global if garray not present garray integer pointer optional output array array real pointer input local array Also output global if garray not present garray real pointer optional output array array double precision pointer input local array Also output global if garray not present garray double precision pointer optional output array subroutine concat_array nval dsize array garray Arguments Description nval integer first index size same on each partition dsize gt integer number of values for each partition array gt integer pointer input local array Also output global if garray not present garray integer pointer optional output array array gt real pointer input local array Also output global if garray not present garray real pointer optional output array gt double precision pointer input local array Also output global if garray not present garray double precision pointer optional output array array Table 18 34 Subroutine concat_array to take local array from each partition and concatenate into global array on partition 1 Ricardo Software December 2009 347 18 USER PROGRAMMING 18 4 USER PROGRAMMABLE ROUTINES 18 4 User Programmable
103. e a a a ee ee 184 111 3 Govermme equations s sg soitaa eS e a ad A a a 184 11 2 Simplified Modelling Equations c e sia ei ee a ee ee Be 189 11 3 Porous Model Types and User Inputs soss secre sos daa A we ee 191 11 34 General porousmodel u s ee ar eni e ai e AE rad de 192 11 3 2 Catalytic Converter air cada dd dd a rd E a 193 11 3 3 General orthotropic model 1 22 4 4 ies o a a ed 193 DVS Radiators ss os siege Gn ee See ge hee de ok SEE aS Rooke ew 194 11 3 5 Heat exchangers o ec e ce Ke ee ew ae A ew ee a a 195 12 MODELLING MULTIPHASE FLOWS 196 12 TAO UCA a gta 6 Sky a A oe ee A ie e kd eM ed Bae ee es 196 12 2 Buler Euler Modelling Equations c 229492 26 ba 2 be de 22 be weds eae as 198 12 3 Multi Phase Mixture and VOF Models o 2 2 00022 ee eee eee 200 123 Mixture EQUALIONS ye oe gee ad we ae A EP wh a we Oe ta ee 200 123 2 VOF quations e a ao od ee ee he eS SO 202 12 3 3 Setting up the homogeneous mixture model o 202 12 4 Phase Change Modelling lt s e soe eke aie eee ORD Re A 203 124 1 Boiling Models cs cc osea e ee ee OA a ee be ee es 203 LALL VECTISS boilineamodels soes v sos ria ad Be es ew 203 Ricardo Software December 2009 vi 124 2 RP boiling model u ass rada as a aa ed be we Be Pw 206 12 4 1 3 Setting up a Boiling Model ocurrir E aae a 208 124 2 Cavitation Models o o o c o c oea 04 ba ca da ed a ea la ad 209 12421 Simshalietal Model oa cewce 6 46 ea
104. ead we Bote oe ee ee 228 Ricardo Software December 2009 xix LIST OF FIGURES LIST OF FIGURES 13 12 RTHERM node as a part of the E A panel in R Desk after opening left and after editing right 230 13 14R Desk boundary condition panel for time dependent boundary data 231 13 15 i 14 1 Control volume and notation s 4655 4 26 6 eee ew Sew 235 14 2 R Desk setup for steady and unsteady simulations o o o 237 14 3 Definition of upstream central and downstream nodes o e 238 14 4 NVD diagram TVD and CBC criteria and characteristics of several convective schemes 240 14 5 R Desk setup for convective schemes and linear solvers o ooo o 241 14 6 Polyhedral control volume around the cell facej o ooo 242 14 7 R Desk setup for 2D or 3D gradient calculation options etc oo o 243 14 8 Control volumes near fluid solid interface ss s coga cas ooo 247 14 9 R Desk setup for pressure correction algorithm o ooo e 252 14 10Schematic representation of interpolative scheme ooo o 254 15 1 Global radiation location within the solver setup tree o oo 258 15 2 R Desk radiation setup panel o o cca eg ea ee 259 15 3 Radiation boundary panel location within solver setup tree ooo o 259 15 4 R Desk radiation boun
105. eae doa ad eee we ae eM 209 12 4 2 2 Zwart Gerber Melamri model o e 210 12 4 2 3 Schnerr Sauer model vio 211 12 4 2 4 Setting up a Cavitation Model o oo e 211 13 BOUNDARY amp INTERFACE CONDITION TYPES 213 13 1 Seting Ups eas ead A aa Rae OG Bee dw A doa ek RR 214 13 2 Velocity Inlet sos s i hee RE A ad pe ee a a we E 216 13 2 1 Setting up a velocity inlet boundary condition 20 217 133 Mass FIO cio ies ee Se a de Gade Hae Ge es Eee alee aw 218 13 3 1 Setting up a mass flow rate boundary condition 4 219 13 4 Pressure Inle Outlet o o s 6442 bake es Ee OR Aa ee eee EE a 220 13 4 1 Setting up a pressure boundary condition o 220 DS Staenation Mle soos a ad a o o A a A As a a 221 13 5 1 Setting up a stagnation inlet boundary condition o o 221 EAS A ik waka deb eS PR ak ee bat E Wa de ae Pk hw 8 222 13 6 1 Setting up a flow outlet boundary condition 0 0 223 13 7 Symmetry Planes eos wooed a Lada E Pa ae ROR DR OR Ge ee 223 13 7 1 Setting up asymmetry boundary condition o o e 224 A EA ae ee eA are WOR Gk ee a ee ee 224 13 8 1 Basic wall Setup 2 2 tods bs 2 dh aie ia bee ee e we ad ae 225 13 8 2 Wall thermal conditions sesos sos 40 4 0 OE we ba aka ee a ed ee 225 13 85 Thm Wall Model lt del ae ee er ee ee Ee eA E 226 13 8 4 Setting up thermal con
106. empirically In some models the modelled transport equation for the turbulent viscosity 4 is solved An example is the Spalart Almaras model cf Spalart and Allmaras 1994 which has become popular in aeronautics Two equation equation models Since Kolmogorov s pioneering work 1941 cf Spalding 1991 the turbulent energy k has been used without exception as the first variable in two equation models k models Kolmogorov proposed the two equation model based on transport equations for the turbulent energy actually two thirds of k and mean frequency k 2 f Although he did not explicitly provide the production term for see Spalding 1991 his model inspired the development of modern k models Saffman 1970 Wilcox 1998 Menter 1994 The variable can be interpreted as the ratio of and k i e the rate of dissipation per unit of turbulent kinetic energy k models Apart from k Harlow and Nakayama 1967 introduced the turbulent dissipation rate k3 2 as the second turbulence scaling variable The choice of e as the second variable has been the most popular it appears in the k equation as the sink term and its exact transport equation can be easily derived from the Navier Stokes equations Unfortunately modelling of the equation is extremely difficult and in the past relied on intuition combined with dimensional consistency coordinate invariance and analogy
107. exponent For air at moderate pressures and temperatures Up 1 716 x 10 kg ms To 273 15 K and n 2 3 The non Newtonian fluids are not supported yet 8 1 2 Energy transport and thermal conductivity The heat conduction process is described by Fourier s law di A a 5 8 5 Xi which states that the heat flux vector is proportional to the temperature gradient The constant of proportionality is the thermal conductivity A With increasing temperature the thermal conductiv ities of dilute low density gases increase and of most liquids decrease The Sutherland formula Equation 8 3 can be used for dilute gases AA T A 8 6 Lit TES where S is the effective temperature and Apo is the conductivity value at reference temperature To 273 15K Metals are generally better heat conductors than non metals The conductivity of most pure metals decreases with the temperature whereas the conductivity of non metals increases Some solid materials fibrous and laminated solids are characterised by the anisotropic thermal conductivity Aj which is in general a tensor quantity the heat flux vector is then given as q A dT dx Materials with the anisotropic conductivity are not yet supported 8 1 3 Mass transport and mass diffusion coefficients Mass transport diffusion occurs in multicomponent flows where individual species move from a region of high concentration to a region of low concentration The concentration or mas
108. file containing information about solution control boundary conditions initial conditions etc All physical units in the file and throughout VECTIS are SI units kg m s K There are no exceptions to this The contents of the solver setup input panel changes dynamically according to the selected branch of the solver setup tree Once the input file is set up the solver can run either from the command line or via R Desk see Section 17 15 The command line interface is as follows vsolve np num grid id debug level project where project is the project name base name of the input file The np option is used to specify the number of processors for a parallel computation The grid option is used to specify grid file format By default this is assumed to be native id 0 ID TYPE 0 Native GRD 1 Universal unv 2 Vectis 3 DAT 3 Nastran nas 4 Star CCM ccm 5 Spider flma 6 CGNS cgns Table 17 1 Grid types supported The debug option is to used help debug any problems that may occur The higher the level the higher the verbosity of output to the screen 278 17 USING SOLVER 17 2 GLOBAL DOMAIN 17 2 Global Domain The Global Domain panel is at the top level of the Solver Setup Tree From within here the parameters that can be controlled are Input Mesh Filename see Figure 7 2 Time Mode see Figure 14 2 File Output Frequencies Frequencies For P
109. from the continuity equation as dp 0 _d at Te gt gt 0 10 17 Considering steady flow and accounting for the isentropic conditions see ideal gas model Equa tion 8 31 and that integration of the momentum gives the Bernoulli equation the relation be tween density and the velocity potential can be established as follows an 1 1 P Prot ve 10 18 2 Hor where Prot and Ho are the stagnation density and stagnation enthalpy respectively They are constant throughout the potential field Clearly potential flows are described by the diffusion type equation where density plays a role of the diffusion coefficient which in turn depends on the magnitude of velocity potential For the incompressible flows density is constant and Equation 10 17 becomes the Laplace equation VECTIS MAX provides a solution of the steady potential flow as a part of the flow initialisation The selection of potential flow initialisation is described in the Rdesk fluid domain setup If the potential flow initialisation is enabled the Potential equation solver in the Equations Solver will be checked and the default control variables displayed as Figure 10 5 illustrates The Under Relaxation Factor is used to under relax density during an iterative solution of the ve locity potential The default value of unity should be always used for incompressible and subsonic flows Its value might be reduced to 0 8 0 9 if there are co
110. fuel thanks filling sloshing crank cases with oil cavitating fuel pumps In terms of physical states a two phase flow can be classified as liquid gas solid gas liquid solid and liquid liquid immiscible liquids In terms of topology Ishii 1975 sub divides multi phase flows into DO dispersed air ice flow oil droplets in water or air gas bubbles in liquid solid particles in gas liquid O separated annular liquid gas flows free surface flows water jets in air O and mixed transitional bubbly droplet annular flows slug churn liquid and gas flows Typical multiphase topology is illustrated in Figure12 1 for flow boiling in a heated tube with a sub cooled inlet The Instantaneous Navier Stokes equations describe both single and multi phase flows This de scription is not restricted to either laminar turbulent regime or dispersed separated topology How ever the resolving of scales associated with the topology of phase interfaces or with the fluid turbulent motion is not yet computationally affordable Instead two basic modelling approaches are employed in practice 196 12 MODELLING MULTIPHASE FLOWS 12 1 INTRODUCTION s J a J J J a ey Churn Slug Bubbly 0 0 1 0 Figure 12 1 Typical multiphase topology flow boiling in a heated tube with a subcool ed inlet and volume fraction distributions for four fields cl continuous liquid cv continuou
111. geometry The result of leaks can be easily visible in Phase 1 as a mixture of triangles with both sides The running time also increases It is suggested to carefully inspect the initial geometry and cap all relatively large holes start the Wrapper with relatively large Feature Resolution Size and gradually reduce it Triangle decimation is on by default The Decimation Threshold Angle is based on the angle between adjacent triangle normals Large values of the Decimation Threshold Angle should not be applied to thin geometries since it can lead to self intersections The default value 2 degrees usually gives satisfactory results Triangle decimation is also based on edge collapsing but for user specified edge length and a high angle tolerance It is recommended to set the decimation distance to be not larger than half the feature resolution size and use a combination of the decimation based on angle and the decimation Ricardo Software December 2009 43 3 GEOMETRY 3 14 GEOMETRY WRAPPING based on distance The deviation decimation edge collapsing technique uses deviation criterion from the wrapped model to the original geometry It can produce more accurate results but requires more time The recommended value for the deviation distance is one tenth of the feature resolution size It is also suggested to use the decimation based on angle and the decimation based on distance prior to using the deviation based decimation Th
112. i i 5 ue LAW gt pk eddy viscosity models ok The hydrostatic head in the above equation Api Pref8 F Fref 10 15 Ricardo Software December 2009 169 10 MODELLING SINGLE PHASE FLOWS 10 3 MODELLING FLUID FLOW features the reference density Pes and position vector of the reference altitude 7 which is cur rently defined as F ef 0 Thus the computed pressure in VECTIS MAX is always relative to the reference pressure but it does not always represent the static pressure ps However when comput ing density of an ideal gas Equations 10 10 10 11 it is assumed that ps p i e the differences between the static and computed pressure are neglected The reference pressure itself and all input pressure values at various boundaries must be specified as absolute pressure values For the gravity driven flows the input values of the computed pressure p are expected and the post processing results will be reported for the solver computed pressure The actual value of the reference pressure is not important in case of incompressible fluid model However if the ideal gas model is used the reference pressure value is important as the density depends on the absolute pressure Paps P Pref The reference pressure should be set to a value for which the computed pressure values are going to be small In practice it is often the average value of the specified inputs at pressure boundaries Closely related to the reference pr
113. id for fluid w momentum integer parameter ipr id for fluid pressure integer parameter ien id for energy integer parameter ics id for species concentration integer parameter ite id for turbulent energy integer parameter ied id for turbulent dissipation rate integer parameter ivit id for turbulent vsicosity integer parameter ips id for passive scalar integer parameter Table 18 6 Variable ids accessible directly through UPRs Ricardo Software December 2009 302 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES 18 3 2 User accessible routines In addition to the user accessible variables a number of user accessible routines UAR are also available These routines can be called from UPRs to exchange information with the solver kernel Each UAR is listed in a table where the purpose of UAR and its arguments are explained Thus a user only needs to look at the table to see which arguments need to be passed to the UAR and what result to expect If the UAR can be called with optional arguments all options are listed as subsections of the table To distinguish types of arguments input arguments are given in italic fonts output arguments in bold and some tables have an input and output column defined explicitly If a UPR is a function its return type is also specified Arguments may also be arrays in which case for example array is a one dimensional array array a two dimensional array etc Here we give a numbe
114. k else do k 1 ndata if xbnd 3 jb lt zinl k and xbnd 3 jb gt zinl k 1 then zfac xbnd 3 jb zinl k 1 zinl k zinl k 1 uzk uz k 1 x 1 zfac uz k xzfac tezk tez k 1 x 1 zfac tez k zfac ub 1 3 jb uzk reg_bval 1 3 ir teb jb tezk exit end if end do end if edb jb cmu mat rproph idens ir ttur idt Steb jb 2 amp eproph ivis ir rbc ibturls ir end do close ifun nullify ub teb xbnd reg_bval case dissipation lIn case of pressure total and mass flow boundary types the user has to supply turbulence length scale instead of actual dissipation The dissipation is also saved in the array fib call get_field idt iget_bnd var_name fib lupr gt start do jb jbndl jbnd2 end do lupr gt end fib gt null Case phase_vol_frac e call get_field idt iget_bnd var_name fib lupr gt start do jb jbndl jbnd2 end do lupr gt end fib gt null ase spec_mass_frac call get_field idt iget_bnd var_name fib lupr gt start do jb jbnd1 jbnd2 end do lupr gt end fib gt null case mass_frac_ps call get_field idt iget_bnd var_name fib lupr gt start do jb jbnd1 jbnd2 end do lupr gt end fib gt null end select Ricardo Software December 2009 18 7 EXAMPLES 361 18 USER PROGRAMMING 18 7 EXAMPLES end if 1 Next usr_obj_name and usr_reg_name
115. lt Channel Re 180 a Channel Re 395 Channel Re 590 10 L y 0 02y 0 0016y M 1 G U 0 38GU i F Gao 2 0859 5 0 1 10 100 y Figure 9 2 A priory comparison of U F y expression 9 49 with DNS data for boundary layer Spalart 1988 and channel Moser et al 1999 flows and the linear relationship Equation 9 41 holds within the viscous sub layer y lt y 2 5 The approximate validity of the log law Equation 9 31 is restricted to a narrow range of data for 15 lt y lt 30 40 Notably classical scaling U F yt shown in Figure 9 1 results with much wider log law region 30 lt yt lt 200 With the help of the blending function G exp Quy u 0 085 Cy 9 48 the DNS data are correlated in terms of star units as follows y 0 02y 1 6 x 10 3y 3 viscous sub layer y lt y id a l 1 G In Ey K K Gy buffer amp log law y gt y id where K 0 38 and y is the intersection point between the wall limiting and fully turbu lent expressions see Equation 9 46 The blended profile in the above equation can also be used within the viscous sub layer However the third order polynomial is introduced instead as it correlates better the DNS data than the blended profile A blending function G exp a y t 0 1 is used to define a unified law depicted in Figure 9 1 Ricardo Software December 2009 160 9 MODELLING TURBUL
116. nee ee e da de eee eA eA eS 345 18 4 User Programmable Routines 2 eco c soca oeu a ee ee 348 184 1 User properties TOULINE s oa san eh bra a el a Me 348 18 4 2 User initialisation routine 40 442 22 28 bee eee a ee ed eas 349 18 4 3 User boundary conditions routine 2 0 ee 350 18 4 4 User SOURCES TOULIME ss dcs ace eS ee be ee pon ko ee Bw Pa ee ee 351 ORicardo Software December 2009 x 164 5 User genero TOUS ai A a al a eS Be 352 18 5 Writing and Compiling UPR s e sorier e eaea a ds a eee a E Ra a 352 18 6 UPR Check Report Messages co da da e eee A a 353 18 Examples s es ss ip se AL A he ek eed Ge A da a A ee da a 353 18 7 1 Example of the user properties routine 2 2 ee ee ee 353 18 7 2 Example of the user initialisation routine 6 4 24 6 be ee ee es 357 18 7 3 Example of the user boundary conditions routine 4 359 18 7 4 Example of the user sources routine es es cosg ee ee 362 18 7 5 Example of the user generic routine 4 cooo eu aaa cea eee eee 364 19 TUTORIALS 367 191 Baste Tutorials ico Bok ce ee BOE Ge or ae da Eek e as ROE Be we ee Re he a 367 19I VEC TUS WOPOW ss 5 ase ad ine A ee ee Se Ge el ee a 367 1942 VECTIS SITUCIUILE o e e e Gee Bo ey ae Ew ow 368 19 1 3 Geometry Preparation s s ak eS eS A A ak a A Pe e 369 19 1 4 Denning the Global Meshi s o sosa gk ee ode DO Re ee Bs 370 191 3 REDES i665 on esi iho od we eB ee OR ee YY BS SG Eg a we a 370
117. needed The mesher vmesh can either be started from the command line or within R desk In this case vmesh will be run from the command line In a terminal or dos command window type the follow ing command vmesh coolant mesh The mesher will write output to the screen This screen output is also written to a file in this case coolant OUT If this file is opened or if the mesher was run from the command line you can Ricardo Software December 2009 422 19 TUTORIALS 19 3 COOLANT FLOW Ricardo VECTIS Phase 1 coolant tri 10 xj Fie Edit View Toolbars Operations Help al B B BI amonenn Check Self Intersecton Options Stitch Mesh Check for Unstitched Setup view Geometry Wrapper Decmate Triangles sce Make Geometry meu Set Model Time Boundary Painting Mesh Lines Hoa Information E UK Refinement MJ o wn mm Options to pass to the mesh generator Tite Comment VECTIS mesh generator input file Add Delete Eat Model FileName coolant tri Refinement Parameters Refinement Depth 2 DEEP Jo FORCE 0 6 Exact fit at sharp edges OK UTISTIS TT TE 20 18 13 58160 Nodes 20 18 13 Model dimensions xmin 0 187500 xmax 0 212000 ymin 0 118101 ymax 0 071992 zmin 0 078000 zmax 0 080000 l Figure 19 70 Using meshlines to control the cell sizes located in the gasket region see a summar
118. no turbulence models iturb_none 0 no modelling laminar DNS iturb_rans 1 Reynolds Averaged Navier Stokes models iturb_les 2 Large Eddy Simulations iturb_des 3 Detached Eddy Simulations For example if turb_par iturb_meth 2 0 then the physical model is laminar DNS for do main 2 O iturb_modf 2 model family belonging to a method Family model identified by iturb_ modf returns one of the following tinvisc 1 inviscid flow idns 0 dns laminar no modelling ievm_zeroeq 1 zero equation algebraic models ievm_oneeq 2 one equation models evm_keps 3 k epsilon models ievm_kom 4 k omega models ievm_v2f 5 Durbin s keps v2 f model iarsm 6 algebraic Reynolds stress models idrsm 7 Differential Reynolds stress models iles 8 Large Eddy Simulation models ides 9 Hybrid RANS LES DES For example if turb_par iturb_modf 2 3 then the family model for domain 2 is the k epsilon model Ricardo Software December 2009 326 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES O iwall_meth 3 near wall method Two options are available iwallf 1 wall modelling with wall functions ilowre 2 low Reynolds number modelling For example ifturb_par 1iwal
119. normal or a normal to 3 points Translation Transformation Panel Ricardo Software December 2009 26 3 GEOMETRY 3 8 TRIANGLE SLICING AND INTERSECTION OPERATIONS 3 8 Triangle Slicing and Intersection Operations Horizontal Slice Button 2 When this operation is selected a horizontal line becomes bound to the mouse pointer When the left mouse button is clicked the line thus defined is used to describe an infinite plane perpen dicular to the plane of the screen This plane is used to slice through the visible i e triangles which are not deactivated by the CHOP function parts of the model in such a way that new nodes are created on the slicing plane where the edges of the triangles intersect the plane The trian gles cut by the plane are divided into smaller triangles and all triangles above the slicing plane are marked These marked triangles can be deleted if desired providing a convenient way of truncating geometry with a plane cut WU QA Triangle Slicing If the model has been chopped and then is saved the user is asked whether to save the whole model or just the visible active part Thus a partial model can be saved Vertical Slice Button 4 This function is as the horizontal chop function but with a vertical line Arbitrary Slice Button 2 When this function is selected no line is bound to the cursor Instead a point on the canvas is selected and then a line is created from t
120. oc eco ba cea ar aaaea e Re 440 19 95 Use Flat Picking to select the required face sce iatera ta ocita a a Ea a ee ee ee 441 19 96Saving the sets tothe GRD file iras ia a ao eide e E A 442 19 97Reloading the GRD fil o ss c saii ee db we ee E E e 442 19 98 Viewing the defined sets on the final GRD file o e oa e sow a w aciou aa a a a a 443 19 99Importing the generated computational mesh into the solver setup 443 Ricardo Software December 2009 xxiv USING THIS MANUAL The latest versions of this manual including pdf version and the release notes are available at Vectis Max Documents 1 1 Contents of This Manual The manual provides all necessary information in order to use VECTIS MAX effectively A brief description of the contents of each chapter is given as follows O Introduction outlines the main features and capabilities of VECTIS MAX A brief guide is also provided which explains how to start run and analyse the VECTIS MAX simulation Geometry related chapter is the description of phasel It deals with the geometry acquisi tion and its preparation for the meshing task The definition of boundary regions open flow boundaries and walls is also described Meshing i e generation of the numerical mesh for single and multi domain simulations is described in this chapter Here the user can learn how to create the VECTIS MAX mesh using the vsolve module or to import non VECTIS MAX mesh
121. of meshfile mesh information fil name of patfile patfile pel ani matrix name of vimfile superpatches information Figure 15 7 Scheme of the radvfm program 15 4 Radvfm Theory 15 4 1 Assembling the view factors matrix To assemble the view factors matrix the direct visibility of each patch to every other patch must be examined If the two super patches can directly see each other the view factors Fij and Fji can be calculated for them Ricardo Software December 2009 264 15 MODELLING RADIATION 15 4 RADVFM THEORY 15 4 2 Calculation of view factor The calculation of view factor between two patches can be done using this simplified formula ae Coba 15 1 r where Fi view factor between super patches i and j B and B angles rad between the position dependent normal vectors to super patches i and j r length of a line connecting the points of evaluation of the normal Aj area m2 of the super patch j The situation is visualized below superpatch j view factor calculation A _ cos f cosp n Fij Figure 15 8 Heating super patch i and heated super patch j in view factor calculation 15 4 3 Calculation of Super patch view factors 15 4 3 1 Hemisphere Base Projection Methods In order to calculate all view factors for an input geometry the Nusselt analog method is used in radvfm For each super patch all patches in the visible half space are projected onto the hemi sphere T
122. of the data from all partitions is formed on partition 1 and then lonly partition 1 writes the output For efficient data exchange a compressed list of values lis exchanged select case icall_pos case icp_beg_run icp_end_iter n_part get_number n_partitions i_part get_number n_current_part n_dom get_number n_domains call get_domain ise_cell iscd iecd call get_grid_geom xyz_cell_c xcc end select select case icall_pos case icp_beg_run Ricardo Software December 2009 364 18 USER PROGRAMMING 18 7 EXAMPLES lbeginning of simulation Isetup data storage and generate compressed data arrays lFind size of compressed data array n 0 do idom 1 n_dom do ic iscd idom iecd idom if abs xcc 1 ic 0 5E0 lt 2 0E 3 n n 1 enddo enddo allocate dsize n_part dsize 0 dsize i_part n number per partition call global_sum dsize n_part gsize sum dsize total number lallocate memory for compressed arrays of of y coordinate and temperature On partition 1 globally sized arrays are created if i_part 1 then global arrays allocate tval gsize allocate xval gsize else locate arrays allocate tval dsize i_part allocate xval dsize i_part endif n 0 Istore y coordinate values do idom 1 n_dom do ic iscd idom iecd idom if abs xcc 1 ic 0 5E0 lt 2 0E 3 then n n 1 xval n xcc 2 ic endif enddo enddo call concat_array dsize xval concatonat
123. ooo o 286 Table of all the variables contained in wall files o ooo o 286 Partitioning controls and corresponding vpre options o o e 292 Mesh join types and corresponding vpre jtype option numbers o 293 Integer and real working precision format in VECTIS MAX 299 ACCESS STOUPS as a e a A o Bade a 299 Various VECTIS MAX ODJeCIS e Sa a aa a aa ta 300 Names of post processing variables srs e acesi e E ee a 300 Parameters accessible directly through UPRS o o e ee eee 301 Variable ids accessible directly through UPRS o ooo e 302 XIV LIST OF TABLES LIST OF TABLES 18 7 Function get_number to get values for main variables numbers defined in the access group TACO MUMDERS 4 5 ra or BG a a ai id 304 18 8 Subroutine get_domain to get values for variables defined in the access group 1ace_domain describing domain structure mainly starting and ending indices of domain objects 306 18 9 Subroutine get_mat to get variables defined in the access group iacc_mat mainly reference material properties and starting and ending indices of material objects 308 18 10Subroutine get_mat continued from previous table o o 309 18 11 Subroutine get_phase to get values for variables defined in the access group iacc_phase mainly start end phase objects
124. or by right clicking the plot in the Plot Tree and selecting view gt properties Make sure that the port GRD model is selected in the Plot Tree then in the Plot Properties panel set the lines to mesh With mesh lines shown the computational grid should look something similar to that shown in Figure 19 34 Now that the computational mesh has been generated we can move onto the next stage setting up the input file for the solver 19 2 10 Solver Setup In order to run the solver the calculation inputs need to be defined This includes the boundary conditions fluid solid properties solver parameters output data and frequency initial conditions and so on This is done using the Solver Setup tree and Solver Setup Input found within an R Desk VECTIS project Firstly open up R desk if it is not already open Then using the project browser select New Project This will open the New project panel select VECTIS from the drop down list Ricardo Software December 2009 395 19 TUTORIALS 19 2 STEADY STATE PORT FLOW ld R Desk 3d1 p2 port GRD 10 x we Fie Edit View Options Window Help la x DAB S props DIS a Eee AO ell x BOB teen lan ls v Solver Setup y Y_ PlatProperties v Solver Setup Tree Solver Control Live Update lv Messages Information Sets Clipping Plots SolverSel Transformation Plot Properties Lighting Steps Line
125. or radial fans in which the flow enters parallel to the axis but leaves in a radial direction ORicardo Software December 2009 273 16 MODELLING FANS 16 3 1D MODEL The total pressure developed by a fan is dependent on the fluid volumetric flow rate In general as the flow rate decreases the fan s total pressure and static will increase The fan s static pressure will reach a maximum value as the flow rate falls to zero Conversely at the other extreme the flow rate will reach a maximum as the fan s static pressure falls to zero The relationship between the fan s pressure and volumetric flow rate is known as its characteristic which is normally provided by the fan manufacturer see Fig 16 6 Fan Pressure Volumetric flow rate m3s 1 Figure 16 6 Typical fan characteristic curve monotonic The fans s characteristic curve can be used during a CFD simulation to look up the flow rate for a given fan pressure The fan pressure is the pressure rise between the inlet and outlet From the flow rate the mean axial velocity can be determined For the mean tangential velocity component we can use the Euler Equation for turbo machinery Eq 16 6 AP par v outlet por v inlet 16 6 where AP is the total pressure rise wr is the impeller blade velocity and v is the fluid tangential velocity Straight Blade Here the tangential and axial velocities are assumed to increase linearly with radial distance fro
126. orthogonality by modifying the second direction vector In the case of non aligned principal and coordinate axes the direction vectors might not be known a priory and the user can use the R Desk geometry tools to determine the direction vectors Among various Porous Material Types General Catalyst Orthotropic and Radiator only the General type as well as Heat Exchanger require specification of the second Direction Vector Y When the second direction vector is not required it is determined by solver as well as the third vector principal axes which is normal to the plane defined by the first two direction vectors The 2nd order resistance tensors specified by user are defined in the principal axes coordinate system and need to be transformed from the principal axes to the solver s global Cartesian axes This is done with the help of transformation matrix Tp This matrix is the transpose of the matrix Pax which is user defined by components of three principal axes vectors i e Tp P Denoting user defined viscous and inertial principal diagonal resistance tensors as R and R respectively they are transformed into the solver s Cartesian axes tensors R Ri and R HR pas R T R T R T Ri TI 11 55 Ricardo Software December 2009 191 11 MODELLING POROUS MEDIA 11 3 POROUS MODEL TYPES AND USER INPUTS cat_monolith Sub domain a Subdomain ID 1 Parent Material 1 Standard Fluid Solid M
127. phase k Us is based on the weighting the instant local velocity U ki with the local density p le Us i PY Ups i gk pf P For single phase turbulent compressible flows the density weighted ensemble Favre averaging results with the RANS equations Similarly ensemble averaging of local instant phasic equations Ishii 1975 Drew 1992 leads to the ensemble mean conservation equations for each phase These equations are coupled through the phase volume fractions amp UK 1 12 1 The Euler Euler equations are presented here for non moving grids and in the framework of eddy viscosity modelling Mass Conservation oap a k ky yk gt aput r 12 2 ot sE dx j P J Here superscript k denotes the k th phase k 1 Npn and al is the phase volume fraction The inter phase mass transfer rate TY occurs for example during evaporation or condensation and needs to be modelled The compatibility condition for the volume fractions need to enforced N ph Ya 1 12 3 k 1 Momentum Conservation serui eta A sate o ene 00 k where the viscous and turbulent stress tensors Th and Tj are given respectively as 2 k k kyk t 2p si 55 6 i i 2p st y Sd 30 kK i 12 5 Ricardo Software December 2009 198 12 MODELLING MULTIPHASE FLOWS 72 2 EULER EULER MODELLING EQUATIONS where the phase mean strain tensor is y _1fauk auf
128. plotted in a 3D canvas eis R Desk 3d1 p1 port tri E 10 x we Fie Edit View Options Window Help 18 xi DBBDA ber gt DIB y AH 95 9 e lr xBMDSPADOB Eb gt 2 Pos sowersempree O OOOO OO Setup Tree Plot Properties AX Lines edges X Faces I Color 1 ME Opacty fos ff IV Auto apply Apply Opening file D RPe training_tutorials V4 Port SSPortFlow PreparedFiles port Extracting Face Set names from file D RPe training_tutori ceramic PreparedFiles port tri Figure 19 29 Visualising the geometry file in R Desk Ricardo Software December 2009 393 19 TUTORIALS 19 2 STEADY STATE PORT FLOW The mesher vmesh can be started from within the VECTIS project Click on the launch mesher button Figure 19 30 lala me Fie Edit View Options Window Help 8 x D A Da Projectfport Y 69 uw 5 959 ll xBODSEt ADE Pets ax E port tri alls Plot pi Canvas 3d1 ide Launch Mesher Dialog 2 x Plots Mesh Input File port mesh a z pul port mes 1 Lines edges sos Faces I Color 1 Opacity 0 5 r IV Auto apply Apply Opening file D RPe training_tutorials V4 Port SSPortFlow PreparedFiles port tri Extracting Face Set names from file D RPe training_tutorials V4 Port SSPortFiow PreparedFile
129. properties and types available o ooo 349 18 37 Subroutine upr_init to modify initial field o oo 349 18 38 Subroutine upr_bnd_cond to modify boundary conditions 350 18 39Subroutine upr_sources to add source terms to equations 2 0 351 18 40Subroutine upr_generic to modify variables in arbitrary fashion 352 18 41 Program position identifiers s a eacee miare gaech wua da a ee ee a 353 19 1 Boundanes for Port Geometry o s ose es ee a e Sk 389 Ricardo Software December 2009 xvi List of Figures 21 4 1 4 2 43 4 4 4 5 4 6 4 7 4 8 4 9 4 10 4 11 4 12 4 13 4 14 4 15 dul S2 a3 Vectis work flow from initial model geometry to viewing CFD results 9 VMESH position is VECTIS MAX system 2 e 79 Start point of the mesh generation 2 144 2444 2 24 baw Bed eo be Pe ba eared 87 Global boxes containing volume y eae e maed ee ee ee 88 Refinement of global BOXES s eor we aes ce a a we a a a Oe ee ee 88 Final grid all polygons are generated lt eo oroso coea ansa ma s pa a a 88 Definition of global refinement depth in Phasel aaau 89 Global box divisions with different depths of refinement soaa 90 Boundary refinement depth 0 blending distance O oaaae 91 Boundary refinement depth 2 blending distance 1 RD 1 option seton 92 Boundary refinement depth 2 ble
130. radsolv the discrete points are the center points of the patches or super patches The equation system then looks like Ricardo Software December 2009 267 15 MODELLING RADIATION 15 5 RADSOLV THEORY MT KT CT RT 0 Co Ro 15 9 where oT A f MT I pe dQ 6 piciViT mass matrix 15 10 Q T T nb Aj dr T T conduction matrix 15 11 j l ij nbz CT h Tar Y hijAijT convection matrix 15 12 T j l nb3 RT f eoT dT 2 e GA T f radiation matrix 15 13 1 nb Q f gdl L qkijAij conduction heat flux matrix 15 14 ne Gel h Tod Daa convection vector 15 15 Es Ro cor dT Y E OA Ta radiation vector 15 16 j The last four equations of the above are correct only if there is an external radiation model and it is not influencing itself With such a situation then these four equations can be merged to one equation which describes the exchange of radiation among surface patch elements nb3 RT R aT Y qrjAij Ragqr 15 17 j l N 6 1 amp x y E v arj Y 8 Fy oT F OT 15 18 j l SJ J j l This can be rewritten into matrix form Rp qr ReT Rpoo 15 19 Equations 15 17 15 18 expand the current equation system of one equation and one unknown to two equations and two unknowns MT KT CT Raqr Q Co 15 20 Ricardo Software December 2009 268 15 MODELLING RADIATION 15 5 RADSOLV THEORY
131. representation of the gasket holes Therefore if the coolant jacket geometry from the engine block cylinder head and gasket holes can be merged to create one solid in the CAD package and this then exported as an STL file the preparation time will be significantly reduced in Phasel Figure 19 63 shows the gasket detail included in this coolant jacket example In this tutorial the geometry is fully stitched ready for meshing The user can now enter the boundary identification section of Phase 1 The purpose of this section is to define different regions of the surface as different boundaries In this example the user needs to distinguish the different regions that will become the walls and inlet outlet boundaries Ricardo Software December 2009 417 19 TUTORIALS 19 3 COOLANT FLOW Ricardo VECTIS Phase 1 coolant tri p 10 xj Fie Edit View Toolbars Operations Help c a a ZACH JAN 2 a ASIA Options Stitch Mesh AAJA A 2 410 G 0 sel de H N g bg A Epi g a gls Ls ES 2 Figure 19 62 The coolant jacket geometry when loaded into Phasel Y Ricardo VECTIS Phase 1 coolant tri 10 x Fie Edit View Toolbars Operations Help aaa masa JU Nada MM Options Stitch mesh JA Fai amp 3 p o 6 El gt lo 3 Pie EN At MEE E 2 elas 9 3 3 3 3 A
132. right porous model from the knowledge of the dimensional Friction Coefficient fort 1 m Rpx 5P fori kg m 11 57 where p denotes the fluid density 11 3 4 Radiators The radiator sub domain type is intended to model heat exchange in automotive radiators where the sub domain fluid represents a hot stream and it is cooled by the cold fluid stream of infinite thermal capacitance This single stream heat exchange model accounts for the fluid sub domain stream explicitly while the second stream is assumed to have a constant temperature T ef specified by the user The flow through a radiator is uni directional and both viscous Rp x and inertial Ri x resistances are specified via Viscous Coeff and Inertial Coeff edit boxes respectively which are shown in Figure 11 3 right The Heat Exchange Mode can be either Heat Transfer Coefficient or Heat Load If the heat load mode is selected then the value of Heat Load the total heat exchange Q should be specified inside the Heat Exchanger Data sub panel see Figure 11 3 When the heat transfer coefficient mode is enabled the heat load Oy will be calculated by the solver In both cases the volume Ricardo Software December 2009 194 11 MODELLING POROUS MEDIA 11 3 POROUS MODEL TYPES AND USER INPUTS Heat Transfer Coefficient Apy should be define as a polynomial function of the superficial velocity magnitude U via six coefficients a_1 to a_6 Thus px will be calc
133. selected to be either a gas or a liquid In the case of liquid simulation the compressibility is set to be incompressible For gases there is a choice of compressibility options Incompressible Fluid Weakly Compressible Fully Compressible Subsonic Ricardo Software December 2009 380 19 TUTORIALS 19 1 BASIC TUTORIAL Fully Compressible Supersonic For the tutorial we specify an incompressible liquid More details of the different options can be found in Setting a phase and its properties Solve all species This determines whether the volume fraction equations for all phases is solved or whether the last phase is taken as the remaining volume fraction after the other phases are solved Property Specification In general there are a number of ways of specifying the properties constant values or relationships In this case the values for water are fixed as constants Density 1012 Viscosity 0 000719 Specific heat 3620 Initial Conditions The initial conditions for the fluid can be specified in the initial conditions panel This allows the user to specify initial flow velocity pressure temperature etc These values will be imposed on the fluid when uniform initial conditions are selected in the FluidDomain panel Output The data to be output to the post file can be selected in the output panel Each of the scalar variable checked will be written to the file 19 1 8 4 Boundary R
134. sensibly amplifying the production of in these regions It also improves predictions of the low turbulence level flows with a large time scale k e Having in mind the simplicity robustness and computational economy the implemented k models and generally all eddy viscosity models are extremely useful tool for industrial flow cal culations provided that the users understand their limitations In this way potentially misleading and wrong results can be detected Ricardo Software December 2009 164 9 MODELLING TURBULENCE 9 7 SETTING A TURBULENCE MODEL 9 7 Setting a Turbulence Model and Wall Treatment For each fluid domain the turbulence modelling inputs have to be selected in the Turbulence Model panel This panel is opened by left click on the Turbulence Model node within the parent fluid domain node see Figure 7 2 By default the Turbulent Modelling Approach is set to a turbulent flow Turbulence Modelling Approach l Reynolds Averaged Navier Stokes Model Turbulence Family K epsilon Models Turbulence Model Standard Near Wall Modelling Low Re number e Wall Functions Variants Scalable Wall Functions Standard Wall Functions Unified Wall Conditions Laminar No Turbulence Modelling RNG Reynolds Averaged Navier Stokes Model Standard Inviscid Flow No Turbulence Models Standard realisable Figure 9 5 R Desk setup Setting turbulence modelling and near wall treatment to
135. separate material will have a unique ID number Multiphase Modelling This allows the user to specify whether the fluid contains a single phase or a homogeneous mixture of a number of different phases Ricardo Software December 2009 402 19 TUTORIALS 19 2 STEADY STATE PORT FLOW Initial Values Initial velocity pressure temperature passive scalar turbulence and species mass fraction for the solution domain at the start time as well as providing some other initial condi tions options Designed by User This option allows initial conditions to be specified using a user rou tine Potential Flow This option will calculate the initial flow field based on the boundary condition data specified Uniform Potential Flow Scale Factor This allows the initial calculated potential field to be scaled This is useful because the calculated potential flow may be much higher than the actual flow as turbu lence and viscous effects are not taken into account Reference Data The reference values should have representative values for the solution Body Force Options to allow the simulation of body force The port flow uses air as the fluid so the default values should be suitable for the simulation In this case the potential solver will be used to initialise the flow field The remaining fluid panels algorithm turbulence model equations amp solver and discretise will be left at their default values Monitoring Points
136. software ricardo com support tutorials vectismax TutorialFiles MeshImport 19 4 1 Introduction VECTIS MAX allows any computational mesh to be solved either the mesh files produced by the VECTIS mesher or third party meshes The purpose of this tutorial is to illustrate importing external third party computational meshes into VECTIS MAX The following items will be covered in this tutorial O Converting external mesh format to native VECTIS MAX GRD file O Defining boundary regions to a computational mesh using the set functionality of R Desk O Importing final computational grid into the soler setup Currently supported mesh files types for import are 1 I Deas unv files here node sets can be used to identify the boundary regions 2 CGNS files Boundary regions should be imported from CGNS files 3 CEDRE files currently these require a ccm extension however these are typically exported as ngeom files 4 Spider flma files 19 4 2 Simple Example In this example case a simple tetrahedral mesh of a tube is imported from an Ideas format unv file The file tube unv can be found following the link above Firstly convert the external mesh file to the native GRD file using vpre This can be done either on the command line or from the R Desk The GRD file can then be read in to R Desk so the the the boundary regions can be defined For GRD file face sets are used to define the boundary regions In the example
137. solver setup tree Check the Show mesh preview toggle and the imported grid file will be displayed may ee ETE Solver Setup E Gobal_domain_1 Global Domain Restart Control Boundary Regions Bnd_Reg_i Wal Boundary Bnd_Reg_2 Wal Boundary Bnd_Reg_3 Wal Boundary interface Regons Sub domsin s Report Regions Solid Domains Figure 19 99 Importing the generated computational mesh into the solver setup ORicardo Software December 2009 443 19 TUTORIALS 19 4 IMPORTING THIRD PARTY MESHES Subsequent selection of the boundaries regions in the solver setup tree will then highlight the corresponding regions on the model in the mesh preview Ricardo Software December 2009 444 Bibliography Absi R and Bennacer R 2006 A New Wall Function for the Turbulent Kinetic Energy k in Y N K Hanjalic and S Jakirlic eds Turbulence Heat and Mass Transfer 5 Begell House Inc 14 4 Baldwin B and Lomax H 1978 Thin Layer Approximation and Algebraic Model for Separated Turbulent Flows AZAA Paper 78 257 9 2 1 Barre S Bonnet J P Gatski T and Sandham N 2002 Compressible High Speed Flows in B Launder and N Sandham eds Closure Strategies for Turbulent and Transitional Flows Cambridge University Press Cambridge 6 4 9 3 Barth T J 1994 Aspects of Unstructured Grids and Finite Volume Solvers for the Euler and Navier Stokes Equations in Computational Fluid Dynamics lecture
138. stresses which are responsible for rates of momentum transport by the turbulence Similar correlations are formed with other flow variables such as p u Eyf c u and T u In case of compressible flows correlations involving fluctuating density and other variables are not negligible A Favre or mass weighted average is then used to avoid the explicit modelling of such correlations Compared to the Reynolds average Equation 6 17 the alternative splitting is introduced sn 0 0p 6 22 where the Favre average is defined as p Po g p 7 6 23 and the following relations hold pe gp 0 po 0 o p are ow y yw 6 24 ory HOY P 97 y pow 6 4 Reynolds Averaged Equations The classical Reynolds averaging is used in incompressible fluid mechanics while Favre averag ing deals with compressible reacting fluids where density fluctuations are significant Applying Reynolds and Favre averaging to the instantaneous equations the incompressible and compressible forms of Reynolds Averaged Navier Stokes RANS equations are obtained respectively see for example Speziale 1996 Vandromme 1993 While both forms are similar a relation between Favre and Reynolds averages remains hidden unless the knowledge of Reynolds averaged density fluctuation correlations is provided see Veynante and Vervisch 2002 The compressible RANS equations Favre averaged are more complex the
139. temperature and turbulent Schmidt number id s from idens 1 to itexc 6 and from idifm 7 to itsch 11 For passive scalar nprops 3 properties are defined ce 66 ce 66 TA O idens_ps 1 density Calculation options ipro_const ipro_mix ipro_igas ipro_ ce ces ce cos poly ipro_user ipro_bouss O idifm_ps 2 mass diffusion coefficient which governs the laminar mass diffusion flux ce 66 cc ce Calculation options ipro_const ipro_poly ipro_user O itsch_ps 3 turbulent Schmidt number which is used to calculate turbulent mass diffusion cc 66 flux for passive scalar Calculation options ipro_const ipro_user To get the list of passive scalar properties calculation options then do the following integer iwp pointer passca_pro_opts Call get_property l_ps_opts passca_pro_opts According to the thermodynamic state of matter the properties of the fluid component can be a function of temperature and pressure The current solver version has built in options to deal with thermally perfect fluids i e the physical properties can be a function of temperature only not of pressure The exception is the density of an ideal gas The user defined properties can be a function of both temperature and pressure The arrays ph_pro_opts 0 nproph 0 n_phases species_pro_opts 0 npr osp 0 n_species an
140. the maximum volume among its neighbours O Minimum wall distance ratio for near wall cells this is an equivalent to the cell aspect ratio Here it represents the ratio of the wall normal distance distance from the cell centre to the centre of the representative wall and an average normal distance of cell neighbours to the same wall Typical values for the maximum face angle minimum volume change and minimum wall distance ratio are 75 0 01 and 0 15 respectively Figure 14 10 Schematic representation of interpolative scheme The Cell Interpolative Scheme CIS presented here excludes poor quality cells from the solution of discretized equations Instead flow variables at these cells are determined from an interpolation of neighbouring cell values and imposed boundary condition values A sketch in Figure 14 10 shows the poor quality cell Z adjacent to the representative wall W with two j 2 neighbours Nj For the purpose of interpolation neighbours are defined as the surrounding cells that are further away from the wall than cell Z The first neighbours those that have common cell faces with the cell J are considered The poor quality neighbouring cell will be a qualifying neighbour only if the cell J does not have any other neighbours There are two stages in the interpolation procedure which have some similarity with the interpola tion algorithm of Kalitzin and laccarino 2003 First neighbour
141. the points of intersection for each triangle and attempts to define a number of new triangles to replace the existing intersecting triangles This is not always successful and is Ricardo Software December 2009 28 3 GEOMETRY 3 9 BOOLEAN VOLUME OPERATIONS likely to fail 1f the geometric tolerances are small which may result in unstitched geometry being created It is then necessary to stitch and repair the geometry where the program has failed to create triangles using the Auto Stitch CapHole and Create Triangle operations Once the new triangles have been created the program will attempt to mark the triangles on each side of the intersection line so the user then has to delete the unwanted geometry and stitch the remaining geometry If several operations are needed it might be necessary to do the check self intersection operation a number of times to keep the intersecting triangles in the model correctly marked This should not be too time consuming because the check self intersection operation only checks active triangles Triangles should not be coplanar The operation works better if the surfaces intersect fully that is the surfaces do not intersect in a number of different places In this case the check self intersection routine will give a patchy result and a number of operations have to be done in succession to generate the intersection geometry 3 9 Boolean Volume Operations The user is able to perform Boolean type Volume
142. the wrapping process with the offset option should be repeated several times It is recommended to use da and dd for element reduction with offset option In case of leak surface offset can be used for quick hole capping The following operations are suggested 1 Wrap the geometry with a coarse fr size without leaks applying a small negative offset distance 2 Merge the wrapped geometry with the original geometry 3 Repeat the wrapping process on the combined original and coarsely wrapped shrunk surface with a fine feature resolution size Ricardo Software December 2009 48 3 GEOMETRY 3 14 GEOMETRY WRAPPING Surface thickness This option produces more accurate positive surface offset up to fr size for thin geometries but requires more CPU resources It can be used to thicken infinitely thin surfaces For more accurate and faster results it is strongly recommended to apply st to some parts of the original model merge thickened parts with the original geometry and run the Wrapper without st mode Distant node smoothing To reduce artificial mesh disturbances near concave or poorly resolved features nodes located rel atively far from the original geometry can be smoothed using distant node smoothing mode This option can be recommended for the majority of geometries Approximation assessment The quality of approximation can be assessed using distances from triangles to the original geom etry Using this mode would take sligh
143. then not included in the radsolv calculation Emissivity This controls the emissivity of the boundary default 0 8 This value ranges from 0 1 Transmissivity This controls the transmissivity of the boundary default 0 0 This value ranges from 0 1 where 0 0 indicates complete opaqueness and 1 0 total transparency NOTE if the boundary is material interface then any other interfaces attached to this transparent material will automatically be updated Superpatch density This is the number of superpatches per boundary default 50 if 1 one superpatch per patch These superpatches are typically a coarser representation of the boundary The lower the value is the quicker but coarser the solution will be Ricardo Software December 2009 260 15 MODELLING RADIATION 15 3 RADIATION SETUP Superpatch Angle Tolerance This is the maximum allowable angle between patches within a superpatch default 40 With this option radprep will construct the super patches for the boundary and try to ensure that the total number of super patches is as near to the defined limit as possible whilst also using the angle limit value to define the maximum angle between the normal vectors of two neighbouring patches i e if two neighbouring patch normal angles differ by more than the angle tolerance value then they will be grouped into different super patches Conduction If this check box is activated then the surface conduction he
144. this is the icon at the bottom Click this icon and then right click on the two parallel red control lines that you wish to measure between The distance is then displayed in the Phasel window This distance can be divided by the required mesh cell size to give the required number of divisions This number of intervals can then be set up for each interval The number of divisions for each section for this example mesh are shown by Figure 19 69 Y Ricardo VECTIS Phase 1 coolant tri 10 x cues mea A N o a x M Options Stitch Mesh a a u o Mesh Lines ada Ema information mr UK Refinement MJ on me SSeS Add Delete Eait Refinement Parameters DEEP jo 2 18 13 Model dimensions xm n 0 187500 xmax 0 212000 ymin 0 118101 ymax 0 071992 zmin 0 078000 zmax 0 080000 PLL Figure 19 69 Example control mesh number of sub divisions to give approximately 5mm size cells The final step for the mesh file setup is to set the boundary refinement level This is done through the Operations gt Mesh File Setup panel The boundary refinement entry should be set to 1 This means that every cell at the surface can be refined by 2 2 in each co ordinate direction The setup for this is shown by Figure 19 70 Save the mesh file as coolant mesh The mesh setup is now complete 19 3 4 Mesh Generation In order to generate the computational mesh the tri and mesh files produced by Phasel are
145. to the screen is also written to the out file at the intervals specified If zero values are used then the output is not written The frequency of post processing file writing is specified in the Post processing File Frequency box Additional output files can be written for monitoring data boundary data report regions and domain data These can be either in ASCII column format or Ricardo SDF binary format The frequency of the two types of file can be chosen independently by entering the required intervals in to the Report Frequency boxes Again an interval of zero will mean that the data is not written The ASCII files include header rows detailing the data in each of the columns Separate files will be written for the different data types Each file adheres to the format projectname lt type gt _ lt runnumber gt Where lt t ype gt is given by Domain data dom Monitoring data mon IO data 10 Wall data wall Arbitrary Surface surf Table 17 3 Naming convention for different output file types The ASCII data files are used by the Live Update utility in R Desk The SDF binary format is written to a single report file named projectname rep_runnumber It contains the data for all the additional output It can be viewed and plotted in R Desk using the SDF File Manager and XY Plot Manager Ricardo Software December 2009 283 17 USING SOLVER 17 7 FLUID PHASE 17 7 Fluid Phase The main pa
146. two adjacent boundary regions are non conformal i e they are not successfully converted into an interface region see Arbitrary Grid Interface In case of conjugate heat transfer the solver will treat these non conformal wall boundaries as adiabatic walls Ricardo Software December 2009 129 MODELLING CONTINUA AND THEIR PROPERTIES Physical properties of fluid or solid materials are very important constituents of the simulation VECTIS MAX defines the general continaum model which supports and manages multiphase multicomponent fluid materials and physical models associated with them The general continuum describes the mixture of phases with each phase consisting of one or more species It is important to understand the difference between the multicomponent phase which is a mixture of species and multiphase mixture The principal assumption behind the definition of a multicomponent phase is that species are mixed at the molecular level In contrast to the multicomponent phase the multi phase mixture contains the single or multicomponent phases which are mixed at the macroscopic level meaning that the corresponding length scales are much larger than the molecular ones Most of physical properties appearing in the transport equations are thermodynamic intensive prop erties They arise from constitutive relations and from the equation of state which are both required to close the system of transport equations 8 1 Constitutive Relations
147. user should have knowledge of the following O Setting up the coolant jacket geometry O Setting up the control mesh for coolant jacket analyses O Setting up the VECTIS MAX solver for a coolant Jacket analysis O In particular how to set up the calculation to model incompressible liquid flows O Monitoring the analysis and convergence using Live Update O Postprocessing the results 19 3 3 Geometry Preparation Phase 1 is the current pre processing package for use with VECTIS MAX Phase 1 is used to read in geometry and process it to a form acceptable to the VECTIS MAX mesh generator Depending on the quality of the initial geometry a certain amount of repair may be needed to form a single closed volume After copying the coolant tri file to your working directory using the hyperlink at the top of this page it can be loaded into the VECTIS MAX pre processor Phasel and should be displayed as shown by Figure 19 62 When performing a VECTIS MAX analysis a closed geometry file of the analysis domain is required For this tutorial this is provided in the form of the coolant tri file For future analyses the geometry can be repaired using the Phasel repair tools However in order to reduce the analysis preparation time it is more efficient to ensure that the STL file being exported from the CAD package is as complete and as closed as possible In particular for coolant analyses it is obviously crucial to get an accurate
148. user to specify initial flow velocity pressure temperature etc ORicardo Software December 2009 405 19 TUTORIALS 19 2 STEADY STATE PORT FLOW Initial Condition h xj X Velocity fo Yveoctyfo Z Velocity fo Pressure 100000 Temperature 293 15 Turbulent Kinetic Energy 0 05 Turbulent Dissipation 0 005 Turbulent Viscosity o 2 2 tO Volume Fraction O y Figure 19 47 Fluid Phase Initial Conditions These values will be imposed on the fluid when uniform initial conditions are selected in the FluidDomainh panel PostProcessing Output The data to be output to the post file can be selected in the output panel Each of the scalar variables checked will be written to the file Boundary Regions Each of the boundary regions can then be specified Region Name A name can be given to each boundary for reference The name can be referenced when using user programming routines Boundary report This specifies whether output data is written to the report files for this boundary Coupled Link Number When running coupled WAVE VECTIS the interface numbers for each cou pled boundary are specified here Boundary Condition Type Next the appropriate boundary type is chosen for each boundary Details of the different boundary conditions can be found in the BOUNDARY CONDITION TYPES chapter The following boundary conditions will be used Boundaries and 2 will be walls with a fixed temperature of 300 degK
149. usr_var_name then if var_name velocity then call get_field idt iget_cell var_name fi_vec lupr gt start do ic icell icel2 end do lupr gt end fi_vec gt null else call get_field idt iget_cell var_name fi lupr gt start call get_grid_geom xyz_cell_c xcell do ic icell icel2 fi ic xcell 1 ic xcell 2 ic xcell 3 ic end do lupr gt end fi gt null xcell gt null endif end if 1 Next usr_obj_name and usr_var_name copy amp edit the above code end subroutine upr_init Ricardo Software December 2009 358 18 USER PROGRAMMING 18 7 EXAMPLES 18 7 3 Example of the user boundary conditions routine User programming routine to specify boundary values profiles for VECTIS 4 solution variables subroutine upr_bnd_cond var_name idt ir jbnd1 jbnd2 This routine is called for each boundary region and each phase lpresent in the corresponding fluid solid material domains after the default initialisation is done by the solver species Modules used imported type definitions use upr implicit none integer iwp intent in idt eke jondl jond2 character len intent in var_name Local Variables real wph pointer real wph pointer integer iwp pointer fib gt null 1l_be gt null character len len_var_name usr_obj_name usr_reg_name iget usr_obj_id integer iwp usr_reg_id 7 jb vic real wph
150. value is 35 It is not recom mended to set a value less than 20 because then the number of generated boundary faces starts to increase rapidly IJK_BLOCK Definition of IJK refinement blocks see section 4 6 2 below It is not recommended to try to manually modify the section MESH_COORDINATES since its format is quite complex and it is easy to make a mistake here The best approach is to let Phasel to write the information down Since the meshfile generated by Phasel must be readable also by the older versions of VECTIS VECTIS 3 it may contain also three keywords which are not used by VMESH PATCH_TYPE VOLUME_DEPTH and OUTPUT 4 4 Command Line Options In this section all switches which can be used with VMESH are described When none of the parameters test rep sep info locate is used normal meshing work is assumed and the given filename is assumed to be the input ascii file vmesh meshfile switches GENERAL SWITCHES h or help or help shows the help v shows the version of the executable elimill all ill cells detected during the meshing task will be removed dd switches on Distance Decimation method which is used after each generation of patches by EF nops switches off Polygon Simplification all patches stay triangular nocs disables Cell Splitting feature fmc b1 b2 bn forces Marching Cubes see section 4 5 method for all boxes intersected with one of the listed boundaries b1 b2 bn However MC
151. variables integer ir j ic mat idt integer iwp pointer Jsbc jsbc 1_bcells real wph pointer dnb real wph ystar non dimensional wall distance get starting amp ending boundary face for each boundary region call get_reg ise_reg_face jsbc jsbc get list of cells adjacent to a boundary face call get_grid_connect 1_bndf_cells 1_bcells get of normal distance from the near boundary cell centre to the boundary face call get_grid_geom bndf_normal_d dnb Calculate non dimensional wall distance for all boundary faces belonging to boundary region ir do j jsbc ir jebc ir ic 1_bcells 3 ystar deln_star idt mat ic 3 dnb 3 end do Ricardo Software December 2009 338 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES function deln_star idt mat ic j deln Arguments Description deln_star real normalised wall distance idt integer data type domain or phase mat integer fluid material id ic gt integer near boundary cell _bndf_cells j integer boundary face or interface ise_reg_face deln real normal distance cell gt boundary bndf_normal_d and interf_normal_d Table 18 29 Function deln_star to calculate non dimensional wall distance 18 3 2 21 local_force This UAR can be used to calculate to calculate viscous and pressure forces at individual boundary faces and interfaces see Table 18 30 Consider an exam
152. vof_mass_flow volume fraction mass flow rate x x Species variables spec_mass_frac species mass fraction x x x x x lam_mass_diff laminar mass diffusion of species x x Passive scalar variables mass_frac_ps mass fraction of passive scalar X X x X x lam_mass_diff_ps laminar mass diffusion of passive X X scalar Table 18 26 List of accessible variables This list provides names or keywords that may be used to get values for transport equations being solved and other variables that are derived from solved transport equations Keywords given under var_name can be used with function get_field to obtain their values Ricardo Software December 2009 333 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES subroutine get_field idt iget var_name fi_out Arguments Description idt gt integer domain type id iget integer takes in iget_cell iget_o iget_oo iget_bnd iget_uif iget_lif iget_face var_name gt character variable name token fi_out gt real pointer can either be a scalar of vector depending on iget the output would be fi at centre fi at centre old fi at centre old old fi at boundary fi at upper interface fi at lower interface and fi at faces respectively fi_out output is a scalar fi_out output is a vector similar to fi but now the result is a vector used for velocity for example subroutine get_field idt var_r
153. 0 2375 3250 4125 5000 Mr 519 fps 62 5 Figure 19 87 Visualising a slice through the gasket holes Select the slice plot in the plottree and then in the data panel select the contour data mass_flux_ across_plane This will show the mass flow distribution through the gasket To find the mass flow through each of the gasket hole passages the extract flux tool can be used Right click on the slice plot in the plottree then select Extract gt Flux and Torque Figure 19 88 Plots 138 E coolant post_001 Plotp1 Hide Canvas 3d Bink Figure 19 88 Extract Flux Options This will bring up the extraction panel Use the polygon option to click around the gasket holes Normally a left click in the canvas is required to define the polygon points Figure 19 89 The mass flux through the selected region will then be written to the information panel Figure 19 90 More information on the general post processing capabilities can be found in the R Desk section of the help Ricardo Software December 2009 436 19 TUTORIALS 19 3 COOLANT FLOW Figure 19 89 Using a polygon to define the region to extract flux data from 2 a xi Flux output kg s Figure 19 90 Flux data written to information panel ORicardo Software December 2009 437 19 TUTORIALS 19 4 IMPORTING THIRD PARTY MESHES 19 4 Importing Third Party Meshes Get the necessary files for the importing third party meshes tutorial http www
154. 0 4 3 Conjugate heat transfer Conjugate Heat Transfer CHT describes coupled heat transfer through adjacent fluid and solid material domains There are numerous engineering applications where detailed CHT analyses are required Typical automotive applications such as cooling of cylinder heads and engine blocks as well as under hood aero thermal management illustrate well the challenges associated with modelling of CHT in complex geometries The total enthalpy equation 10 7 or total energy equation 10 26 is solved in terms of temperature in a fully implicit manner over the entire global domain see also the multi domain example in Figure 7 1 The conformal meshes at material interfaces should be provided with the help of arbitrary grid interface tool The discretisation of diffusion fluxes at interfaces is dealt with in the section Discretisation of the energy equation There are no special inputs related to the setting of CHT problems Based on the supplied multi domain structure i e when there is at least one fluid and one solid domain the solver will activate the CHT solution 10 5 Modelling Mass Transfer Mass transfer is analogous to heat transfer As heat is transferred from regions of high temperature to regions of low temperature the mass of one species travels from regions of high concentration to regions of low concentration By its nature mass transfer occurs within multi component phase continuum which is defined as mixtu
155. 009 145 8 MODELLING CONTINUA 8 6 SELECTING CONTINUUM AND ITS PROPERTIES The refreshed Solver Setup Tree will present the Species node with the above WAVE species see Figure 8 11 right Left clicking on the Properties node of the selected Fluid Species opens the Species Properties panel similar to the one shown in Figures 8 4 and 8 10 Here the phase properties labelled with the Mixture option can be specified in the same way as for any general species Density and specific heat will show the Inactive option Passive scalar and its properties Passive scalars are similar to species however they do not affect properties of the parent phase The Fluid Phase Setup panel contains the Define Passive Scalar check box as shown in Figure 8 12 left Phase_1 Fluid Phase Phi 1 Phase Name ase_ Initial Condition 4 Mixture Of Species Option Multi Component Phase Output Species Phase Property Filename none H Species_1 Fluid Species Properties Phase ID 1 Output Phase Type Species_2 Fluid Species Properties e Gas Liquid Output Passive Scalars Compressibiity Incompressitie Fi al v Solve All Species Cupin Passive_Scalar_2 Passive Scalar y Define Passive Scalar Output Figure 8 12 R Desk setup Adding passive scalars Left click on this box adds the Passive Scalars node to the Fluid Phase node in the Solver Setup Tree Figure 8 12 right Only one passive scalar is ad
156. 1 0 1 5 05 TVD 0 5 b CBC Figure 14 4 NVD diagram TVD and CBC criteria and characteristics of several convective schemes As Figure 14 4 shows the basic schemes such as CDS LUDS and QUICK are linear and there fore not always bounded A general form of these schemes applicable to the uniform grids is derived in Gaskell and Lau 1988 Taking into account the grid spacing for the situation shown in Figure 14 3 the unbounded limiter can be obtained as Pj gp aj gu 0 r 14 25 1 1 so fi I ff su 5 fi 147 14 26 The parameter a defines a family of the second and third order accurate schemes For example a 0 gives the QUICK scheme and 0 5f 1 f7 gives the LUDS scheme 14 1 3 2 Setting up the convective scheme A number of convective schemes are also available through R Desk GUI setup Under each Fluid Domain node in the Solver Setup Tree panel left click on Equations amp Solver node and the Equations Solver panel is displayed to the right For each transport equation being solved for the given Fluid Domain one of four available convective schemes can be selected from the ListBox Convective Scheme UDS CDS MINMOD and SMART The value of the Blending Factor and Relaxation Factor can also be specified The panel also includes an option for User Defined Sources see upr_sources An example for Momentum Equations is given in Figure 14 5 The solver part in Figure 14 5 in discussed in
157. 10 and Figure 8 11 The Species modelling equation in the Equations Solver panel Figure 10 1 will be also activated l e Ns The computed values of mass fractions c have to satisfy the compatibility condition ci 1 This condition can be enforced by either solving mass fractions for all species or solving for Nsp 1 species and calculate the mass fraction of the last species as Nsp 1 CN 1 Y ci 10 30 i l The FluidPhase panel shown in Figure 10 3 contains the Solve All Species CheckBox By default the box is not checked i e Ns 1 species will be solved for 10 5 1 Passive scalar Transport of a passive scalar is governed by the same equation as standard species As the attribute passive suggests the thermo physical properties of the parent fluid phase are not affected by the passive scalar properties i e by their mass fractions Section Passive scalar and its properties should be consulted about setting of a passive scalar Ricardo Software December 2009 181 11 MODELLING POROUS MEDIA A porous medium is a three dimensional region occupied by continuum which comprises both fluid material and fine scale solid structure The structure and its interstices pores through which fluid permeates are usually too small to be resolved by computational mesh Exhaust catalysts packed columns heat exchangers filters flow distributors and tube bundles are some of porous media application examples
158. 12 31 cond With amp defining the reduced thermal equilibrium void fraction amp is the measure of the departure of the current status with respect to the reduced thermal equilibrium which is con sidered as the driving force to return to the reduced thermal equilibrium state tevap and tcona in Equations 12 30 and 12 31 are the timescales for to return amp through the processes of evaporation or condensation In the saturated region this is defined as n a eq 12 32 wag Sag where Pz and py are the densities of vapour and liquid x y is the vapour mass fraction under the thermal equilibrium condition which can be calculated according to the enthalpy of the mixture h the saturated liquid enthalpy Asy and the latent heat of vaporisation A fa h hsf o hfg 12 33 When n 1 amp becomes the value in thermal equilibrium state In current work n 1 1 to take effects of thermal non equilibrium and non zero drift velocity between two phases into account In sub cooled domain amp is simply set to zero The time scale for evaporation feyap is set to a constant in the current study whilst the time scale of condensation tcona 1s a function of degree of sub cooling Tsat teond K con 12 34 AAT Ricardo Software December 2009 204 12 MODELLING MULTIPHASE FLOWS 12 4 PHASE CHANGE MODELLING where Kong is a constant Tsat the saturation temperature and AT the degree of sub cooling
159. 139 R Desk setup Setting up a fluid phase oo csoc soes ee 140 R Desk setup Setting properties of the single component phase 141 A material property in terms of piecewise linear amp polynomial functions 142 R Desk setup Property calculation options Piecewise linear 0 2 142 R Desk setup Property calculation options Polynomial Piecewise polynomial 143 R Desk setup Property calculation options Sutherland formula 144 R Desk setup Defining the multi component phase in terms of species 144 R Desk setup Setting properties of individual species o o o 145 R Desk setup Setting properties of WAVE components Species o 145 R Desk setup Adding passive scalars o o ee 146 R Desk setup Setting properties of passive scalars o ooo o 147 Velocity distribution in the inner wall region for channel and boundary layer flows with zero pressure gradients s e da wta ar Gs Soe Al e we 156 A priory comparison of U F y expression 9 49 with DNS data for boundary layer Spalart 1988 and channel Moser et al 1999 flows o o o 160 A priory comparison of wall functions for near wall turbulent energy production left and its dissipation rate right Eqs 9 52 and 9 55 9 56 respectively with DNS data for channel flows Moserct al 1999 s
160. 16e 5_wp t ic 273 15 1 5 273 154111 t 4c 111 end do nullify t tb ti tli case conductivity case specific_heat case gas_constant case lam_mass_diffusion end select lupr gt end 1 end if Next usr_obj_name and select case pro_name copy edit the above code end subroutine upr_properties The above example is generalised to handle any multi species phase and multi domain for that matter case E g for the case where there are 2 phases amp 4 species air methane water methane this case will only set properties for the air species component To set properties for more than one component phase repeat code block and change us r_obj_name accordingly The properties set using the case construct are density and laminar viscosity Within the density case block the temperature amp pressure fields are retrieved from the solver along with the molecular weight Prior to calling get__field it is first necessary to determine the domain type id idt This is done via the 2 calls call get_id mass_pressure ieq iget_eq Get equation index ieq from equation name land object type iget_ieq idp eq_idt ieq iph iget_phase Get domain type id from ieq and phase id iph land object type iget_phase The object type passed to get __idis the transport equation identifier i yet_eq see Table 18 23 An alternative way of determining eq via get _
161. 19 1 6 Generating the Grid File lt o ee aj erste weed cedars Be A ee one ge Aad a 372 19 1 7 Importing the Grid Piles seoa hace da ed beeen keane ha bd babes 373 IONS SONET SENP A 373 19 1 8 1 Solution Conttol s s a ios ria ee EER a eS 374 19 1 6 2 Fl id Domain 2 2 6 6 6 wd ene ta A AS Re er 377 POSS Fluid Phase cc satie ena aa E a e Beards es 380 19 1 8 4 Bo dafy Repions 400 24 404 ba ela a a Xe a 381 191 9 Grid Preparation ido a de ek e Ea a eka ee a eee we 383 19 1 10 Running thesolver oi AE eed e da Oe EOE eA eG 383 PO MT Maver Updaters co ga be sd ee Se ee EGE ew we eS e 384 19 1 12 POst processing oia ae ee wee er gen eh a a ea dt ge Mee 384 19 1121 2D plane see gasie ae A EP aw A a we a a a 385 19 1 13 Potential Plow Solvers os c rior radar a eed as 385 19 2 Steady State Pont FIOW 00m rd Ge oe So a a a we Re ee 387 192 1 Introducir dis aka bale ede PS ee Ses eee es 387 ORicardo Software December 2009 xi 192 2 AIMS 6 inset te eh ho ee A A A a Be a 387 19 2 3 Geometry Preparation ora eared oh ME eee ek ard ee 387 192 4 Selecting Boundaries iuris ba ed bye p ee ha hd babes 388 19 2 5 Boundary Description lt s sssr emie eas Ha Oe A Oe Ra ee we ee 389 19 2 6 Mesh Specilication rasa asas e Re SARE RRA Ree eS 8 390 19 2 7 IK Refinement eos cd a a A a eo 391 19 28 Boundary Refinement ss 20 054 2s vd woe aa a da 392 19 29 Mesh Generation as tara Ba ea de de Bee edd aa 393 TO 2 10 Solver Sep coco ar
162. 2 4 2 Cavitation Models The modelling of cavitation if very significant in engineering designs Cavitation can have a pro found effect on many engineering systems and its careful simulation is crucial Some of the exam ples where cavitation plays an important role include fuel injectors hydrostatic bearings marine propellers and many others When cavitation occurs on a device it affect performance such as load symmetry noise vibration reduced flow rate etc and may cause physical damage to the device Thus physical understanding of cavitation and implementation are very important to industrial applications so that cavitation can be eliminated during the design stages The cavitation model presented here is a multiphase homogenous mixing of the liquid and vapour phases and the slip velocity between phases is neglected The onset of cavitation occurrence is associated with the saturation pressure in the liquid at a given temperature which causes cavitation bubbles to generate These cavitation bubbles then grow and eventually collapse in a short time to release huge amount of energy associated with very high temperature and pressure on very small surfaces Cavitation was first studied by Lord Rayleigh in the 19th century Rayleigh 1917 More sophis ticated models that can be applied to a vast number of cavitating problems have been developed recently O Volume of Fraction Equation Recall equation 12 2 for a multi fluid homogeneous
163. 227 238 9 2 1 Singhal A Athavale M Li H and Jiang Y 2002 Mathematical basis and validation of the full cavitation model Journal of Fluids Engineering vol 124 pp 617 624 12 4 2 1 Slattery J C 1999 Advanced Transport Phenomena Cambridge University Press Cambridge 11 1 1 Sleijpen G and Fokkema D R 1993 BiCGstab for Linear Equations Involving Unsummetric Matrices with Complex Spectrum Electronic Trans Numer Anal vol 1 pp 11 32 14 2 6 Smagorinsky J 1963 General Circulation Experiments with the Primitive Equations I The Basic Experi ment Mon Weather Rev vol 91 pp 99 164 9 2 1 Spalart P 1988 Direct simulation of a turbulent boundary layer up to Reg 1410 Journal of Fluid Mechan ics vol 187 pp 61 98 document 9 4 4 9 2 Spalart P R and Allmaras S R 1994 A One Equation Model for Aerodynamic Flows La Recherche Aerospatiale vol 1 pp 5 21 9 2 1 Spalding D B 1991 Kolmogorov s Two Equation Model of Turbulence Proc Roy Soc London A vol 434 pp 211 216 9 2 1 Speziale C G 1991 Analytical Methods for the Development of Reynolds Stress Closures in Turbulence Annual Review of Fluid Mechanics vol 23 pp 107 157 9 2 1 Speziale C G 1996 Modeling of Turbulent Transport Equations in T Gatski M Hussaini and J Lumley eds Simulation and Modeling of Turbulent Flows Oxford University Press Oxford 6 4 6 4 9 3 9 3 1 Sweby P K 1984 High Re
164. 295 295 296 298 18 3 1 User accessible variables o sumo dc a Be A a Bw 298 139 3 2 Wseraccessible routines s ao oe rei ei eee sears hh dl Gre hae geld da Geared les o 303 16 321 get number 422 4444 44254 6424 id hae ea id bee 304 183 22 Set domain s a Hw Sh HUH ew a GS a we ae Ge 305 18 32 Sl Mai asada Rod Ra RS eee Soka ae ex 307 18 324 get phase eile i ud as eh a A 310 1893 23 Pelos peces sie a RN Rw bv ee ES ae 313 18 320 PeP esta Mie db week ee ie ee A we ae eek a 314 18 32 SELIG ic eee hk chb eed bad eda ee dake PA ed de 314 18 3 28 Pet propertyO cba aia ee E edi eA ed ee Bad bee eA A 318 18 329 get end LEO ek se a Ge a hae a ee hw OG dow Be A ogee doe 322 13 3210 get gid Connect 0 a ace ir a a a aa 323 18 32 TI REDO ca a as EP ae ms 4 Oe ee Oe Ee aa a e 326 1353212 get run ClO sia ridad db a SH 327 132 13 Set MAME a rd e Ge a a a rs alee es 329 1532 14 Set mame STO uf eee A dd A a i 329 DO ME es oy ok ord ee A a Sk HERS BEAMS ES BO 330 183 216 get parent SAN 331 18 AOU ri daa a AAA A 331 18 3218 Bet MM obo A PAS ES GES LARA A a BE EH 332 18 3 2 19 BELLA a a aa id ls eee dia AA BOR Ee de a 334 18 32 20 deln star oe sew sara aa da aa 337 18 3 2 21 local TOTES 0 a a a eh ad ee E E ANNA 339 18 3 2 22 USER DOS ci tna ares has ce eds He MI eo od GO Be wr ay es 341 13 3 2 23 global SUM iia ba LAO eh OE OR a di 343 18 3 2 24 global _max global_MidO 2 iaee bed ee ee ee a 343 132 23 conca amay O 466 bare A A
165. 3 Figure 19 67 Re positioning the external red control lines For coolant analyses the geometry scale of the main coolant jacket to the gasket geometry is usually significantly different Therefore it is good practice to use red control lines to control the mesh cell sizes which are located in the gasket Create two new mesh lines positioned at the height extents of the gasket Also define the number of intervals between these lines as 2 This is shown by Figure 19 68 Y Ricardo VECTIS Phase 1 coolant tri joj x Fie Edit View Toolbars Operations Help gage gage i N e E forbes of Cols neren Number of cells 2 Length of cells m 0 001472 zmin 0 078000 zmax 0 080000 ORDERIN 20 22 46 IG TRIANGLE NORMALS 20 22 46 FINISHED ORDERING TRIANGLE Ni S DONE 20 22 46 WRITING MODEL DATA TO D RPe treining_tutoriels V4 coclant coolant trs 3 Figure 19 68 Using meshlines to control the cell sizes located in the gasket region The divisions for the other sections between the red control lines can now be defined For this example the computational control mesh cell size will be set as 5mm x 5mm x 5mm For future analyses it is up to the user to determine what size cells to use Typical cell sizes are 2mm to 4mm Ricardo Software December 2009 42 19 TUTORIALS 19 3 COOLANT FLOW The spacing distance between the red lines can be determined by using the measurement icon in the Mesh Setup Toolbar
166. 3 1 X ERROR 1505 Tying patches routine cannot break the required edge in box 53463 in these IJK positions 58 31 3 SUCCESSFULLY DONE BUT WARNINGS APPEARED CHECK MESH QUALITY Normally the solver will run successfully with these non fatal error however occasionally these problematic areas can hinder convergence The mesher output indicates the IJK regions where the errors occur These can be visualised in Phasel and any obvious problematic geometry can be rectified by the user Load the coolant tri and coolant mesh file into Phasel The tool chop triangles by IJK number quickly allows the problematic geometry to be visualised Enter in the location of the first IJK region 42 31 3 19 71 Switch on the visualisation of surfaces and lines to make things easier to see In the option tab check on show surfaces and set outline colours to mono It may be necessary to grow the chopped area to visualise exactly what the geometry looks like 19 72 It can be seen that there is a very thin triangle which creates an overlap highlighted in green The easiest way to rectify this is to delete the triangle and its neighbours and create new triangles with a better aspect ration 19 73 Use the chop triangles by IJK number tool again to visualise the region 58 31 3 A small imperfection can be seen in the surface Ricardo Software December 2009 424 19 TUTORIALS 19 3 COOLANT FLOW
167. 333 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES To loop over cells for domain dom 1 then use real Up V W do i ic_s dom ic_e dom u vel_cell 1 1i lx velocity component v vel_cell 2 i ly velocity component w vel_cell 3 1 lz velocity component end do Indices can also be passed when the call to get__field is made as integer icl ic2 icl ic_s dom ic2 ic_e dom call get_field idt iget_cell velocity vel_cell 1 3 icl ic2 One can also use Fortran 95 2003 ubouna 1bounad explained earlier 18 3 2 19 get_grad This UAR can be used to obtain the gradient of scalar or vector solution variables The user is repsonsible for allocating memory for the gradient array The first part of the Table 18 28 deals with scalar gradient and the second part with vector gradient a Obtaining a scalar gradient To get the gradient of temperature then do the following integer dom idt n_cells real wph pointer g_temperature dom 1 Idomain 1 used as an example Get temperature domain index idt eq_idt ifmas idom iget_dom Allocate gradient array n_cells get_number n_cells allocate g_temperature 3 n_cells get gradient of temperature call get_grad idom idt g_temperature temperature The call to get_grad can also be made as call get_grad idom idt gfi_out g_temperature amp Var_name temperature where pointers are assigned directly
168. 4 5 BASIC SCHEME OF VMESH 10 GENERATE CELLS ON BOUNDARY BOXES For all boxes 1t is decided whether the box is intersected by the surface fully inner or outer The optimal order of processing boundary boxes is found smaller boxes need to be processed first to be always sure that when processing a box that has more than one neighbour the neighbours are all done already Than all boxes intersected by the boundary are looped and their faces are generated in these steps POLYGON GENERATION PART A1 Generate boundary faces patches If a sharp edge is detected in the box or if there is more than one boundary conditions or if one of the edges of the box is intersected more than once Exact Fit method of generation of patches is used It means that the triangles intersecting the box are broken to get the triangles exactly cutting the box In the other cases there is just a simple cut Marching Cubes method is used there are 14 basic patterns how to create triangles on the intersections of edges A2 Polygon Simplification This technique simplifies the patch structure storing boundary faces as polygons is more memory efficient than keeping triangles A3 Generate inner faces if neighbours are already processed Loops all velocity locations of the box and on those locations where patches of the two adjacent boxes were already generated from both sides the patches are tied not to have gaps between boundary faces and polygons of inner faces are
169. 401 x 0 076178 y Distance 0 012768 offset x 0 012766 y 0 Node co ordinate display Interrogate Line Button EN Once this command has been selected the user can select a line on the model and the line number will be displayed In addition the line end nodes and the line s adjacent triangle numbers are displayed together with the line length and the direction vector the line defines Ricardo Software December 2009 37 3 GEOMETRY 3 12 TRIANGLE INTERROGATION OPERATIONS Line12943 Length0 002364 offsetsdx 0 002364 dy 0 000005 d2 0 000011 Interrogate Line display Interrogate Triangle Button Al Once this command has been selected the user can select a triangle on the model and the triangle number will be displayed In addition the triangle s line numbers line end nodes and the triangle s adjacent triangle num bers are displayed together with the triangles boundary number and the triangle s normal vector components Triangle3587 BoundaryHumber4 HormalYector x 0 004768 y 0 581374 2 0 813623 Interrogate Triangle display Interrogate Boundary Marked Region Button Ricardo Software December 2009 38 3 GEOMETRY 3 13 HINTS FOR MANUAL STITCHING Once this command has been selected the user can select a triangle in the model and the triangle number will be displayed In addition the triangle s boundary number the surface area of the boundary region and the centroid of the boundary regi
170. 6 0D bu giving ou 0 and 0D 1 It is not difficult to prove that the inequalities 14 14 and 14 15 are simultaneously satisfied if the following relations hold lt 1 8 O gt 4c if 0 lt c lt 1 0 0c if oc lt 0 or dc gt 1 These are CBC constraints derived earlier by Gaskell and Lau Gaskell and Lau 1988 In the case of the monotonic profile characterised by p c c y gt 1 we might prefer to bound the 14 17 Ricardo Software December 2009 238 14 NUMERICAL SOLUTION 14 1 FINITE VOLUME DISCRETISATION cell face value by the extrapolated downstream value 5 2 c y rather than by the actual value Op If this constraint expressed as CMe c1 or SE lt a 14 18 c du Oc is added to the previous inequalities 14 14 and 14 15 the TVD criteria are obtained 051 8 OS 26c amp gt c if 0 lt Gc lt 1 his oj Qc if Oc lt 0 or Oc 2 1 For arbitrary unstructured grids the upstream node is not known However an imaginary upstream cell can be defined in such a way that the vector identity CU CD is satisfied and that an imaginary face between C and U is placed at the same distance from C as the considered face j Figure 14 3 In this case the centred difference p dy is simply Vdc 2d which yields the reconstructed value 7 as 07 bn V c 2d 14 20 The above reconstructed value need to be bounded by the values at neighbouring cells which surround the upstream
171. 7 1 Simple multi domain simulation example Reg 2 Outlet 8 Wall 10 Wall computational domain 7 2 Boundary Regions Different physical boundary conditions are distinguished by Boundary Condition Types and they are applicable to different parts of the domain boundary Therefore material domain boundaries are sub divided into non overlapping boundary regions according to the boundary condition type which will be applied later A boundary region is described by a set of not necessarily adjacent boundary faces Boundary regions are created directly on the domain geometry before the mesh generation see Geometry In the future it will be possible to create and manipulate boundary regions after the mesh generation 7 3 Interface Regions Interface Regions are implicitly defined at fluid solid and solid solid material interfaces For every pair of adjacent material domains there is one interface region With reference to the solution of transport equations the interface region can be treated as explicit or implicit In case of the solution of energy equation over coupled domains conjugate heat transfer the corresponding interface regions are defined as implicit For other equations they are explicit i e they act as the standard wall boundary regions Ricardo Software December 2009 125 7 MODELLING SPATIAL DOMAINS 7 4 ORDERING OF DOMAIN COMPONENTS 7 4 Ordering of Domain Components VECTIS MAX domain structure is built up by indexin
172. 7 Running the solver Now that the inp file has been created and the grid file has been processed the solver can be run Again this will be run from the cmd window by typing the following vsolve port inp The solver will now start to run 19 3 8 Live Update Using the live update is a utility in R Desk allows a simulation to be monitored whilst it is running Within the VECTIS session in R Desk open the Live Update and XY plotmanager panels if they are not already open Then open an xy canvas using the new XY canvas button In the Live Update panel browse to the directory where the simulation is running The available data files are presented in the Files window The data files available are determined by the ascii files selected in the reporting section of the 19 3 5 1 in the solver setup tree Select the residual data entry for the appropriate run The data available in the file is shown in the Value window Drag the data from this panel in pairs Iter_Num Vs Residual to the XY canvas It will then be plotted Each curve on the plot will appear in the XYplotmanager In this panel the properties of the Ricardo Software December 2009 432 19 TUTORIALS 19 3 COOLANT FLOW individual curves and of the XY plot can be modified Reading VECTIS 3 input file Al O boundaries are mass flow setting first outlet to pressure Figure 19 82 Live Update Additional the scale can be changed to a logarithmic type by right cl
173. 8 3 ACCESSING SOLVER VARIABLES real wph pointer g_temperature 35 amp cell temperature tb amp boundary temperature ti amp upper interface temperature ee LEC lower interface temperature get gradient of temperature call get_grad idom idt gfi_out g_temperature amp fi_c t fi_b tb fi_ui ti fi_li tli b Obtaining a vector gradient To get the gradient of velocity components then integer 2 dom idt real wph pointer gx_vel amp gradient of x component gy_vel amp gradient of y component 1gz_vel Igradient of z component dom 1 domain 1 used as an example Get velocity domain index idt eq_idt ifmom idom iget_dom get gradient of velocity call get_grad idom idt gx_vel gy_vel gz_vel var_name velocity It is important to note the following information related to var_name 1 if var_name is supplied then there is no need to pass optional arguments such as fi_c fi_b fi_ui and fi_li The routine will automatically find the values of these optional arguments 2 if var_name is not supplied and for example we want the gradient of a user defined variable then fi_c should be supplied at least 3 if arguments fi_b fi_ui and fi_li are not passed then the routine will use near boundary cell values as boundary values 18 3 2 20 deln_star This UAR can be used to calculate non dimensional wall distance based on the wall function approach see Table 18
174. 8 7 EXAMPLES else iobj iph iget iget_phase end if lGet associated object id with this object usr_obj_name call get_id usr_obj_name usr_obj_id iget Only perform block below if have right object id if iobj usr_obj_id then Get property index call get_id pro_name ipro iget_ph_pro Determine property to be calculated for select case pro_name case density lupr gt start Calculate density for ideal non isotropic gas flow Pressure field is required call get_id mass_pressure ieq iget_eq Get index of pressure eq idp eq_idt ieq iph iget_phase Get pressure domain index Get pressure field NOTE returned are RELATIVE pressure values var_name pressure call get_field idp scal_field var_name amp fi_c p fi_b pb fi_ui pi fi_li pli Get reference pressure call get_mat r_pressure p_ref Get phase molecular weight and calculate gas constant Call get_property r_phase_values rph_val gc 8315 0_wp rph_val imolw iobj If temperature field is required Call get_id energy ieq iget_eq Get index of energy equation ide eq_idt ieq iph iget_phase lGet energy temp domain type Get temperature field var_name temperature call get_field ide scal_field var_name amp fi_c t fi_b tb fi_ui ti fi_li tli Define property values as a function of temperature pressure or both temperature and pressure fi_pro f T p_absolute do ic icell icel2
175. 94 Muzaferija 1994 which minimises the sum of squares of the differences op i O Fp i assuming the linear distribution between neighbouring nodes P and P 7 op Vop F Fp 14 28 To compute the face gradient the Gauss formula 14 27 can be applied to the control volume constructed around the face j Figure 14 6 This control volume is made of triangular faces A and Ai defined by the face vertices k and k 1 for the last vertex k 1 1 and neighbouring nodes P and P If the vertex values are reconstructed from the variable values and gradients at adjacent cells P and Pj the derivation outlined in Przulj and Basara 2002 leads to the following expression for the face gradient gt gt A _ VO DATA or 0 V6 d VO fiVOr 1 fi Vor 14 29 q 14 1 4 1 Setting up dimensionality and gradient calculation options gt From the Solver Setup Tree under Fluid Domain left click on Discretise This opens up the Discretise panel as seen in Figure 14 7 The Dimensionality of the problem can be specified ORicardo Software December 2009 242 14 NUMERICAL SOLUTION 14 1 FINITE VOLUME DISCRETISATION Selecting the Two dimensional Flow RadioButton will set the flow to be 2 Dimensional whereas the Three dimensional Flow will set the flow to be 3 Dimensional Cell gradient calculation methods are available for each fluid and solid domains In the Discretisation Setup panel under Gra
176. ANGLE MARKING OPERATIONS Cutaway Showing Volume Union Result 3 9 2 Boolean Operation Batch Mode The boolean operations are supported in batch mode in VECTIS version 3 9 0 onwards To run Phasel in batch mode the command line arguments are phasel bou bos boil boundappend boundmerge tolr tolrncy filenamel filename2 filename_result where bou is the boolean operation UNION default bos is the boolean operation SUBTRACTION boi is the boolean operation INTERSECTION boundappend performs the boolean operation with the boundaries and parts of the second geometry appended after the boundaries and parts of the first geometry default boundmerge performs the boolean operation with the boundaries and parts of second geometry will be merged with the boundaries and parts of the first geometry tolr tolrncy performs the boolean operation using the defined value for the merging tolerance of near nodes The default setting is AUTO which means that the program will automatically detect what the merge tolerance should be when loading the geometry files It is possible to use the batch mode capability in to perform consecutive boolean operations by simply having a script that performs the boolean operations where the file generated from the preceding operation is used as the input for the next operation For example phasel bou boundaappend A tri B tri AB tri phasel bou boundaappend AB tri C tri ABC tri
177. Approach PhD dissertation University of London 14 1 4 Nakayama A 2008 Theory of Porous Media and Its Numerical Applications to Engineering Problems in Proceedings CHT 08 International Symp Advances in Computational Heat Transfer Marrakech Morocco 11 1 3 11 1 3 Nakayama A and Kuwahara F 1999 A Macroscopic Turbulence Model for Flow in a Porous Medium ASME Journal of Fluids Engineering vol 121 pp 427 433 11 1 3 Patankar S V 1980 Numerical Heat Transfer and Fluid Flow McGraw Hill New York 10 2 1 14 14 1 5 4 14 1 6 14 2 6 Pedras M H J and de Lemos M J S 2001 Macroscopic Turbulence Modeling for Incompressible Flow Through Undeformable Porous Media Int J Heat Mass Transfer vol 44 pp 1081 1093 11 1 3 Popovac M and Hanjalic K 2007 Compound Wall Treatment for RANS Computations for Complex Tur bulent Flows and Heat Transfer Flow Turbulence and Combustion vol 78 no 2 pp 177 202 9 4 4 9 6 Przulj V 2009 Pragmatic Wall Treatment for RANS Simulations on Cartesian Cut Cell Grids in K Hanjalic and Y Nagano eds 6th Int Symp Turbulence Heat and Mass Transfer Rome Italy 9 4 4 Przulj V and Basara B 2001 Bounded Convection Schemes for Unstructured Grids AZAA paper 2001 2593 14 1 3 Przulj V and Basara B 2002 A SIMPLE Based Control Volume Method for Compressible Flows on Arbitrary Grids AZAA 2002 3289 14 14 1 4 Ranz W E and Marshall W R 1952 Ev
178. Averaging Kluwer Academic Publishers Dordrecht 11 1 1 11 1 1 Wilcox D 1998 Turbulence Modeling for CFD 2nd ed DCW Industries Inc La Canada California 6 4 9 2 1 Wolfshtein M 1969 The Velocity and Temperature Distribution in One Dimensional Flow with Turbulence Augmentation and Pressure Gradient Int J Heat Mass Transfer vol 12 pp 301 318 9 4 2 9 4 4 Yakhot V Orszag S A Thangam S Gatski T B and Speziale C G 1992 Development of Turbulence Models for Shear Flows by a Double Expansion Technique Physics of Fluids A vol 4 pp 1510 1520 9 2 1 9 3 2 Yakhot V and Orszag S 1986 Renormalization Group Analysis of Turbulence I Basic Theory Journal of Scientific Computing vol 1 pp 3 50 9 2 1 9 3 2 Yang Z and Shih T 1993 New Time Scale Based k Model for Near Wall Turbulence AJAA Journal vol 31 no 7 pp 1191 1198 9 3 9 5 Yao Y F Savill A M Sandham N D and Dawes W N 2002 Simulation and Modelling of Turbulent Trailing Edge Flow Flow Turbulence and Combustion vol 68 pp 313 333 9 5 Zwart P Gerber A and Belamri T 2004 A two phase flow model for predicting cavitation dynamics in ICMF2004 Yokohoma Japan 12 4 2 2 Ricardo Software December 2009 450
179. Boundary 3 will be a static pressure boundary set to 100000 Pa inflow and with a temperature of 300 deg K Boundary 4 will be a static pressure boundary set to 93133 Pa outflow and with a temperature of 300 deg K Ricardo Software December 2009 406 19 TUTORIALS Bnd_Reg_1 Wall Boundary y 19 2 STEADY STATE PORT FLOW Bnd_Reg_3 Pressure Boundary E Region Name Bnd_Reg_1 Region Name End _Reg_3 Region ID 1 Region 1D 3 Material ID I Boundary Report Coupled Link Number 0 Material 10 a sti s sSOSOSCisS T Boundary Report Coupled Link Number o Boundary Condition Type a Boundary Condition Type Presue O a Thermal Condition Op Prescrbed temperature Thermal Condition Op Given StaticPressure oollu U Boundary Setting Juniform vales SCS Boundary Setting Uniform values SCS Roughness Height o Inflow Outfow info Roughness Constant 0 5 eg Bnd_Reg_4 Pressure Boundary U U O O y o fo fo SS Heat Flux ec negon Men Bnd_Reg_4 Temperature ET O M VeeS Material 1D a HTC TO Temperature SN Lal EE Coupled Link Mumber o pRadiation one ron Emissivity fo Temperate fo Thermal Condition Op Given Static Pressure T Thin wal Boundary Setting Uniform Values m NEE AN Figure 19 48 Setting up the boundary regions Report Regions Report Regions can be specified to allow data to be extracted on arbitrarily defined surfaces The surfaces are
180. DELLING CONTINUA 8 4 PROPERTIES OF MULTICOMPONENT PHASE Modelling of fully compressible or high Mach flows Ma U a U v KRT gt 0 3 often exploits the properties of ideal gases with constant specific heats In this case using Equations 8 12 and 8 30 one can describe the isentropic process between states 1 and 2 s1 s2 K const by following relations Th mye T 2 Pp m y es or constant 8 31 Ti E T Api p iia It is also convenient to work with properties evaluated at the stagnation state The stagnation state is attained when a flowing fluid with U h T p p decelerates to zero velocity isentropically The corresponding properties are designated as stagnation or total properties The total enthalpy and temperature are then defined as U p U hrot h 32 tot gt e p gt 8 32 hio U igs T 8 33 Cp 2Cp In case of an ideal gas the total temperature and pressure are given respectively k 1 leat 2 8 34 P To K K 1 K 1 K K 1 Pioi Ha Prot p 1 Ma 8 35 p T 2 For an incompressible fluid p const the isentropic process toward the stagnation state results with U2 Trot T and Prot PrP gt 8 36 8 4 Properties of Multicomponent Phase The physical properties of the multicomponent fluid phase mixture of species depend on the relative amounts of the components usually specified in terms of mass fractions concentrations
181. December 2009 188 11 MODELLING POROUS MEDIA 11 2 SIMPLIFIED MODELLING EQUATIONS where the body force production is given by Equation 9 18 while the production due to mean macroscopic flow reads d Wi d Wi P uu T 11 44 ko p j Ox ij Ox and the macroscopic turbulent stress 7 j is defined by Equation 11 18 The extra terms appear ing in the k equation Sk ex are associated with spatial deviations of volume averaged velocity and turbulent kinetic energy 9 ET 0U Ssa Fe EU yU U T 11 45 These extra terms are usually modelled together Examples of specific models can be found in Nakayama 2008 de Lemos and Pedras 2001 O Dissipation of turbulent kinetic energy Ce1Pk Ce2p Ce3Pp T a 9 z PE gt ypeU o OYE YCeaP ESkk du ue de Se ex 11 46 The extra terms which are derived from the presence of porous medium are as follows S k ex o a ee e aU k 1 dE Se ex Ce ex amp U j 7C E E A Ox p EU j YCeap de aa V A OX The current practice is to model the above terms together see proposals of Nakayama 2008 de Lemos and Pedras 2001 njdA 11 47 11 2 Simplified Modelling Equations The current implementation of porous modelling is based on the simplified governing equations with the superficial velocity U being used instead of the intrinsic one U For all other variables the intrinsic volume averages are emplo
182. Domain Each of the solid parts may consist of many different materials Material A solid domain may consist of a number of different materials each of which can have different material properties Figure 19 2 VECTIS Solver Structure This first example shows how to set up and run a basic flow analysis It will assume a basic famil larity with performing CFD calculations Ricardo Software December 2009 368 19 TUTORIALS 19 1 BASIC TUTORIAL 19 1 3 Geometry Preparation Phase 1 is the current pre processing package for use with VECTIS Phase 1 is used to read in geometry and process it to a form acceptable to the VECTIS mesh generator Depending on the quality of the initial geometry a certain amount of repair may be needed to form a single closed volume Once this is complete Phase 1 is then used to identify regions of the geometry as different boundaries for the CFD calculation Finally the global mesh and other control parameters for the mesher are defined Further information can be found in the chapter GEOMETRY VI Ricardo VECTIS Phase 1 tube tri O Xx File Edit View Toolbars Operations Help E a a al 13120 0 10 MN 9 a Y Y Options Stiteh uE Lights 4 Visualize Leak Paths _ Fast Rotate Move F Show Surfaces J Flat Shading Flip Surface Display Show Mesh Setup J Number cells 43d Subdivide IJK Block Display v Off Solo All Outline Colours Off wv Multi wv Mono Highlight J
183. E y Velocity y Pressure v Temperature y Mass Fraction y Turbulent Energy y Turbulent Dissipation y Turbulent Viscosity v Volume Fraction y Density Y Laminar Viscosity Y Thermal Conductivity v Specific Heat Gas Constant Thermal Expansion Enthalpy Mach Number Figure 17 9 Postprocessing Output Panel Postprocessing Output o K Y Mass Fraction Mass Diffusion Thermal Diffusion Figure 17 10 Species Output Panel Iter No Area_TOT Temp_AVE Mflow_SUM Density _AVE S Pres_AVE T Pres_AVE Vee ee ER A E m 2 K kg s kg m 3 Pa Pa 1 1 000000E 01 2 931500E 02 1 188574E 00 1 188574E 00 1 008974E 05 1 009574E 05 2 1 000000E 01 2 931500E 02 1 188574E 00 1 188443E 00 1 006558E 05 1 007158E 05 3 1 000000E 01 2 931500E 02 1 188574E 00 1 188546E 00 1 002759E 05 1 003359E 05 4 1 000000E 01 2 931500E 02 1 188574E 00 1 188568E 00 1 000825E 05 1 001426E 05 5 1 000000E 01 2 931500E 02 1 188574E 00 1 188569E 00 9 999433E 04 1 000544E 05 6 1 000000E 01 2 931500E 02 1 188574E 00 1 188570E 00 1 000057E 05 1 000657E 05 El 1 000000E 01 2 931500E 02 1 188574E 00 1 188570E 00 1 000045E 05 1 000646E 05 8 1 000000E 01 2 931500E 02 1 188574E 00 1 188571E 00 1 000064E 05 1 000665E 05 The meaning of the variables are shown in Table 17 4 The _AVE values are area averaged 17 USING SOLVER values and are calculated thus Iter_No Time Tstep Area_TOT Temp_AVE Mflow_SUM Density_AVE S Pres_AVE T Pres_AVE
184. ENCE 9 4 NEAR WALL MODELLING O Temperature and mass fraction of species For the entire wall region the unified expressions presented for the standard wall functions are employed see Equations 9 43 9 44 O Turbulent energy production Following the standard wall function practice the near wall pro duction of turbulent energy is calculated using the velocity gradient obtained from the Equa tion 9 49 In addition the constant shear stress T across the near wall region is assumed Bearing in mind that k k y the velocity gradient can be written as OU _ O U Uz U dy Ty ELO y 9 50 J Jy T 105y 9 50 1 _ 9U yok 0U U Fi y y Syke ay z 9 51 Neglecting the normal stress production the production of k becomes U z P Ty E H F y y F O y 9 52 where T Tw H 9U dy denotes the turbulent near wall shear stress The exact form of the function F Equation 9 51 is not used currently Instead a form of F y y 9U dy is assumed To compensate for neglected k y derivative in Fj the coefficient amp in the blending function 9 48 has been re tuned with the help of DNS data to a value of o 0 075 Also K has been set to 0 37 and the final piecemeal function is given as gt 1 0 008y 0 005y y lt y Fily f 1 6 Ky G x 1 0 09y 0 075In Ey 10 yzy O As can be seen in Figure 9 3 left the expression for normalised turbu
185. FORMATION Opening fie D RPe training_tutonais V4 mesh_import tube GRD 11 02 28 INFORMATION Closing file D RPe training_tutorials V4 mesh_import tube GRD Figure 19 94 Pick options available for set editing Single Pick In this mode faces are picked one at a time To pick a face press Shift Left Click and Ricardo Software December 2009 440 19 TUTORIALS 19 4 IMPORTING THIRD PARTY MESHES when the mouse is released the face under the cursor will be picked shown by colouring it red and added to the set Paint Pick In this mode faces are picked as if they are being painted over To pick a face press Shift Left Click and then drag the cursor across the model As soon as the cursor moves over an unselected face it will be selected and added to the current set Flood Pick In this mode faces are picked within the bounds of an existing selection e g if we draw a circle around the circumference of our port we can then flood pick the inner faces To flood pick press Shift Left Click and then drag the cursor across the model as soon as the cursor moves over an unselected face it will be added to the current set Flat Pick In this mode all faces on the same surface are picked up to an edge where the edge angle is defined in the Preferences dialog To flat pick press Shift Left Click and when the mouse is released all faces on the same surface under the cursor will be added to the current set Polygon This mode involves t
186. Factor 1 Max Number Of Interations 20 Max Normalised Residual 0 0001 Potential Solver Solver Type Bi cg Stabilized Symmetric Conjugate Gradient Number Of Inner Solver Iterations 1000 Solver Tolerance 0 0001 Preconditioning Incomplete LU level0 4 Number Of Preconditioning Iterations ILU Preconditioning Options For Parallel Runs Mixed 4 Figure 10 5 R Desk setup Setting the solution of the velocity potential in Equations Solver panel Turbulence Model amp Turbulence Modelling Approach Laminar No Turbulence Modelling Turbulence Family DNS Laminar No Modelling Turbulence Model Figure 10 6 R Desk setup Selecting laminar flow regime in the Turbulence Model panel 10 3 7 Steady and unsteady flows In principle fluid flow characteristics are always time dependent i e the flows in nature are un steady transient Many flows however can be defined as steady in the statistical sense meaning that flow variables do not change significantly with time A selection of the steady or unsteady simulation is performed within Global Domain panel where the Steady or Unsteady time mode can be specified Figure 10 7 Ricardo Software December 2009 174 10 MODELLING SINGLE PHASE FLOWS 10 3 MODELLING FLUID FLOW Global_domain_1 Global Domain a E Time Mode Steady Unsteady 4 D Fig
187. Figure 19 98 Once the boundaries have been defined correctly run vpre on the GRD file This will reorder the Ricardo Software December 2009 442 19 TUTORIALS 19 4 IMPORTING THIRD PARTY MESHES 5 x MR File Edt View Options Window Help lal BOS prosa ER viene y le 20 26 gt eh Sr gt gt 60 x gt 2 gt 11 04 21 DEBUG e e me orcos 11 04 21 INFORMATION F auto apply Apply Opening fie D RPe training_tutorials V4 mesh_import tube GRD 11 04 21 INFORMATION PootPropertes J Data se f Closing file D RPe training_tutorials V4 mesh_import tube GRD gt Figure 19 98 Viewing the defined sets on the final GRD file faces according to the new boundary region definitions Once this has been done then the mesh can be imported as normal to set up the inp file ina VECTIS project Other Notes In the case of multi domain simulations Each domain should be converted sepa rately The boundaries and interface regions for each domain then needs to be identified before joining all the domains using vpre in the standard way 19 4 2 3 Importing grid file data into solver setup Now that the boundaries have been defined the computational grid can be imported into the solver setup In a VECTIS session in R Desk click on solver setup in the Solver setup tree Browse to the GRD generated previously Click on extract and then OK The boundaries defined in the file will then be read and added to the
188. For internal points setting it is recommended to wrap the geometry use Phasel Chop Area Create triangle Split Line and Measure Node tools Decimation angle To reduce the number of output triangles angle based decimation can be used It collapses edge preserving normal angles between old and new triangles at user specified interval Recommended value for the threshold is 2 3 degrees The maximum allowed da is 10 degrees Please note that for thin geometries self intersection can occur after decimation Decimation distance This decimation is also based on edge collapsing but for user specified edge length and a high angle tolerance It is recommended to set dd not larger than fr 2 and use a combination of da and dd Deviation distance This edge collapsing technique uses deviation criterion from the wrapped model to the original geometry It can produce more accurate results but requires more time The recommended value for deviation distance dv is fr 10 It is also suggested to use da and dd prior deviation based decimation Offset distance This feature enables an offset to be applied to the wrapped geometry The offset is based on moving the triangles along their normal direction by a defined distance If the input value is zero no offset occurs In case of the positive value the wrapped surfaces grows If the offset distance is negative it shrinks The allowed range is fr fr To obtain a larger offset
189. Gh cae Se on aa a 294 18 1 Solver flow chart and the corresponding UPR stages upr_ oo 297 POW VECTIS WORNOW it tga ods sek rr dnd on Meo ay he Sterns seg E 367 19 2 VECTIS Solver Structure op 4 4 ois a i a hella a ao Ge aha MSE a 368 19 3 Geometry Preparation and Boundary Painting 2 02 004 369 19 4 Global Mesh Dennition sor 24h ade ae Ee aE ord ae eee eo eA E 370 19 5 Options to manipulate mesh lines o ee a e a a e e e 371 Ricardo Software December 2009 xxi LIST OF FIGURES LIST OF FIGURES 19 06 REDES Layout narra amp ate oh ee Aa A a ae 371 19 7 Launching the mesh generator in R Desk o o e eee eee 372 19 8 Importing the mesh into R Desk Solver Setup o e 373 19 9 Selecting a boundary in the solver setup input tree o o o 374 19 10 The solvet setup input tree eo cria abe ga Bako hae aa a a 374 19 11 The Global Domain Panel cosas ta a a de 375 19 12 The Timebase Pamel as o a dws bed ek Ra es es ads 375 19 13 The Output Panl sni iie A wl ok Bie dhe a boa A ee A we a A de a 376 19 14 The Restart Panel io s cor hake a ehhh aed eh ed eee bake ehh bes 378 19 15 The Fluid Domain Panel oia a ee E edi A ee aed be a oS 379 19 16The Solver Setup Tree a 645 aa Seka gawd a da ba whe da OG wae ee ae eared a 380 19 17 Vhe Pluid Phase Panel c i a sts g ee we a a A ee a 382 19 18 initial Conditions s idos Be we Bo ee ae A EP
190. Gi Diz 10 6 where the turbulent mass diffusion coefficient is defined by Equation 9 10 The section in troducing mass transport highlights the role of the molecular mass diffusion coefficient Y Its value should be specified for each species see setting of fluid phase properties O Energy conservation _ ap aw 5 PH ere pH U U OL a a Ck Ho 5 Unsteady Term Pressure Change Convection Due to Species Diffusion EA qi a U a tj pUgj Pqu pf 10 7 dx qjv qj Or Ox ij T tij PUg j PQ PSiVi Heat Fluxes a Work of Stresses Pressure Work Source Body Force Work Turb Energy The above equation also outlines the physical meaning of various terms Laminar and turbulent heat fluxes represent energy transport due to conduction The work of viscous and turbulent Ricardo Software December 2009 167 10 MODELLING SINGLE PHASE FLOWS 10 2 EQUATION OF STATE MODELS stresses 1s also known as viscous dissipation The source term might include the heat of chem ical reaction or any user defined source In conjunction with eddy viscosity k e modelling the sum of laminar and turbulent heat fluxes is evaluated as OT qj 45 hero deff A A 10 8 J The turbulent heat flux is modelled according to Equation 9 9 and the turbulent thermal con ductivity A is given by Equation 9 10 The molecular thermal conductivity A is a fluid phase or species physical property which is specified as explained in
191. I file will be named as projectname intf_runnumber For more details see File Output section O For turbulent flow simulations where the wall roughness effects are not negligible the user should provide values for the Roughness Height in m and Roughness Constant Contact Resistance Similarly to the Thin Wall Model the thermal resistance caused by the thickness of a thin wall placed along an interface can be taken into account via thin interface model To enable the thin interface model the Contact Resistance box should be checked The Contact Resistance Data group box will appear and the user should enter the following data Ricardo Software December 2009 232 13 BOUNDARY amp INTERFACE CONDITION TYPES 13 10 INTERFACE CONDITIONS 5 Xx Interface_1 Interface Region Interface Name Interface_1 Neighbour Material I Higher Interface Report Roughness Height 0 Roughness Constant 0 5 Contact Resistance Contact Resistance Data Thickness 0 002 Conductivity 0 5 Heat Source 1e 07 Figure 13 15 Interface Region panel in R Desk O Thickness this is the interface wall thickness in m O Conductivity thermal conductivity of the interface wall material W mK O Heat Source this is a heat generation rate in W m7 The thin interface wall can be placed along an interface between the fluid and solid material or between solid and solid material
192. ION OF BOXES 4 attempts to split a volume failed 16 67 of all attempts 4 attempts gave incorrect number of volumes O attempts gave volumes with too low quality so the undo was applied Cell quality Number of correct boundary cells 6704 93 27 NO PROBLEMS with negative volumes NO PROBLEMS with gaps There were 484 small volumes they are deleted now 6 73 Number of cells which had to be deactivated 484 6 73 Total mesh volume 1 49073e 09 u3 6 40121e 10 u3 2648 inner cells 8 50607e 10 u3 6704 boundary cells If there were problems with generation of a cell warning appears here Problems like this are nearly always linked to topological problems on the input surface See section 4 7 which summarizes how to detect the problematic IJK spot 13 WRITING THE MESH FILE Finally the auxiliary files aux are read the output arrays are assembled and the output gridfile GRD is written down Then the aux files are deleted 4 6 Generation of Boxes This section describes how the mesh generator creates boxes in which future cells are constructed It includes an explanation for the user to understand the different options available in the input file The process is illustrated in Fig 4 2 to Fig 4 5 in 2 D though it actually operates in 3 D 4 6 1 Box Generation Procedure The starting point is the surface definition and the set of mesh line positions defined in Phase 1 see Figure 4 2 VME
193. IS configuration file name vectis cfg file can be placed in the working directory the users home directory UNIX LINUX only or the VECTIS config directory and is read at program start Configuration file for Vectis fpath to logo files png format LOGO_FOR_WHITE_BACKGROUND path to logo file LOGO_FOR_BLACK_BACKGROUND path to logo file it it it phasel and phase6 view options Ricardo Software December 2009 76 3 GEOMETRY MODEL CANVAS_LOW_REFRE phasel view options PHASE1_FAST_RENDER ON PHASE1 SURFACES ON O SH_RATE ON ON OFF ON OFF N OFF PHASE1_SURFACES_FLAT OFF ON OFF PHASE1_NUMBER_CELLS O PHASE1_3D_SUBDIVIDE O PHASE JK_BLOCK SOLO PHASE UTLINES MULTI PHASE GHLIGHT_PARTS S S PHASE1_SURFACES_FLIP N 3 OFF ON OFF N ON OFF N ON OFF OFF SOLO ALL OFF MONO MULTI ON ON OFF ON ON OFF PHASE O H PHASE1_HIGHLIGHT_HOLES H H PHASE PHASE1 MESH _VIEW_FACES DARIES BOUNDARI PHASE1_ MESH VIEW BOUN Ricardo Software December 2009 ON ON OFF GHLIGHT_SHARP_EDGES ON ON OFF GHLIGHT_HOLE_NODES OFF ON OFF ES DOMAI 3 19 THE VECTIS CFG FILE NS of MESHING 4 1 Introduction VECTIS MAX solver can read several formats of grids generated in
194. Its Properties s ses es usadia a mit ea a ai m ee 138 8 6 1 Setting a multiphase mixture a ca ca dce ka oi ea e k a ee 139 8 6 2 Setting a phase and its properties lt ssw c seee aci aca ue a a a a a e e 139 9 MODELLING TURBULENCE 148 Oil Introducir i a e ee h ahl 148 9 2 Overview of Turbulence Models for RANS o e 00200 eas 149 9 2 1 Eddy viscosity formulation 46 6 64 66 See eee ee eee Bee 150 9 3 Linear Two Equationk Models soco 2244 omia ba oe Ped es be Pewee eae dou 152 9 3 1 Standard k model sis asi oon eg Wo Go wee Be we ee ew 153 O32 RNG R Cmodel o roe ote he a eo we De OR oe eB eii 154 9 3 3 Standard k model with the realisable time scale bound TSB 154 O34 Thek model cociicients 60 5 40 ook Geb we wa ewe we ee PE a ed 154 Oa Near Wall Modeling cios da ee ER a a a Oe a 154 9 4 1 Background logarithmiclaw s sa sarei 0 20 02 pe eee ee ee es 155 9 42 standard Wall Tuncuions s 1 sarene dene Rew ee ee A wo ee a 157 O43 Scalable wall TUNCLONS Ss 6 pana ei gis ep ards Be A ee ee BO ae de 159 9 44 Enhanced unified wall functions 2 160 9 5 Low Reynolds number modelling e e 162 9 6 Inherent Limitations of K Models ovas haha rara A bebe eae 163 9 7 Setting a Turbulence Models ccs esis ek ew a a a a A a 165 10 MODELLING SINGLE PHASE FLOWS 166 10 1 Governing Equations s s e w enoda eae AE ord A ees 166 10 2 E
195. LOW 19 2 9 Mesh Generation In order to generate the computational mesh only the tri and mesh files are needed The mesher vmesh can either be started from the command line or within R desk In this case R desk will be used Start the R Desk GUI either by typing rdesk at the command prompt or on Windows through the start menu Start gt Programs gt Ricardo gt R Desk By default R Desk will open ina VIEWER project which allows general visualisation of results for numerous Ricardo Software products however in this case we require a VECTIS project to give access to the mesh generation and solver setup menus There close the current default VIEWER project File gt Close Project and open a new project File gt Open gt Project The new project dialog box will appear Here select VECTIS from the combo box and name your project port axl Type K vecris E RTherm Name port ER Vie OK Cancel Figure 19 28 The New Project dialog box Note The default project type can be changed in the preference panel Options gt Preferences gt Miscellaneous Alternatively from the command line the project type can be specified using the project switch for example rdesk project vectis The geometry file can be opened and visualised File gt Open gt File then browse to the saved port tri file The geometry should open and then be
196. La LZ oO COO This example defines 3 plane cuts through a mesh If the number of partitions resulting from the user defined planes was less than the number of partitions required then further decomposition is performed on each user defined partition Additional automatic checks are performed by vpre These include checking the partition and material interface boundaries do not coincide in the case of multi material grid files If they do the partition boundaries are adjusted This is done by looping over partitions in order of size so as to minimize the degree of load imbalancing and expanding these where necessary across material interfaces In order to improve pre conditioning in vsolve each partition in decreasing order of neighbourli ness is coloured according to rule no two colours are allowed to be neighbours The partition numbers are then renumbered according to this colouring For example Ricardo Software December 2009 107 5 READING amp MANIPULATING MESHES 5 3 MESH JOINING Parallel interface Material interface Figure 5 1 Picture of 2 materials partitions with slight coincidence Green arrow shows how partition interface is moved from smaller 1 to larger partition 2 Figure 5 2 Example of colouring for 7 partitions 5 3 Mesh Joining This section describes the way in which a multi material grid file can be created from coalescing separate single material grid files
197. Launder and N Sandham eds Closure Strategies for Turbulent and Transitional Flows Cambridge University Press Cam bridge 9 2 1 George A and Liu J 1981 Computer Solution of Large Sparse Positive Definite Matrices Prentice Hal 14 2 6 George W K 2007 Is There a Universal Log Law for Turbulent Wall Bounded Flows Phil Trans Royal Soc A vol 365 pp 789 806 Hanjalic K 1970 Two Dimensional Flow in an Axisymmetric Channel ph D Thesis University of London 92 1 Hanjalic K 1994 Advanced Turbulence Closure Models A View on the Current Status and Future Prospects Int J Heat amp Fluid Flow vol 15 pp 178 203 9 3 1 9 6 Hanjalic K 2005 Will RANS Survive LES A View of Perspectives Journal of Fluids Engineering vol 127 pp 831 839 9 1 Hanjalic K and Launder B E 1972 A Reynolds Stress Model of Turbulence and Its Application to Thin Shear Flows Journal of Fluid Mechanics vol 52 pp 609 638 9 2 1 Harlow F and Nakayama 1967 Turbulence Transport Equations Physics of Fluids vol 10 pp 2323 2332 F2 Hinze J O 1975 Turbulence 2nd ed McGraw Hill New York 6 3 1 ORicardo Software December 2009 446 19 TUTORIALS 19 4 IMPORTING THIRD PARTY MESHES Hirt C and Nicholls B 1981 Volume of Fluid VOF Method for Dynamics of Free Boundaries J Comput Phys vol 39 pp 201 221 12 3 Huang P Coleman G and Bradshaw P 1995 Compressible Turbulent Channel Flows
198. MAX object name subroutine get_name_list Iame get Input Output iget integer Iname cha pointer iget_dom iget_mat iget_phase iget_specs iget_ps iget_breg iget_ireg list of domain names list of material names list of phase names list of species names list of passive scalar names list of boundary regions list of interface regions Table 18 22 Subroutine get_name_list Get a list of names Iname for each VECTIS MAX object 18 3 2 15 get_id This UAR is used to retrieve the solver object id see Table 18 23 In subroutine get_name we showed how to obtain a solver object name by providing its id This section deals with the case whereby we specify the solver object name to obtain the solver object id To obtain the id of a phase with name water do the following character len ph_name call get_id ph_name idt iget_phase Ricardo Software December 2009 330 18 USER PROGRAMMING Now idt contains the id of phase with name water 18 3 ACCESSING SOLVER VARIABLES In order to make use of options get_eq iget_ph_pro iget_sp_pro and iget_ps_pro refer to sub routine get_name for relevant names subroutine get_id cname idt iget Input Output cname cha iget integer idt integer domain name iget_dom domain id material name iget_mat material id phase name iget_phase phase id species name i
199. METRY WRAPPING fi Geometry Wrapper Input Parameters X m Wrapper Input Parameters Panel File m Wrapper Parameters Feature Resolution Size m Jo 000606063 Decimation Threshold Angle degs 2a Decimation Threshold Distance rm Jo 000303032 Decimation Deviation Distance m 6 060638 005 Surface Offset Magnitude m OO Surface Thickness m fo Apply Distant Node Smoothing Yes X Approximation Assessment No v Geometry Wrapper Input Panel Wrapper input parameters can be specified either in the panel or can be read from a file by select ing the File radio button When the File radio button is selected the wrapper parameter input filename can be specified using the input parameter file browser Refer to the Geometry Wrapper Input File section and Appendix D for more information regarding the wrapper parameter file The main input parameter is the Feature Resolution Size It corresponds to the size of minimal base mesh surrounding the initial geometry The lower this parameter the higher the resolution that can be achieved The default value in the menu corresponds to 128 divisions or smallest Cartesian cells in the largest direction of the model The initial geometry can be any set of triangles To simplify the geometry correctly holes in the initial geometry should be no larger than Feature Resolution Size Larger holes may cause leaks of the Cartesian mesh inside the
200. Note that optional arguments that correspond to fi_c fi_b fi_ ui and fi_li need to be declared as pointers before using them similar to g_temperature used in the example above Ricardo Software December 2009 334 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES Momentum variables var_name Meaning Availability cell bnd uif lif face 00 velocity velocity x x x x x density density x x x x x eff_viscosity effective viscosity Uy L X X x x lam_viscosity laminar viscosity x mass_flow_rate mass flow rate x x gas_constant gas constant x drho_dp derivative of density over pressure x constant temperature Pressure variables pressure pressure x x x x x tot_pressure total pressure X tot_temperature total temperature x sat_pressure saturation pressure x Turbulence variables turb_energy turbulent kinetic energy x x x X x dissipation dissipation rate of turbulent kinetic x x x x x energy Energy variables temperature temperature X X X X X tot_enth_energy total enthalpy or total energy x x x X x lam_conductivity laminar conductivity x x specific_heat specific heat x x heat_flux heat flux x x nuc_heat_flux boiling heat flux x x sat_temperature saturation temperature X latent_heat latent heat of evaporation x surface_tension surface tension xX Volume fraction variables phase_vol_frac phase volume fraction x x x X x
201. Nucleate boiling model Two nucleate boiling models VECTIS3 and RPI designed to predict boiling possibly taking place in engine cooling passages are provided Both models belong to the homogeneous mixture family Surface to Surface Radiation The surface to surface radiation module applicable to fluid domains and containing the radprep and radvfm sub modules is directly coupled to the VECTIS MAX solution of the energy equation User Programming A powerful set of User Programmable Routines UPR Fortran 95 2003 have been implemented that allow direct access into the solver kernel data structure s These functions allow the user to perform many tasks such as setting up boundary conditions initial conditions material properties or adding source terms Ricardo Software December 2009 F 2 INTRODUCTION 2 1 MAIN FEATURES AND CAPABILITIES Other User Tools The user can monitor and or interact with the solution process in a number of ways Monitors Flow variable values can be monitored at any number of user specified locations Alpha numerical reports These reports containing typically variable average values are avail able for fluid solid domains boundary and interface regions and for user defined arbitrary sur face O Live update The Live Update from R Desk can be used to monitor the solution conver gence The equations residuals or averaged variables over boundary regions or over domains are
202. Numerical Method for Coupled Fluid Flow Heat Transfer and Stress Analysis Using Unstructured Moving Meshes with Cells of Arbitrary Topology Comp Methods Appl Mech Eng vol 125 pp 235 255 14 1 3 Drew D 1992 Analytical Modelling of Two Phase Flow Boiling Heat Transfer Elsevier Science Publ 12 2 Durbin P 1991 Near Wall Turbulence Closure Modeling without Dumping Functions Theoretical and Computational Fluid Dynamics vol 3 no 1 pp 1 13 9 2 1 9 3 9 4 1 9 5 Durbin P A 1996 On the k e Stagnation Point Anomaly Int J Heat and Fluid Flow vol 17 no 1 pp 89 90 9 3 3 9 5 9 5 Durbin P A 2009 Limiters and Wall Treatments in Applied Turbulence Modelling Fluid Dynamics Re search vol 41 pp 1 18 9 5 Esch T and Menter F R 2003 Heat Transfer Predictions Based on Two Equation Turbulence Models with Advanced Wall Treatment in Y K Hanjalic and M Tummers eds Turbulence Heat and Mass Transfer 4 Begell House Inc pp 663 640 9 4 3 Ferziger J and Peric M 1997 Computational Methods for Fluid Dynamics Springer Berlin 14 1 3 14 1 5 1 14 2 2 14 2 3 Gaskell P H and Lau A K 1988 Curvature Compensated Convective Transport SMART A New Bound edness Preserving Transport Algorithm International Journal for Numerical Methods in Fluids vol 8 pp 617 641 14 1 3 1 14 1 3 1 14 1 3 1 Gatski T B and Rumsey C L 2002 Linear and Nonlinear Eddy Viscosity Models in B
203. OLVER VARIABLES integer iwp pointer nc_faces call get_grid_connect n_cell_faces nc_faces For the given cell ic nc_faces 1 ic returns a number of faces whose normal vectors point out of the cell upper cell faces while nc_faces 2 ic gives a number of faces with normals pointing into the cell lower cell faces To get a list of faces that enclose each cell then integer iwp pointer cell_faces integer iwp pointer ISA call get_grid_connect l_cell_faces cell_faces call get_grid_connect isa_cell_faces isa_cf For the cell ic with jc 1 nc_faces 1 ic upper faces their indices are obtained as j cell_faces 1 isa_cf 1 ic jc Similarly lower cell faces are given as J cell_faces 2 isa_cf 2 ic jc where jc 1 nc_faces 2 ic Arrays isa_cf 1 ic and isa_cf 2 ic return starting addresses for upper and lower faces in array cell_faces respectively subroutine get_grid_connect var_name integer pointer ival Input Output var_name cha ival integer pointer L_bndf_cells list of cells adjacent to a boundary face ival 1 n_bnd_faces n_face_verts number of face vertices for each face ival 1 n_internal_face isa_face_verts initial starting address for face vertices ival 1 n_internal_face l_face_verts list of face vertices for all internal faces ival 1 nafve n_bndf_verts number of boundary face vertices ival 1 n_bnd_faces isa_bndf_verts initial add
204. PING OPERATIONS 3 11 Triangle Chopping Operations Chop Area Button EA This operation is a useful feature for processing large models in that it enables the user to select a portion of a model for viewing thus making the response time of dynamic manipulation faster The increase in manipulation speed gained by using Chop Area can be dramatic and its use 1s recommended for all but the smallest of models The perimeter of a polygon may be drawn about the model by clicking the left mouse button Two clicks in the same place will end the drawing process and close the polygon The model will then be chopped so that only those triangles that lie completely inside the polygon are shown The model can not be moved during chopping Pressing the escape key during chopping cancels the function IJK Chop Button EH Selecting this command pops up the IJK Chop panel UK Chop X Model IJK Chop Mesh Input File meshsetup e Import Mesh Set up File Min Max Limit E BB jie Y O A el E tel OK Cancel y Ve This panel can be used to import an IJK block mesh set up file and specify the IJK block extents to which the model should be chopped This makes all the triangles visible that have at least one node within the specified IJK block slice Ricardo Software December 2009 34 3 GEOMETRY 3 11 TRIANGLE CHOPPING OPERATIONS Chop Selected Boundary Button ES This makes all the triangles invisible on
205. PU 1 Us In Eyb 9 31 In Eyp or UP tog _ pKuUg 32 EY aa If the wall moves with velocity U then the velocity Up in all the above equations should be replaced by the difference of velocities parallel to the wall surface AUp Aup 0 0p 0 0p 5 5 9 3 where Oy Up fy denotes the magnitude of velocity normal to the wall and 7 is the outside wall surface normal Using similarities between the momentum and energy transfer and neglecting the viscous heating the normalised temperature given as Ty Tp PcpUk dw can be also approximated with a log law expression However it is commonly expressed in terms of the normalised velocity U r 9 34 Pr Tag Pr ur 22 035 The on going debate cf George 2007 questions existence of the universal log law for wall bounded flows ORicardo Software December 2009 157 9 MODELLING TURBULENCE 9 4 NEAR WALL MODELLING The function Py represents the viscous thermal resistance and according to Jayatillaka 1969 it reads i Pr Py 9 24 1 7 The distribution of species c f y in the fully turbulent layer is described analogously to the temperature P 0 28exp 0 0077 9 36 t de Gee 9 37 Jiw Ci log SC y P 9 38 where J is the diffusion flux of species i at the wall the function P Sc Sc is calculated from Equation 9 36 by replacing molecular and turbulent Prandtl n
206. R Desk distance and angle measuring tools can be used After loading an RTH file and creating a transformation matrix all Set Names will appear in the drop down list of Extracted Sets Figure 13 13 right At this stage Boundary ID corresponding to each Set Name will not be assigned In the next step the user can revisit the first considered wall boundary This time RTHERM Set Ricardo Software December 2009 229 13 BOUNDARY amp INTERFACE CONDITION TYPES 13 9 TIME DEPENDENT DATA RTHERM eE RTHERM a Thermal Conditions File Thermal Conditions File RTH File UNSET__ Browse Update from File RTH File _IL4 RTH Browse Update from File Show Preview Show Preview Transformation Matrix Transformation Matrix 1 0 0 0 1 0 0 0 0 1 0 0 0 1 0 0 0 0 1 0 0 0 1 0 0 0 0 1 0 04699 0 00650 0 08850 1 Extracted Sets Extracted Sets Boundary ID w Set Name Boundary ID v Set Name E 18 port CYL_1 VAL_11 27 cylinder CYL_1 1D_headGasFa 28 cylinder CYL_2 1D_headGasFa 29 cylinder CYL_3 1D_headGasFa 30 cylinder CYL_4 1D_headGasFa f 34 cylinder CYL_1 1D_headGasFa 35 cylinder CYL_2 1D_headGasPa Figure 13 13 RTHERM panel in R Desk after opening left and after editing right Selection list will contain all R Therm sets imported from the RTH file After selecting a correct R Therm set to be associated with the current wall boundary the Region ID number will be ass
207. Ra q RecT Rpoo 15 21 Non linear solver The Newton Raphson method is used as the non linear solver which uses the following equation system This method is based on incomplete Taylor s expansion This expansion is used in the equation 36 N d F 15 22 N di Ad F 15 23 l _ ON N d Ad N d glaad 15 24 oN Jd Ad F N d 15 25 Kr d Ad F N d 15 26 with the solution updated using dl d Ad 15 27 Conduction The surface conduction model allows for the radiation solver to calculate the conduction between neighbouring super patches based on their defined thickness and thermal properties Each bound ary set and therefore super patch has a defined thickness so that along with the super patch area the super patch mass is then known The planar conduction is solved as a 2 dimensional heat transfer problem with a uniform thickness for each boundary which implies that the surface conduction represents thin components such as heat shields Surface Tangential Conduction Matrix The conduction matrix can be expressed as nb TaT I Ap KT fk dl ki T T 15 28 r r al fa i Ajj is the absolue distance of line C Dp rc p rij is the absolue distance of line AB rag ki is the conductivity T T are the temperatures i represents the i th super patch j represents the j th neighbour super patch An example of the surface conduction model results for a si
208. Routines The set of user programmable routines UPR subroutine upr_properties subroutine upr_generic subroutine upr_init subroutine upr_bnd_cond subroutine upr_sources all have a predefined interface argument list which will be detailed below The calling sequence for these routines was previously illustrated in Section 18 2 Generally each routine is called for each phase species and in turn for for each changeable variable property field etc In order to restrict the user supplied routine to apply to specific phases materials species conditional blocks if case blocks are normally used to surround variable changing code 18 4 1 User properties routine subroutine upr_properties pro_name mat iph isp icell icel2 jondl jbnd2 jiul jiu2 jill jil2 fi_c fi_b fi_ui fi_li Arguments Description pro_name char property name see Table 18 36 mat integer material id iph gt integer phase id isp gt integer species id icell integer starting cell index icel2 integer ending cell index jon integer starting boundary face index jbn2 integer ending boundary face index jiul integer starting upper interface index jiu2 integer ending upper interface index jill integer starting lower interface index jil2 integer ending lower interface index fi_cl real optional scalar variable cell values fi_b g
209. SH calculates which global boxes have any volume inside the model see Figure Figure 4 2 Start point of the mesh generation 4 3 As shown in Figure 4 4 the program then subdivides refines boxes intersected by the input Ricardo Software December 2009 87 4 MESHING 4 6 GENERATION OF BOXES Figure 4 4 Refinement of global boxes surface to produce an accurate fit to the original shape Figure 4 5 visualizes the final generated grid The boundary boxes are cut by the surface and polygons of boundary faces are generated here in 2D the boundary faces appear to be segments Figure 4 5 Final grid all polygons are generated Ricardo Software December 2009 88 4 MESHING 4 6 GENERATION OF BOXES 4 6 2 Parameters controlling generation of boxes There are three ways how to affect refinement of global boxes 1 Global refinement depth refers to how many times a cell can be split into two This option can be set in the input ascii meshfile see description of REFINEMENT_DEPTH in section 4 3 Figure 4 6 shows the panel in Phasel which can be used for setting the global refinement depth Refinement depth RD defines the maximum possible division of global box to subcells which is 289x282 Therefore a refinement depth of 1 allows boxes to be split into at most 2x2x2 refined cells and a depth of 2 allows boxes to be divided into at most 4x4x4 refined cells A depth of 0 does not permit any refinement Boxes with diff
210. Section 14 2 6 Ricardo Software December 2009 240 14 NUMERICAL SOLUTION 14 1 FINITE VOLUME DISCRETISATION Equations Solver EJ Momentum Mass Energy Species Turbulent Energy Turbulent Dissipation Turbulent Viscosity Volume Fraction Potential Passive Scalar ISIA KEKR Momentum Equations MINMOD SMART Convective Scheme Blending Factor _ User Defined Sources Momentum Solver Solver Type e Bi cg Stabilized y Symmetric Conjugate Gradient P Incomplete LU level0 Incomplete LU level1 Jacoby Modified LU level0 Modified Stone s SIP Number Of Inner Solver Iterations Solver Tolerance Preconditioning Incomplete LU level0 Number Of Preconditioning Iterations ao ILU Preconditioning Options For Parallel Runs Figure 14 5 R Desk setup for convective schemes and linear solvers Mixed Global Ricardo Software December 2009 24 14 NUMERICAL SOLUTION 14 1 FINITE VOLUME DISCRETISATION 14 1 4 Gradient computation Two different methods are identified for computing the cell gradient Gauss based and least square based Gradient based on the Gauss divergence theorem is calculated as ae Vop Y OA 14 27 Vp jal Cell gradient can also be calculated by using linear least square minimisation method Barth Figure 14 6 Polyhedral control volume around the cell face j 19
211. Solver Setup tree opens on the left hand side and Solver Setup Input should open on the right File Edit View Options Window Help oj xi D 70 gt pa DE olen xBOB tSsenm Bae gt 2 Solver Setup Tree ax Solver Setup E Global_domain_1 Global Domain Restart Control Timebase put id_1 Fluid Domain Monitoring Points Ou E Flui Materials And Boundaries From Mesh File Algorithm Turbulence Model Extract E Phase_1 Fluid Phase Initial Condition IT Show Mesh Preview Postprocessing Output tions amp Solver Equal EX Bnd_Reg_1 Inlet Given Velocity Boundary Phase_1 Boundary Phase Interface Regions Report Regions Solid Domains camara gt AS Figure 19 36 Default panel layout for VECTIS Project The structure of the Solver Setup Tree reflects the general structure of calculation Global domain containing general options for the calculation timebase etc Fluid domains containing data for phases species and boundaries Solid domains containing materials and boundaries The contents of the Solver Setup Input panel will change dynamically corresponding to the selected entry in the solver setup tree Click on the first entry in the tree Solver Setup The Solver Setup Input will show options to import the computational grid Ricardo Software Decem
212. Values Ss Value 1 15 Viscosity Option Constant Values Value 1 983e 05 Inactive Constant Values Mixture Pexponenio n Polynomial f T Power Law f T Sutherland Law f T User Subroutine Inviscid Flow Piecewise Linear f T Figure 8 4 R Desk setup Setting properties of the single component phase O piecewise linear functions and O piecewise polynomial functions Piecewise linear functions A material property can be described using piecewise linear functions where each function represents a linear variation of a property and in general form is defined as T On Pri n en T Tn 8 47 n 1 n Figure 8 5 represents a material property in terms of piecewise linear functions The number of points in this example is 3 and the number of intervals is 2 If the temperature value T is in interval T Tn 1 then the piecewise linear function is that given in Equation 8 47 If the temperature value T is in interval T 1 742 then Equation 8 47 takes the form On 2 On 1 T T 8 48 T2 D 1 8 48 0 T n 1 T Ricardo Software December 2009 14 8 MODELLING CONTINUA 8 6 SELECTING CONTINUUM AND ITS PROPERTIES On Th Tny Iny TP Figure 8 5 A material property in terms of piecewise linear amp polynomial functions Viscosity Option Piecewise Linear f T 4 Number of Points 2 l Point 1 Temperature 273 15 Viscosity
213. Weakly Compressible Gas Density may vary with temperature but the change of the com pressibility coefficient is relatively small Pressure changes relative to the reference pressure are small In terms of the Mach number weakly compressible fluid is broadly defined for Ma lt 0 3 Fully Compressible Subsonic 0 3 lt Ma lt 1 Fully Compressible Supersonic Ma gt 1 Single component phase The lower part of the Fluid Phase Setup panel provides several sub panels to specify all required thermo physical properties namely Density Viscosity Conductivity Specific Heat Molecular Weight and Thermal Expansion Coefficient The sub panels for setting the density and viscosity are depicted in Figure 8 4 top For each property there is the Option list box which displays calculation options available for the considered property For example left click on the Viscosity Option will display the viscosity calculation options given in Figure 8 4 bottom Constant property values are the default options and these values are entered in the corresponding Value line box For properties dependent on the temperature 7 various functions f T are available Property calculation options are summarised in Table 8 1 Piecewise functions A property can be specified in terms of the following piecewise functions Ricardo Software December 2009 140 8 MODELLING CONTINUA 8 6 SELECTING CONTINUUM AND ITS PROPERTIES Density Option Constant
214. Y 3 17 MESH SET UP VIEW OPTIONS 3 17 2 Control Mesh Setup Suggestions The global mesh is used to define the required base cell sizes through out the model An example control mesh is shown below An example control mesh defining the global cell size The green lines are the global cell divisions and the red lines are fixed control lines Note that two red lines must be completely outside the model extents in each direction The reproduction of the surface detail is achieved by either ensuring that the global cells are small enough to capture the local geometry or by the use of surface cell refinement Local IJK refine ment regions can be setup in the detailed core flow locations to improve the accuracy of the CFD calculation The following suggestions are provided to improve the mesh setup Cell Connectivity Cell connectivity is concerned with how the cells are linked to each other between refinement regions The best practice is to ensure that the cell connectivity never exceeds 2 cells connected to Ricardo Software December 2009 71 3 GEOMETRY 3 17 MESH SET UP VIEW OPTIONS 1 cell as shown by figure one below 2 cells connected to 1 cell gives o connectivity 2 1 Cell connectivity Ensuring that the cell connectivity never exceeds 2 1 reduces the inaccuracy due to numerical error and artificial viscosity This is best achieved when using the IJK refinement by having overlap regions of ideally at least t
215. _phases mixture_opts calculation options for mixture of species 11 1 n_phases ise_specs starting species ending species 11 1 n_fluid_phases 12 1 n_fluid_phases ise_ps starting passive scalar ending passive scalar i1 1 n_fluid_phases 12 1 n_fluid_phases ise_phase_bnd_reg index of starting boundary region index of ending boundary region 11 1 n_phases 12 1 n_phases isa_phase_bnd_reg starting allocation addresses for phase variables at boundary regions 11 1 n_phases ise_phase_interf_ index of starting interface region index of ending interface region reg 11 1 n_phases 12 1 n_phases isa_phase_interf_ starting allocation addresses for reg phase variables at interface regions 11 1 n_phases Table 18 11 Subroutine get_phase to get values for variables defined in the access group iacc_ phase mainly start end phase objects and phase properties Ricardo Software December 2009 312 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES 18 3 2 5 get _species This UAR is used to retrieve information related to species mainly starting and ending boundary or interface regions see Table 18 12 Species are constituents of fluid phases They are indexed respecting the order in which they appear in the ordered set of fluid phases In order to store the region wise species values array addresses or species regions are assigned to the first species and then subsequently for o
216. a ee Ba Oe ee 398 Ricardo Software December 2009 xxii LIST OF FIGURES LIST OF FIGURES 19 38Importing the mesh into R Desk Solver Setup o o e 398 19 39Grid file imported into R De sk aer a ds Ri ed we a 399 19 40Solver setup input tree after grid import 00 20 00 eee eee eee 400 19 41 The Global Domam Panel s s s a 4 ac ea a ees area OA a a ee a ed 400 19 42 The Timebase Panel 2 3 44 4 4 6 0 pa rra da oa ESE RS Pe eS a 400 19 43 The Solution Control Panel si cs aaie i oua a a a we Rw 401 19 44 The Pluid Domain Panel ss s e each o a a a A ee a 404 19 45The Monitoring pint panel 2 e gaa ea eee 404 19 46Fluid Phase Options crasas cada a ee a i a 405 19 47 Fluid Phase Initial Conditions s socs moens a a E ed a be a SS 406 19 48Setting up the boundary regions ee iaa 407 19 49Cut the model in the location of the arbitrary surface is required 408 19 50 Delete the marked area be gos m ane do e are ee OR ee a yy 408 19 51 Model cut at location of arbitrary surface 2 a 409 19 52Using the cap hole tool to fill the Open edge sc s ss mesoa o o o 409 19 53Chop outihe required SUITE eme ek ee RRA AA we a 410 19 54 Einal Arbitrary Surface c s s 2 4 2 e at a Gr Re ee Be 410 19 SS Add TENOR ROM oo oid ke a Bee A A A Bl ae ee 411 19 56Specity triangle filename s so uoa 4444 48 24 ds 44 ee bee He A De oe wee eae 411 19 S Launch Y pre
217. ae for the solution variables at arbitrary placed near wall cells Three types of wall functions are available in VECTIS MAX O Standard or conventional O Scalable and Unified or enhanced wall functions To understand them better some physical aspects of the near wall region including a logarithmic law of the wall are introduced 9 4 1 Background logarithmic law The near wall region is characterised by high velocity gradients and dominant molecular effects In essence the wall modifies the mean flow and the turbulence in its vicinity through viscous as well as non viscous effects Viscous effects reduce the velocity fluctuations parallel to the wall Non viscous effects are due to the kinematic blocking of the velocity fluctuations normal to the wall cf Durbin 1991 Also the fluctuating pressure field is modified by the presence of the wall The friction velocity Uz and velocity scale Uz defined as Ue ft qe 9 24 Ty 1s the wall shear stress are commonly used as wall normalisation variables Thus the velocity parallel to the wall a non moving wall is assumed and the normal distance from the wall denoted as y can be normalised either by Uz or Ug Normalisation by Uz gives the wall or plus units U _ Pury Ut q u 9 25 Ricardo Software December 2009 55 9 MODELLING TURBULENCE 9 4 NEAR WALL MODELLING while normalisation with the k based velocity scale U gives
218. ain times by clicking on the Add button and entering the time and corresponding filename The filename is exactly as specified here no project name or run number are added for timed restart files See figure Fig 17 3 Restart Writing Restart File Frequency 1000 Time Filename 10 flowBlock Delete 20 flowBlock Delete Add Figure 17 3 GroupBox for Restart Writing 17 4 Fluid Domain The fluiddomain panel defines domain material name material id and multiphase modelling type see Figure 8 1 Initial values can be set as Uniform User Defined or Potential Flow For User Defined the user must write and compile a user function see Section 18 4 2 The Poten tial Flow option uses the solution from a simplified model irrotational flow assumed which is initially run prior to the main solver This is normally very quick If necessary the user can scale down between 0 1 the estimated inflow velocities via the Potential Flow Scale Factor LineEdit Initial Values Designed By User Potential Flow Uniform Potential Flow Scale Factor 1 Figure 17 4 GroupBox for Initial Values The Reference Pressure Location GroupBox is for setting the reference pressure cell or more conveniently the xyz coordinate See below Other reference quantities that can be set for this fluid domain are shown in Fig 17 6 Setting of Body Force options has been described in the section dealing with modelling bu
219. ained in the section dealing with Setting a phase and its properties with the help of Figure 8 3 Figure 10 3 shows the selection of multi component phase under the Mixture of Species Option If a multi component phase is selected then the Solver Setup Tree will initially show two Fluid Species nodes 10 3 3 Incompressible and compressible flow The fluid compressibility is discussed in the section Equation of state amp thermodynamic properties its mathematical definition is given by Equation 8 13 The above Figure 10 3 also depicts the list of Compressibility options when the Gas is chosen as the Phase Type If the fluid phase is liquid then the density changes caused by pressure changes in Equation 8 13 are negligible and the fluid is considered as incompressible The selection of the phase type gas or liquid and a compressibility option determines the mod elling of equation of state Ricardo Software December 2009 17 10 MODELLING SINGLE PHASE FLOWS Phase_1 Fluid Phase Phase Name Mixture Of Species Option Phase Property Filename Phase ID Phase Type e Gas Compressibility Solve All Species Define Passive Scalar Phase_1 Multi Component Phase none Liquid Incompressible Weakly Compressible Fully Compressible Sub sonic Fully Compressible Super sonic 10 3 MODELLING FLUID FLOW Y Figure 10 3 R Desk setup Setting a type of a fluid phase a
220. al Note that the gen eral fluid material can be either single phase or multi phase fluid and each fluid phase can have more components species Fluid Sub Domain This type is a part of the fluid domain It can be used to activate certain features such as modelling of porous media heat exchangers or fans Modelling of rotational effects is not yet supported Mathematically domains are defined by the corresponding volume meshes while the domain boundaries and interfaces are described by the boundary surface meshes and the interface meshes respectively Thus the domain boundary is a collection of boundary faces and the interface is a collection of internal faces which are shared by two cells belonging to different materials Prior to meshing the CAD geometry defines the domain boundaries and interfaces The domain boundary can not be shared between two material domains In general the actual solution domain for the given transport equation consists of one or more material domains As the solver operates on the non overlapping cells it is important that the domain and sub domain interfaces are represented by the union of internal cell faces which are defined by identical vertices common to the neighbouring domain sub domain cells Such domain interfaces and associated meshes will be referred as conformal The non conformal meshes can be converted into conformal ones by using arbitrary grid interface tool Simple examples of multi domain
221. al Domain in the solver setup tree Once a file is selected the data available to be plotted is shown in the Value window The different data can be dragged onto a xy plot canvas The xy plot can then be further modified using the XY plot manager 19 1 12 Post processing Once the simulation converges it will stop and write restart and post processing files Open a 3D canvas Open the post file in R Desk using the File gt Open gt File menu option It should appear in the plot tree The plots from the tree can be dragged onto a 3D canvas The attributes of the plot can be modified Different data can be shown on the plot by selecting it in the data panel ORicardo Software December 2009 384 19 TUTORIALS 19 1 BASIC TUTORIAL 19 1 12 1 2D plane A 2D slice can be made of a plot Right click the plot in the tree and select slice The slice is then defined in the slice definition panel The slice can be previewed on the plot if it is present in a 3D canvas Once the slice is in the correct position click apply gt Open an additional 3D canvas These can be tiled using the window gt Tile slice plot into the second canvas options Drag the The data plotted on the slice can be changed using the data panel 19 1 13 Potential Flow Solver We will now examine the potential flow solver option This option will calculate the initial flow field based on the boundary condition data specified Open the
222. alculation options for variables describing fluid solid properties species_pro_opts 0 nprosp 0 n_species andpassca_pro_opts 0 nprops 0 n_ ps The physical property can be inactive ipro_inact 0 constant ipro_const 1 or solu tion dependent i e a function of temperature and or pressure Note that properties of the multi component phase are usually solution dependent even for the constant properties of constituent species For such properties the ipro_mix 2 calculation option should be used and phase prop erties will be calculated from Equation 8 39 However if it is desirable to specify the composition independent property not a function of mass fractions of the multicomponent fluid this property will have the calculation option different from ipro_mix The ideal gas model identified by ipro_igas 3 option can be used in many practical situations to closely approximate properties of compressible gas flow Typically it is used to calculate the density and specific heat The temperature dependence for any property can be expressed by a general polynomial function option ipro_poly 4 see Equation 8 26 For some properties such as the molecular viscosity it is sometimes more appropriate to use the power law form ipro_power 5 see Equation 8 4 Additional option available for the calculation of viscosity and thermal conductivity in terms of temperature is Sutherland formula for gases ip
223. along the Direction Vector X axes 11 3 2 Catalytic converter The flow through the catalytic converter is assumed to be in the direction of the first principal axes i e in the Direction Vector X The flow direction is enforced in the solver by using large resistance values in other two principal directions Also the resistance is modelled according to Darcy s viscous model where the pressure drop depends linearly on the local velocity magnitude Thus the inertial resistance is neglected The viscous resistance tensor Ry is then derived from the Catalyst Data depicted in Figure 11 2 This data include specification of Internal Area per Channel Porosity 1 Porous Material Type Catalyst el Principal Axes Direction Vector X X 0 998377 Y 0 031748 Z 0 047271 Catalyst Data Internal Area per Channel m 2 m 0 004 Number of Channels per Frontal Area 1 m 2 826000 Friction Factor 12 Channel Diameter m 0 001 Heat Exchange Mode None Figure 11 2 R Desk Sub domain panel User inputs for the catalyst porous model Number of Channels per Frontal Area Friction Factor and Channel Diameter Based on the above data the viscous resistance in the flow direction is calculated as ANefcll yDe where Ac Ne fe and De signify internal area per unit length of a single catalyst pore channel number of channels per frontal catalyst area non dimensional friction coefficient and hy
224. ame The defined triangles are mapped to the solver mesh Each triangle is successively divided along its longest edge until one of the following criteria are reached All nodes of a triangle are contained within the same cell All nodes of triangle are greater than or less that the mesh extents i e triangle is outside mesh Longest edge is less than the user specified edge tolerance The input file should now be complete for the analysis Save the file as port inp file gt save as 19 2 11 Grid Preparation Vpre prepares the computational grid for use in the solver It can be started either from a cmd window or from R desk In a command window type vpre port GRD to run the vpre mesh pre processor From R Desk the Launch Vectis Pre button can be used to start vpre Click on the button to open the Launch vpre dialog box Insert the name of the computational grid file in the mesh file box Ricardo Software December 2009 411 19 TUTORIALS 19 2 STEADY STATE PORT FLOW Figure 19 57 Figure 19 57 Launch Vpre vpre can also be used to re partition the mesh for parallel calculations 19 2 12 Running the solver Now that the inp file has been created and the grid file has been processed the solver can be run Again this can either be run from the cmd window or from R Desk From the cmd window type vsolve port inp From R Desk the Launch Vectis Solver button can be us
225. ameter iget_ph_pro fluid solid phase properties integer parameter iget_sp_pro species properties integer parameter iget_pS_pro passive scalar properties integer parameter Table 18 3 Various VECTIS MAX objects Post processing variables Table 18 4 provides the user with names of pre post processing vari ables Post processing variables Variable Description Type i_cell cells integer parameter i_face internal faces integer parameter i_bnd boundary faces integer parameter i_intp higher fluid solid interface integer parameter 1_intm lower fluid solid interface integer parameter Table 18 4 Names of post processing variables Ricardo Software December 2009 300 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES Other user accessible variables Other parameters available directly to the user are given in Table 18 5 This lists for example different boundary condition types global data identifiers related to transport equations and variables etc Some constants and numbers Variable Description Type small 1 0E 30_wph real parameter zero 0 0_wph real parameter one 1 0_wph real parameter Material id s Variable Description Type len_mat_name length of the material name 24 integer parameter Argument pointer for array Ibc ibct boundary condition type integer parameter Types of boundaries ibin
226. and line from which Phase 1 is launched will be opened If the file to be loaded is a VECTIS Triangle file of either format the choice of whether to replace any existing model with the new model or to merge the new with the old will be given Merge with existing model or replace existing model Merging or replacing 3 4 2 ASCII Triangle Files Once Phase 1 has determined that an ASCII Triangle file is to be imported the program will prompt for the location of the associated sets and types files P thang wil esting pode or opaca amis poda erce ASCII Triangle file import Either or both of these text fields may be left blank if required 3 4 3 VDA File Triangulation Phase 1 reads VDA FS files these can be read and written by most CAD systems The infor mation of interest to VECTIS in these files is the polynomial surfaces which may be trimmed or untrimmed The geometrical elements handled by Phase 1 are SURF Polynomial surfaces CONS Trim lines defined on a SURF element FACE Trimmed surfaces defined by a SURF and one or more CONS elements Ricardo Software December 2009 16 3 GEOMETRY 3 5 GENERAL VIEW OPTIONS The user is referred to the VDA standard documentation for details of how these elements are defined VECTIS converts these surfaces to a triangulated approximation with a user specified accuracy which ensures that all triangle vertices lie on the original surface and all triangle si
227. and initial conditions have to be specified Note that the above conservation equations also describe the multi phase flow i e they describe the flow of each un coupled phase For this interface conditions providing the mass momentum and energy balances at phase interfaces are required However due to a wide range of time and length scales associated with the turbulence the current computer power enables the full numerical solution Direct Numerical Simulation DNS of the closed instantaneous equations for very simplified flow cases A practical and well established alternative to DNS is the solution of the mean flow or averaged equations which are obtained from the instantaneous equations through some kind of the averaging or filtering process The mean flow equations require turbulence models in order to account for the turbulence effects rather than to simulate turbulence directly 6 3 1 Reynolds and Favre averaging Following classical Reynolds averaging Hinze 1975 any instantaneous flow variable n is de composed into the mean and fluctuating part respectively as follows 09 0 6 17 Various averaging procedures can be used to define the mean A general type of averaging is the ensemble averaging and it can be applied to any kind of flow The ensemble average of random functions of space and time is defined as the arithmetic mean over many macroscopically identical realisations of x t see for example Landahl an
228. and mass transfer O Modelling Porous Media chapter deals with three dimensional regions fluid sub domains occupied by continuum which comprises both fluid material and fine scale solid structure It contains theoretical background relevant to the volume averaged porous media equations and description of currently available porous types and their definition by the user O Modelling Multiphase Flows chapter introduces the physical aspects and modelling approaches for multi phase flows It describes in details multi fluid Euler Euler and single fluid Mixture and VOF models and how to use the currently available homogeneous mixture model boiling and cavitation models O Boundary Condition and Interface Types available in VECTIS MAX are described in this chap ter Their physical aspects are explained as well as their definition O Numerical Solution of the governing equations is described here The user inputs required at the different stages of an iterative solution procedure are explained O Modelling Radiation describes the surface to surface radiation module its individual programs radvfm radsolv and its interface to VECTIS MAX In particular attention is paid to the mesh file and radiation setup O Modelling Fans chapter presents pragmatic and computationally inexpensive modelling practice for fluid flows through fans Two models labelled as 1D and sub domain are available and both require the specification of
229. and phase properties o o 312 18 12Subroutine get_species to get values for variables defined in the access group iacc_species mainly start end boundary interface regions for Species o o 313 18 13Subroutine get_ps to get values for variables defined in the access group iacc_ps mainly start end boundary interface regions for passive scalars o ooo o 314 18 14Subroutine get_reg to get integer variables defined in the access group iacc_region 318 18 15Calculation options for variables describing fluid solid properties 321 18 16Subroutine get_property to get values for variables defined in the access group iacc_pro mainly phase species passive scalars calculation options and reference properties 322 18 17Subroutine get_grid_geom to get grid geometry variables defined in access group iacc_ SHO SEO arica dadas ad a a a a od 324 18 18Subroutine get_grid_connect to get grid connectivity variables defined in access group ACC Smd CONNECT as a o o ea le a A e N 325 18 19Subroutine get_turb to get turbulence variables defined in access group 1ace_turb 328 18 20Subroutine get_run_ctrl to get run control variables defined in access group iacc_run_ctrl 328 18 21 Subroutine get_name to get VECTIS MAX object name oo o 330 18 22Subroutine get_name_list Get a list of name
230. ank var_name fi_c fi_b fi_ui fi_li fi_c_o fi_c_oo fi_f var_id Arguments Description var_rank gt integer rank of the variable array var_rank if var_name is a scalar rank is one var_rank if var_name is a vector rank is two fi_cQ gt real pointer optional cell variable associated with var_name fi_c if var_name is a scalar fi_c if var_name is a vector fi bO gt real pointer optional boundary variable fi_b if var_name is a scalar le fi_b if var_name is a vector fi_ui gt real pointer optional variable at upper material interface fi_ui if var_name is a scalar fi_ui if var_name is a vector fi_liQ gt real pointer optional variable at lower mat interface fi_li if var_name is a scalar fi_li if var_name is a vector fi c_00 gt real pointer optional cell variable old fi_c_o if var_name is a scalar fi_c_o if var_name is a vector fi_c_oo gt real pointer optional cell variable old old fi_c_oo if var_name is a scalar fi_c_oo if var_name is a vector fi f0 gt real pointer optional variable at internal face h L l4 Ed fi_f if var_name is a scalar fi_f if var_name is a vector var_id gt integer optional variable index Table 18 27 Subroutine get_field to get field values either for a scalar or a vector Ricardo Software December 2009 336 18 USER PROGRAMMING 1
231. another 3D canvas and data from the last step is selected in the step panel Ricardo Software December 2009 385 19 TUTORIALS 19 1 BASIC TUTORIAL File Edit View Options Window Help p A D A Poeavecnsi Bo EJ white gt x AOAS amp B Live Update B o y Browse for run influence of initial conditions selection fome 19 inlet pressure Files 4 Domain VO 102 Run 1 1103 Run 1 Run 2 influence of initial conditions selection outlet mass flow Refresh Available Files influence of initial conditions selection outlet mass flow La tter_No Area_TOT Vel_AVE Temp_AVE Mflow_ SUM Density_AVE TPres_AVE Update Interval ms po EMO Figure 19 21 Comparison of IO mass flow convergence with different flow initialisation Ricardo Software December 2009 386 19 TUTORIALS 19 2 STEADY STATE PORT FLOW 19 2 Steady State Port Flow Get the necessary files for the port flow tutorial http www software ricardo com support tutorials vectismax TutorialFiles SsSPortF low 19 2 1 Introduction The purpose of this tutorial is to illustrate how to set up and solve a steady state intake port flow calculation The geometry represents a test rig and has a plenum chamber intake port and valve 19 2 2 Aims Upon completion of this tutorial the user should be able to perform the following O Define IJK mesh refinement blocks and use boundary refinement DO Activate the St
232. ant to specify a particular boundary condition profile or modify the solver param eters directly For this reason a powerful set of User Programmable Routines UPR have been implemented into VECTIS MAX solver to allows direct access to most of the solver data structure which in turn extends the capability of the VECTIS MAX solver This enables a user to write and compile their own code that runs with the solver as a shared object enabling the user to modify or add new features to the solver The user may perform many tasks such as setting up boundary conditions initial conditions mate rial properties or output results for a specific part of the domain They can also be used to specify additional sources for the transport equations solved by VECTIS MAX To use the UPR capabilities of VECTIS MAX the user should have knowledge of CFD and knowl edge of Fortran 95 2003 programming language which is used to write user programming code It should be noted however that even though UPRs significantly enhance the VECTIS MAX solver capability it is not possible to deal with every CFD problem using UPRs User routines can be utilised at different stages during the simulation process The five base user routines are described in section 18 2 These five routines act as wrappers for the user written code to interact with the solver The solver variables that can be accessed in the user programmable routines are described in sec tion 18 3 The set of U
233. aporation from drops Heat Transfer vol 48 pp 142 180 12 4 1 2 Ricardo Software December 2009 448 19 TUTORIALS 19 4 IMPORTING THIRD PARTY MESHES Rayleigh L 1917 On the pressure developed in a liquid during the collapse of a spherical cavity Phil Mag vol 34 pp 94 98 12 4 2 Rhie C M and Chow W L 1983 Numerical Study of the Turbulent Flow Past an Airfoil with Trailing Edge Separation AZAA Journal vol 21 pp 1525 1532 14 2 2 Rodi W Mansour N M and Michelassi V 1993 One Equation Near wall Turbulence Modeling With the Aid of Direct Simulation Data ASME Journal of Fluids Engineering vol 115 pp 196 205 9 4 4 Rohsenow W M 1952 A Method of Correlation Heat Transfer Data for Surface Boiling of Liquid Trans ASME vol 72 pp 1025 1032 12 4 1 1 Rung T Lubcke H M and Thiele F 2000 Universal Wall Boundary Conditions for Turbulence Transport Models Zetschrift fur Angewandte Mathematik und Mechanik pp 1756 1758 9 4 2 9 4 4 Saffman P G 1970 A Model for Inhomogeneous Turbulent Flow Proc Roy Soc London A vol 317 pp 417 433 9 2 1 Schnerr G and Suaer J 2001 Physical and numerical modeling of unsteady cavitation dynamics in ICMF2001 New Orleans USA 12 4 2 3 Shih T H Liou W W Shabbir A Yang Z and Zhu J 1995 A New k Eddy Viscosity Model for High Reynolds Number Turbulent Flows Model Development and Validation Computers Fluids vol 24 no 3 pp
234. ar also contains some viewing option commands that are likely to only be used when setting up mesh lines These are the top four buttons In order they are a X Y Projection a X Z Projection a Z Y Projection Ricardo Software December 2009 62 3 GEOMETRY 3 16 MESH SETUP ol 3D View Only in the 3D view option may the model be manipulated however mesh lines may only be edited in the first 3 2D projections As the first button is automatically selected when the mesh set up mode is entered the model is put into the X Y projection ready for mesh lines to be added The mesh is defined by a series of main mesh lines in three directions with the spaces between these lines being divided evenly into any desired number of cells The main lines are shown in red and intermediate lines are shown in green In the 3 D view intermediate lines are not shown unless requested from the View Options panel The positioning of the main mesh lines and the number of divisions between main lines controls the cell density in different parts of the model For any particular problem the actual mesh density used will be governed by the requirement to resolve the flow accurately and limited by the maximum problem size that can be analysed on the hardware available In setting up the mesh the user should be aware that VECTIS requires a halo of external cells around the active or internal cells in the calculation To guarantee that this is achieved ther
235. aries should be checked 4 10 Grid Data Structure 4 10 1 Grid components In VECTIS MAX both the grid connectivity and geometrical data are defined and maintained for the single computational mesh generated over multiple fluid and solid domains The global domain is composed of fluid and solid domains The computational mesh is obtained by dividing the global domain in a number of non overlapping finite control volumes CV or cells which constitute the computational mesh The computational mesh can be described in terms of see Figure 4 12 left Boundary y 2 S Domaini Domain2 Domain3 Vertices Inter faces Cell Global domain Figure 4 12 Grid objects Boundary face E n a Ta f Figure 4 13 Control volume and notation Ricardo Software December 2009 99 4 MESHING 4 10 GRID DATA STRUCTURE O vertices nodes O edges O faces and cells Each of these grid objects need to have an index label and need to be defined in terms of other objects with exception of vertices Information provided in order to identify and connect each grid object adjacent to the given object are commonly called connectivity data Vertices Vn n 1 N are basic grid objects since they define physically the numerical grid through their position vectors 7 xz 1 where xz k 1 2 3 are Cartesian coordinates The position vectors are defined with respect to a fix
236. arity of the energy Equation 6 15 is that the convective and unsteady terms are written in terms of total enthalpy energy whereas the diffusion flux is naturally written in terms of temper ature gradient However the energy equation should be discretised in terms of the same dependent variable i e either total enthalpy or temperature It is straightforward to express the temperature gradient as a function of total enthalpy and other variables but it is more desirable to define boundary conditions in terms of temperature An effec tive discretisation technique which describes both diffusion and boundary conditions in terms of temperature has been designed by applying the deferred correction approach To this end we can write the diffusion flux D T in terms of temperature according to Equation 14 33 and add with opposite signs two equivalent diffusion terms discretised in terms of total enthalpy A n 1 Taj Hp Hp Aj dj A n DT r Hp H 14 38 T dr P Hp Dj where superscripts n and n 1 denote current and new iteration steps respectively The first total enthalpy term is now treated implicitly while remaining terms having values at the current iteration are treated explicitly In case of conjugate heat transfer however solving the total enthalpy in terms of temperature has a number of advantages Murthy and Mathur 1998 Notably temperature is continuous across fluid solid inter
237. ary faces for each boundary region The call to this subroutine can be made in four different ways a to d as reflected by subsections of Table 18 14 Note that some variables used in this table and some other tables are for illustrative purpose only These variables parameters are nte number of transport equations currently nte 14 Ricardo Software December 2009 314 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES nbcop number of integer control parameters to deal with boundary conditions currently nbcop 8 nbcopr number of real boundary condition control parameters currently nbcopr 9 O nspreg number of species boundary regions npsreg number of passive boundary scalar regions O nphreg number of phase boundary regions nafve number of addresses for internal face vertices nabve number of addresses for boundary face vertices nproph number of phase properties currently nproph 6 nprosp number of species properties currently nprosp 1 1 nprops number of passive scalar properties currently nprops 3 nturb_par number of turbulence control parameters nturb_par 4 nbtype number of boundary types currently nbtype 15 Other variables given in plain text including those in Table 18 15 provide optional arguments with option value and need to be declared before using them a T
238. ary faces must be generated here This problem might be caused by unclean geometry near the box Ricardo Software December 2009 96 4 MESHING 4 9 WARNINGS AND ERRORS ERROR 1505 Tying patches routine cannot break the required edge A logical error occured in routine of tying patches ERROR 1506 Tying patches routine cannot move the required node A logical error occured in routine of tying patches ERROR 1507 The closed loop cannot be gathered for box X ii jj J kk K bx box This error is the most common symptom of problematic topology of input triangles Check the triangulated surface in the global box I J K ERROR 1508 Processing of command line cannot find any cell index for the switch viewc A fatal problem occured when command line option viewc was processed The help section 4 4 should be consulted ERROR 1509 Processing of command line cannot find correct number of parameters for the switch viewcgrp A fatal problem occured when command line option viewcgrp was processed The help section 4 4 should be consulted ERROR 1510 Processing of command line cannot find any boundary number for the switch viewb A fatal problem occured when command line option viewb was processed The help section 4 4 should be consulted ERROR 1511 Unknown switch switch encountered An unknown switch was encountered when pro cessing the command line ERROR 1512 No input ascii file was specified The meshing task is ex
239. ary triangles are added to by the boundary painting functions described below Add Delete Boundary The Add boundary function adds a boundary definition without assigning triangles to it The Delete Boundary function deletes the triangles on the selected boundary and sets its type to wall Show Hide All Boundaries The Show Al function switches all the boundaries to active and displays them if they have not been chopped and the Hide All function switches all the boundaries to inactive so that the boundaries will no longer be displayed Toggle Boundaries The Toggle function switches all the active boundaries to inactive and so they will no longer be displayed and switches all the inactive boundaries to active and displays them if they have not been chopped Compress Boundaries The Compress function removes all boundary definitions which have no triangles assigned to them Boundary Painting The Paint Line and Paint Face buttons operate in much the same way as the Mark Line and Mark Face stitching operations Instead of marking the triangles however these functions change the Ricardo Software December 2009 56 3 GEOMETRY 3 15 BOUNDARY PROCESSING boundary that the triangles are assigned to This results in the colour of the triangles changing to that of their new boundary Paint All assigns all the active triangles to the currently selected boundary Face painting is also controlled by an ed
240. at Mass Transfer vol 24 pp 1025 1032 12 4 1 2 Khosla P K and Rubin S G 1974 A Diagonally Dominant Second Order Accurate Implicit Scheme Com puters amp Fluids vol 2 pp 207 209 14 1 5 2 Kim S E and Choudhury D 1995 A Near Wall Treatment Using Wall Function Sensitized to Pressure Gradient in ASME Symp Separated and Complex Flows FED Vol 217 pp 273 280 9 4 2 Kunz R Siebert B Cope W Foster N Antal S and Ettorre S 1998 A Coupled Phasic Exchange Algorithm for Three Dimensional Multi Field Analysis of Heated Flows with Mass Transfer Computers amp Fluids vol 27 p 741 12 2 Kurul N and Podowski M 1990 Multidimensional effects in forced convection subcooled boiling in Pro ceedings of the 9th International Heat Transfer Conference Jerusalem Israel 1 BO 04 pp 21 26 12 4 1 2 Lahey R and Drew D 2001 The Analysis of Two Phase Flow and Heat Transfer Using a Multidimensional Four Field Two Fluid Model Nuclear Engineering and Design vol 204 pp 29 44 document 12 1 Lamb H 1932 Hydrodynamics 6th ed University Press Cambridge 9 1 Landahl M and Mollo Christensen E 1986 Turbulence and Random Processes in Fluid Mechanics Cam bridge University Press New York 6 3 1 Launder B E and Spalding D B 1974 The Numerical Computation of Turbulent Flows Computer Methods in Applied Mechanics and Engineering vol 3 pp 269 289 9 2 1 9 4 2 Ricardo Software
241. at flux for this boundaries super patches will be solved The following thermal property data should be defined Thickness This is the thickness of the super patches m Default 107m This is the material density kgm Heat Capacity Material heat capacity Jkg K Conductivity Material heat conductivity Wm K Once the radfile has been generated this is automatically done after saving the input file it is then necessary to generate other auxillary files required by the radiation solver module This is currently done manually via the command line See below 15 3 1 Running of radprep The radprep program is used to generate the patfile which contains the superpatch connectivity The radprep module can be run in two ways radprep options see below radfile Available options Ricardo Software December 2009 261 15 MODELLING RADIATION 15 3 RADIATION SETUP nosc Connectivity for surface conduction will not be computed notc Connectivity for through conduction will not be computed sn Superpatches s normals will be switched to opposite direction for find ing connectivity for through conduction minangle Angle Minimum view angle for finding through conduction connectivity Default minangle 80 0 degrees maxangle Angle Maximum view angle for finding through conduction connectivity Default maxangle 89 0 degrees ntcs Number Number of connection to store for through conduction connectivity s
242. aterial Heat Exchanger Porous Subdomain Name cat_monolith Subdomain Type Porous Turbulence Model Porous Standard C Turbulent Intensity Dissipation Length Scale m 0 01 Porosity 1 Porous Material Type General General Turbulent Data Principal Axes Catalyst Orthotropic Direction Vector X ENR x 0 998377 Y 0 031748 z 0 047271 Direction Vector Y xlo Y 1 z 0 Resistance Tensor Components Viscous kg s m 3 xx 1500 yy 100000 zz 100000 Inertial kg m 4 xx 25 yy 10000 zz 10000 Heat Load Heat Transfer Coefficient Heat Exchange Mode Figure 11 1 R Desk Sub domain panel User inputs for general porous media modelling 11 3 1 General porous model If a general porous material or heat exchanger has been selected by the user the diagonal elements of Viscous and Inertial resistance tensors Rj and Ri respectively need to be provided for each principal direction xx yy and zz see sub panel Resistance Tensor Components in Figure 11 1 Large resistance values in certain direction 50 00 to 100 000 for viscous and 10 000 for inertial resistance will suppress the flow in that direction For example values in Figure 11 1 the Ricardo Software December 2009 192 11 MODELLING POROUS MEDIA 11 3 POROUS MODEL TYPES AND USER INPUTS flow will be directed along the first principal axes 1 e
243. ation rate may be needed Then all required fields cell values can be obtained by using the get_field access routine In general additional sources can be expressed linearly by src src_l src_2 fi where src_1 is the phi fi independent term and src_2 fi is phi dependent part In order to improve stability src_2 can be added to the source explicit or to ap implicit depending on to the sign src_fi 1 ic src_fi 1 ic min 0 src_2 fi ic ap_fi ic ap_fi ic max 0 src_2 Ricardo Software December 2009 351 18 USER PROGRAMMING 18 5 WRITING AND COMPILING UPR For the phi independent part src_1 if it is negative always it can be added to ap by arti ficially introducing a dependence on phi by scaling it with 1 fi_old ic from the previous iteration src_fi 1 ic src_fi 1 ic max 0 src_1 ap_fi ic ap_fi ic min 0 src_1 fi 1c See Section 18 7 4 for example In case of momentum equations all velocity components have the same central coefficient ap ap_ uQ ap_vQ ap_w Therefore the user should try to redefine negative source term src_2 to be equal for all velocity components If this is not possible the negative source term s are added to src_fi Note The source terms used above src src_1 src_2 must have the units physical dimensions equal to the units of mass flow rate kg sec times the units of variable phi For example the units will be mass mass volume fraction source gt kg
244. ations The exception is the energy equation for which all the interfaces are treated in an implicit manner i e no specific boundary types are required 13 1 Setting Up Boundary Condition Types When starting anew VECTIS MAX project in R Desk a Boundary region is created within a Fluid Domain by default as Bnd_Reg_1 Inlet Given Velocity Boundary The initial Solver Setup Tree is given in Figure 7 2 top left Figure This boundary region contains Phase_1 Boundary Phase definitions that is also added by default R Desk Setup allows the user to add or delete boundary regions For example the user can add a new boundary region by right clicking on the current boundary region and then selecting Add Boundary Region which is the first option in the panel triggered by right clicking When there are 2 or more boundary regions R Desk creates a new node called Boundary Regions which contains all boundary regions defined for that fluid domain Multiple boundary regions can be added by left clicking on Boundary Regions A panel is displayed to the right that contains a button Add Child Boundary Regions Left clicking this button adds more boundary regions and gives default names A boundary region can also be removed by right clicking on the relevant boundary region and selecting for example the Delete Bnd_Reg_1 Inlet Given Velocity Boundary option The automatic extraction of boundary and interface regions described in Section 7 5 is currently a si
245. ations gt e so e e sece ca moose e ad ees 27 3 8 1 The Intersect Slice Operations cociososs sa eds da D oa Ba ee eae se 28 3 9 Bool an Volume Operationss os a s a a hee ek ae e aci 29 3 91 GULinteracive mode e s ao es a dd E e ai 29 3 9 2 Boolean Operation Batch Mode 220 0 ca 0456 bee wee eee 31 3 10 Triangle Marking Operations ss 05 enega dai ae a a a ee a a 31 3 11 Triangle Chopping Operations s s ase a ee ee RR Re ee ee ha E a 34 3 12 Triangle Interrogation Operations lt os e soa soe eR ee Be 37 3 13 Hints for Manual Stuchine Ssa 6660 0 tae ea aw ed ee wae dala eee a 39 3 13 11 Triansle Redcon s s e 264 4 clea a te wh we ce ee a ee hoe we ea 40 3 13 2 Self Intersection Checki g idad edora eG aS ROE aa ha daa 40 3 14 Geometry Wrapping s os lt sa sam died ale as da ee ROR e ee ee 41 3 14 1 GUI Interactive Mode oscars sea da dhe oo de de aw ee ee Es a 42 3 14 2 Batch Modes a a veh ad E s Se wR ee ad ee ww Re A 46 3 14 33 Geometry Wrapper Input File a aie Se a ee 49 3 15 Boundary Process 4 0 44 2 24 Rb CREE RLS EMER ORE EERE REE LES 51 3 15 1 Partand Boundary Definition s ss s s wae dees adm amado ee 51 3 152 Boundary Processing so sag a AE ee e de e E A Od 55 3 15 3 Boundary defined refinement setup e 57 3 15 4 Saving the boundary refinement settings s oss ss es a coa woa doa 59 3 15 5 Boundary Refinement examples ee eee eee eee 60 3 16 Meshi setip 4 2
246. aviour can be changed in the preferences panel The attributes of the plot can be modified in the Plot Properties panel Different data can be shown on the plot by selecting it in the Data panel Ricardo Software December 2009 413 19 TUTORIALS 19 2 STEADY STATE PORT FLOW lis R Desk 3d1 p1 port post_001 gt 101 x we Fie Edit View Options Window Help la x A 3 A lllrroraf ea viewer 1 7 Plots ax portpost_DO1 13D El lega Plot p1 Canvas 3d1 Ly w Data ax Contour vector B Elements a conductivity density mol_viscosity specific_heat static_pressure temperature total_enthalpy turb_dissipation turb_energy u_velocity y IV Auto apply Apply jo x _PlotProperties _ Data sets J x Scales ax B Colour Map vec Connections O Colour ar 2 FY Elements static_ 14 155 88281 1753371239 Reset 4 I Auto apply Apply A Set Range Auto apply Apply os a Command gt A Figure 19 60 Pressure variation at the model surface 19 2 14 1 2D plane A 2D slice can be made of a plot Right click the plot in the plottree and select slice The slice is then defined in the slice definition panel If the plot is present in a 3D canvas the slice will be previewed Once the slice is in the correct position click apply if auto apply is not active Then open an additional 3D canvas
247. be used Refer to the Radprep theory section for information about how the best connections are determined 15 3 2 Running of radvfm After the radprep program has completed the view factor matrix should then be created using the following command radv m r s v h radfile where Ricardo Software December 2009 263 15 MODELLING RADIATION 15 4 RADVFM THEORY y Shows version of the program If this parameter occurs among arguments the other parameters are omitted s Allows to set the number of used pixels in hemispere The asterisk stands for 1 5 numbers For example if s3 is used the number of pixels on the hemisphere diameter will be 300 3x100 This means that the hemisphere will be placed on the square 300x300 pixels The level 3 is default T Reverses the boundary orientation used only when a trifile is used as the input geom etry h A help similar to the information above is shown The radvfm program does the following 1 Reads the radfile and determines the following information O name of mesh file or tri file O name of patfile O name of vfmfile which should be created 2 Reads the phase4 mesh file or tri file 3 Reads the super patch file patfile 4 Allocates the necessary memory space 5 Loops over all the super patches and calculates the rows of the view factors matrix These rows are directly written to vfmfile using the Ricardo SDF file format radfile meshfile name
248. ber 2009 397 19 TUTORIALS 19 2 STEADY STATE PORT FLOW Solver Setup Tree xj i Postprocessing Output User Function Radiation Global_domain_1 Global Domain Restart Control Timebase Output Fluid_1 Fluid Domain Monitoring Points Algorithm Turbulence Model Discretise Phase_1 Fluid Phase Initial Condition Postprocessing Output Equations amp Solver Bnd_Reg_1 Inlet Given Velocity Boundary Phase_1 Boundary Phase Interface Regions Report Regions Solid Domains Figure 19 37 The solver setup input tree 19 2 10 1 Importing the Grid File The materials and boundaries defined in a grid file can be imported into the solver setup This will populate the solversetuptree with the correct number of materials and boundaries A GRD file needs to be entered into the Filename box Clicking on the browse button will open a dialog box and allow the relevant grid file to be selected Once a file has been selected The name will appear in the Filename box Next click the extract button A dialog box pops up and displays the materials found in the imported grid file ni R Desk File Edit View Options Window Help B DB amp Proiealvecns B w e QQ x Eea AS 131 AutoPreview portGRD ul ES Solver Setup E Global_domain_1 Global Domain Restart Cont
249. by left click on the Fluid Phase node The top Phase Name edit box can be used to enter the phase name As Figure 8 3 top shows the single component phase is the default selection Clicking on the Mixture of Species Option box a panel pops up Figure 8 3 bottom left where three phase models can be selected single component phase multi component phase and a phase belonging to the WAVE data base The Phase Type can be selected by using the Gas or Liquid radio button Ifthe Gas is selected then the phase compressibility should be defined by clicking on the Compressibility list Ricardo Software December 2009 139 8 MODELLING CONTINUA 8 6 SELECTING CONTINUUM AND ITS PROPERTIES Phase Name air Mixture Of Species Option Single Component Phase Phase Property Filename none LZ Y Phase ID f1 Phase Type e Gas Liquid Compressibility Incompressible Fluid Solve All Species Define Passive Scalar Single Component Phase AS Multi Component Phase Fully Compressible Sub sonic 0 3 lt Ma lt 1 Wave Properties Fully Compressible Super sonic Ma gt 1 Figure 8 3 R Desk setup Setting up a fluid phase box The pop up panel shown in Figure 8 3 bottom right will contain four options to describe the fluid phase in terms of compressibility see Equation 8 13 1 e in terms of Mach number Incompressible Fluid Density is constant compressibility coefficient 0
250. ce thus allowing the user to traverse a slice through a model The default rotation increment produced by pressing a function key is 45 degrees The sensitivity can be adjusted with the numeric keypad keys as follows Key 0 45 degrees Key 1 1 degree Key 2 5 degrees Key 3 15 degrees The sizes of the translations and zooming scale accordingly The Esc key can be used to cancel any current commands requested from any of the tool panels This may also be done using the cancel command button located on the button bar Holding down the CONTROL key and pressing r resets the model view This can also be achieved using the view reset button located on the button bar In the reset view operation the active part of the model is re centred and the axis orientation is reset to an orthogonal state 3 4 Model Import 3 4 1 Model Import Phase 1 currently understands the following file formats O ASCII VECTIS Triangle file before version 3 2 2 O binary VECTIS Triangle file version 3 2 2 and later VDA File STL File both binary and ASCII O Phase 4 output mesh file viewing only Ricardo Software December 2009 15 3 GEOMETRY 3 4 MODEL IMPORT O VECTIS mesh input file will be associated with the current model All of these file formats may be imported by selecting the Open option from the File menu Addi tionally on those platforms that support a command line a filename specified upon the comm
251. ce value of viscosity for phase iph subroutine get_property var_name ival Input Output var_name cha ival integer pointer l_phase_opts property calculation options for phases ival 1 nproph 1 n_phases l_species_opts property calculation options for species ival 1 nprosp 1 n_species l_ps_opts property calculation options for passive scalars ival 1 nprops 1 n_ps subroutine get_property var_name rval Input Output var_name cha rval real pointer r_phase_values reference values for phase properties rval 1 nproph 1 n_phases r_species_values reference values for species properties rval 1 nprosp 1 n_species r_ps_values reference values for passive scalar properties rval 1 nprops 1 n_ps Table 18 16 Subroutine get_property to get values for variables defined in the access group iacc_ pro mainly phase species passive scalars calculation options and reference properties 18 3 2 9 get _grid_geom This UAR is used to retrieve information related to grid objects There are two section in Ta ble 18 17 and are explained here a b a To get the cell volume for all CV then real wph pointer cell_volumes call get_grid_geom cell_vol cell_volumes Ricardo Software December 2009 322 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES Array cell_volumes 1 n_cells contains the cell volumes for all cells n_cells To get the uppe
252. ch time step icp_end_iter end of each iteration icp_end_time end of each time step icp_end_run end of simulation Table 18 41 Program position identifiers be included to provide access to the user accessible routines variables UARs UAVs Compilation of UPRs can be done by running the umake utility e g umake upr_bound f90 which would create a shared object libupr so Multiple source files can be specified on the command line Options to umake include f name of compiler program default ifort compver compiler version 32 run in 32 bit mode 64 run in 64 bit mode 0 output file name default libupr so Current compiler support is Intel amp gfortran GNU Compiler linker flags can be set via the FFLAGS and LDFLAGS environment variables 18 6 UPR Check Report Messages Error messages generated during the compilation of a UPR typically arise from a missing use upr statement or some inconsistency in the way a routine UAR is called Any error messages generated during the running of an UPR are usually quite self explanatory Typically error warning messages may arise from trying to retrieve or set non available fields properties for a given material phase etc NOTE Certain arrays that are directly accessible should only be read For example the boundary conditions array can be retrieved via a Fortran 95 2003 pointer and modified illegally causing unexpected results Currently there are
253. checked the GRD file will be displayed in a 3D canvas Then when entries in the Solver Setup Tree are selected the relevant domains materials or boundaries regions of the grid are highlighted checked the GRD file will be displayed in a 3D canvas Then when entries in the Solver Setup Tree are selected the relevant domains materials or boundaries regions of the grid are highlighted 19 2 10 2 Global Domain Click on the global domain entry in the Solver Setup Tree The global domain panel allows general simulation parameters to be set Domain Name A name can be given to each global domain in the calculation For example Coolant or Cylinder Head In this case we will call the domain port Input Mesh Filename The solver will default to using a computational grid file with the name pro jectname GRD A different grid file can be specified here otherwise the projectname is taken as the inp file prefix Again here we will use the GRD file that was previously generated normally port GRD ORicardo Software December 2009 399 19 TUTORIALS 19 2 STEADY STATE PORT FLOW Solver Setup Tree El Postprocessing Output User Function Radiation Global_domain_1 Global Domain Restart Control Timebase Output E Fluid_1 Fluid Domain Monitoring Points Algorithm Turbulence Model Discretise Phase_1 Fluid Phase Initial Condition Postprocessing Output Equations amp Solver B Boundary Regions Bnd_Reg_1
254. cific criteria see below for example Section 18 7 2 18 4 3 User boundary conditions routine subroutine upr_bnd_cond var_name idt ir jbnd1 jbnd2 Arguments Description var_name char name of field that can be modified idt integer domain type id ir integer boundary region id icell integer starting boundary id for ir icel2 integer ending boundary id for ir Table 18 38 Subroutine upr_bnd_cond to modify boundary conditions This routine is called for each boundary region and each phase species present in the correspond ing fluid solid material domains after the default initialisation is done by the solver See Sec tion 18 7 3 for example The order in which the variables var_name are passed to this routine is the following temperature heat_flux velocity pressure turb_energy dissipation phase_vol_ frac spec_mass_frac and mass_frac_ps In order to modify specific variables a conditional block block e g case structure should be set up The domain type index idt is associated with the mem ory address of var_name field For a fluid phase equation it is equal to the negative phase index and for the single phase fluid mixture of fluid phases or solid domain it coincides with the corre sponding domain indices In case of species mass fraction equation idt is equal to species indices In addition the user needs to check that the current domain type index idt and boundary regio
255. cording to Kenning and Victor 1981 as follows i Cpt PiAT sat Pehfg The quenching heat transfer coefficient 0 in relation 12 59 is calculated as 0 24 Caf MPC 12 62 T where A is the conductivity of the liquid and C is a constant with a value of 0 8 12 61 O Evaporation rate from the wall Mass flow rate from the wall is now calculated using relation 12 54 for the nucleate boiling heat flux Gnuc dnuc T 12 63 hfg Cpf AT sub i Ricardo Software December 2009 207 12 MODELLING MULTIPHASE FLOWS 12 4 PHASE CHANGE MODELLING 12 4 1 3 Setting up a Boiling Model Two phases need to be defined so that boiling model can be used the liquid and vapour phases It is recommended that the compressibility for the vapour phase should be set to weakly compressible To set up a boiling model select the appropriate Fluid Domain from the Solver Setup Tree This opens up the Fluid Domain set up panel Under Multiphase Modelling select the Mixture option see section 8 6 1 As a result a Multiphase option is added to the Solver Setup Tree Left clicking this option opens the Multiphase panel Under Mixture Model Option select Boiling Models as in Figure 12 3 A number of fields become active Either RPI model or VECTIS3 model can be selected The effect of evaporation condensation and boiling heat flux can either be increased or decreased by entering the appropriate values for Evaporation Condensation and Heat Flux
256. d Mollo Christensen 1986 Blouse lim LY GMa 6 18 M M m 0 where represents the ensemble average and gm xx t is the m th realisation of b xk t For the fluid flows the macroscopically identical realisation means that statistically independent flows are exposed to the same set of initial and boundary conditions Obviously the ensemble averaged quantity may be time dependent In the statistically steady turbulence the ensemble average is the same as the time average lim f ol Xpt 6 19 T 0 T In the case of a periodic flow with a period of T it is convenient to use the periodic phase averaging over Mp periods Part 0 x1 0 Jim y x t mT 6 20 P Mo mo 0 where is now the instance corresponding to the particular periodic phase 0 The phase averaged quantity is not averaged over the periodic phase but at the particular phase The following averaging rules apply Tennekes and Lumley 1986 Pet 059 9 ov 0 9 W YW 7 6 21 Ricardo Software December 2009 118 6 SOLVER FUNDAMENTALS 6 4 REYNOLDS AVERAGED EQUATIONS Considering Equation 6 17 splits instantaneous motion represented by the velocity field Oil xg t into an organised ensemble averaged part U U and stochastic turbulent part u The corresponding one point second moments uu see the last term in Equation 6 21 are rarely zero They are known as the kinematic Reynolds
257. d domain and porous or heat exchanger models will not be applied The heat exchanger type is provided for modelling of the heat transfer between two fluid streams and may or may not have porous structure The porous sub domain may or may not involve heat exchange a radiator always involves heat exchange which is determined by selecting options from the Heat Exchange Mode list box shown at bottom of the sub domain panel If the porous turbulence model is selected then the Turbulent Intensity and Dissipation Length Scale values should be supplied inside the Turbulent Data sub panel In this case the turbulence kinetic energy and its dissipation rate will be calculated algebraically from Equation 11 54 The next step in defining both porous and heat exchanger sub domain models is to specify the volume Porosity y and Principal Axes or eigenvectors of the resistance tensor given by Equation 11 20 Note that the viscous Rij and inertial KR j tensors defined along principal axes have only three diagonal coefficients The basic approach is to specify two direction vectors in a Cartesian coordinate system 1 e the components X Y and Z forthe Direction Vector X and Direction Vector Y see the Principal Axes sub panel Whilst these two direction vectors have to be orthogonal they need not to be aligned with the domain Cartesian system or mesh lines If the user fail to specify two normal direction vectors the solver will ensure their
258. d expression is obtained by using a blending model of Kader 1981 VIS Tp e Ti te TF 9 43 where Kader s blending function is given as _ 0 01 Pry5y 9 44 1 5Pr3y5 The wall heat flux is given by Equation 9 39 with the wall thermal conductivity calculated from Equation 9 40 Mass fraction and wall diffusion flux of species Formulae used for the temperature and heat flux apply also to species if the Prandtl numbers are replaced by corresponding Schmidt numbers Turbulent energy production The k equation is solved for the near wall cells with zero value and zero diffusion flux at the wall Its production within the log law layer can be calculated with respect to the velocity distribution given by Equation 9 31 t dU on gt yt Pop i P dy J mp JPE Ye 9 45 0 yp lt Ye where the turbulent stress is approximated by the wall shear stress Tp Tw The constant shear stress T can be assumed across the entire near wall region The value of yz 2k eri 222 733 corresponds to the intersection of equations defining the dissipation rate in the log law layer Ejog Equation 9 29 and in the wall limit amp js 3 4 3 2 Cu k 2uk Elog EP Evis B gt 9 46 Kyp PYp O Turbulent energy dissipation rate In the near wall cells the dissipation rate is set explicitly to 1 pk y eg Ep flog A 1Rexp 5 9 47 The above expression covering the entire wall
259. d input file Figure 19 80 Boundary Regions Each of the boundary regions specifications should have been imported from the old input file As noted previously the second mass flow boundary defined in the original input file has been changed to a pressure boundary In these cases it is important to verify that the conditions being simulated still correspond to the physical boundary conditions The only modification that should be necessary for the boundary regions is to check the Boundary Ricardo Software December 2009 430 19 TUTORIALS 19 3 COOLANT FLOW Phase_1 Fluid Phase 218 xi Figure 19 80 Fluid Phase Options Report option for the inlet and outlet boundaries See Figure 19 81 The following boundary conditions will be used Boundary 1 will be a wall with a fixed temperature of 373 deg K Boundary 2 will be a specified mass flow boundary with a rate of 2 02 kg s with a temperature of 373 deg K Boundary 3 will be a static pressure boundary set to 100000 Pa outflow and with a temperature of 373 deg K The input file should now be complete for the analysis Save the file as coolant inp file gt save as On important thing to note is that VECTIS MAX uses a lower case inp extension for the solver input file 19 3 6 Grid Preparation Vpre prepares the computational grid for use in the solver It can be started either from a cmd window or from R desk In this case w
260. d passca_pro_opts 0 nprops 0 n_ps control evaluation of the phase species and passive scalar properties respectively For example the density ipro idens of a phase iph will be calculated according to the option flag returned by ph_ pro_opts ipro iph If ph_pro_opts idens iph 6 ipro_suth then density is calculated according to Sutherland law A number of control parameters have been introduced in order to cover a wide range of calcula tion options which can be assigned to the arrays ph_pro_opts 0 nproph 0 n_phases Ricardo Software December 2009 320 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES Calculation options for fluid solid properties Property type Description ipro_inact 0 inactive property ipro_const 1 constant properties ipro_mix 2 mixture of fluid species ipro_igas 3 ideal gas or ideal gas mixture ipro_poly 4 temperature dependent polynomial ipro_power 5 temperature dependent power law ipro_suth 6 temperature dependent Sutherland law ipro_user 7 supplied by user subroutine ipro_expr 8 expression supplied by user ipro_invis 9 inviscid fluid for viscosity ipro_nonwt 10 non Newtonian fluid ipro_bouss 11 Boussinesq approx for density ipro_isgas 12 isentropic gas ipro_wave 13 mixture properties from WAVE ipro_expo 14 temperature dependent exponential Table 18 15 C
261. d_Reg_1 Wall Boundary Radiation Interface Regions Giuh damaine Figure 15 3 Radiation boundary panel location within solver setup tree The radiation panel only appears for wall boundaries ORicardo Software December 2009 259 15 MODELLING RADIATION 15 3 RADIATION SETUP Radiation 7 G x Radi Emissivity 0 8000000000 Transmissivity 0 0000000000 Superpatch Density 50 PORNOS Superpatch Angle Tolerance 40 0000000000 X Conduction Conduction Thickness 0 0010000000 Conductivity 1 9000000000 Specific Heat Capacity 560 0000000000 al alp aly alo Density 7658 0000000000 Figure 15 4 R Desk radiation boundary panel The Radiation S2S check box activates a particular radiation boundary If a boundary is acti vated then radprep radvfm and the radation solver will perform the relevant calculations for the patches and super patches belonging to this boundary set If it is inactive and radprep and radvfm are run then the boundary will not be included in the super patch construction and view factor calculation respectively and hence this boundary can not be included in the radiation calculation without re running radprep and radvfm with it activated If a boundary is set as active and radprep and radvfm are run so that the boundaries super patches and view factors are calculated the bound ary can then be set as inactive before running vsolve so that this boundary is
262. dary panel e 260 15 5 Scheme ofthe radprep program ico ask gee a ee aed A 262 15 6 Figure showing the definition of the normal conduction minimum and maximum f search A a ea tae BOR a GE ee GS we Se we ai Dox ai en Gh oh ee ag Gt we a aC oe N 263 13 7 Scheme of the radvim prosrams ops eae a dha A war ee OR ee ee e a a 264 15 8 Heating super patch i and heated super patch j in view factor calculation 265 15 9 Figures showing the Hemisphere projection used in the radvfm calculation 266 15 10Example results for the surface conduction model showing the patch temperature 270 16 1 Adding afan 5444 bs da GRAS a a ee ee HG Ge Ge eS aR ad 271 16 2 Fan panels e224 a 2a et a Bi ht AO we ee ewe ee WG ow ee 272 16 3 Velocity vectors at fans outlet sos saans ae ew eed Bl a bo wo a ew 274 164 Top view of anaxial fan oo oscar ce ea he ee te dee RRS wee eas 275 16 5 Front view of an axial fam Shrouded 2 240 464 wi a ea ee a da 275 16 6 Typical fan characteristic curve monotonic o s lt se esa a ao sa eee ee ee 276 Ricardo Software December 2009 XX LIST OF FIGURES LIST OF FIGURES 16 7 Axial fan representation acc o A AA A a A ee we ew 277 16 8 Radial f n representation y cos econ ek ea a ea eared Be ee de BR ead ee E 277 16 9 Outflow vector specification for radial fan setup o o a 277 17 1 GroupBox for Frequencies For Printing Data into Project Files
263. de midpoints lie within the given tolerance of the surface The user is prompted for the triangulation tolerance The default is 0 1 mm Tiangulation Tolerance Enter Trianguiabon Tolerance Tolerance 0 1 rum TOK a VDA Triangulation Tolerance Reading and triangulating the file may take several minutes for a large model 3 4 4 Merging Once a file has been read into Phase 1 the nodes in the model are merged to eliminate duplicate or very close nodes The program prompts the user for the tolerance for this merging The default tolerance is such that the merging will eliminate no triangles i e it is less than the shortest triangle side in the model Tnangle Merging Tolerance Enter Merge Tolerance Cancel for no Merging Merge Tolerance 3 81192e 005 Cae Node Merging The triangle merging process also links the imported triangles into a surface Therefore it is recommended that the OK button is chosen from this dialog even if no triangles will be eliminated 3 5 General View Options Once a model has been loaded into Phase 1 the way in which it is viewed may be modified The viewing options appear to the left of the drawing canvas when the application first starts up Ricardo Software December 2009 EF 3 GEOMETRY 3 5 GENERAL VIEW OPTIONS El Selecting Options from the View menu will bring this panel back up Alternatively the user can select the view tab or the view options but
264. ded by this action Additional ones can be added by right click on the Passive Scalar node s For more than one passive scalar the right click will display a pop up panel with an option to delete the current scalar To access property definition of passive scalars the user should left click on the Passive Scalar node This action opens the Passive Scalar Properties panel Figure 8 13 Ricardo Software December 2009 146 8 MODELLING CONTINUA 8 6 SELECTING CONTINUUM AND ITS PROPERTIES Solver Setup Input Nm Passive Scalar Name Passive_Scalar_1 Option Constant Values al Initial Uniform Value 0 5 al Value 1 Option Constant Values Turbulent Schmidt Number Value 1 189 Option Constantvalues 5 Value 1 Figure 8 13 R Desk setup Setting properties of passive scalars In this panel Passive Scalar Name can be edited and Initial Uniform Value specified The required physical properties are Density Mass Diffusion coefficient and Turbulent Schmidt Number These properties should be defined in a similar way as for the standard species Ricardo Software December 2009 147 MODELLING TURBULENCE 9 1 Introduction Turbulence is commonly found in all fluid flows in nature and engineering It is the chief out standing difficulty of our subject cf Lamb 1932 and the last unsolved problem of classical physics attributed to R Feynman Bradshaw 1994 p
265. del is used if a liquid or an in compressible or weakly compressible gas phase has been specified A selection of the fully com pressible sub sonic or super sonic gas implies the use of an ideal gas model Considering density calculation options all the options which define density with constant or temperature dependent Boussinesq exponential and polynomial values can be used with the incompressible model This also include the ideal gas option where now the constant reference pressure value is used to cal culate density Pref P RT 10 12 If the equation of state is described by an ideal gas model then the ideal gas option should be selected as a density calculation option 10 2 1 Reference and solver working pressure The momentum conservation see for example Equation 10 2 contains the pressure gradient d p xi whose accurate predictions are affected by the numerical round off errors In order to reduce the round off errors especially for incompressible and compressible low Mach number flows the concept of reference pressure Pref is used to define the relative static pressure ps Ps Pabs Pref 10 13 which in turn defines the solver computed or working pressure The solver working pressure p represents the pressure which is computed during solution of the resolved flow equations Consid ering simulations involving the gravity buoyancy force it is given as is 5 uv U laminar flows iS p fres
266. delling can be illustrated by work of Kunz et al 1998 They analysed boiling flow in a vertical coolant duct by using the four field system similar to that shown in Figure 12 1 Employing the high Reynolds k e model for two continuous phases and single pressure model p p the 25 unknown variables were solved for 4 U H ak 2 k e p About 30 inter phase models were required Ricardo Software December 2009 199 12 MODELLING MULTIPHASE FLOWS 3 MULTI PHASE MIXTURE AND VOF MODELS 12 3 Multi Phase Mixture and VOF Models The multi fluid model often referred to as the full Eulerian model is theoretically well advanced model but also very complex Alternative and the simpler Mixture and Volume of Fluid VOF models are based on the single fluid approach The physical assumption behind these models is that all phases share the same pressure temperature and turbulent quantities In case of the Mixture model the phases can move at different velocities relative to the mixture velocity by introducing the concept of slip velocities While the Mixture model deals with interpenetrating phases the VOF model originally devel oped by Hirt and Nicholls 1981 is applicable to immiscible fluids ie to the capturing of liquid gas interfaces free surface flows jet breakup large gas bubbles in a liquid dam break liquid motion It is a front capturing method in a sense that the volume fraction of a phase liquid a takes the f
267. designated as a fan Fan Modelling This is either Subdomain or 1D Fan Type 271 16 MODELLING FANS 16 1 INTRODUCTION AND OVERVIEW Fan Modelling Sub domain Fan Type Axial Inlet Outlet Speed rpm 2100 Fan Axis Fan Centre X 1 22375 Blade Type Blade Angle Blade Tip Radius 0 2225 Hub Radius Under Relaxation Factor Minimum Flow Rate Pressure Volume Flow Rate Figure 16 2 Fan panel This is either Axial or Radial 1D Fan Model ONLY Inlet Outlet These are the boundary or subdomain interface IDs depending on the model type selected Speed ORicardo Software December 2009 272 16 MODELLING FANS 16 2 SUBDOMAIN MODEL This is the fan speed RPM Fan Axis This is the fan s axis In the case of an axial fan this will correspond to the inlet amp outlet axis For radial fans this will define the inlet axis Fan Centre This is the fan s centre Blade Type This is either Straight or Twisted 1D model ONLY Blade Angle Subdomain model ONLY This is the angle the blade makes with the fan s axis Blade Tip Radius This is the radial extent of the fan Hub Radius This is the radius of the fan s hub Under relaxation Factor This is the under relaxation factor Minimum Flow Rate Subdomain model ONLY This is sets a lower limit on fan s flow rate Pressure Volume Flow Rate This is a table used to store the fan s characteristic curve
268. df TP FP db Tb TP y di Fi Fp 4 9 where rp F and 7 are the position vectors at the centres of the boundary face and interface respec tively Finally normal distances between centres of near boundary cells and adjacent boundary faces are used to implement various boundary conditions They are defined as bn Az dn 1A5 4 10 Note that the angle between the distance vector d y and the face area vector Ay is a measure of the grid non orthogonality 4 11 Mesh Import 4 11 1 Mesh requirement The Vectis solver requires face to node and face to cell connections often called face address ing and also cell to face connections The connectivity data can be then described by O Number of vertices number of internal and boundary faces and number of cells Ricardo Software December 2009 102 4 MESHING 4 12 MESH QUALITY CHECKS O List of coordinates for all vertices O List of vertices that form a face for all faces face to node connectivity O List of two cells for a boundary face one cell adjacent to a face for all faces face to cell connectivity O List of faces that enclose a cell for all cells cell to face connectivity 4 11 2 Importing a grid data file The grid connectivity defined above is the one used inside the VECTIS MAX solver Obviously 1t does not provide association of grid data with other data structure objects such as domains and bounda
269. dient Options two methods are available Gauss Based and Least Square Method as seen in Figure 14 7 Gradient can be limited by selecting Limit Gradient CheckBox from the same panel Cross diffusion can also be limited by selecting Limit Cross Diffusion CheckBox Discretise o K Dimensionality e Three dimensional Flow Two dimensional Flow Gradient Options Gauss Based e Least Square Method Limit Gradient Pressure Switch 0 1 Limit Cross Diffusion Enhance Stability v Viscous Heating Terms In Energy Equations Bad Cell Treatment All Variables Except 1st Pressure Correction Y Solve All Phases Off All Variables Except 1st Pressure Correction All Variables Except Pressure Correction Figure 14 7 R Desk setup for 2D or 3D gradient calculation options etc 14 1 5 Discretisation of the generic equation 14 1 5 1 Transient term First the unsteady term in 14 4 and 14 8 which is mainly used for time dependent calculation is integrated over each time interval Ar by either the first order accurate Euler or second order accurate three time level scheme see Ferziger and Peric 1997 Both schemes are implicit and unconditionally stable When the first order Euler scheme is applied then the transient term is discretised as d vo pVO p 7 PV oP aa 14 30 Ricardo Software December 2009 243 14 NUMERICAL SOLUTION 14 1 FINITE VOLUME DISCRETISATION When the three time level scheme is a
270. different mesh generators VECTIS MAX offers its own mesh generator called VMESH which works automatically and pro duces a locally refined Cartesian mesh The Figure 4 1 shows position of the mesher in VECTIS MAX system and its communication with the other modules Two files containing the input infor mation need to be supplied to the mesher trifile and meshfile Both these files can be generated by the preprocessor Phase1 Trifile its usual extension is tri is a SDF file containing a com pletely closed fully connected triangulated surface Meshfile its usual extension is tri is an ascii file containing the meshing control parameters The format of this file will be described in section 4 3 The mesher produces final gridfile its usual extension is GRD which contains the generated grid In OUT file there is a copy of the messages printed to the screen Also VMESH temporarily creates several aux files which are removed at the end of the mesher s run VMESH also contains tools which can help the user to automatically correct certain portion of problems in the input triangulated surface These tools will be described in section 4 7 4 2 Howto Run VMESH VMESH is a console application To start it from the command line type vmesh meshfile switches The format of the input ascii meshfile and possible switches are described in two following sec tions VMESH produces gridfile whose name is derived from the name of the meshfi
271. ding distance 2 1 Remove gaps and intersections This work should be done in the preprocessor Phasel How ever VMESH offers a possibility to automatically detect and repair overlapped and intersected triangles The detection is activated when sep option is used see section 4 4 The problem atic parts will then be painted as a new boundary If the user wishes to automatically repair as many problems as possible s he can use the command line option rep The algorithm detects all problematic triangles removes them and tries to fill the gaps automatically When the gap cannot be easily capped the triangles are locally restored such a problem needs to be resolved manually later Since these automatic reparations might be dangerous in some cases they can change geometric features VMESH also offers a possibility to visualize the automatically re paired parts of the geometry Usage of rep sep causes painting of all new triangles as a new boundary so as the user can visually check correctness of the performed changes 2 Run the mesher When the geometry is sufficiently clean the mesher can be started VMESH performs a simple check of overlapped triangles can be switched off by not If there are any their indicies are printed and the mesher continues Usually even though there are some overlapped triangles detected the mesher has no problems to correctly generate the grid 3 Check whether the generated grid is correct If there is a ser
272. discretisation A number of discretisation methods exist and the VECTIS MAX is build around the Cell Centred discretisation method and is explained in this chapter Considering a non moving control volume in Figure 14 1 the second order accurate approxima tion of the generic transport equation 14 1 leads to the following balance equation written for the cell P with the volume V nf nf Z v i Lee Lo Y sAr 53 Ve 14 4 where convection and diffusion fluxes C and Dj are defined as C mj0j D rovo 4 14 5 Boundary face e fs Ag Figure 14 1 Control volume and notation Ricardo Software December 2009 235 14 NUMERICAL SOLUTION 14 1 FINITE VOLUME DISCRETISATION In the above equation is a flow variable and I represents its isotropic diffusion coefficient The outward face area vector is denoted as A Akik where FA 1s the Cartesian unit vector The convective and diffusion fluxes through the face j are C and Dj respectively the source terms s and s are related to the cell volume and cell face respectively ny is the number of cell faces Mass fluxes m are defined as mj pjU Aj 14 6 and they must satisfy the integral mass balance over each cell d gi q PVP t Y 0 14 7 j 1 14 1 2 Time Integration Time integration is needed in unsteady fluid flow problems The semi discrete general transport equation 14 4 is written in the form of
273. displayed against the solver iterations time steps O Run time solution control Run time control of the solver allows controls such as changing the solution end time changing output file writing frequencies or saving immediately the post processing file Interface to WAVE It is possible to run coupled co simulations with the Ricardo WAVE 1D gas flow product where the CFD inflow and or outflow boundary conditions are supplied from a WAVE calculation Non flow boundaries can either be specified as symmetry boundaries or walls with fixed heat flux or temperature 2 1 3 Pre and post processing R Desk All pre and post processing utilities for CFD analysis are integrated into the new Ricardo GUI R Desk This product introduces many new capabilities including multiple viewports and combined display of CFD results and structural analysis results together Further details of the features are described in the R Desk Help Ricardo Software December 2009 8 2 INTRODUCTION 2 2 USING VECTIS MAX BRIEF GUIDE 2 2 Using VECTIS MAX Brief Guide The following section provides a brief guide to the setting up process from start to finish for a typical CFD simulation within VECTIS Figure 2 1 illustrates this process M ripley R Desk es paration Joining Parallel solver setup decompostion Boundary Condtions Figure 2 1 Vectis work flow from initial model geometry to viewing CFD results The work flow using
274. ditions via R Therm module oo o 227 13 9 Time Dependent Datani a a a ad a ia hw 230 13 10 Interface Conditions ei cos tee eo a e ee ee OR E a 232 14 NUMERICAL SOLUTION 234 14 1 Finite Volume Disctetisation s os coa race edi ee ea EG ee a e 234 141 1 Generic transportequanlom sta e dee a ed Bs ew 234 Ricardo Software December 2009 vii 14 1 2 Time Intesration cga eane Be O a A a ee Pw 236 14 1 2 1 Setting up a steady or unsteady flow o o 236 14 1 3 Interpolative schemes for cell face values o oo 237 141 31 Bound dnessicntetia e s s sess sawa eaa asas ee ae ed 237 14 1 3 2 Setting up the convective scheme o socia sacara oo e 240 14 1 4 Gradient computation as s do a Kd EA Se a we a 242 14 1 4 1 Setting up dimensionality and gradient calculation options 242 14 1 5 Discretisation of the generic equation o e e ee eee 243 14 1 5 1 Transient terfi 242404 h4G4 4 c4 004 Fo babe esha dieu 243 14 1 5 2 Convection term 4 8 0 2 che be ao edi A te ek Bad be eA EO 244 14 1 5 3 DifuSiomterm oo se aw uPA a ws Ba dae OG doa BG dw eae des 244 14 LSA SOURCE TEMM o c osad Sos wk oA ee wl ee ea wee 245 14 1 6 Discretisation of the energy equation o s e aoea ce tec need a E ee eee 245 14 1 6 1 Diffusion flux at interfaces for conjugate heat transfer 246 14 2 SIMPLE based solution procedure e 248 14 2 1
275. draulic of a channel respectively The molecular viscosity of fluid is denoted as u Rox kg ms 11 56 11 3 3 General orthotropic model The general orthotropic model also enforces uni directional flow in direction of the first principal axis Figure 11 3 left The viscous resistance is neglected and the inertial resistance is calculated Ricardo Software December 2009 193 11 MODELLING POROUS MEDIA 11 3 POROUS MODEL TYPES AND USER INPUTS Porosity 1 Porosity 1 Porous Material Type Orthotropic Porous Material Type Radiator Principal Axes Principal Axes Direction Vector X Direction Vector X x 0 998377 Y 0081748 z o047271 X 0 998377 Y 0 031748 z 0 047271 Friction Coefficient 1 m 247 5 oo Vi ff kg ma 819 7 Inertial kg ms4 247 5 Heat Exchange Mode Heat Load ae ti i PET een E y Heat Exchange Mode Heat Transfer Coefficient a_l o a_2 23303 a_3 E 885 Heat Transfer Coefficients a4 0 as o a_6 0 a_1 0 a2 23303 a_3 1885 Heat Exchanger Data a_4 0 a_5 0 a_6 0 Heat Load W 0 Heat Exchanger Data Heat Exchanger Temperature K 385 Heat Exchanger Temperature K 385 Minimum Temperature Difference K s Minimum Temperature Difference K 5 Figure 11 3 R Desk Sub domain panel User inputs for the orthotropic left and radiator
276. e Refinement button in the Boundary Painting panel in phasel as shown below Ricardo Software December 2009 57 3 GEOMETRY 3 15 BOUNDARY PROCESSING loxi File Edit view Toolbars Operations Help Auto Stitch Ctrl U Check Self Intersection Check For Unstitched Decimate Triangles Make Geometry Set Model Time Boundary Painting ij Boundary Painting xj Boundaries Show Ti Type Refine Yes 1617 Inlet Outle No 71 Vall 4 78 Wall No 5 94 Wall No Delete Boundary 1 Show All Hide All Toggle Compress Reduce Paint All Auto Paint Paint Line Motion Info Y Boundary Refinement Specification for bounday2 _ lt gt Delete Refinement depth at boundary EE Refinement blending distance jl Refinement Blending Blend to boundary depth Blend to boundary depth 1 Auto Paint Angle as Refinement Specification Destination Save specification in triangle file Save specification in mesh file aa ee Wi Opening the boundary refinement panel This brings up the following panel shown to the left The Refinement Depth specifies the refinement depth to be used in the global cells which are next to a particular universal boundary The Blending Distance specifies an integer which is used to control how the refinement at the boundary blends into the refinement level of the surround
277. e Surface Offset Magnitude feature enables an offset to be applied to the wrapped geometry The offset is based on moving the triangles along their normal direction by a defined distance If the input value is zero no offset occurs In case of the positive value the wrapped surfaces grows If the offset distance is negative it shrinks The allowed range is fr fr To obtain a larger offset the wrapping process with the offset option should be repeated several times It is recommended to use the decimation angle and decimation distance option for element reduction when using the offset option In the case of leak occurring during a wrapping process the surface offset can be used for quick hole capping The following operations are suggested Wrap the geometry with a coarse fr size without leaks applying a small negative offset distance Merge the wrapped geometry with the original geometry Repeat the wrapping process on the combined original and coarsely wrapped shrunk surface with a fine feature resolution size The Surface Thickness option can be used to wrap infinitely thin surfaces 1 e shells It works by creating a thickness for infinitely thin surfaces For more accurate and faster results it is strongly recommended to apply the surface thickening option to the parts of the original model that are infinitely thin as a separate wrapping process then merge the thickened parts with the original geometry and run the Wrapper witho
278. e 119 7 MODELLING SPATIAL DOMAINS 123 7 1 Mult Domam Approach 4 6 se a estea ga OE ee e a a 123 Ta Boundary REDIONS soson ha ha a a OE RE Ae bee RAE SESS 125 wo Interface Regions o soies 20 4 BMG ord ek ee Adee WS RE RA ee ee eon 125 4 Ordering of Domain Components ssas s eean A Bal eee ew 126 7 3 Creating a Domain Structure is ae a eras al BO We ee oh SR Redd 126 8 MODELLING CONTINUA 130 S Constitutive Relations s ea o y 242 do e ea ek a E eo en 130 8 1 1 Momentum transport in fluids and molecular viscosity 130 8 1 2 Energy transport and thermal conductivity o o e e 131 8 1 3 Mass transport and mass diffusion coefficients o o 131 92 EBquauon Of State Voir ad bed ek aed be e 132 8 2 1 Coefficients of expansion and compressibility o o 133 8 22 SPENE MEAL oi alg oy a A get go a Re at ies ate oh 133 8 2 3 Specific thermal energy and enthalpy o oo oteak au 134 8 3 Thermally Perfect Fluids lt i secs a so aoea dad e e a e da 134 8 3 1 Incompressible substance model o s s ep s scs woe dosd ae oo e 135 8 3 2 Ideal gas modele s a doo d ip a E a A A 135 8 4 Properties of Multicomponent Phase o e a e e ee 136 SH Reacting MIXTE ROWS o s aiaia a ir E a AA bo ee A 137 Ricardo Software December 2009 iv 5 Properties of Multiphase Mixture sos ae a a a A le we Bae a 137 8 6 Selecting Continuum and
279. e an ewe es ee we a ee a Ewe bee BS a Ow ae 161 Characteristic flow regions for a circular cylinder a steady separation b unsteady separation vortex shedding oda de be ta ee ee OS A A A ee Ge Yow 164 R Desk setup Setting turbulence modelling and near wall treatment top modelling ap proach options left bottom and k model family variants right bottom 165 R Desk setup List of equations solved for in VECTIS MAX 168 R Desk setup Selecting dimensionality of the simulation in Di R Desk setup Setting a type of a fluid phase and its compressibility in Fluid Phase 172 Ricardo Software December 2009 xviii LIST OF FIGURES LIST OF FIGURES 10 4 R Desk setup 10 5 R Desk setup 10 6 R Desk setup 10 7 R Desk setup Setting a steady or unsteady simulation in the Glo 175 10 8 R Desk setup Setting body force buoyance fluid model in Fluid Domain 176 10 9 R Desk setup 178 10 10R Desk setup Inclusion of viscous heating in the solution of the energy equation 179 10 11R Desk setup Setting solution of the energy equation for solid domains in the E A A E E Hpk SOR deem his She needy Maree E 180 11 1 R Desk Su 192 11 2 R Desk Sub domain 193 FLS 194 12 1 Typical multiphase topology flow boiling in a heated tube with a subcool ed inlet and volume fraction distributions for four fields cl continuous liquid cv continuous vapour dl dispersed liquid dv
280. e angle between the blade s normal and its direction of motion as shown in Fig 16 4 Eq 16 2 can then be plugged into Eq 16 1 to give Fo PfanVaxialA or el Vaxial tan B 16 3 where P fan is the average fan density Assuming the total force to act along the blade s normal we can calculate the axial component as shown in Eq 16 4 Faxial Fotan B 16 4 The cartesian forces Fy y experienced by a given point in the plane of the fan can be resolved by using its radial position r and corresponding x y components Xqist Ydist See Eq 16 5 and Fig 16 5 F Z apoi rL Xdist E e 16 5 rL These linear forces are then assimilated as momentum sources In addition these forces are scaled by a source factor in such a way that the fan s operating point is eventually found Ricardo Software December 2009 274 16 MODELLING FANS 16 3 ID MODEL tf Outlet Blade end Direction of motion tf Inlet Figure 16 4 Top view of an axial fan Shroud Direction of Hub Figure 16 5 Front view of an axial fan shrouded 16 3 1D model The model represents a fan using a pair of coupled inlet outlet boundaries so that the body of the fan itself lies outside the VECTIS flow domain and the flow generated by the fan is imparted via the inlet outlets The model can be used to simulate axial fans in which the flow enters and leaves parallel to the direction of the fan s axis
281. e are a number of factors that affect stability and numerical accuracy A number of criteria can be used to assess the quality of a computational mesh The following checks are performed in VECTIS MAX see Figure 4 15 for mesh checks 3 4 5 and 6 1 Negative cell volume If this is the case then the user is warned and the appropriate warning message displayed 2 Cell enclosure If the sum of cell face vectors for each cell is not zero or very close to zero 10710 then the cell is not completely enclosed by its faces 3 Wall distance ratio for near wall cells it represents the ratio of the wall normal distance and an average normal distance of cell neighbours to the same wall 4 Non orthogonality angle between the face vector A f and the distance vector d connecting two neighbour cells lt 75 This introduces the cross diffusion term and modifies the discretisation of the diffusion 5 Warp angle angle between surface normals of the triangulated cell faces lt 50 Ricardo Software December 2009 104 4 MESHING 4 12 MESH QUALITY CHECKS 6 Volume change ratio ratio between a cell volume and the maximum volume among its neigh bours lt 2 Criteria 3 4 and 6 are used to identify poor quality cells The user is informed of the total number of poor quality cells xxxxxChecking quality of cells there are 20 poor quality cells The way such cells are treated makes the subject of section Poor Quality Cell Treatment
282. e boundaries of the geometry are surrounded by the Cartesian Ricardo Software December 2009 4 3 GEOMETRY 3 14 GEOMETRY WRAPPING cells base mesh having no intersections with the geometry External cell faces of the base mesh are projected onto the boundaries And after that the surface triangles are decimated in order to reduce their number Using this technique small holes in the initial geometry over lapping self crossings are allowed and geometry triangular connectivity is not important The geometry wrapper can be used in the GUI interactive mode or in a batch mode with the wrapper settings defined in the GUI or from a parameter file NOTE that some features may only be available by using the parameter input file method since the GUI may not support them yet For example in VECTIS 3 9 0 the leak detection mode can only be used with the input file option 3 14 1 GUI Interactive Mode The Wrapper is invoked by selecting the phasel Operations gt Geometry Wrapper menu item from the which displays the wrapper input parameter panel Ricardo Software December 2009 Operations Help Auto Stitch Ctrl U Harmonize Normals Check Self Intersection Check for Unstitched Geometry Wrapper Decimate Triangles Make Geometry Set Model Time Boundary Painting Mesh File Setup Generate Mesh Slice Mesh View Radiation Setup 42 3 GEOMETRY 3 14 GEO
283. e can write AU pY AUN pYy Un 14 76 where AU p and AUy p are given as AU p Uy p Ui p and AUy p Uw p Un p respectively The above equation defines the wall parallel velocity ratio Q y as _ AUip _ Un YU AUNp n pi YU Qiu 14 77 The direction of the parallel velocity at i is taken to be that of the parallel velocity at the neighbour node N For the wall normal velocity the parabolic profile is physically sound 5 i ae AU n AUN n 5 with AU n w 0 fi AUN n w 2 Un Ti 14 78 N The resulting velocity vector at intermediate nodes can be expressed as U 0 UN 1 Qu Ow oiu 81 31 y ny 14 79 A similar procedure employing the wall effective conductivity A Equation 9 40 and assuming the constant heat flux qy across the near wall region leads to an expression for the temperature T Qir Ty 1 Qi 7 Tw 14 80 where 7 denotes the temperature ratio _ TIw T _ 6 AwPrn YNT Tw Ty Oy APr YT Qi T 14 81 and the dimensionless temperature distribution T should be known in terms of the Y see Equa tion 9 43 Ricardo Software December 2009 255 14 NUMERICAL SOLUTION 14 5 PARALLELISATION The main difficulty here is how to determine the turbulent kinetic energy k at intermediate nodes in order to evaluate the non dimensional wall distances Y in particular 1f the value of k at these nodes is around the maximum A crude appr
284. e eh ee ees 4 11 1 Meshrequirememt 2 64 244M wd a a OE ewe eee ee eo 4 11 2 Importing a grid data file eee ait eae ge ape eR en Be ee ee e 4 12 Mesh Quality Checks 4 0 5 s 4 ais amp ed baie alge lee woah amp eal ak e als 5 READING amp MANIPULATING MESHES 5A Introduction e soirs a ie A ee ar eee ar eee asd ew RL A a a worn 5 2 Meshi Partitioning somo ios See Bo al ae A ae ha we ee Oe A ae a 5 3 Me shJoming i arica ies ye eda Pha EG Pe eae eR ees ae 5 3 1 Sub d0maim JOM NE s a s eee e we a a a ee e E SA Restarting Partitions e os 224 4 da4 bea eee oe hak ee ea a as 6 SOLVER FUNDAMENTALS Ricardo Software December 2009 78 78 78 79 80 84 87 87 89 91 94 95 99 99 101 102 102 103 104 106 106 107 108 110 110 112 iii 61 IMFOdUCUON e s ss a oh a A A OB he a Be Bea 112 6 2 Continuum Conservation Equations o s e cae ew ee ee ee e 114 6 2 Terminology 2 24 4644 42464 26424 05 deeb eae de Rah Eo bd 114 6 2 2 Instantaneous equations under eee ae ae ee wo 115 6 2 2 1 Mass Conservation os 246 4494 oP SHOR DEDEDE S Owe ES a 115 62 2 2 Momentum Conservation 240 4 curada ee es 116 6 2 2 3 Enerey Conserva s sa eae Sr de ane eR ee aa A a ae 116 6 2 2 4 Spaceconservationlaw oo is eo cree o eee ee ee ee 117 6 3 Closure Problem and Averaging o o cacca eooni 54 ronda a ee 117 6al Reynolds and Eavre averaging xl dad Bod he eae ee E 118 6 4 Reynolds Averaged Equations e
285. e green mesh subdivision lines that may also be defined in mesh set up mode to appear in a 3d view In Mesh Set up Mode the IJK Block Display options can be used to switch off the IJK Block display off display the current IJK Block only solo or display all the IJK Blocks all User Defined Mesh Set up The mesh set up view options can also be set in the Vectis cfg file which is read at program start 3 17 1 IJK Refinement Blocks A panel is provided on the mesh set up panel for the setting of IJK blocks that control the level of mesh refinement in different parts of the mesh To use this panel select the mesh tab or the mesh set up button on the button bar or select the IJK Navigation option from the Operations menu m WK Refinement mj n n ooo Add Delete Edit Refinement Parameters DEEP fo FORCE 0 IJK Refinement block settings The Add button on this panel allows a new IJK block to be created When the left mouse button is clicked on the canvas a rectangle is drawn between the point chosen and the current cursor position Clicking the left mouse button again defines the extent of the IJK block To completely define the IJK block a different 2d view must be selected and the extent of the block in the third dimension defined This may be done by clicking the Edit button and dragging out the extent of the IJK block in exactly the same way as before The current IJK block is shown in yel
286. e interfacial area between fluid and solid inside a REV and n is the unit outward vector normal to A The above surface integral introduces into volume averaged equa tions additional dispersion terms which describe microscopic interactions between fluid and solid phases The transport theorem relates temporal derivatives UN one dg dy ot 7 f Pida or SE e EN n nuda 11 8 where U is the fluid velocity vector at fluid solid interface For non moving interfaces the surface integral vanishes 11 1 2 Double decomposition concept Note that volume averaging can be applied to the instantaneous quantity f 0 0 6 6 or to the ensemble or Favre averaged quantity Also the ensemble Favre averag ing can be applied to the volume averaged quantity In this way one can arrive at the double decomposition concept for an instantaneous variable as shown by de Lemos 2005 d 9 9 e ee erie 11 9 9 9 0 6 0 0 11 10 In the case of the rigid porous medium the above averaging operators commute 6 0 6 6 0 6 11 11 11 1 3 Governing equations Macroscopic equations describing turbulent flow in a porous medium can be derived in two ways either applying ensemble Favre averaging and then volume averaging or using a reverse order of these averaging procedure As long as the averaging operators commute the final form of the mass momentum and energy equations will not be a
287. e local arrays to global array on partition 1 if i_part gt 1 deallocate xval free local arrays for y coordinate case icp_beg_time lbeginning of each time step case icp_end_iter lend of each iteration store temperature value for each partition n 0 do idom 1 n_dom idt eq_idt iene idom iget_dom call get_field idt iget_cell temperature tfield do ic iscd idom iecd idom if abs xcc 1 ic 0 5E0 lt 2 0E 3 then n n 1 tval n tfield ic endif enddo enddo call concat_array dsize tval concatonate local arrays to global array on partition 1 partition 1 writes data to file if i_part 1 then call get_run_ctrl project_name proj_name call get_run_ctrl proj_run_number proj_num filename proj_name 1 len_trim proj_name UPR proj_num open unit iupl file filename status replace amp access sequential form formatted do n 1 gsize write iupl 2e18 8 xval n tval n enddo close iupl endif ORicardo Software December 2009 365 18 USER PROGRAMMING case icp_end_time lend of timestep case icp_end_run lend of simulation free memory deallocate tval deallocate dsize if i_part 1 deallocate xval end select end subroutine upr_generic Ricardo Software December 2009 18 7 EXAMPLES 366 19 TUTORIALS These tutorials contain simple examples chosen to help demonstrate the typical usage of the soft ware whilst attempting to
288. e options This tool compares boundary faces of common parts of the two given meshes and performs some actions to ensure their conformality actions like movement of node removal of node split of edge removal of edge removal of face or split of face An example of meshing of a multidomain case consisting of two parts internal and external can look like this vmesh internal mesh int 12 1 3 vmesh external mesh int 12 13 L2 vmesh conform internal GRD 1 13 external GRD 12 13 When these commands are completed new gridfiles internal_conform GRD and external_con form GRD should be created The boundaries forming the interface of the two grids should be conformal 4 9 Warnings and Errors There is a list of warnings and errors which can occur during the run of VMESH The majority of them can appear only under very special circumstances Great deal of them are caused by problems in the input geometry If it happens that a problem occurs but the geometry in the reported global box is without flaws it might point to a problem in the algorithm In such a case supplying the problematic geometry to Ricardo Support Centre is recommended ERROR 1000 Fatal problem with IN OUT test There is a problem with in out test The input geom etry is probably not clean enough ERROR 1001 IN OUT test inconsistency detected There is a problem with in out test The input geometry is probably not clean enough ERROR 1002 Fa
289. e shall run it from a command window Type the following command to run the tivpre mesh pre processor Ricardo Software December 2009 431 19 TUTORIALS 19 3 COOLANT FLOW Bnd_Reg_1 Wall Boundary 28 x bnd_reg_2 Mass Flow Rate Boundary 2181 x Bnd_Reg_3 Pressure Boundary 718 x Region Name End_Reg_1 Region Name Bnd_Reg_2 Region Name Bnd_Reg_3 Regon 1D 1 Regon 1 2 Region 10 3 Material ID E Material ID 1 Materia I E I Boundary Report F Boundary Report IV Boundary Report Coupled Link Number 0 Coupled Link Number 0 Coupled Link Number 0 Boundary Condition Type wal y Boundary Condition Type Mass Flow Rate Boundary Condition Type Pressure Boundary Condition Op Prescibed Temperature z Boundary Condition Op Normal velocity Scaied 2 Boundary Setting Unform voues a Boundary Setting Urform values X Boundary Setting Unform values Roughness Height 0 Mass fiow Rate 2 02 ln Inflow Outfiow Outfiow a Roughness Constant 0 5 ielocity Direction Wal Velocity s pp pp z x fo y o z o Phase_1 Boundary Phase 2 a xj Hess ro JO y La Phase_1 Boundary Phase EEE mco Temperature o Pressure 100000 Radeto Temperature 373 nen A naaa Turbulent Intensity 0 1 0 0005 F Thin wal Figure 19 81 Boundary region settings vpre coolant GRD vpre 1s also be used to re partition the mesh for parallel calculations using the np flag 19 3
290. e should always be a complete global mesh cell beyond the limits of the model in all six directions Asx Asy and A z Mesh lines can only be defined while the viewing mode is one of the three orthogonal views To add a new horizontal or vertical main line click on the button with the appropriate single line El insert a vertical mesh line El insert a horizontal mesh line A line of the required type will become bound to the cursor and clipped at the limits of the existing mesh Once the mesh line is approximately in the right position clicking the left mouse button will place the mesh line Multiple mesh lines may be placed in this manner To exactly place a mesh line once it is on the canvas click on the Place Mesh Line button El and type in the exact co ordinate To specify the number of divisions between two main lines the appropriate subdivision button should be selected EJ create vertical sub divisions create horizontal sub divisions When the mouse cursor is moved over the model a shaded area will appear between the main mesh lines that bound the cursor When the left mouse button is clicked a panel will appear to allow the number of subdivided cells that should appear in that shaded area as shown in the figure to the left Ricardo Software December 2009 63 3 GEOMETRY 3 17 MESH SET UP VIEW OPTIONS Number of Cells Between Lines Number of cells 4 Lenath of cells fo 00881667 Cancel Wi
291. e type option for universal boundaries 2 and 3 should be set to Inlet Outlet by right clicking the mouse on the row type entry for each boundary The final setup is shown by Figure 4 1 Use the Paint Face tool to mark the Inlet Outlet Boundaries 2 Click on the type box to toggle to the correct boundary type After painting the boundaries save the geometry file as coolant tri The next stage is to setup the control mesh for the coolant jacket analysis Activate the Mesh Setup toolbar as shown by Figure 19 66 Phasel automatically positions the two external red control lines around the geometry In general it can be beneficial to re position these lines closer the the model extents as shown by Figure 19 67 This should ensure that the aspect ratio of the cells adjacent to these lines is as good as possible This can be done using the movement icon or by creating new red control lines and deleting the original red lines Use the mesh division lines to position the red control lines closer to the model extents and add the gasket control lines Ricardo Software December 2009 419 19 TUTORIALS 19 3 COOLANT FLOW Y Ricardo VECTIS Phase 1 coolant tri 101 xi Fie Edit View Toobars Operations Help Baga seal A N a x Yl Options Stitch mesh Triangles Al a Al lola VIE Parts Boundaries Show Thi Type Refine 1 Ye 16472 N E gt ARA 2j
292. eady State solver O Add Monitoring Points Use arbitrary surfaces The user should also be aware of boundary identification and the input parameters used for the solution of the simulation 19 2 3 Geometry Preparation Phase 1 is the current pre processing package for use with VECTIS MAX Phase 1 is used to read in geometry and process it to a form acceptable to the VECTIS MAX mesh generator Depending on the quality of the initial geometry a certain amount of repair may be needed to form a single closed volume Once this is complete Phase 1 is then used to identify regions of the geometry as different boundaries for the CFD calculation Finally the global mesh and other control parame ters for the mesher are defined Further information can be found in the GEOMETRY chapter In this tutorial the geometry is fully stitched so no further manual repair is necessary The user can now enter the boundary identification section of Phase 1 The purpose of this section is to define different regions of the surface as different boundaries In this example the user needs to distinguish the different regions that will become the walls and inlet outlet boundaries Ricardo Software December 2009 387 19 TUTORIALS 19 2 STEADY STATE PORT FLOW 19 2 4 Selecting Boundaries The boundaries can be selected either automatically or manually For a geometry as simple as this it is recommended that the boundaries be defined manually To do th
293. eal precision are defined Both these precision format UAV can be used directly from UPRs Precision format Variable Description Type iwp wph integer working precision 32 bits real working double precision integer parameter integer parameter Table 18 1 Integer and real working precision format in VECTIS MAX Access groups All user programming routines are classified into nacc_grp groups Hence each access group name in Table 18 2 identifies a UAR These access group names are not used in UPRs but are mentioned here to illustrate the structure of user programming in VECTIS MAX For example iacc_numbers is associated with the UAR get_number Access groups Name Description Type len_var_name len_warn iacc_number iacc_domain iacc_mat iacc_phase iacc_spec 1acc_ps iacc_reg iacc_property iacc_grid_geom iacc_grid_connect iacc_turb iacc_run_ctrl iacc_field lacc_grp_name access variable name length 32 length of warning message 256 main numbers domain variables material variables phase variables species variables passive scalar variables boundary amp interface regions access properties grid geometry variables grid connectivity variables turbulence constants run control variables solution variable fields lists names of access groups currently there are 14 access groups integer parameter integer parameter integer paramet
294. earch Default ntcs 3 V Shows version of this program h Shows this help When radprep is run in this way the program does the following 1 Reads the radfile and finds following information O name of mesh file O name of patfile which should be created O universal boundary numbers for the boundaries which should be taken into account for radiation and requirements for the super patch generation for each of these boundaries 2 Reads the VECTIS phase4 mesh file 3 Reads the phase4 mesh surface patch connectivity file CON confile If the confile does not exist radprep determines the connectivity of the phase4 mesh surface patches and generates the confile 4 Assembles the required number of super patches for each boundary as defined in the radfile 5 Assembles the connectivity information for the created super patches 6 Determines the surface conduction connectivity 7 Determines the normal conduction connectivity 8 Writes the patfile radfile meshfile name of meshfile mesh infomation name of confile o name of patfile confile patfile boundaries affected stored superpatches by radiation required connectivity information number of supepatches foreach boundary RADPREP Figure 15 5 Scheme of the radprep program See below for the description of minimum angle and maximum angle Ricardo Software December 2009 262 15 MODELLING RADIATION 15 3 RADIATION SETUP NOTE that t
295. eat flux due to species diffusion in Equations 10 26 10 7 is not currently implemented Some other terms on the right hand side of the above equations can be neglected in some flow situations Typically pressure work and the work of stresses i e viscous dissipation or heating term are often negligible in incompressible flows However VECTIS MAX always account for all terms except for the viscous heating term This term should be included when the Brinkman number Br _ HU 10 27 TAT oj eee Br is close or above one Ricardo Software December 2009 178 10 MODELLING SINGLE PHASE FLOWS 10 4 MODELLING HEAT TRANSFER Discretise o K Limit Gradient HA Pressure Switch 0 1 v Limit Cross Diffusion Enhance Stability Viscous Heating Terms In Energy Equations Bad Cell Treatment All Variables Except Pressure Correction v Solve All Phases l Figure 10 10 R Desk setup Inclusion of viscous heating in the solution of the energy equation The inclusion of the viscous heating term is done via Discretise panel Figure 10 10 where the Viscous Heating Term in Energy Equation can be ticked on 10 4 2 Heat transfer in solids In solid regions the heat conduction is the only mode of heat transfer and the total energy equa tion 10 26 is reduced to the simplified internal energy form dpe oT a a P9v 10 28 where internal energy of a solid is defined as e Fr CyT Currently t
296. ed Cartesian coordinate system x y z as it is shown in Figure 14 1 which illustrates a general control volume An edge is a staright line which is formed by connecting two neighbouring vertices The line connecting vertices V and V2 in Figure 4 14 represents and edge A face is defined in terms of vertices and forms a plane in 3D In 2D a face is simply an edge In Figure 14 1 a face is formed by connecting vertices V1 V2 V3 and V4 A cell is defined by a number of faces that forms a closed volume in 3D A number of cell shapes are given in Figure 4 14 A face which is shared by two adjacent cells is an internal face otherwise it is a boundary face Thus a face is usually almost planar simple connected polygon which can be uniquely defined by the corresponding list of vertices In this list the face vertices are ordered to obey the right hand orientation with respect to the face normal vector A f In other words the vertices are ordered in the counter clockwise direction when seen from the direction which is opposite to the face normal In VECTIS the face normal is always directed from a cell with a lower index to the cell with a higher index with reference to the above figure j lt 1 for the face f In case of boundary faces normals always point out of the adjacent cell The building blocks of the computational domain are O global domain O fluid solid domains DO interfaces and boundaries The global
297. ed arguments e g call get_field idt vect_field velocity fi_c v fi_c_o vold would return velocity amp velocity old from previous time step at cell centres into user defined pointers v amp vold respectively The preceding example pulls back pressure values for cells p boundaries pb upper pi amp lower pli material interfaces This is followed by calls to get the reference pressure and molecular weight The call to get_property rphase_values rph_val returns the property array in rph_val which can be indexed using the property id s as listed in Table 18 5 Once the temperature is found the density at the cells are modified thus do ic icell icel2 fi_c ic p_ref mat p ic gcex t ic end do Where icell icel2 which are passed down into this routine represent the starting and ending cell index for this material domain mat The following 3 if blocks for this density case block modify the boundary and material interface values in a similar way The boundary and interface arrays are optional passed down depending on the variable Property boundary values only density conductivity specific heat capacity amp laminar mass diffusion multi species can be changed For material interface values only the density can be modified 18 7 2 Example of the user initialisation routine Ricardo Software December 2009 357 18 USER PROGRAMMING 18 7 EXAMPLES Description User programming
298. ed to start vsolve Click on the button to open the Launch vpre dialog box Insert the run directory and the name of the inp file Fig ure 19 58 2 amp AB Feo Fe Launch Solver Dialog ax Number of Processors 1 Run Hosts localhost 2 om Figure 19 58 Launch Vsolve The solver will start 19 2 13 Live Update Live update is a utility in R Desk that allows a simulation to be monitored whilst it is running Firstly open a xy canvas using the new XY canvas button Then in the Live Update panel browse Ricardo Software December 2009 412 19 TUTORIALS 19 2 STEADY STATE PORT FLOW to the directory where the simulation is running The available data files are presented in the Files window The data files available are determined by the ascii files selected in the reporting section of the 19 2 10 2Global Domain in the solver setup tree Once a file is selected the data available to be plotted is shown in the Value window The different data can be dragged onto a xy plot canvas The xy plot can then be further modified using the XY plot manager In this case the in plane swirl can be plotted for the arbitrary surface Click on the arbitrary surface in the files panel then drag the Iter_no entry onto an XY canvas followed by the Z Ang_Mom entry This will show how the swirling motion converges throughout the simulation In this case as the port is symmetric then the swirl around the cylinder axis
299. eft to right vmesh vpre amp vsolve The mesher El launch dialog Fig 17 20 allows the user to set the mesher input file we Launch Mesher Dialog x Mesh Input File Jusr2 scratch BOIL_CASES ANNULAR_nightly Launch Cancel Figure 17 20 Mesher launcher dialog box The vpre H launch dialog Fig 17 21 provides a GUI interface to the vpre utility Each element corresponds to the vpre arguments Sec 5 1 The mesh file is non optional The Partitioning GroupBox contains controls for generating parallel grid files Control Vpre Option Number of Processors np Metis Partitioning Method meth Repartition Restart Files rest Table 17 6 Partitioning controls and corresponding vpre options The Arbitrary Grid Interface GroupBox contains controls for joining multiple grid files multi domain E g for joining 3 grid files together the user would enter 2 then lt ENTER gt in the Number of Additional Mesh Files field and then the corresponding file names The Mesh Join Type corresponds to command line meth option The output file option is used to override the default output file name COALESCED GRD when joining meshes Ricardo Software December 2009 292 17 USING SOLVER 17 16 WAVE VECTIS CO SIMULATION F Launch Pre Dialog 2o x Mesh File Istore cappo pm RADIATION_VECTIS4 STAR_TESTHluid1 GRO A Partitioning Number of Proce
300. egions Each of the boundary regions can then be specified Region Name A name can be given to each boundary for reference The name can be referenced when using user programming routines Boundary report This specifies whether output data is written to the report files for this boundary Coupled Link Number When running coupled WAVE VECTIS the interface numbers for each cou pled boundary are specified here Boundary Condition Type Next the appropriate boundary type is chosen for each boundary Ricardo Software December 2009 38 19 TUTORIALS 19 1 BASIC TUTORIAL Figure 19 17 The Fluid Phase Panel Details of the different boundary conditions can be found in the BOUNDARY CONDITION TYPES chapter Here we will set Boundary 1 to be a wall with a fixed temperature of 450 degK Boundary 2 to a mass flow boundary with a flow in of 0 1kg s an initial pressure of 1 001 bar and temperature of 300 degK Boundary 3 will be a static pressure boundary set to 1 bar outflow and with a temperature of ORicardo Software December 2009 382 19 TUTORIALS 19 1 BASIC TUTORIAL Initial Condition 138 x Xey o Yeosyfo 1 o S Pressure 100000 Temperature 293 15 Turbulent Kinetic Energy 1 Turbulent Disspa on 1 ti C SsS Turbulent Viscosity 0 s i S S S S S Volume Fraction 0 Figure 19 18 Initial Conditions Postprocessing Output 2 a xi Figure 19
301. either fluid or solid material whose behaviour is governed by transport equations expressing basic conservation laws of physics for O Mass Momentum Newton s second law O Energy First law of thermodynamics and Entropy Second law of thermodynamics The spatial domain occupied by the fluid or solid continuum is defined as material domain if the continuum does not exchange mass with other material domains Material domains are separated from the external environment by domain boundaries and from other material domains by domain interfaces Our knowledge about physical conditions at domain boundaries and interfaces should determine the behaviour of the continuum within a material domain For the time dependent problems the continuum behaviour will also depend on the initial conditions Considering an individual trans port equation so called numerical boundary interface conditions need to be imposed at domain boundaries and interfaces These are fixed variable values Dirichlet conditions zero or specified gradient values Neumann conditions and mixed Robin conditions At the continuum level it is useful to introduce physical boundary or interface conditions along boundaries and interfaces Their role is to define and if necessary update in a physically correct way the numerical bound ary conditions for each transport equation Typical examples of physical boundary conditions are walls symmetry planes flow i
302. elected in the tree and is shown in the preview window 19 1 8 Solver Setup The input to the solver is specified using the solver setup tree and solver setup input panels The structure of the tree reflects the general structure of calculation Global domain containing general options for the calculation timebase etc Fluid domains containing data for phases species and boundaries Solid domains for materials and boundaries Ricardo Software December 2009 373 19 TUTORIALS 19 1 BASIC TUTORIAL k R Desk 3d1 AutoPreview tube GRD FE Fle Edit View Optons Window Heb la x IDO te Es vice E eaae x E2 E Global_domain_1 Global Domain Restart Control Timebase Output Material zo 1 Radaton El Fluid _1 Fid Domain T Boundary Report Monitoring Points Algorithm Turbulence Model Coupled Link Number 0 Discretse El Phase_1 Fluid Phase Boundary Condition Type inlet Given eloaty z Boundary Setting Junform vates z Bnd_Reg_1 Wall Bo Phase_1 Bnd_Reg_3 Inlet G Phase_1 Bound Interface Regions Sub domans Report Regions file D RPe training_tutorials V4 basic_tutorial tub e GRD Figure 19 9 Selecting a boundary in the solver setup input tree The contents of the in solver setup input panel will change corresponding to the selected entry in the solver setup tree Solver Setup Tree 28 xi E Solver Setup E Global_domain_1 Global Domain Restar
303. emperature tli gt null low interface temperature real wph pointer se pl gt null amp cell pressure ypb gt null amp bnd pressure ypi gt null 8 lupper interface pressure pli gt null low interface pressure real wp pointer p_ref gt null amp mat domain reference press vph_val gt null reference phase propty vals integer iwp pointer n_li gt null amp number of low interfaces ia_li gt null amp start address for low intf il_li gt null list of low interfaces real wph te gE gas constant character len 1en_var_name usr_obj_name user selected name for either lphase or species object integer iwp iget amp flag to get phase species index ziobj amp phase or species object index ripro amp index of property usr_obj_id lindex of phase or species object corresponding to upr_obj_name integer iwp ide amp domain type index for energy idp amp domain type index for pressure ieq amp equation index nic amp cell index rib amp bnd face index HL TA Imaterial interface indices character len len_var_name var_name Ivariable field name usr_obj_name Comment line below to diasble this routine usr_obj_name air if usr_obj_name return Find if the phase or species object is passed to if isp gt 0 then iobj isp iget iget_specs Ricardo Software December 2009 354 18 USER PROGRAMMING 1
304. en used this value eclipses the value set in EDGE_THRESHOLD which has the same meaning see section 4 3 ps_ncang degangle sets allowed level of nonconvexness of output polygons generated by polygon simplification routine inner angles of polygons can be 180 ps_ncang The default value for ps_ncang is 5 0 TEST GRID test tests all cells in the generated grid indexes of cells with the worst problems are reported In order to use this command the grid needs to be generated already The supplied filename is expected to be the name of the gridfile verbosetest tests all cells in the generated grid the indexes of all problematic cells are reported it is also possible to combine this switch with locate command see below verbosetest locate meshfile passes the name of the input ascii meshfile so as the verbosetest could report not only indexes of the problematic cells but also their IJK information BASIC VISUALIZATION TOOLS The proper tools for visualization of the mesh are imple mented in R Build However VMESH also contains some features which allow simple vi sualization of cells Zoom rotate and pan can be controlled in the same way as in Phasel There are several hotkeys which can be used switches off on visualization of lines of faces k shows vertices which are scalable by pressing and s switches on off light n toggles additional lines if drawn viewc CellInx reads VECTIS MAX grid
305. ent energy production due to body forces In order to close the compressible RANS equations models are required for Turbulent mass diffusion flux J A p Reynolds stress tensor tj and turbulent heat flux q Turbulent mass flux u Turbulent dissipation rate O Diffusion Dz and the pressure dilatation term p du Ox in the turbulent kinetic energy Equa tion 6 40 Production of turbulent kinetic energy by body force Pp O Additional terms in the total enthalpy equation related to turbulent transport of species da He The turbulent mass flux ul and associated compressibility terms in the turbulent energy equation including the pressure dilatation correlation p du Ox do not appear in the incompressible RANS equations Modelling of RANS equations is dealt with in the turbulence modelling part of the manual In the following sections and generally throughout this manual the Reynolds ensemble and Favre averaging operators and respectively will be used only for averages of fluctuating correlations All flow variables without operators will be considered as averaged mean flow quantities or instantaneous quantities in case of laminar flow where there is no turbulence Ricardo Software December 2009 122 MODELLING SPATIAL DOMAINS The solver design enables simulations of momentum mass and energy transport in multiple fluid or solid domains which are parts of the sing
306. equired node is not visible in the current view of the model the view may be adjusted with the mouse or keyboard as described above Delete Triangle Button Al Once this operation has been selected pressing the left mouse button with the crosshair cursor over a triangle will cause that triangle to be highlighted with a bold red fill This triangle is now the deletion candidate If the left button is clicked a second time over the same triangle the triangle will then be deleted If the escape key is pressed or if a different triangle or no triangle at all is selected the triangle deletion mode is exited and the Phase 1 returns to the Manipulate View state If a triangle is deleted Phase 1 remains in the triangle deletion state and further triangles may be highlighted and deleted Join Triangles Button This operation allows the user to merge a number of nodes together The user may select the centre point of a circular region by clicking the left mouse button Then the radius of the circle is defined by moving the mouse And approximation of the circle remains bound to the mouse cursor while it moves Clicking a second time fixes the size of the circle and performs the join operation Any nodes inside the circle and within a limited distance in the direction perpendicular to the screen are merged together to give a single point with the average co ordinate of all merged points This function should be used with care The user should onl
307. er integer parameter integer parameter integer parameter integer parameter integer parameter integer parameter integer parameter integer parameter integer parameter integer parameter integer parameter integer parameter character 1 nacc_grp Table 18 2 Access groups Ricardo Software December 2009 299 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES Various VECTIS MAX objects variable values can be obtained for various objects A list of these objects available in VECTIS MAX are given in Table 18 3 Various VECTIS MAX objects Variable Description Type iget_cell cell values integer parameter iget_bnd boundary face values integer parameter iget_uif upper material interface integer parameter iget_lif lower material interface integer parameter iget_face face values integer parameter iget_o cell old values integer parameter iget_00 cell old old values integer parameter iget_dom fluid solid domain integer parameter iget_mat fluid solid material integer parameter iget_phase phase integer parameter iget_specs species integer parameter iget_ps passive scalars integer parameter iget_breg boundary region integer parameter iget_ireg interface region integer parameter iget_part partition parallel run integer parameter iget_eq transport equation integer parameter iget_bctype boundary condition type integer parameter iget_mat_pro material properties integer par
308. er Residuals Information Figure 19 78 Output Options Output Here set the postprocessing frequency to a high value as data is only required at the end of the simulation Also set the Ascii report frequency to 1 so that the files are available for use with live update Additionally check the residual data options so that the residual values are written to ascii files and the rep sdf file Figure 19 78 Ricardo Software December 2009 429 19 TUTORIALS 19 3 COOLANT FLOW Fluid Domain In this case the calculation will consist of one single phase fluid domain which is liquid to represent the coolant The potential flow initialisation will be used to set up the initial conditions The reference data for density viscosity and specific heat should be changed to match the constant values found in the Fluid Phase setup which was imported from the old input file Fluid_1 Fluid Domain 18 Reference Pressure 100000 Reference Temperature 293 15 Reference Gas Constant 237 0 Reference Area 1 Boussnesq Approach Fixed Disabled Gravity Enabled Standard Approach Figure 19 79 The Fluid Domain Panel The remaining fluid panels algorithm turbulence model equations amp solver and discretise will be left at their default values Fluid Phase The fluid phase panel contains data for each fluid phase In this case the properties for the single phase should have been imported from the ol
309. erent refinement levels are illustrated in Figure 4 7 KJ YECTIS Phase1 3 6 pipej tri 0 x File Edit View Toolbars Operations Help Auto Stitch Ctrl U Check Self Intersection Check for Unstitched Decimate Triangles Make Geometry Set Model Time Boundary Painting Mesh File Setup DK Refinement Blocks Generate Mesh Slice Mesh View iY Mesh Generator Options Options to pass to the mesh generator Title Comment VECTIS mesh generator input file Model File Name pipej tri Refinement Depth 2 El le Exact fit at sharp edges gt Cancel Ll Figure 4 6 Definition of global refinement depth in Phasel 2 IJK refinement block allows to set different refinement level to a rectangular block of global boxes The block can be defined within Phasel1 the information is then stored to input meshfile Ricardo Software December 2009 89 4 MESHING 4 6 GENERATION OF BOXES Figure 4 7 Global box divisions with different depths of refinement The format is following IJK_BLOCK is ie js je ks ke ideep iforce For each this block defined as IS IE JS JE KS KE where S stands for start and E stands for end values IDEEP and IFORCE are defined IDEEP is maximum allowed refinement level IFORCE is a forced refinement level the global box is going to be split without testin
310. ergy containing eddies which are close to the largest eddies extract energy from the mean flow and start turbulence cascade where energy is transferred towards smaller and smaller eddies until final dissipation by viscosity The length and time scales of these large energetic eddies can be defined in terms of turbulence kinetic energy k and its dissipation rate e ee re 9 2 k 3 2 Ed 148 9 MODELLING TURBULENCE 9 2 OVERVIEW OF TURBULENCE MODELS FOR RANS The time and length scales are combined to define the characteristic Kolmogorov and turbulent velocity scales vg Ve 1 4 and v K12 respectively In addition the turbulence Reynolds number can be defined as E k 24 E pk SE Note that the turbulence Reynolds number based on the Kolmogorov scales is Reg 1 9 3 Re gt From the relation 4 Re it is clear that the range of scales increases with the turbulence Reynolds number or with the mean flow Reynolds number Re as Re Re l L and k 2 U One can estimate a minimum number of computational cells as Nee x 0 AN Re At present the DNS can be used for flows in simple geometries with Re up to 10 Another route which still requires too much computational efforts but the next best alternative to DNS is Large Eddy Simulation LES The LES resolves the large scale eddy motion in space and time while the sub scale motion defined in terms of numerical mesh requires statistical m
311. ers the flow resistance as for the general porous media The next section presents porous media modelling theoretical background including the governing equations Then currently supported porous model types are described as well as the required user inputs for each model type 182 11 MODELLING POROUS MEDIA 11 1 THEORETICAL BACKGROUND 11 1 Theoretical Background A fluid flow through a permeable structure is described by instantaneous Navier Stokes equations In practice ensemble averaging resulting with the Reynolds averaged equations RANS is em ployed to address the turbulence problem However the existence of the fine porous structure makes even RANS simulations computationally impractical and further volume averaging at an affordable macroscopic scale which accounts for the porous structure 1s required 11 1 1 Volume averaging procedure In the Volume Averaging Theory VAT see for example Whitaker 1999 Slattery 1999 grid elements control volumes contain both fluid and solid phases i e small scale solid structures and fluid pores are not resolved with a numerical grid Instead the numerical grid often associated with a Representative Elementary Volume REV should be fine enough to resolve macroscopic flow and temperature fields One can expect that governing equations describing these macroscopic fields account for interactions between fluid and solid parts within a REV Considering a general variable
312. ersection border depth the meaning of which is described below Clicking on OK starts the intersection check which may take some time for a large model A progress indicator shows how many triangles have been processed An example of the operation of the function is illustrated in the following figures which show a model consisting of an intersecting torus and sphere Ricardo Software December 2009 40 3 GEOMETRY 3 14 GEOMETRY WRAPPING Sphere and torus before intersection check After the intersection check the intersecting triangles are highlighted in red All other triangles are deactivated except for a context border which allows the user to recognize the self intersecting sections more easily The depth of this border is controlled by the intersection border depth entered before performing the check 3 14 Geometry Wrapping The VECTIS Phase 1 geometry wrapper generates an external triangular surface mesh around an existing geometry model simplifying the details and removing internal triangles It can be used as a clean up tool for reasonable size geometries However the user should be aware that deviations from the initial geometries can occur The wrapping process is controlled by several input parameters A new surface mesh is generated automatically however some manual repair can be required The wrapping algorithm is based on projection mesh generation methods The input geometry is placed in a cubic box domain Th
313. es The mesh structure connectivity data used by the solver as well as mesh quality checks are presented Reading and Manipulating meshes chapter describes usual pre processing tasks needed for the solver run This includes mesh partitioning in case of parallel runs enabling partitioning for the solver restarts and creating conformal meshes at fluid solid interfaces for multi domain CHT simulations All these tasks are done with the vpre module Solver Fundamentals contains mathematical background continuum conservation equations their closure problem and averaging leading to the Reynolds averaged equations Modelling Spatial Domains chapter provides information about multi domain structure and how to create this structure Modelling Continua and Their Properties describes thermo physical properties of fluid and solid continua including single and multi component phases The calculation options for various properties the corresponding equations and setting of the properties are provided 1 PREFACE 1 2 OTHER MANUALS O Modelling Turbulence chapter introduces the turbulence problem gives an overview of the modelling approaches and describes currently implemented models and how to select them O Modelling Single Phase Flows describes the available single phase physical models and their selection set up Separate sections present modelling of fluid flows heat transfer in fluids and solid as well as conjugate heat transfer
314. essure is the reference pressure location Its relevance depends on the existence of pressure boundaries in the considered fluid domain O If there are no pressure boundaries then the pressure value at one cell the reference pressure location must be fixed to the reference value in order to obtain the unique solution Patankar 1980 O If the pressure boundaries exist the reference pressure location is not relevant any more as the pressure level is set by specified boundary values Apart from the reference pressure the specification of the reference temperature T ef and density Pref 18 important for simulations involving the gravity force Setting of all reference variable values including the reference pressure location is done via fluid domain panel Figure 17 6 10 3 Modelling Fluid Flow Flow of any fluid phase can be modelled in accordance with various flow types The next sections explain different flow models and how to select them 10 3 1 Two and Three Dimensional Flows Naturally the fluid flows are three dimensional If the geometry boundary and initial conditions lead to the numerical solution where flow variables changes in one fixed direction are not signif icant such a flow can be considered as two dimensional VECTIS MAX supports planar flows taking place in the x y coordinate plane Mathematically the flow variables are functions of x y coordinates and time t f x y t The velocity component in the
315. esult and the remaining unstitched areas must be dealt with manually The tools used in this manual stitching process are available from the stitching toolbar This toolbar will appear when any model file is opened by the GUI All of the operations available from this panel may be reversed by use of the Undo option from the Edit menu or by the undo button e located on the button bar Phase 1 keeps a list of operations that can be reversed discarding the list whenever something is done that can not be reversed The Undo option is grayed out when there is nothing that can be undone The space available to store the information necessary to undo operations is limited and therefore the number of steps that can be reversed is limited 3 7 Triangle Creation Operations Create Triangle Button a On selecting this operation the cursor will change to a crosshair and instructions will be written to the information area If three different nodes are selected by clicking upon them and if these three nodes are valid vertices for a new triangle a new triangle will be created using these nodes Ricardo Software December 2009 22 3 GEOMETRY 3 7 TRIANGLE CREATION OPERATIONS The only limit to the number of triangles that may be created one after another is the number of nodes available to create valid triangles with While choosing nodes to form the vertices of the new triangle the nodes chosen will be marked with a white dot If a r
316. et_ps 18 3 ACCESSING SOLVER VARIABLES This UAR is used to retrieve information related to passive scalars mainly starting and ending boundary or interface regions see Table 18 13 Passive scalars are defined within fluid phases They are indexed in a similar way to species To obtain starting and ending addresses for boundary region wise values of passive scalar vari ables do the following integer iwp pointer illy 22002 call get_ps ise_ps_bnd_reg il i2 subroutine get_ps var_name il 12 Input Output var_name cha i1 integer pointer optional i2 integer pointer optional ise_ps_bnd_reg isa_ps_bnd_reg ise_ps_interf_reg isa_ps_interf_reg index of starting boundary region 11 1 n_ps starting allocation addresses for species at boundary regions 11 1 n_ps index of starting interface region il 1 n_ps starting allocation addresses for species at interface regions il 1 n_ps index of ending boundary region i2 1 n_ps index of ending interface region 12 1 n_ps Table 18 13 Subroutine get_ps to get values for variables defined in the access group lacc_ps mainly start end boundary interface regions for passive scalars 18 3 2 7 get_reg This UAR is used to retrieve information related to boundary regions species and passive scalar regions boundary conditions as well as starting amp ending indices of bound
317. ew Options Window Hep la x BrE S je 200 jac eh Ora Figure 19 96 Saving the sets to the GRD file One thing to note is that face sets are non exclusive in that a face be defined to belong to more than one face set However this is not the case for the boundary regions on a mesh If a face is found to belong to a number of different boundary sets the highest numbered boundary region will take precedence To avoid confusion it is recommended that this situation is avoided Use the deselect function of the set editing tools to remove unwanted faces from boundary regions Additionally all faces should belong to a boundary region and should not be left unpainted iz R Desk 301 p1 tube GRD 10 xj IFR Fie Edt View Options Window Hep la xi DAB S freee Some je 200r ac gt eh 8 7 180 gt gt gt CONE TE 1 03 42 WARNING boundary set Boundary_3 as Boundary_3 NOTE ti wil oleo affect e cotati boundary sets Reload jad the file to see these changes 1 03 44 INFORMATION Opening fie D RPejtraning_tutorials V4 mesh_mport tube GRD 1 03 44 INFORMATION Closing fle D RPetraining_tutorials V4 mesh_import tube GRD Figure 19 97 Reloading the GRD file Refresh the saved GRD file to visualise the the final boundary region definitions Right Click on the file name in the plot tree and select Refresh Figure 19 97 Select the different face sets to check that they are defined correctly
318. external convective heat transfer external radiation heat transfer and combination of external convective and radiation heat transfer In conjunction with any of the above thermal conditions the wall thickness can be taken into account using thin wall model O R Therm The wall thermal conditions can be imported via R Therm model i e using R Therm files Interface conditions For fluid flow equations with exception of the energy an interface con dition type is the same as the wall boundary type For the energy equation all interfaces are treated in an implicit manner i e no specific boundary types are required Basic Physical Models VECTIS MAX provides a range of physical models for the single or multi component phase in terms of Equation of state and compressibility Incompressible substance liquids gases and solids and ideal gas model are available All speed flow flows incompressible or weakly compressible subsonic and supersonic can be simulated OD Dimensionality Two and three dimensional simulations can be performed O Single or multi component phase For either single phase or multi phase fluid flow the fluid phase can be selected as a single component or multi component The multi component phase is a mixture of species Ricardo Software December 2009 6 2 INTRODUCTION 2 1 MAIN FEATURES AND CAPABILITIES Flow regimes Both steady and unsteady simulations of invi
319. f Ti P njdA E PUn 11 15 The last term on the right hand side of the above equation often called inertial dispersion is associated with spatial fluctuations of the ensemble mean velocities For gravity driven flows the double averaged p p and ensemble mean p pressure in the above equations represent the modified pressure which includes the reference buoyancy force see Equation 10 23 The macroscopic intrinsic and superficial viscous stresses 7 and 7 y Are defined as q 1 AS Ss ls n y 2u sy Sm y 1 Tj Tij Y Ti 2u si Sin 11 16 where the macroscopic intrinsic and superficial strain tensors S and S respectively read J ij a 1 fay ayu E dW Skk Sij SN 2 Ox ar Ox Si 5 Y and Sir Ox Skk y In the case of linear k models the turbulent stress tensors T jis given by Equation 9 4 After volume averaging its macroscopic counterpart becomes 2 a 1 Tij pU U My su wei 3 where Uy signifies the macroscopic turbulent viscosity It can be defined similarly to the mi croscopic turbulent viscosity see Equation 9 14 where now k u u u k and e denote the macroscopic volume averaged turbulent kinetic energy and its dissipation rate Ricardo Software December 2009 185 11 MODELLING POROUS MEDIA 11 1 THEORETICAL BACKGROUND Considering the isotropic porous medium closure for the porous resistance force Eq
320. f boundary conditions at the inlets and outlets In the spirit of SIM PLE we need to express the boundary mass flux and its correction as a function of the pressure and pressure correction respectively The mass flux corrections lead to the boundary pressure cor rection coefficient aj and upon solution of the pressure correction equation the boundary pressure and mass flux are corrected as pp p Opp tity m al pi Subsonic Inlet The total or stagnation boundary condition is valid for inlets Thus the total pressure p total temperature T and the flow direction i Up Up are defined and the pressure is extrapolated from the inside From the following isentropic relations y I Pt Pb 1 dete 5 Me 14 66 1 T T 1 ih imei 14 67 one calculates the boundary Mach number Ma and temperature Tp respectively Then the mag nitude of the velocity vector can be obtained from Equation 14 65 as gt 2YR T Y 5 HH 0 0 7 1 14 68 Differentiating the above equation with respect to pressureDemirdzic et al 1993 the mass flux correction at the boundary si p ip A 9 U 0p can be derived as follows 2y 1 R T y 1 I mau o SA b b 1 Assuming pj Pp the pressure correction coefficient aj can be easily defined from the above equation Ricardo Software December 2009 251 14 NUMERICAL SOLUTION 14 2 SIMPLE BASED SOLUTION PROCEDURE Supersonic Inlet If the flow is supers
321. f time and space Phase The physical state of continuum can be either fluid gas or liquid or solid and a phase is the thermodynamic definition for these states of matter see for example Moran and Shapiro 1992 More precisely the term phase refers to homogeneity in both physical structure of matter and chemical composition A phase can consist of one or more chemical components For example gases can be mixed in any proportion to make a single gas phase Only certain liquids e g alcohol and water can be mixed to form a single liquid phase while other such as oil and water are not miscible and form two liquid phases Note that in some cases the concept of a phase used in CFD can have a broader sense than the one used in thermodynamics Pure substance Another thermodynamics term is a pure substance which contains one or more chemical components with uniform and fixed chemical composition An example is Air which 1s the mixture of Nitrogen Oxygen and other gases Note that a pure substance can exist in more than one phase for instance liquid water and water vapour form two phases Ricardo Software December 2009 114 6 SOLVER FUNDAMENTALS 6 2 CONTINUUM CONSERVATION EQUATIONS Multi component phase A phase which has more than one pure substance will be called the multi component phase Species Here and generally in CFD a fluid phase component is often referred as species having the same meaning as the pure compressible substance
322. face which is not the case with total enthalpy Total enthalpy experiences the jump discontinuity which need to be addressed when calculating the diffusion flux To use temperature as the primary variable we need to express the convective flux Equation 14 32 in terms of Ricardo Software December 2009 245 14 NUMERICAL SOLUTION 14 1 FINITE VOLUME DISCRETISATION temperature In this equation the second term which represents contribution of the high order convective schemes should remain as it is i e defined in terms of total enthalpy The first term is based on the upwind scheme UDS and this term contributes towards coefficients of the discretised equation Considering an iterative solution procedure it can be re arranged as l n n cups m 0 4 Tp 4 8 d pint 14 39 P a5 p a5 p Recognisably a p and a p are convection coefficients associated with values of temperature at neighbouring nodes P and P With the help of the discrete continuity Equation 14 7 and following Patankar 1980 corresponding contributions to the central coefficients can be derived as l Hp n l Hp ap p max 1i1 0 7 ap p max va 0 32 14 40 Regarding the unsteady term d d d UU i pvH T PVcpT d pv gt 1 14 41 the first term containing temperature is treated implicitly while the second term is absorbed ex plicitly into source Similar treatment as in the abo
323. ffected by order of averaging Thus we can start with the governing equations of resolved flow Eqs 10 1 10 2 10 5 and 10 7 and perform spatial averaging over a representative control volume Considering a rigid porous medium non moving control volumes with the no slip condition on the fluid solid interface and denoting double averaged variables without double average notations 6 conservation equations describing porous media can be derived as Ricardo Software December 2009 184 11 MODELLING POROUS MEDIA 11 1 THEORETICAL BACKGROUND O Mass conservation J J J YP a U ma U j 11 12 I Tag V Tega Wt 11 12 The last term in the above equation represents mass dispersion This term is neglected as spa tial fluctuations of the ensemble mean density can be considered small in comparison to the volume averaged ensemble mean density i e P P p P Neglecting mass disper sion effectively eliminates a large number of density related terms arising from the volume averaging procedure Consequently the volume averaged ideal gas law will keep the original functional form p P Ry PT Re D F Re PT Re D F RgPT 0113 O Momentum conservation 0 E 2 ypU ypUU IP p fit vy t 4 hp 11 14 Ot dx ETA xj where fp represents the resistance force per unit volume to flow in the porous medium This force is given by d fo gt
324. fi_c ic p_ref mat p ic gext ic end do Boundary values if present fi_b then do jb jbndl jbnd2 fi_b jb p_ref mat pb jb gc xtb jb end do end if Upper interface values if present fi_ui then do ji jiul jiu2 fi_ui ji p_ref mat pi ji gc ti ji end do end if if present fi_li then To define low interface values we need a number of low interfaces n_li mat initial address ia_li mat the list of low interfaces 11_1i call get_mat n_low_interface n_li call get_mat isa_low_interface ia_li call get_mat 1_low_interface 1_1i do 3j1 1 n_1i mat ji 1_1i ia_1li mat 31 fi_li ji p_ref mat p1i 31 gex xt1i 31 Ricardo Software December 2009 355 18 USER PROGRAMMING 18 7 EXAMPLES end do nullify n_li ia_li 1_li end if nullify p pb pi pli nullify p_ref rph_val nullify t tb ti tli case lam_viscosity Boundary and interface values are not modifiable If temperature field is required call get_id energy ieq iget_eq Get index of energy equation ide eq_idt ieq iph iget_phase lGet energy temp domain type Get temperature field var_name temperature call get_field ide scal_field var_name amp fi_c t fi_b tb fi_ui ti fi_li tli Define property values as a function of temperature pressure or both temperature and pressure fi_pro f T p_absolute do ic icell icel2 fi_c ic 1 7
325. field call get_field idt iget iget_bnd var_name eff_viscosity amp fi_out vis_b Get phase domain index idt eq_idt ifvf idom iget_phase Get boundary volume fraction call get_field idt iget iget_bnd var_name phase_vol_frac amp fi_out vol_frac_b get boundary face surface vector call get_grid_geom var_name bndf_surf_v bfs_vec get of normal distance from the near boundary cell centre to the boundary face call get_grid_geom bndf_normal_d dnb get starting amp ending boundary face for each boundary region call get_reg ise_reg_face jsbc jsbc Calculate local pressure and viscous forces for all boundary faces belonging to boundary region ir do j jsbc ir jebc ir ic 1_bcells 3 call local_force j ic pb j vel_cell ic vel_bnd j amp vis_b j vof_frac_b j bfs_vec j dnb j fpre fvis area end do Ricardo Software December 2009 340 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES subroutine local_force lj ic pj uc uj vitj vofj sj deln fpre fvis area Arguments Description j integer boundary face or interface ic integer near boundary cell pj real boundary pressure uc 1 3 real cell velocity uj 1 3 real boundary interface velocity vitj real boundary effective viscosity vofj real boundary volume fraction sj 1 3 real boundary interface surface vector deln gt real normal dista
326. file opens a GLUT window and visualizes the cell CellInx in this case the given filename is interpreted as the name of the VECTIS MAX grid viewcgrp Ncells Cell1 Cell2 reads VECTIS MAX grid file opens a GLUT window and visual izes Ncells number of cell CellInx in this case the given filename is interpreted as the name of the VECTIS MAX grid viewb BouToVis reads VECTIS MAX grid file opens a GLUT window and visualizes the boundary faces on boundary BouToVis in this case the given filename is interpreted as the name of the VECTIS MAX grid viewbgrp NBouToVis BouToVis1 BouToVis2 reads VECTIS MAX grid file opens a GLUT window and visualizes boundary faces on NBouToVis number of boundaries BouToVisl BouToVis2 in this case the given filename is interpreted as the name of the VECTIS MAX grid view reads VECTIS MAX grid file opens a GLUT window and visualizes all boundary faces in this case the given filename is interpreted as the name of the VECTIS MAX grid Ricardo Software December 2009 82 4 MESHING 4 4 COMMAND LINE OPTIONS viewijk IS IE JS JE KS KE the geometry in the specified I J K area is visualized For example it is possible to run vmesh test mesh viewijk 2 2 54 54 25 25 to see the situation in the global cell 2 54 25 When n is pressed the global box is drawn as a wire model CHECK AND REPAIR INPUT GEOMETRY rep repair this switch activates an iterative healing subroutine which attempts
327. for Phase1 Fluid Phase then Boundary Passive Scalar node will be added A boundary condition type and the region attributes can be changed by left clicking on the Bnd_ Reg and as a result the Inlet Given Velocity Boundary panel is displayed to the right as in Figure 13 2 left For each boundary region the following attributes are displayed O Region Name can be any given meaningful name O Region ID each region has a unique ID by which it can be identified O Material ID each region belongs to a particular material O Boundary Report this option is used to output results related to boundary regions For example if a boundary is a wall then using this option will output wall force coefficients to ASCII files with extension wall O Coupled Link Number for 1D 3D coupled simulation with WAVE code used for Fluid Domains only O Boundary Condition Type O Boundary Setting Ricardo Software December 2009 215 13 BOUNDARY amp INTERFACE CONDITION TYPES 13 2 VELOCITY INLET inlet Inlet Given Velocity Boundary a amp Region ID 1 Material ID 1 Inlet Given Velocity Y Boundary Report Qutlet Coupled Link Number 0 Symmetry Boundary Condition Type Inlet Given Velocity Wall Pressure Boundary Setting Uniform Values s Stagnation Uniform Values By User RTHERM Time Dependent Figure 13 2 Boundary condition setup panel and boundary condition types
328. for channel and boundary layer flows with zero pressure gradient Ricardo Software December 2009 156 9 MODELLING TURBULENCE 9 4 NEAR WALL MODELLING The log law appears to be valid for simple boundary layer flows in local equilibrium that is when the production of the turbulent kinetic energy balances its dissipation P pe In terms of Prandtl s mixing length theory the log law can be obtained by assuming that the turbulence length scale in Equation 9 7 is proportional to the normal distance from the wall y Ky For a turbulent layer in local equilibrium the following relations can be derived say at point P within the turbulent layer 2 1 4 1 2 Ur Ur Cy kp orkp 9 28 VCu U3 ELp e LT EE 9 29 Kyp Kyp My pKUryp peT E yp 9 30 9 4 2 Standard wall functions Launder and Spalding 1974 re defined the log law in terms of the velocity scale Ug star units as follows Ur KU k This formulation relaxes a condition where T 0 and for U U 1 the universal log law Equa tion 9 27 is recovered The ratio U U can be seen as a non equilibrium index that changes the slope of the velocity profile in the turbulent layer cf Kim and Choudhury 1995 It can also takes into account some non equilibrium departures i e when the production of k is not in balance with its dissipation e Thus the above equation is used to calculate explicitly the wall shear stress pUp
329. g labelling its components with consecu tive numbers For every component global domain fluid solid domain material domain boundary and interface region the corresponding ordered set is created in the following way Global Domain Global domain is labelled with the index 1 Fluid Solid Domains Domain indices are first assigned to the fluid domains and then indexing continues with solid domains Material Domains Fluid material domains are ordered first in the same way as the fluid do mains Then solid material domains are indexed in the order which follows appearance of solid materials in the ordered set of solid domains Sub Domains Indexing of sub domains is done in the order as they appear in the ordered set of material domains Boundary Regions Indexing of boundary regions is done in the order as they appear in the ordered set of material domains Interface Regions Ordering of interface regions is based on the convention that an interface region is assigned to the material domain which has lower index than its neighbour Thus indexing is performed in the order which follows appearance of interface regions as one loops through the list of material domains An example of ordered domain components is shown in Figure 7 1 7 5 Creating a Domain Structure Domain components are created through the process of mesh importation The non partitioned mesh produced by the pre processing module vpre needs to be used Such a mesh wil
330. g whether the input surface intersects it or not For example if a box has IFORCE 1 and IDEEP 2 it will first be split into a 2x2x2 set of boxes and refinement down another level will then be applied only if the input surface inter sects the box Thus it is possible to effectively have a finer mesh localised to a particular area e g the valve bridge region in a coolant flow which does not extend throughout the mesh in the x y and z directions The forced refinement should be used with care level 1 produces 8 boxes from each global box level 2 produces 64 level 3 produces 512 Similarly using too large a level of IDEEP refinement should be avoided 3 Boundary refinement specifies the refinement depth which is to be used in the global boxes which are intersected by a particular boundary With this type of refinement three variables can be set for a boundary Refinement depth at boundary refinement level prescribed to the boundary Refinement blending distance specifies an integer value which is used to control how the refinement at the boundary blends into the refinement level of the surrounding boxes Blending is achieved by giving the boxes at the boundary a forced refinement level which is less than or equal to the specified refinement depth and propagating away from the boundary in layers of successively lower forced refinement The blend distance defines how many layers of boxes there are to be at each forced refinement level
331. ge angle that is set at the bottom of the panel This angle is a threshold value for the dihedral angle between the normals of adjacent triangles Any edge with an angle greater than this is considered to be a sharp edge A face is then delimited not only by triangles of a different colour but also by sharp edges This makes the painting of the end face of a pipe for example much simpler Selecting an appropriate edge angle and clicking on a triangle on the end face can mark the face The face is then identified by the program as bounded by the sharp edges without needing to chop the model or mark lines Boundary Reduction The Reduce button will cause non adjacent boundaries to be merged into one bringing down the total number of boundaries This operation will use all available boundaries not just those that are currently active Auto Paint This button paints boundaries automatically on the entire active part of the model using the spec ified edge angle to define the limits of faces The user is asked to confirm that the operation be continued or cancelled at this point In a complex model this may introduce a large number of boundaries Using the Reduce button can bring the number of boundaries back down 3 15 3 Boundary defined refinement setup VECTIS allows the user to control local mesh refinement based on boundary number The mesh refinement depth for cells adjacent to a given boundary can be specified by selecting th
332. generated If any generated inner face is concave it is split to convex parts here CELL ASSEMBLING PART B1 Find a complete box Find whether there is a box that has all its polygons already gener ated If there is no such a box continue with a new box go to Al B2 Distinguish separated volumes Find connectivity of faces and paint the polygons to find separated volumes in the box B3 Cell splitting On each separated volume test whether there are concave features If there are split the cell B4 Delete small cell Test cell volume and if it is too small remove it inner faces need to become boundary faces of neighbouring cells B5 Save faces The generated boundary and inner faces are saved to auxiliary files aux to be retrieved back later when assembling the final grid Continue with B1 11 FINISH GENERATION OF COMMON FACES The velocity locations of those cells which are fully inner need to be processed to generate their inner faces 12 PRINT STATISTICAL DATA OF GENERATED MESH Nov the statistics of the generated grid is printed Here is an example of such a report Number of generated cells 9816 boundary 7168 internal 2648 Patching method Number of cells processed by Marching Cubes 6076 84 53 Number of cells processed by Exact Fit 608 8 46 Number of volumes broken by Cell Splitting 20 0 28 of boundary cells Ricardo Software December 2009 86 4 MESHING 4 6 GENERAT
333. get_specs species id passive scalar name iget_ps passive scalar id boundary region name iget_breg boundary region id interface region name iget_ireg interface region id equation name iget_eq equation id boundary condition type name iget_bctype boundary condition type id phase property name iget_ph_pro phase property id species property name iget_sp_pro species property id passive scalar property name iget_ps_pro passive scalar property id Table 18 23 Subroutine get_id to get VECTIS MAX objects id 18 3 2 16 get_parent This UAR can be used to obtain parent identifier and optionally a name for the vectis object idt described by the identifier iget see Table 18 24 To obtain the parent id of a phase with idt do the following integer iwp ee idt parent_id parent_name character len idt 1 phase id call get_parent idt iget_phase parent_id parent_name Note that parent_name is optional A similar approach is used for other options of iget 18 3 2 17 eq_idt This UAR can be used to return identifier for derived data types related to the given equation eq and for the object obj domain phase etc specified by the get see Table 18 25 Ricardo Software December 2009 331 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES subroutine get_parent idt iget par_id par_name Input Output idt integer iget integer par_id integer par_name cha
334. gles to it and assigns it a default part name in the form part _ If the Delete Part function is selected the user is prompted to delete both the part and the triangles in the part or to delete just the triangles in the selected part Cut Paste Part Selecting Cut selects a part to cut which can subsequently be pasted into another part using the part pop up menu The Paste menu item becomes selectable after either a boundary or a part has been cut The cut item Boundary or Part can then be pasted into a part using the Paste operation Chop Unchop Part Selecting Chop Part and Unchop Part makes the part Active or Inactive respectively Users can optionally chop and unchop the children of parts and make whole branches of the model active or inactive in one operation Ricardo Software December 2009 52 3 GEOMETRY 3 15 BOUNDARY PROCESSING Add Boundary Adds a new boundary to the current part with no triangles assigned to it Part Properties Selecting Part Properties pops up the Part Properties Panel which can be used to change part names and add or edit comments for the parts Part Properties Panel Yy Part Properties X Part Navigation lt gt PartName Engine Part Comment 2 01tr 6 Part properties menu Selecting a part with the Right Mouse Button causes the part pop up menu to be displayed Model Parts a 1 Chamber W 2 Body H O 3 Engine E 1 walt Add ondaa Dele
335. hat point to the cursor When the left button is clicked a second time the line between the two selected points is used to create an infinite plane in the manner of the two previous functions Ricardo Software December 2009 27 3 GEOMETRY 3 8 TRIANGLE SLICING AND INTERSECTION OPERATIONS Intersect Slice Selected Button El This function can be used to slice a selected surface along a line of intersecting triangles Intersect Slice Both Button E This function can be used to slice two intersecting surfaces along the line of intersecting triangles 3 8 1 The Intersect Slice Operations The functions currently require the user to have used the Check Self Intersection function of phasel to mark the triangles that intersect one another This marks the intersecting triangles in red If the user then wants to slice the intersecting triangles along the line of intersection the Intersect Slice selected or Intersect Slice Both buttons can be activated to slice the triangles of either the selected surface or both surfaces along the line of intersection respectively Slicing Along an Intersection The program currently only calculates the intersections over a continuous area of marked inter secting geometry If the program can t find an adjacent triangle that intersects it stops looking for other triangles to slice This means several operations may be needed to slice the triangles in one boundary set The program calculates
336. he initial fields In the absence of the realistic initial fields a simple approach could be to obtain the initial fields by performing a steady run using boundary conditions which are those corresponding to the start time as specified within each boundary region data file The current contents of boundary region files restrict the specification of time dependent data to the single phase flow where the phase can be defined as a single component or as mixture of the wave species see Setting a phase Also the boundary region types are supposed to be either mass flow or pressure boundaries 13 10 Setting Up Interface Conditions For fluid flow equations with exception of the energy an interface condition type is the same as the wall boundary type Thus the interface condition setup is similar to the Basic wall setup In order to open the Interface Region panel an Interface Region should be selected mouse left click from the solver Interface Regions sub tree see for example Figure 13 12 A typi cal Interface Region panel is depicted in Figure 13 15 Here the following interface attributes can be changed O Interface name a meaningful name should be entered Interface Report if an ASCII file report is wanted then this box should be checked By default the interface report is written to the SDF binary rep file projectname rep_runnumber where interface variable names in the file contain INTF string The ASCI
337. he isotropic thermal conduc tivity A must be used However the solid solution domain can have an arbitrary number of solid materials see modelling of spatial domains and also the multi domain example in Figure 7 1 In case of multi material domain the grids along material interfaces must be conformal To achieve this the arbitrary grid interface tool is available The solid energy equation is always solved in terms of temperature as a primary variable To set up the solution of the energy equation in solids it should be activated in the solid Equations Solver panel see Figure 10 11 In addition User Defined Sources can be included In case of constant solid properties the finite volume discretisation leads to a system of linear equations with constant coefficients which should converge in one outer iteration if the grid non orthogonality is absent For such cases attempts should be made to use optimal i e maximum under relaxation factors one or very close to one In the presence of poor quality cells the user should try to run cases without the poor quality cell treatment Ricardo Software December 2009 179 10 MODELLING SINGLE PHASE FLOWS 10 5 MODELLING MASS TRANSFER Equations Solver a x RE a Energy Equations Relaxation Factor 0 8 User Defined Sources vr 4 M Figure 10 11 R Desk setup Setting solution of the energy equation for solid domains in the Equations Solver panel 1
338. he radius of the hemisphere is unity Firstly each patch is spherically projected to the hemisphere Secondly the projection is further cylindrically projected to the base of the hemi sphere The form factor is calculated as the area projected on the base of the hemisphere divided by the area of the whole base of the hemisphere The pictures below show the described princi pal of the Nusselt analog The first picture shows a 2D situation the second picture illustrates a situation in 3D Ricardo Software December 2009 203 15 MODELLING RADIATION 15 5 RADSOLV THEORY Figure 15 9 Figures showing the Hemisphere projection used in the radvfm calculation The hemisphere is based on a square formed by pixels The pixels outside the hemisphere are not used The relevant pixels are coloured by projected patches when objects which are nearer eclipse the objects which are far When all patches are projected the number of pixels painted by the same colour are counted View factor of processed super patch i and an other super patch j can be counted as _ AN where Nj is number of pixels painted by the colour of the super patch j D is number of pixels on the hemisphere diameter the length in pixels of the side of the square onto which is the hemisphere placed 15 5 Radsolv Theory Heat Transfer Theory Ricardo Software December 2009 266 15 MODELLING RADIATION 15 5 RADSOLV THEORY The governing equation for the heat transfe
339. he universal boundary regions that are setup in the VECTIS phasel pre processor should be identified so that one boundaries triangles are connected and not separate from each other So there should not be separate clusters of triangles that are identified as the same boundary number that are not connected and hence all triangles belonging to one boundary number should be connected See below for the description of minimum angle and maximum angle Minimum Angle and Maximum Search Angle and number of connection settings The minangle angle and maxangle angle switches are used to set the minimum and maximum search angles for the normal conduction connectivity calculation The default values for the minimum search angle that will be used if it is not set using the command option is 80 degrees and the maximum search angle is the minimum angle 5 degrees with a maximum limit of 89 degrees The definition of the search angles are shown by the figure below Upper surface B Figure 15 6 Figure showing the definition of the normal conduction minimum and maximum search angles The ntcs Number switch that can be used at the command line allows for the definition of the maximum number of super patches that an individual super patch can connect to for the normal conduction By default the number of connections is set to 3 Therefore if the number of connec tions is set to 3 the best three normal connections for each super patch will
340. hermal Conductivity Iv Specific Heat Figure 17 12 Postprocessing Output Panel Solid Domain The Solid Material panel is similar to the FluidPhase panel except it looks like Solid_Material_2 Solid Material Material Name Material ID Reference Temperature Density Thermal Conductivity Specific Heat Density Thermal Conductivity Solid_Material_2 10 293 15 Constant Values Constant Values Constant Values 2707 204 Specific Heat 8960 Figure 17 13 Solid Material Panel The Initial Conditions panel contains justa Temperature LineEdit Initial Condition a e Temperature 293 15 Figure 17 14 Initial Conditions Panel Solid Domain Ricardo Software December 2009 288 17 USING SOLVER 17 12 VECTIS FILES 17 12 VECTIS Files This section describes the various files used by VECTIS MAX All the files written out adhere to the format projectname lt type gt _ lt runnumber gt where lt type gt reflects the file type and lt runnumber gt is a 3 digit number e g 005 gt run 5 The alphanumeric output files were explained in Table 17 3 Other files include the out file which contains screen output For parallel runs there are also parallel output files in each parallel sub directory These contain simulation preamble specific to each parallel grid file e g grid size etc The post file is an SDF file
341. i eda dace ee bdo eb de 395 19 210 1 Importing the Grid File o scce be Se ed a Bd oe Ee A SS 398 19 2 10 2 Global Domaine css 5 2 ace coe a ko Ba w Pe de OS be Be dk oe are dvs 399 TIMeDase osas a te an ee Be ae hy Gad do a ad ee E 400 OUMU oe E eh aR eo we we arte N 401 Fluid Domai cord 6456 444 64 e bie Pee od aA 402 192 11 Grid Preparato sa s we e eee de we ee ae ee ee ee a ee 411 19 2 12 Running the solyer macia eS eS OO A a ae hae Sa eee 412 19 213 Live Update aes nr O aad eK ee ode De a ee Ba 412 19 24 POSt processin Ss asi ecg a Re RO a ad oe a ew 413 19 2141 2D planets us asia a ela ete A ce a eal 414 192 15 initialisation Tips carisma dd eA ROE ee RA ad bes 414 19 3 Coolant FIOW incida Aa Ga eed ie ee Ee ATA BOR ee ee ee 416 193 1 TntroduchOn a 4 4 44 4 2 eG ge ag a Backs Ba oe Pe oe he dake ke eee 416 1D 3 2 DAMS A eed Se Ae a Ae he Ce T es 417 19 33 Geometry Preparation vit et ake we Se A eek ad es 417 19 3 3 1 Selecting Boundaries 2 a ca caora 4 bee bee eee EE es 419 19 3 4 Mesh Generation s o at oan oa CE we be ek ea alee deed wee 422 19 35 SOWER SEMIN aai a eo id ae E tte de ee a eS ed 427 19 3 5 1 Global Domiaim iker dh Ce Eee eee hes 428 TIMEDASE i et o a ee ee aa a ee 429 QU PUL 4 0664 deseada a Sie eee ke wk we doe ad 429 Fluid Domain 2 42 44 6455444404404 05 bee od ah 430 19 3 60 Grid Pr paration s aoon a ele ee Pe ee pon ko be Ba PA ee Oe E 431 ORicardo Software December 2009 xii 19 37 R
342. icardo Software December 2009 39 3 GEOMETRY 3 13 HINTS FOR MANUAL STITCHING NL The bold line marks a The overlapping trian New triangles are created set of unstitched trian gles are deleted to fill in the gaps gle edges The triangles to the upper left overlap those to the lower right 3 13 1 Triangle Reduction In some cases it may be desirable to reduce the number of triangles in a model by removing unnecessary detail This can help speed up manipulation of the model on the screen as well as saving memory and runtime in subsequent phases of VECTIS Selecting Decimate Triangles from the Operations menu performs this operation This function works by collapsing short edges in the model to a single point The user is prompted to enter the limiting triangle edge length below which collapses should be performed Each line collapse operation removes two triangles The algorithm is constrained so that it does not alter the topology of the model and so that sharp edges are preserved The triangle reduction function operates only on the active part of the model 3 13 2 Self Intersection Checking This function performs checks to see if any of the triangles in the currently active part of the model intersect each other It is useful to identify bad geometry which may not be handled properly in the mesher The function is selected from the operations menu This brings up a confirmation panel with a field in which to enter the int
343. ice to neglect the underlined term in Equation 14 58 and the contribution to the mass flux correction from the underlined term in Equation 14 57 The lat ter is associated with non orthogonal grids With these approximations and mass flux definitions my M5 i the mass balance equation 14 7 is reduced to d nf T Cor ppVe 11 Sh 14 60 j 1 gt From the above equation the pressure correction equation is derived ny dppp Y ap Pp Sm 14 61 j 1 which has a similar form as a general discretized equation 14 48 The distinctive features of the pressure correction equation are discussed inDemirdzic et al 1993 Ferziger and Peric 1997 After solving the algebraic equations for p before that p is initialised zero everywhere the cell velocities pressure and mass fluxes are corrected as discussed above The pressure is corrected only by a fraction of p pp pp Qpp p where 0 is typically 0 1 0 2 The effect of the grid non orthogonality the last term in Equation 14 57 can be now taken into account by poner the second correction stepFerziger and Peric 1997 with Up U pt U pt ue P PP Pp Pp pp mj m n t mi The new corrections for the face velocity and mass flux are obtained from Equations 14 57 and 14 58 as follows 2 Vp A Vp Q E pr pb _ 2 ap Apd NI aP m i 07 Aj piU A 14 63 where the second density correction p7 is given by oe 14 59 with p
344. ich Phase 1 expects Outline Colours The Outline Colours option draws an outline around each of the triangles defined in the model These outlines can be switched off or displayed as mono coloured lines blue or multi coloured lines boundary colours Highlight Parts The Highlight Parts toggle can be used to identify parts defined in the part tree shown in the Part Boundary Panel When part highlighting is switched on selecting a part in the part tree with the mouse cursor will cause the part to be highlighted in red Highlight Holes and Sharp Edges The Highlight Holes and Highlight Sharp Edges toggles highlight potential defects in the model Holes are indicated by outlining the edges of the triangles that border them in red The finished model must be free from holes Sharp triangles are defined as adjacent triangle pairs that have an angle of more than 170 degrees between their normals The edge that these triangles share is outlined in green Ricardo Software December 2009 19 3 GEOMETRY 3 5 GENERAL VIEW OPTIONS Number Nodes Number Nodes actually annotates nodes triangles and edges that border a hole in the surface with a unique reference number as shown below The numbers of the nodes are shown in red those of the triangles in green and the numbers associated with the lines are in yellow Config File The view options can also be set in the Vectis cfg file which is read at program start
345. ich is intended to run quickly and so the end time will be set to 500 iterations Since the analysis is steady state the only Postprocessing file required is at the analysis end time or when a converged solution has been achieved In this instance an existing set up from a VECTIS 3 calculation will be imported to set up the majority of the input Open R Desk select File gt Open and browse to the coolant_V3 INP file This will import the file and open a new VECTIS session in rdesk The fluid properties timebase settings and boundary condition specifications should be read in from the old input file One thing to note is that in the VECTIS MAX solver it is recommended to have at least one non mass flow boundary to ensure the the boundary conditions are posed correctly In the original VECTIS 3 set up both the boundaries were specified as mass flow As such when the input file is imported the first outlet mass flow boundary is changed to a pressure boundary as indicated in the messages area Additionally the GRD file can be imported This will allow the boundary definitions imported from the existing V3 input file to be verified against the boundaries defined in the GRD file Click on Solver Setup found at the top of the Solver Setup Tree and select the coolant GRD file in the file browser Check mesh preview and select import Figure 19 76 Now when the boundary regions are selected int he Solver Setup Tree they should be highl
346. icking the relevant scale Figure 19 83 Looking at the residual values it can be seen that the majority have stablised after approximately 100 iterations Figure 19 83 The residual Values for the coolant simulation Additionally plotting the massflow rate on the outlet pressure boundary shows the the convergence of the simulations Figure 19 84 ORicardo Software December 2009 433 19 TUTORIALS 19 3 COOLANT FLOW Outlet Massflow N fu N n uo N a N o a H Fa al pa coe ae A E ropas E io u Massflow rate kg s io E e a 18 T SS ey 300 Iteration Figure 19 84 Massflow at the outlet pressure boundary 19 3 9 Post processing The simulation should reach the convergence criteria after around 400 iterations at which point the simulation will stop and write out post and restart files Open the post file in R Desk using the File gt Open menu option It will appear in the plottree By default the plot will be added to the active 3D canvas Note this behaviour can be changed in the preferences panel The attributes of the plot can be modified in the Plot Properties panel Different data can be shown on the plot by selecting it in the Data panel as shown in Figure 19 85 lol x FE Fle Edit View Optons Window Heb DOB b gt eis e 20 6 gt Elements absolute_pr Max 1 26 tax 1 2 z Se 04 1et05 105e 05 Llet05
347. id is to simply look up the identifier as listed in Table 18 5 global data identifiers for transport equations The call to eq_idt returns the domain type id domain or phase from the equation identifier iget__eq phase id iph and the object type iget_phase The object type argument 3 sets the context for the object id argument 2 If the domain type id doesn t exist eq_idt returns zero e g for a 2 domain 2 material fluid solid case Ricardo Software December 2009 356 18 USER PROGRAMMING 18 7 EXAMPLES idt eq_idt ifmom 2 iget_mat Get domain type id from ieq and phase id iph would return zero as momentum equation is not solved on material 2 solid The var_name passed to get_field can be any one of those listed in Table 18 26 There are two ways of calling get_field see Table 18 27 get_field idt iget var_name fi get_field idt rank var_name fi_c fi_b fi_ui fi_li fi_c_o fi_c_oo fi_f fi_n indx The first method is for retrieving a single field of specified type iget e g cell or boundary etc see Table 18 27 for valid codes The second calling method allows the user to optionally pull back many field types in one go If the variable var_name is a scalar then rank should be rank 1 array use parameter scal_field If variable is a vector then rank should be rank 2 array use parameter vect_field As all the field types are optional and some can be missing it is recommended to used nam
348. id lindex of phase species object corresponding to usr_obj_name integer iwp cs le cell index real wph ES sre amp calculated scalar src term y src_vel 1 3 calculated momentum src term real wph lt 2 fi old real wph ft tol real wph pointer o den MO gt null cell centres lupr gt start Select a phase species name usr_obj_name lupr gt end lusr_ob3j_name air if usr_obj_name return Find phase species index from the selected usr_obj_name ORicardo Software December 2009 362 18 USER PROGRAMMING 18 7 EXAMPLES if isp gt 0 then iobj isp call get_id usr_obj_name usr_obj_id iget_specs else iobj iph call get_id usr_obj_name usr_obj_id iget_phase end if if iobj usr_obj_id then Default values for sources src_vel 0 src 0 Get equation name call get_name ieq eq_name iget_eq select case eq_name 1 case momentum lupr gt start Get velocity field if required var_name velocity call get_field idt iget_cell var_name u Define the source term vector src_vel 1 3 per unit volume land if appropriate a negative source term src per unit volume lequal for all velocity components u v w src src_u Ssrce_v Src_w do ic icell icel2 src Isrc_vel 1 Isrc_vel 2 Isrc_vel 3 src_fi 1 3 ic src_fi 1 3 ic src_vel 1 3 x vol_ph ic ap_fi ic ap_fi ic src vol_ph ic end do nullify u lupr gt end case mass_pressure
349. ides the Solve All Phases option box to select whether to solve all phase equations or not The default option is not i e the option box is not active Ricardo Software December 2009 202 12 MODELLING MULTIPHASE FLOWS 12 4 PHASE CHANGE MODELLING Equations Solver amp A y Turbulent Dissipation a Turbulent Viscosity 7 Volume Fraction Potential Dancin Ceaniar v Volume Fraction Equations Relaxation Factor 0 8 Convective Scheme UDS 4 gt Blending Factor 1 User Defined Sources Figure 12 2 R Desk setup Setting solution of the volume fraction equation 12 4 Phase Change Modelling In automotive industry two types of phase changes are modelled boiling and cavitation 12 4 1 Boiling Models Boiling occurs when the liquid is heated to its saturation temperature boiling temperature which results in rapid vaporisation of the liquid Boiling models discussed in this section are homogenous boiling models and are designed to calculate boiling flow possibly taking place in engine cooling passages Two boiling models are presented VECTIS3 model and RPI model Both models use the homogenous multiphase mixture equations whereby the liquid and vapour phases are solved individually but other fields like temperature velocities etc are shared between each phase Recall the mass source I from equation 12 2 Then if i quid and e are the mass source of the liquid and vapour phases respectivel
350. ighted in the 3d canvas Ricardo Software December 2009 427 19 TUTORIALS 19 3 COOLANT FLOW Reading VECTIS 3 input fie All 10 boundaries are mass fiow setting frst outlet to pressure Figure 19 75 Importing VECTIS 3 solver setup m R Desk 3d1 AutoPreview coolant GRD o Heb D B aje 3 60 3 ta Ee A Q Ele Solver Setup Global_domain_1 Global Domain Restart Control Timebase Output El Flud_1 Fluid Domain itoring Points Boundary Regions Bnd_Reg_1 Well Boundary Bnd_Reg_2 Wall Boundary Bnd_Reg_3 Wall Boundary Interface Regions domains Report Regions Figure 19 76 Importing the computational grid into the solver setup Now we will run quickly through the Solver Setup Tree More detailed information on all of the panels in the tree can be found in the manual or in previous tutorials 19 3 5 1 Global Domain Click on the global domain entry in the Solver Setup Tree The global domain panel allows general simulation parameters to be set Figure 19 77 Input Mesh Filename The solver will default to using a computational grid file with the name pro jectname GRD where the projectname is taken as the inp file prefix A different grid file can ORicardo Software December 2009 428 19 TUTORIALS 19 3 COOLANT FLOW Global_domain_1 Global Domain 218 xi Domain Name Global_domain_1 r Input Mesh Filename WM Set Filename coolant GRO v Co
351. igned to the selected R Therm set and will be visible in a drop down list of Extracted Sets in the RTHERM panel see Figure 13 11 and Figure 13 13 right respectively In a similar way other wall boundaries can be associated with the corresponding R Therm sets 13 9 Setting Time Dependent Boundary Conditions Unsteady simulations can be performed in conjunction with time dependent boundary values The specification of time dependent boundary variables is currently provided via ASCII files As Fig ure 13 14 illustrates for each boundary region having Boundary Setting set to Time Dependent option a separate file should be prepared The corresponding file name need to be entered into Region File name line box Ricardo Software December 2009 230 13 BOUNDARY amp INTERFACE CONDITION TYPES 13 9 TIME DEPENDENT DATA Region Name Cyl1 txt Region ID 3 Material IC Y Boundary Report Coupled Link Number 0 Boundary Condition Type Mass Flow Rate Boundary Condition Op Normal Velocity Scaled Boundary Setting Time Dependent Region File Name Cyl1 txt Massflow Rate 0 000799 In Velocity Directior Figure 13 14 R Desk boundary condition panel for time dependent boundary data File names are arbitrary and their format follows a VECTIS3 practise Thus there are 12 entries for each time step and these entries can be specified using one or two lines The meaning of entries is as follows
352. ilable for mass flow rate Normal Velocity Scaled which is the default selection and Velocity Vector Scaled When Velocity Vector Scaled is selected then the Velocity Direction is enabled as in Figure 13 4 right where Cartesian velocity components should be entered ORicardo Software December 2009 219 13 BOUNDARY amp INTERFACE CONDITION TYPES 173 4 PRESSURE INLET OUTLET Phase species and passive scalar boundary values are also relevant in the mass flow rate boundary type Species and passive scalar boundary set up is done in a similar way as for the Inlet Given Velocity Boundary condition type To set up the phase boundary values left click on the appropriate phase under Boundary Phases within Mass Flow Rate Boundary region in the Solver Setup Tree This opens up the Boundary Phase setup panel similar to Boundary Phase panel in Figure 13 3 top right figure The Pressure an estimated absolute value Temperature and Volume Fraction can be entered into the InputBoxes In case of a turbulent flow Turbulent Intensity and Turbulent Length should be specified Compared to the Velocity Inlet type the X Velocity Y Velocity and Z Velocity components are not displayed as they are not relevant for the mass flow rate boundary type 13 4 Pressure Inlet Outlet This boundary condition type is based on the known or assumed pressure distribution at bound aries The constant static pressure or boundary averaged pressure can be specified at both in
353. ill be used to add a refinement block Firstly click the Add button and draw out a box by pressing and holding the left mouse button to include the area required If desired this can be done using more than one IJK box however this is not strictly necessary Having done this change views to another 2D view using the Mesh View toolbar Currently the IJK block defined in the previous view will be shown by a highlighted region and should be lined up along one edge of the mesh Click Edit from the IJK Block Navigation window and draw another box by pressing and holding the left mouse button so as to include the cells required for the refinement region Repeat for any other blocks This should locate the IJK boxes in the correct location in 3D space This can be checked by activating the 3D view from the Mesh View toolbar and turning on 3D sub division and Show IJK block regions in the Viewing Options panel The IJK block setup should be similar to that shown in Figure 19 26 ioixi File Edit View Toolbars Operations Help o a m S E e E 1 N 28 YM Options Stitch Mesh Figure 19 26 Setting IJK refinement Next the required refinement depth for each IJK block must be set Using the left and right buttons from the IJK Block Navigation window select the IJK block you require and set the DEEP and Ricardo Software December 2009 391 19
354. in R Desk Left click on Boundary Condition Type will open up the ListBox that contains all boundary con dition types as shown in Figure 13 2 right Similarly left click on Boundary Setting displays currently available setting options which are depicted at the bottom of the above figure These options are O provision of region wise Uniform Values O specification of boundary values By User which implies the use of User boundary conditions routine O definition of wall thermal conditions by using Ricardo s R Therm module option RTHERM The R Therm module generates thermal conditions at various wall boundaries which represent typical engine components used in conjugate heat transfer simulations O provision of Time Dependent boundary values Depending on the selected boundary condition type the definition of additional attributes might be required The following sub sections describe inputs relevant to a specific boundary condition type 13 2 Velocity Inlet The fluid velocity vector and other scalars with exception of pressure are prescribed at the inlet The velocity inlet should not be used for sub sonic compressible flows ORicardo Software December 2009 216 13 BOUNDARY amp INTERFACE CONDITION TYPES 13 2 VELOCITY INLET Empirical relations are used to estimate the turbulent kinetic energy in absence of experimental data The knowledge of the relative free stream turbulence intensity can be used to estimate the tu
355. in equation of k 6 40 one can assume that its E gra dient drives the turbulent transport by triple product of velocity fluctuations pk u u pul pul yl a 2 The same assumption is physically unsound for the transport by pressure fluctuations Badian 1994 Since it appears that the turbulent transport by the velocity fluctuations is dominant one modelling of both terms by a gradient transport hypothesis 2 e a E My ok a au pk i p 5 an da 9 15 Ricardo Software December 2009 152 9 MODELLING TURBULENCE 9 3 LINEAR TWO EQUATION K MODELS can be considered as an intentional approximation A non dimensional constant Oog 1 is the effective Prandtl number for the diffusion of k Turbulent kinetic energy production by shear is calculated as OU 2 2 P 1 i s 25 OK 9 16 where S represents the strain tensor invariant S 4 28 Si 9 17 Modelling of the production by body force P depends on the character of the force The gravita tional force f g is frequently encountered and for flows driven by thermal or concentration buoy ancy the production term P becomes p fil gip u u Following the standard eddy diffusivity approach the unknown correlation p u is modelled by the gradient of mean flow density He DO P 9 18 where Prp is usually taken to be the same as for the temperature i e Prp Pr The compress ibility correction terms in the k eq
356. ine FindNdInOut problem with evaluation of in out status of a velocity location Location There is a problem with in out test on the velocity location Location The input geometry is probably not clean enough ERROR 1206 Generation of faces incorrect in out test detected on location location There is a problem with in out test The input geometry is probably not clean enough ERROR 1207 Cannot read the grid file filename ERROR 1211 Distance Decimation routine not enough memory ERROR 1212 Polygon Simplification routine not enough memory ERROR 1213 Incorrect version of meshfile detected There is an incorrect version written in the meshfile The first version valid in VECTIS MAX is VECTIS_MESH_INPUT_V1 100 ERROR 1500 Problem with generation of face on velocity location location On the given velocity location there is a problem with generation of internal face This problem might be caused by unclean geometry near the velocity location ERROR 1501 Problem with processing program arguments A fatal problem occured when com mand line options were read The help section 4 4 should be consulted ERROR 1502 Trifile filename cannot be read ERROR 1503 Incorrectly placed meshlines The meshlines read from the meshfile do not correspond with the geometry read from the trifile ERROR 1504 No patches generated in box box despite the fact there are different in out statuses The box has different in out statuses therefore some bound
357. ing cells The blend distance defines how many layers of cells there are to be at each forced refinement level when stepping back from the forced boundary refinement to the global mesh cell size Blending is achieved by giving the cells at the boundary a forced refinement level which is less Ricardo Software December 2009 58 3 GEOMETRY 3 15 BOUNDARY PROCESSING than or equal to the specified refinement depth and propagating away from the boundary in layers of successively lower forced refinement The blend distance defines how many layers of cells there are to be at each forced refinement level Boundary refinement is applied after IJK refinement blocks i e it will override IJK refinement blocks However the forced refinement level and refinement depth are only ever increased by boundary refinement specifications never decreased Refinement Specification for boundary 1 Refinement depth at boundary 2 Refinement blending distance A Refinement Blending t Blend to boundary depth Blend to boundary depth 1 m Refinement Specification Destination f Save specification in triangle file C Save specification in mesh file Apply Cancel Wi 3 15 4 Saving the boundary refinement settings The boundary refinement specifications can be saved in the triangle file or in the mesh input file This is controlled by the second radio button on the panel shown left Where to save the specificati
358. ing internal face 12 1 n_mat_doms ending high interface 12 1 n_mat_doms Table 18 10 Subroutine get_mat continued from previous table Ricardo Software December 2009 309 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES 18 3 2 4 get_phase This UAR is used to retrieve information related to phases see Table 18 11 To establish the type of phases used then do the following integer iwp pointer ph_type Call get_phase phase_type ph_type Array ph_type may contain one of the following entries O gas 1 O liquid 2 O solid 3 If ph_t ype 1 1 indicates that phase 1 is a gas The compressibility options can be obtained as integer iwp pointer ph_comp call get_phase phase_compress ph_comp The compressibility of each fluid solid phase is defined by array ph_comp 0 n_ phases The following parameters identifying a fluid solid type in terms of compressibility i e Mach number can be assigned to ph_comp O incompres 0 incompressible fluid phase or solid material density is constant compress ibility coefficient 6 0 O iwcompres 1 weakly compressible fluids density may vary with temperature but the change of compressibility coefficients is relatively small This definition includes liquids and gases where the gases pressure changes relative to the reference pressure are small In terms of the Mach number
359. ing nodes are projected normally onto an interpolation plane parallel to the wall and passing through the cell The state of vari ables at new intermediate plane locations i is constructed directly from the states of qualifying neighbouring cells N and wall nodes W j 1 n ny being a number of qualifying neighbours In the next step an interpolation based on the inverse distance weighting factors is designed to Ricardo Software December 2009 254 14 NUMERICAL SOLUTION 14 4 POOR QUALITY CELL TREATMENT calculate variables at the cell ny 1 ny 1 t E alt with 0 51 2 53 14 74 j l lj j 1 ij 1 where i represents the distance between intermediate nodes i and the cell 7 The type of interpolation used to obtain values at the interpolation plane locations depends on the near wall behaviour of the considered variable First we assume that flow variables at wall nodes W have the same values as variables at the wall node W Considering the intermediate node i subscript j is omitted the velocity vector is split into components parallel and normal to the wall U p and ng respectively U Dip Uin with i n a ir MI i p ed U i n 14 75 Similar decomposition is used for adjacent nodes N and W The parallel velocity can be calculated from the universal near wall profile Equation assuming the constant wall shear stress T Employing the effective wall viscosity Uw Equation 9 40 w
360. ing the following integer iwp pointer il 12 call get_mat ise_phase il i2 where il and 12 will have 1 12 1 2 for material 1 3 12 2 3 for material 2 In a similar way to subroutine get_domain the upper bounds for arrays rval i1 and i2 can be obtained using function get number Ricardo Software December 2009 307 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES subroutine get_mat var_name rval Input Output var_name cha rval real pointer r_temperature reference temperature rval 1 n_mat_doms r_pressure reference pressure rval 1 n_mat_doms r_gas_const reference gas constant rval 1 n_mat_doms r_area reference area rval 1 n_mat_doms r_length reference length rval 1 n_mat_doms r_velocity reference velocity rval 1 n_mat_doms r_density reference density rval 1 n_mat_doms r_viscosity reference viscosity rval 1 n_mat_doms r_spec_heat reference specific heat rval 1 n_mat_doms r_spec_heat_ratio reference specific heat ratio rval 1 n_mat_doms Table 18 9 Subroutine get_mat to get variables defined in the access group iacc_mat mainly reference material properties and starting and ending indices of material objects Ricardo Software December 2009 308 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES subroutine get_mat var_name il 12 Input Output var_name cha i1 integer
361. ins that take part in a multidomain calculation are meshed separately and than a tool that ensures conformal faces on interfaces needs to be applied The basic requirements for this meshing are these 1 Meshlines of all parts must be exactly the same All the complementar meshes need to use ex actly the same set of meshfiles In order to check whether this condition is fulfilled cmpb command line option can be used If the problems are small the mesher can try to repair it when asked by cmpbw option For more details see description of these two command line options in 4 4 2 The used refinement must be exactly the same It is essential to have the same refinement set in both geometries for the global boxes intersected by interfaces The same refinement depth should be set for both geometries also boundary refinement depth on interfaces must be the same When IJK refinement blocks or boundary refinement depth with blending distance are used in a geometry it must be set carefully so as the boxes created after refinement are exactly the same on the interfaces 3 Quality of input triangles The normal meshing process can usually cope with small flaws in the input geometry imperfections like folded triangles or triangles of nearly zero area However for multidomain simulations it is recommended to pay higher attention to the quality of trian gles defining the interface The better quality of interfaces the higher probability that the tool maki
362. int 1 fluid2 mesh vmesh int 2 3 solidl mesh Next additional mesh processing is preformed using the conformrew option This is to ensure these boundary regions are made as conformal as possible These commands are repeated for each pair of meshes to be joined vmesh conformrew fluidl GRD 1 solidl GRD 2 vmesh conformrew fluid2 GRD 1 solidl GRD 3 Note as an alternative to the above vmesh conform fluidl GRD 1 solidl GRD 2 vmesh conform fluid2 GRD 1 solidl_conform GRD 3 The first conform would produce two new files fluidl_conform GRD and solid1_con form GRD The next conform would create the files fluidl_conform GRD and solid1_ conform_conform GRD This method is safer 1 e doesn t overwrite original meshes but a little less convenient to type For the final stage the three meshes are joined using the vpre utility remembering to place the fluid domains first vpre jtype 0 fluidl GRD fluid2 GRD solidl1l GRD This would produce a multi material file By default this is called COALESCED GRD unless the ORicardo Software December 2009 109 5 READING amp MANIPULATING MESHES 5 4 RESTARTING PARTITIONS o filename option is specified All the boundary faces that are successfully joined are converted to interfaces internal faces along a material interface region Any boundary faces not joined are stored in the grid file as boundaries When this multi domain grid is imported into R Desk these unj
363. interfaces whose normal vectors point out of the given material domain Similarly iget_lif corresponds to lower interface values These are fi values at cell faces along material interfaces whose normal vectors point into the given material domain b The second method is used to obtain field values at one or more grid object Thus a number of optional arguments are available in the second part of Table 18 27 Example Obtain pressure field at cell centre boundary faces lower interface and upper interface firstly obtain domain type idt for pressure and then obtain pressure values integer idt dom dom is domain id integer scal_field 1 rank of variable here it is a scalar real wph pointer p amp cell pressure rpb amp boundary pressure pi amp upper interface pressure ypli lower interface pressure scal_field 1 Iscalar rank dom 1 domain 1 used as an example Get pressure domain index idt eq_idt ifmas idom iget_dom get pressure values Call get_field idt scal_field pressure fi_c p fi_b pb fi_ui pi fi_li pli It is important to note that the starting amp ending indices of field values can be obtained using get_domain subroutine In the above example for velocity field the indices of starting amp ending cell for all domains can be obtained as integer iwp pointer ic_s ic_e call get_domain ise_cell ic_s ic_e Ricardo Software December 2009
364. introduce the user to some of the many numerous features available The latest version of the VECTIS MAX manuals and tutorials can be found online on the Ri cardo Software website http www software ricardo com support manuals vectismax Check here for updates and also PDF and XHTML versions 19 1 Basic Tutorial Get the necessary files for the port flow tutorial http www software ricardo com support tutorials vectismax TutorialFiles Basic 19 1 1 VECTIS Workflow The VECTIS workflow consists of several stages illustrated in Figure 19 1 Figure 19 1 VECTIS workflow Geometry Import and Preparation Phase 1 Mesh Generation vmesh 367 19 TUTORIALS 19 1 BASIC TUTORIAL Mesh Preparation vpre Solver Setup R Desk CFD Solver vsolve Post processing R Desk 19 1 2 VECTIS Structure Vsolve is a general CFD solver allowing multi material multi phase and multi species calculations to be considered The general structure is illustrated below in 19 2 Global Domain The global domain considers the entire computation and can consist of a number of fluid and solid parts Fluid Domain Each of the fluid domains represents a liquid or gas or a mixture of both within the calculation Fluid Phase Fluids may consist of single or multiple phases which can be liquid or gases or a mixture of both Species Components Each of the fluid phases may consist of a single or multiple species Solid
365. ion of state p p T p i e dp 9p 0T pdT 9p 0dp rdp and dh de d p p one can get e dp _ B Cp Cy r and OT 4 Z 8 19 8 2 3 Specific thermal energy and enthalpy With the help of Equations 8 12 and considering first e e T v and s s T v and then h h T p s s T p the following thermodynamic relationships for the differential internal energy and enthalpy are obtained respectively dp dp B dp c dT P dT T IS 2 de c dT p 32 p7 cd 0 E y 8 20 1 dp dp dp dh cpdT 1 T cpdT 1 BT 8 21 P p OT i p P p Changes in specific internal energy and enthalpy between two states 1 and 2 are then given as Th P2 B dp eo e1 ay cydT r r _ 8 22 Ti p J p Th p2 dp heh a sir p tl 8 23 Ti p P 8 3 Thermally Perfect Fluids Most of common gases and liquids can be considered thermally perfect in the sense that internal energy and thermal enthalpy depend on the temperature only Apart from phases declared to have WAVE properties only thermally perfect fluids can be used in the current version Selecting the state 1 as reference with e T ef A T ef 0 the constitutive relations for internal energy and enthalpy see Equations 8 22 and 8 23 can be approximated as T 1 T e T f odT O T T ale rep G T c T dT 8 24 Tref f 0 T 1 T h T f cpdT p T T Cp Trep Tress TT 7 f p T aT 8 25 Tref i 0 In the above eq
366. ious flaw in the input geometry the mesher can stop its run with error ERROR 1507 The closed loop cannot be gathered for box Box ii I jj J kk K bx box In the case of smaller problems the grid is correctly generated but when the test test is per formed run vmesh geometry GRD test problematic cells are reported 4 If problems find IJK repair geometry and run again If the generated gridfile cannot be used 1t is necessary to find the I J and K of the global box and check the part of the geometry inside 1t In the case when the grid was not generated because of ERROR 1507 the IJK is already reported If the problem was detected by running test option the global boxes can be found by using verbosetest locate option For example when the mesher is run as Ricardo Software December 2009 93 4 MESHING 4 8 MESHING FOR MULTIDOMAIN SIMULATIONS vmesh port GRD verbosetest locate port mesh indicies of all problematic cells in port GRD grid are gathered and corresponding IJK boxes are found according the information about meshlines stored in the passed meshfile port mesh The positions are found according geometrical position of vertices of the problematic cells Then user should check triangles in all reported global boxes 4 8 Meshing for Multidomain Simulations In comparison with normal meshing task for one domain only preparation of conformal meshes for a multidomain simulation requires some additional effort All doma
367. is use a combination of the Chop and Mark Face commands Y Ricardo VECTIS Phase 1 port tri 10 x File Edit View Toolbars Operations Help aa wale Aloe JN aa MM Options Stitch mesn r Triangles _ __ AJA Al _ A E Y Parts Bose Show Tr a 2 Plie mia e O Add Boundary Delete Boundary 9 S 3 3 2 5 3 Show All Hide All E Toggle Compress yay gt a Reduce Paint All se r Paint Face Auto Paint Vlek nye KAE E Show All A Paint Line Motion A a Refinement Motion Info erm A aa TR Auto Paint Angle 45 0 153870 0 193000 ymin 0 020499 ymax 0 270500 zmin 0 036690 zmax 0 326520 Figure 19 22 Using the chop geometry command The boundaries can be selected either manually or automatically To manually select the bound aries first open the Boundary Painting Tab in the Operations menu Then select with the left mouse button the colour required by clicking on the number of the boundary required Once a colour has been selected the row that contains that colour will turn yellow Then either the MARK LINE or a combination of the CHOP and MARK FACE commands can be used to select the different bound aries The automatic method works on separating surfaces that are at angles greater to each other tha
368. iscreet levels The whole used space is divided to 4 2 billions of discrete levels on which new nodes can be created For meshlines a bit rougher grid is used they can be shifted to each 8 th level Also vertices of input triangles are shifted to each 8 th discreet level however those levels lie exactly in the middle of the levels used for the meshlines 4 RAPID DETECTION OF OVERLAPPED TRIANGLES Defaultly a rapid algorithm for de tection of overlapped triangles is run here If there are some problematic triangles warning 1800 is printed together with the list of indexes of the triangles If some problematic triangles are detected the mesher will not stop Usually VMESH automatically overcomes small prob lems in the input triangulated surfaces However if it happens that the final gridfile contains cells with low quality test option detects problems the user should try to improve the flaws in the surface and run VMESH again See more information in the section 4 7 5 PREPARE SHOEBOXES Shoeboxing is a system heavily used in several Ricardo products that helps to quickly limit number of elements that need to be taken into account when intersection tests are performed The 3D space is divided to NI x NJ x NK boxes so called shoeboxes Then for each input triangle all shoeboxes that are intersected by it are found The index of the triangle is remembered by all found shoeboxes This information will help later when intersections of a segme
369. it advisable to preprocess the mesh file with vpre in order to eliminate any topological gaps within the boundaries For the case of multi domain meshes this should be repeated for each single domain grid prior to coalescing stage See below for examples Single domain vpre close boundary gaps meshl GRD Multi domain 3 domains vpre close boundary gaps meshl GRD vpre close boundary gaps mesh2 GRD vpre close boundary gaps mesh3 GRD vpre jtype 0 mesh1 GRD mesh2 GRD mesh3 GRD o multidom GRD 257 15 MODELLING RADIATION 15 3 RADIATION SETUP 15 3 Radiation Setup In order to run VECTIS MAX with radiation modelling on it is first necessary to set up various files required by the radiation modules The first stage involves preparing a radfile required by the two modules radprep and radvfm and later the radiation solver within vsolve This radfile is a simple text file which is automatically created upon saving the main solver input file if radiation 1s activated The radiation panel can be navigated to from within the Global Domain panel Fig 15 1 This panel Fig 15 2 is used to set up global quantities pertaining to a radiation problem Global_domain_1 Global Domain Restart Control Timebase Output Fluid_1 Fluid Domain Figure 15 1 Global radiation location within the solver setup tree The Radiation Modelling pull down menu currently has only two options none and Surface to surface The Radiati
370. ithin a new project the Solver Setup Tree will be refreshed and show anew RTHERM node within Global Domain sub tree Figure 13 12 Solver Setup Tree cfd 2vortx vp V4TLIB CHT_HOND 3 Solver Setup al Global_domain_1 Global Domain Restart Control Timebase Output Fluid_1 Fluid Domain Monitoring Points Algorithm Turbulence Model Discretise Phase_1 Fluid Phase Equations amp Solver Boundary Regions Interface Regions Interface_94 Interface Region Fin Interface Region Fin Interface Region E Figure 13 12 RTHERM node as a part of the Solver Setup Tree After selection of the RTHERM node an empty RTHERM panel will be displayed as shown in Figure 13 13 left Assuming that a file containing thermal conditions has been created earlier using R Therm module the user should be able to Browse and load the required RTH File Note that a RTH file has extension RTH The Show Preview can be used to display R Therm geometry or mesh As the solver mesh might be created within a different Cartesian coordinate system than the R Therm mesh Transforma tion Matrix need to be edited This matrix is then used to transform VECTIS MAX coordinates more precisely boundary face centres into R Therm coordinates In order to obtain the correct transformation matrix the user can preview both R Therm sets and solver boundaries and establish required coordinate system translation or rotation For this
371. itial Conditions Post File J Unesr Equations Solver Residuals Information Figure 19 13 The Output Panel Reporting General output Here the frequency of data reporting output is chosen Firstly the monitoring and convergence data written to the screen is also written to the out file at the intervals specified If zero values are used then the output is not written The frequency of post processing file writing is specified in the Post processing File Frequency box Additional output Additional output file can be written for monitoring data boundary data report regions and domain data These can be either in ASCII column format or Ricardo SDF binary format The frequency of the two types of file can be chosen independently by entering the required intervals in to the Report Frequency boxes Again an interval of zero will mean that the data is not written The ASCII files include header rows detailing the data in each of the columns Separate files are written for the different data types Domain Monitoring IO and Wall If reporting is selected for the relevant boundary Arbitrary Surface Residual data can also be written to and ascii file and to the report file by selecting the relevant check boxes Ricardo Software December 2009 376 19 TUTORIALS 19 1 BASIC TUTORIAL The ASCII data files are used by the Live Update utility in R Desk The SDF binary format is written to a single report file named pr
372. j Fp Tp i e its cosine is given as cos 6 Aj dj Aj dj 14 34 Very large face angles above 75 if they are greater than 90 the mesh is in principle invalid will at least slow down the solution convergence A simple measure would be to neglect the cross diffusion above the certain face angle A better compromise between accuracy and numer ical stability is achieved by limiting the cross diffusion by the normal diffusion The following expression describes this limiting ID D Djemin 17 14 35 De Ricardo Software December 2009 244 14 NUMERICAL SOLUTION 14 1 FINITE VOLUME DISCRETISATION where y 1 is the cross diffusion limiting factor In our experience the use of the alternative definition for the normal diffusion in the above equation A D I V9 sd 14 36 jn hj tj ee TO Aj dj improves convergence properties The same limiting can be applied to boundary cross diffusion fluxes 14 1 5 4 Source term Surface and volume sources in Equation 14 4 are calculated explicitly using the mid point rule When the volume source term S f depends on it is linearised implicitly see Patankar 1980 as Sh S Soop 14 37 The constraints s gt 0 and S gt 0 are applied to Equation 14 37 This ensures that diagonal dominance remains positive and also it is increased when the term S is added 14 1 6 Discretisation of the energy equation A peculi
373. k Skk Ps 9 6 By using different arguments analogy with the kinetic theory of gases dimensional analysis and phenomenological models cf Speziale 1991 the turbulent viscosity is defined in terms of characteristic turbulence length time and velocity scales P Ps b Ly Pr n Pyrrhi 9 7 Turbulent fluxes for species and heat can be formulated in a similar way by assuming an analogy between molecular and turbulent diffusion see Equations 8 9 and 8 5 Hampi ME 9 8 J Ty poe J with Y and A being turbulent eddy diffusivity for i th species and turbulent thermal conduc tivity respectively The turbulent mass diffusivity and thermal conductivity are usually defined in terms of turbulent viscosity Ly LC p E poe 9 10 Pis Sci A Pr 9 10 where Sc is the turbulent Schmidt Prandtl number for mass transport and Pr denotes turbulent Prandtl number for thermal transport The above equation implies an analogy between the mo mentum species and mass transfer It is now possible to express the total diffusion fluxes laminar plus turbulent through so called effective diffusivity values Hef f HU Lr Dies f Di Pir Nef f 2 9 11 Ricardo Software December 2009 150 9 MODELLING TURBULENCE 9 2 OVERVIEW OF TURBULENCE MODELS FOR RANS In Equation 9 4 the relation between Reynolds stresses and the mean strain tensor i e the mean velocity gradient is linear and the corresponding model
374. l have to O describe the whole computational domain consisting of all spatial material domains and their sub domains O be re ordered through all material domains and be conformal across material and sub domain interfaces o and contain information about boundary regions and material sub domains When a new project is started in R Desk the Solver Setup Tree shown in Figure 7 2 will contain by default the one fluid domain material and one boundary region The Solver Setup Input panel 1s then used to select the relevant grid file by using the Browse button Once a file has been selected its name will appear in the Filename box Clicking on the Extract button the Grid Extract Dialog pops up and displays a number of materials material domains found in the imported grid file Note that these materials have already been ordered as explained in the section about ordering of domain Ricardo Software December 2009 126 7 MODELLING SPATIAL DOMAINS 7 5 CREATING A DOMAIN STRUCTURE Solver Setup Tree O Solver Setup Postprocessing Output User Function Global_domain_1 Global Domain Restart Control Timebase Output Radiation Fluid_1 Fluid Domain Monitoring Points Solver Setup Input e x Algorithm Turbulence Model Discretise Phase_1 Fluid Phase Initial Condition Extract Materials And Boundaries From Mesh File Filename Postprocessing Output Equations amp Solver VHEADLAMP headiamp GRD
375. l inlet boundary given velocity integer parameter vector ibout outlet boundary integer parameter ibsym symmetry boundary integer parameter ibwal wall boundary no slip slip integer parameter ibpre pressure static total average integer parameter ibtot total stagnation boundary integer parameter ibmas inlet boundary given mass flow integer parameter rate Argument pointers for real array rbc ibturin turbulent intensity integer parameter ibturls turbulent length scale integer parameter Global data identifiers related to transport equations ieq ifmom id for fluid momentum equations integer parameter ifmas id for continuity mass pressure integer parameter iene id for energy equation integer parameter ifcs id for concentration of species integer parameter ifte id for turbulent energy equation integer parameter ifed id for dissipation of turbulent integer parameter energy ifvf id for volume of fraction integer parameter ifps id for passive scalar calculation integer parameter Property identifiers idens density id integer parameter ivis viscosity id integer parameter Table 18 5 Parameters accessible directly through UPRs Ricardo Software December 2009 301 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES Global data identifiers related to variables Variable Description Type iu id for fluid u momentum integer parameter iv id for fluid v momentum integer parameter iw
376. l is employed then the above thermal conditions are applicable to the external outer side of a thin wall while the internal inner side is in the contact with either fluid or solid material domain To enable the thin wall model the Thin Wall box should be checked Then the following Thin Ricardo Software December 2009 226 13 BOUNDARY amp INTERFACE CONDITION TYPES 13 8 WALL Prescribed Temperature Combined Convection amp Radiation Given Heat Flux External Convection External Radiation Convection HTC 0 Temperature 0 Radiation Emissivity 0 Temperature 0 Thin Wall Thin Wall Data Thickness 0 Conductivity 0 Heat Source 0 Figure 13 10 Setting up wall thermal conditions in R Desk drop down list of all conditions top and input variables for the Combined Convection amp Radiation in conjunction with Thin Wall bottom Wall Data should be entered O Thickness this is the wall thickness in m O Conductivity thermal conductivity of the wall material W mK O Heat Source this is a heat generation rate in W me Note that the thin wall model will be inactive if its thickness is set to zero 13 8 4 Setting up thermal conditions via R Therm module The user can import wall thermal conditions from a RTH file Ricardo THermal when the Bound ary Setting option is set to RTHERM Figure 13 11 The thermal condition RTH file must be cre ated in advance
377. l_meth 2 1 then near wall method used for domain 2 is wall modelling with wall functions O iwall_modl 4 near wall model from a given method Three options are available iwallf_std 1 standard wall functions iwallf_sca 2 scalable wall functions iwallf_uwb 3 unified wall boundary conditions For example if turb_par wall_modl 2 2 then the wall function type used for domain 2 is the scalable wall function b To get Cy or any other turbulence constant then do the following real wph pointer c_mu call get_turb cmu c_mu Array c_mu now contains Cy values for each domain c Note the use of a 2D array in eps_const and prandt1_number in Table 18 19 18 3 2 12 get_run_ctrl This UAR is used to retrieve information related to run control variables This UAR can be called in 2 ways represented with 2 sections in Table 18 20 a The following example integer iwp curr_iter call get_run_ctrl current_iter curr_iter 1 returns the current iteration curr_iter b The get the name of the current project that is running then character len proj_name call get_run_ctrl project_name proj_name Variable pro j_name now contains the name of the project that is running Ricardo Software December 2009 327 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES subroutine get_turb l var_name ival Input Output
378. le GRD extension is added e g when port mesh is used as the input file the name of the output file will be port GRD 78 4 MESHING 4 3 SETTING UP THE INPUT FILE mesh tri BnD UT aux possibility to visualize GRD i 4 x 4 Figure 4 1 VMESH position is VECTIS MAX system 4 3 Setting Up the Input File VMESH is non interactive and is controlled by the input file and command line options The input file can be generated by Phasel see MESH SETUP section of Phase 1 Phase 1 writes the mesh line coordinates to the file together with the default values for other parameters The user needs to edit this file if values other than the defaults are required for any parameter The principal parameters for the user to check are the following MODEL_NAME The name of the SDF file trifile produced by Phase 1 This file contains the ge ometry of the model the boundary information produced by the boundary identification section of Phase 1 and the boundary types information REFINEMENT_DEPTH The number of times a cell can be subdivided or refined see section 4 6 2 below default value is 2 EDGE_THRESHOLD angle The sharp edge criterion it is a threshold parameter in degrees Ricardo Software December 2009 79 4 MESHING 4 4 COMMAND LINE OPTIONS When angle between normal vectors of two adjacent triangles is greater then this threshold the common edge will be considered as sharp feature The default
379. le computational domain In order to achieve this the solver has a hierarchical domain structure comprising material and fluid solid domains O domain boundaries and domain interfaces The domain boundaries and interfaces are further decomposed into O boundary regions and D interface regions These regions are used to implement physical boundary and interface conditions Note that the word domain is often used to refer to any type of domain 7 1 Multi Domain Approach The domain hierarchy is outlined below 123 7 MODELLING SPATIAL DOMAINS 7 1 MULTI DOMAIN APPROACH Computational Domain This is the top level domain which can be composed of any number of fluid and or solid domains Depending on the solution of en ergy equation contiguous fluid and solid domains are grouped into a global domain Currently the computational and global domain is the same entity Global Domain The global domain represents the solution domain for the energy equation if this equation needs to be solved implicitly over more than one fluid or solid domain Fluid Solid Domain The fluid domain is the same entity as fluid material domain while the solid domain can consist of one or more solid material domains The fluid domains must be separated by solid domains and vice versa Material Domain The material domain is the 3D region of space filled in with the same continuum general fluid or solid materi
380. le in R desk File gt Open gt File then browse to port GRD This will be Ricardo Software December 2009 394 19 TUTORIALS 19 2 STEADY STATE PORT FLOW opened into the 3D canvas in the same way that the triangle file was If both plots are found in the same canvas you may want to remove the port tri plot from the canvas Right click on the relevant canvas name found under the plot port tri in the plottree and select remove See Figure 19 32 iis R Desk 3d1 2 plots loi x poo i Sd o e lax DAB a oo l 160 8 a 592 A O ell IxBMDEBAPIORA BO gt z Plots E port tri A EAE E port GRD Remove Es dla Plot p2 Canvas 3d1 Plots Plot Properties ax Lines none z Faces Color 1 Opacity 0 5 J y JW Auto apply Apply Number of Polyhedron Elements read 181027 Read 181027 elements read 181027 elements 634271 faces 272398 nodes extents 0 15387 lt x lt 0 193 0 020499 lt y lt 0 2705 0 03669 lt z lt 0 32652 Plot Properties command gt oo A a Figure 19 32 Removing a plot from a canvas To visualise the computational cells set the line attribute for the plot in the Plot Properties panel to Mesh If the PlotProperties panel is not open it can be opened by either through the fol lowing menu View gt Plot Properties right clicking the header bar of a currently open panel Figure 19 33
381. lent energy production P defined as Pt Pyu t2 1 Fi y FiQ shows satisfactory agreement with the DNS data for channel flows 4 TTTTT T T T TT TTTT T L I T a e a i T T TTITIT T TT TTTT 7 T T CITITI T TITIN H 4 L L o DNS Channel Re 180 J 0 25 DNS Ch nnel Re 180 a DNS Channel Re 395 a I o DNS Channel Re 395 fax 3 a DNS Channel Re 590 Bi a DNS Channel ld F Wolfshtein 1969 sde 02m Present blend y gt 25 Popovac amp Hanjalic 2007 foots J L a BENS b A 1 Present blend y lt 2 5 e ob lid 1333 E e _ ax 0 15 a Presenta E po A H Presentb 4 Ed o 7 0 1 A 1 0 05 a L g7 l se 0 PA il L 0 a ta 1 viol i 1 tt uul 1 1 1ttit 0 1 1 10 100 y y Figure 9 3 A priory comparison of wall functions for near wall turbulent energy production left and its dissipation rate right Eqs 9 52 and 9 55 9 56 respectively with DNS data for channel flows Moser et al 1999 Ricardo Software December 2009 161 9 MODELLING TURBULENCE 9 5 LOW REYNOLDS NUMBER MODELLING O Turbulent energy dissipation rate The blending function 9 48 with amp a is also employed to define the turbulence energy dissipation rate e E Gevis 1 G Elog 9 54 where the viscous wall limit and log law fully turbulent values denoted as Evis and Ejgg respectively are given by Equation 9 46 Note that the y value
382. ll get_grid_connect l_face_verts internal_face_verts call get_grid_connect isa_face_verts isa_fv Array internal_face_verts now contains a list of face vertices for all internal faces For the face j with k 1 f_verts j vertices the particular vertex is given as kv internal_ face_verts isa_fv j k where array isa_fv provides the starting address in the array internal_face_verts which provides the list of face vertices To get a list of vertices for each boundary face then Ricardo Software December 2009 323 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES subroutine get_grid_geom var_name rval Input Output var_name cha rval real pointer cell_vol cell volume rval 1 n_cells bndf_normal_d normal distance from the near boundary cell centre to the boundary face rval 1 n_bnd_faces face_weight weighting interpolation factor rval 1 n_internal_face subroutine get_grid_geom var_name rval Input Output var_name cha rval real pointer xyz_vert x y z coordinates of vertices rval 1 3 1 n_vertices xyz_face_c position vector at cell face centre rval 1 3 1 n_internal_face xyz_bndf_c position vector at boundary face centre rval 1 3 1 n_bnd_faces xyz_cell_c cell centre coordinates rval 1 3 1 n_cells face_surf_v cell face surface vector rval 1 3 1 n_internal_face bndf_surf_v boundary face surface vector rval 1 3 1 n_bnd_faces face_dist_v distance vecto
383. lobal_max mach_max Global minimum global_min is used in a similar way b Get global maximum minimum for vectors real wp pointer vec Ivector variable integer n Itotal entries in vec l1 n calculate vec for each partition Get global maximum values over all partitions call global_max vec n Ricardo Software December 2009 344 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES subroutine global_max svalue subroutine global_min svalue Arguments Description svalue global max min svalue character max min svalue integer max min svalue real max min svalue double precision max min subroutine global_max values nvalues subroutine global_min values nvalues Arguments Description values gt values 1 nvalues values 1 nvalues gt values 1 nvalues values 1 nvalues nvalues gt global max min character max min integer max min real max min double precision max min integer number of elements for array values Table 18 33 Subroutine global_max Subroutine global_min to get max min over partitions 18 3 2 25 concat_array This UAR can be used to take local array from each partition and concatenate into global array on partition 1 only order by partition number see Table 18 34 Output can be optional global array or local array on partition 1 The first part of the table deal
384. low and the current block may be cycled through the available blocks by using the VCR style controls or the slider on the panel The Delete button will delete the current IJK block and the allowed and forced refinement depths for the current IJK block are set via the panel as shown in the Figure The refinement level of an IJK block is shown by the colour of the block on the canvas A depth of 0 is shown in red of 1 is shown in magenta and a refinement depth of 2 the default is shown in cyan Examples of all of these and of the current block are shown in the Figure Ricardo Software December 2009 66 3 GEOMETRY 3 17 MESH SET UP VIEW OPTIONS Note that the blocks are defined in XYZ co ordinates so that they do not move if the number of mesh lines is changed IJK Block Setup Example of how to define an IJK refinement block Local refinement with in the mesh structure can also be set up using IJK refinement blocks This are set up using the panel highlighted in red below Ricardo Software December 2009 67 3 GEOMETRY 3 17 MESH SET UP VIEW OPTIONS o phaser sa pipese Jaz AG aS wi sole _ x ola yy The process for setting up an IJK refinement block is shown below Ricardo Software December 2009 68 3 GEOMETRY 3 17 MESH SET UP VIEW OPTIONS Joc Refinement Blacks x 1 Define the DEEP maximum surface refinement level and FORCE internal cell refinement level values
385. m the fan centre As a consequence the outflow is characterised by a constant flow angle Twisted Blade For this blade design the tangential velocity is assumed to vary inversely with radial distance from the fan centre The axial velocity field is assumed to be uniform Axial Fans Axial fans are represented by two boundaries which are normal to the fan s axis see Fig 16 7 No boundary conditions need to supplied elsewhere i e from within the boundary panel for these boundaries Radial Fans The figure below Fig 16 8 illustrates a the geometric representation of the radial model For radial fans uniform velocity distributions are assumed at the inlet and outlet The axial com ponent of the flow at the outlet is determined from the Outlet Flow Vector as specified in the fan Ricardo Software December 2009 276 16 MODELLING FANS 16 3 ID MODEL T outlet a Inlet Figure 16 7 Axial fan representation Inlet Figure 16 8 Radial fan representation panel see Fig 16 9 Outlet Flow Vector Figure 16 9 Outflow vector specification for radial fan setup Ricardo Software December 2009 277 17 USING SOLVER 17 1 Introduction The setting up and running of the solver is explained more fully within this chapter As illustrated before this can all be done from within R Desk using the Solver Setup Tree available in a VECTIS project The input file for the solver is a formatted text
386. me advanced modelling features not available in the above linear two equation models It includes Non linear and Algebraic Reynolds stress models Some critical deficiencies of the linear k models can be rectified by defining non linear relation between Reynolds stresses and the mean velocity gradients see Equation 9 12 Elliptic relaxation models Durbin 1991 developed the Elliptic relaxation model referred to as the k v model which is probably the most successful EVM model for near wall turbulence modelling Apart from k and equations two additional equations are solved the equation for wall normal Reynolds stress v and for the elliptic relaxation function 9 3 Linear Two Equation k e Models Together with the mean flow equations 6 28 6 31 transport equations for the turbulent kinetic energy k and its dissipation rate e are solved in the k models Replacing the time and length scale in Equation 9 7 by turbulent scales from Equation 9 2 the turbulent viscosity can be defined as A My PCyfukT T max con E 9 14 where Cy and C are model coefficients and fy is the near wall dumping function employed in conjunction with low Reynolds number models In the above equation for 7 the turbulence time scale k g is bounded from below by the Kolmogorov time scale T C U p as proposed by Durbin 1991 Yang and Shih 1993 In order to close the diffusion term Dx
387. might be dangerous in the situations when the surface intersects an edge of the box more than once Correct tying of patches is not assured in such cases fef b1 b2 bn similar option to fmc but Exact Fit see section 4 5 is forced instead of Marching Cubes Ricardo Software December 2009 80 4 MESHING 4 4 COMMAND LINE OPTIONS info reads the given gridfile and prints its basic information 0 name the output will be written to files name GRD and name OUT instead of files with names derived from the input meshfile REFINEMENT RELATED SWITCHES blendcontrol this switch can be used in cases when there are close surfaces with different blend ing and the user wants the blending to spread in one direction only so as the surface B is not unnecessarily divided because of propagation of refinement from the surface A usage of this switch can decrease the number of generated cell but generation of boxes might be 10 slower DELETING OF SMALL CELLS smallio ratio sets the ratio SMALL_CELL_VOLUME BOX_VOLUME which defines the maxi mum volume of cell which should be considered as small for all open flow boundaries input and output or cyclic boundaries the default value is 1 0 x 107 smallint ratio sets the ratio SMALL_CELL_VOLUME BOX_VOLUME which defines the maxi mum volume of cell which should be considered as small for all boundaries that form interface between domains the default value is 1 0 x 1076 smallb bouinx ra
388. mixture model where the mass conservation for each phase is given as dakpk where superscript k denotes the k th phase and a is the phase volume fraction The mass source I in relation 12 28 is given by f 12 65 evap cond where ris and i ond Yepresent the evaporation and condensation rates These phase change rates for cavitation are derived from Rayleigh Plesset and are presented here for a number of models 12 4 2 1 Singhal et al Model The cavitation model presented here builds on the full cavitation model proposed in Singhal et al 2002 Ricardo Software December 2009 209 12 MODELLING MULTIPHASE FLOWS 12 4 PHASE CHANGE MODELLING The vapour mass fraction and fluid density is given by 1 y 1 4 ES P Py Pl 12 66 where P Py are the density of liquid and vapour respectively Similarly fy represent the mass fraction of vapour The vapour volume fraction is derived as ee ee 12 67 Py The phase change rates E and TE q are approximated as follows Vk 2P P Pro Devap Ce P gt 1 Jin 12 68 k Vk 2P P ne PyQ Lond Ce o PiPv E Pi p where Cevap and Cond are empirical constants and their recommended values are 0 02 and 0 01 respectively O is the surface tension of the liquid The phase change threshold pressure P in rela tion 12 68 and 12 69 is estimated by taking into account the effect of turbulence on cavitating flow
389. mple and efficient method to import already defined boundary and interface regions from the numerical grid A boundary and interface region sub trees a part of Solver Setup Tree listing typical boundary types for the Fluid Domain are illustrated in Figure 13 1 The Solid Domain is described by either Wall Boundary or Symmetry Boundary In case of the Multi Component Phase R Desk adds by default the node Boundary Species for the following boundary condition types Inlet Given Velocity Mass Flow Rate Pressure and Stagnation types The node Boundary Species contains all species defined for each Fluid Phase in the Solver Setup Tree see also Figure 8 3 Ricardo Software December 2009 214 13 BOUNDARY amp INTERFACE CONDITION TYPES 13 1 SETTING UP RTHERM Fluid_1 Fluid Domain Monitoring Points Algorithm Turbulence Model Discretise Phase_1 Fluid Phase Equations amp Solver Boundary Regions outlet Pressure Boundary Phase_1 Boundary Phase Boundary Species Species_1 Boundary Spe Species_2 Boundary Spe Phase_1 Boundary Phase Boundary Species Species_1 Boundary Spe Species_2 Boundary Spe head Wall Boundary block Wall Boundary inlet_pipe Wall Boundary Interface Regions Interface_ 94 Interface Region Fin Interface Region Figure 13 1 Boundary and interface region nodes in the R Desk Solver Setup Tree Similarly if passive scalars are defined
390. mple test where one patch is hotter than its neighbours is shown below Ricardo Software December 2009 269 15 MODELLING RADIATION 15 5 RADSOLV THEORY Figure 15 10 Example results for the surface conduction model showing the patch temperature ORicardo Software December 2009 270 16 MODELLING FANS 16 1 Introduction And Overview There are currently two fan models implemented within VECTIS MAX These are Subdomain and 1D Both models require the specification of a pair of inlet outlet boundaries subdomain interfaces No conventional I O data should be entered for these boundaries instead a new fan should be created as shown in Fig 16 1 Phase_1 Fluid Phase Initial Condition Postprocessing Output Equations amp Solver Fan Model Bnd_Reg_1 Wall Boundary Radiation Expand All Interface Regions Sub domains L Report Regions Add Fan Model Collapse All Figure 16 1 Adding a fan In general the fan geometry can be simplistic and so the whole fan assembly can be represented by a simple model geometry e g cylinder shape which has the appropriate inlet and outlet areas that correspond to those of the real fan geometry Fan parameters The various fan panel parameters in Fig 16 2 are explained below Some of these are only relevent to one of the model types as indicated Subdomain ID This is relevant when the Fan Modelling is set to Subdomain It specifies which subdomain is
391. mport tube GRD 11 01 42 INFORMATION Closing fie D RPe trairing_tutorials V4 mesh_import tube GRD Figure 19 92 Creating a new face set Ricardo Software December 2009 439 19 TUTORIALS 19 4 IMPORTING THIRD PARTY MESHES 19 4 2 2 Editing Sets Once the face set has been created then the boundary faces that will make up the region should be painted This is done by editing the face set Either right click the set name and select edit or select the set and then press the edit button The set being edited will be written in red and the edit options will be displayed in the sets panel Figure 19 93 Different options are available for editing sets faces can be selected deselected or toggled L R Desk 341 p1 tube GRD TES FR File Edt View Options Window Hep 18x E Project FEV J A B ll Mewe 1 jJe 20s ae gt ehn Sr 11 02 02 DEBUG deleting plot p1 64 11 02 02 DEBUG processing plot p1 11 02 02 INFORMATION fie D Opening RPeftraring_tutonals V4 mesh_import tube GRD 11 02 02 INFORMATION Closing fie D RPe training_tutorials V4 mesh_import tube GRD Figure 19 93 Editing the face set Different picking options are available for editing sets Figure 19 94 L R Desk 341 p1 tube GRD lB x MR File Edt View Options Window Help l8 x ADS pde a _ Je 2 jaj Jeh 8 3 M gt x H 11 02 28 DEBUG deleting plot p1 68 11 02 28 DEBUG processing plot p1 11 02 28 IN
392. n ir is relevant Note that a boundary condition type and corresponding options associated with the specified region should not be modified i e user accessible variables l_reg_cond and l_reg_opts available via get_reg should not be modified Also there is no need to select and modify region variables with the symmetry or outlet boundary condition type The user specified static pressure must be relative to the reference pressure while the total stagnation pressure is always absolute Important note In case of pressure total pressure and mass flow boundary types the user has to supply turbulence intensity and length scale boundary fields instead of actual turbulence energy and dissipation fields respectively They are temporarily stored in the same arrays used for the turbulence energy and dissipation Ricardo Software December 2009 350 18 USER PROGRAMMING 18 4 USER PROGRAMMABLE ROUTINES subroutine upr_sources mat ieq idt iph isp icell icel2 vol_ph ap_fi src_fi Arguments Description mat char index of fluid solid material domain leq gt integer equation index idt integer domain type index iph integer fluid solid phase index isp gt integer fluid species index icell integer starting boundary id for ir icel2 integer ending boundary id for ir vol_ph real cell volume occupied by phase ap_fi real output central coefficient of the discretised equation src_fi
393. n R Desk 13 8 Wall Either smooth or rough impermeable walls are assumed For real fluids the fluid in contact with the wall has the same velocity as the wall This no slip condition is enforced by specifying the wall velocity components In case of the inviscid fluid slip wall the wall shear stress is zero Consequently the velocity normal to the wall must be zero and the tangential component is equal to its counterpart at the near wall cell centre These are actually the symmetry plane conditions which means that the slip wall can be modelled using the symmetry plane condition type Note however that the symmetry conditions will be enforced for all equations solved for In case of heat transfer the wall represents a solid material at which various thermal conditions can be specified such as the prescribed temperature or heat flux As the convective fluxes are zero at the wall only the diffusion fluxes need to be considered in order to predict the wall shear stress temperature or heat flux distribution For this a general expression for the discretised boundary flux Equation 14 73 is employed see Implementation of boundary conditions The wall diffusion coefficients used in the above equation are defined by Equation 9 40 They are valid for both laminar and turbulent flow simulations For the latter case these are effective wall turbulent coefficients based on the wall function approach see Section Near Wall Modelling for more detail
394. n System Steering Committee http www gre nasa gov WWW cgns charter index html O HDFS SZIP The HDF Group University of Illinois http www hdfgroup org HDF5 Ricardo Software December 2009 3 INTRODUCTION VECTIS MAX is a next generation computational fluid dynamics CFD software product devel oped by Ricardo Itis a general purpose tool for solving advanced 3D industrial fluid flow and heat transfer problems with particular attention given to the requirements of automotive applications The product contains a fully automatic Cut Cartesian mesh generator a multi domain solver capa ble of running on arbitrary unstructured meshes and an advanced Graphical User Interface GUI to prepare geometry and display simulation results Typically the multi domain features are used to simulate Conjugate Heat Transfer CHT in automotive applications such as cooling of cylinder heads and engine blocks 2 1 Main Features and Capabilities An overview of the VECTIS MAX capabilities is given below by considering the three key soft ware modules 2 1 1 Meshing vmesh O Rapid and automatic vmesh generator The vmesh program is a fully automatic hexahedral mesh generator producing Cut Cartesian meshes using a hybrid scheme of Exact Fit and Marching Cubes The ability to mesh easily complex CAD geometries without resorting to the boundary surface grid generation is the main advantage of vmesh The program has a number of impr
395. n is solved in terms of total enthalpy or energy the temperature field is extracted according to the definition of total enthalpy energy Equation 10 9 or 10 25 During iterative solution the new temperature is under relaxed by using the pressure under relaxation Thus when solving the energy equation in fluid domains the user can specify the solution in terms of either Enthalpy or Temperature by activating the corresponding RadioBoxes in the Equations Solver panel shown in Figure 10 9 The default Under relaxation Factor Convective Equations Solver a e y Momentum a y Mass Energy Species I Turbhulant Een nra s Energy Equations Option Enthalpy Temperature Relaxation Factor 0 8 Convective Scheme UDS 4 Blending Factor 1 User Defined Sources ha Figure 10 9 R Desk setup Setting solution of the energy equation in the Equations Solver panel Scheme and Blending Factor for convective schemes are also shown The selection of a convective scheme is discussed in the section Interpolative schemes for cell face values see Figure 14 5 With reference to the same figure the setting of parameters which control the linear equation solver s can be found in the section Linear solvers If there is a User Defined Source it should be activated in this panel Note that higher values of the under relaxation factor should be tried 0 9 for the solution in terms of temperature The h
396. n is specified in the Auto Paint Angle option the default being 45 degrees Having specified an angle selecting the Auto Paint button divides the various surfaces on the model into separate colours The Reduce command can then be used to consolidate on the number of colours used and organise these so that the minimum number of colours are used whilst still separating surfaces Once the surfaces have been separated the user may need to reorganise the sequence and or con tents of each colour so that items such as inlet outlet boundaries are uniquely defined For this example the boundaries should be painted as described in Table 19 1 It is possible to drag and drop triangles from one boundary to another by clicking and dragging on the number of triangles in that boundary For this example there are four boundaries the inlet defined as the back of the plenum chamber i e opposite the inlet port the outlet defined as the bottom of the cylinder the valve used for boundary refinement and the rest which are walls Ricardo Software December 2009 388 19 TUTORIALS 19 2 5 Boundary Description The figure below shows the definition of the inlet outlet and main wall boundaries Figure 19 23 Inlet outlet and main wall boundaries 19 2 STEADY STATE PORT FLOW Use the Chop geometry tool along with the Mark Face tool to define the back of the valves as a separate boundary as shown in the figure below It is not st
397. n of engine components in the RTH file and in the solver x cylinder number y port valve number HF heat flux EHT external heat transfer R Therm module defines the engine component geometry in terms of a regular relatively coarse finite element mesh Thermal data fields heat flux external heat transfer coefficient and tempera ture are then generated based on the Ricardo s empirical data for each component These fields are described by thermal values at nodes of the R Therm finite element mesh As the VECTIS MAX boundary mesh describing engine components which are identical to those used in R Therm module is unstructured irregular the solver thermal boundary values defined at boundary face Ricardo Software December 2009 228 13 BOUNDARY amp INTERFACE CONDITION TYPES 13 8 WALL centres are generated from the R Therm nodal values using a linear interpolation method The interpolation is required only if the thermal field is not uniform as it is a case for cylinder liners and flame faces To perform an interpolation or assign uniform thermal values at solver s wall boundaries the user has to associate the wall boundary marked with the RTHERM setting flag with an R Therm set from the RTH file If the user sets a new project and has already created a Domain Structure Figure 7 2 the Set list of the RTHERM Set Selection shown in Figure 13 11 will be initially empty After selecting the RTHERM option for the first time w
398. n their incompressible counterparts The Reynolds decomposition is used for density and pressure P P p P p p 6 25 while the Favre decomposition is applied to velocities mass fractions temperature total enthalpy U ul G 4 c T T T A H H 6 26 and other variables appearing in the convective terms The Reynolds decomposition is generally employed for viscous stresses mass diffusion and heat fluxes a Tij Tij Hijo Jenj Ves j Jaj Ti T tai 6 27 Ricardo Software December 2009 119 6 SOLVER FUNDAMENTALS 6 4 REYNOLDS AVERAGED EQUATIONS as there are no obvious advantages in using the Favre decomposition Huang et al 1995 How ever some care should be taken to satisfy that the average of fluctuating parts is zero for example T i 0 Employing the above combined Reynolds and Favre decompositions ensemble averaging of the instantaneous equations results with the RANS equations Pa P B v 0 6 28 2 p a Bei 5 Uei a Tej Jb 7 5c 6 29 p S gt pu 0 U P Pfit x ti Ti 6 30 do d ix OP O A Eee 5 PH T ju P Si Yes Fr ag WAM ga L Tast ah 6 31 J re a a qe ld Tij T Tiju aul pr BUg j p FG ful 4 J The unknown quantities appear in the species momentum and total enthalpy equations O Turbulent mass diffusion flux Ji j pelt pel 6 32 O Turbulent momentum flux Reynolds stress tensor t Dy ny
399. n with species indices in Table 18 12 see subroutine get_ species as integer ir region index integer is species index real 23 bval integer iwp pointer is_sp_bndreg ie_sp_bndreg integer iwp pointer a_sp_bndreg get index of starting amp ending boundary region for species call get_species ise_specs_bnd_reg is_sp_bndreg ie_sp_bndreg get the starting allocation addresses for species at boundary regions call get_species isa_specs_bnd_reg a_sp_bndreg do ir is_sp_bndreg is ie_sp_bndreg is bval sp_val irt ta_sp_bndreg is end do Ricardo Software December 2009 316 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES Now bval represents the boundary value at region ir for species with index is Data for option specs_flux are used in a similar way to specs_value For passive scalar option ps_ value and ps_f lux index pointer from Table 18 13 are used Additional boundary region data real type variables are provided by addresses in the array 1_ reg_value 0 nbcopr 0 n_regions For the considered region ir and each address the array 1_reg_value stores the following region wise data ibflo 1 mass flow rate 1_reg_value ibflo ir O ibmachv 2 boundary region Mach number 1_reg_value ibmach ir ibptot 3 total pressure 1_reg_value Ibptot ir O ibttot 4 total stagnation temperature 1_reg_
400. nce cell gt face fpre 1 3 gt real pressure force fvis 1 3 gt real viscous force area gt gt real boundary face area Table 18 30 Subroutine local_force to calculate viscous and pressure forces at individual boundary faces amp interfaces 18 3 2 22 user_post This UAR can be used to register user post processing data to stored at the user selected post processing write frequency see Table 18 31 This can be done initially e g within upr_init or within an if block if using upr_generic Consider the following example Ricardo Software December 2009 34 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES subroutine upr_generic id icall_pos integer iwp intent in id amp global domain id 1 icall_pos calling position with solver real wp pointer save upr_data Pointer to user data integer iwp oe de Domain type id integer iwp d_type Data type e g cell based character len x d_name Data base name integer iwp pointer ic_s amp Cell start id s for each domain rie eii Cell end id s for each domain integer iwp ee iel 102 if icall_pos icp beg_run then lGet domain cell indices get_domain ise_cell ic_s ic_e Allocate upr_data for domain 1 and leave allocated d_type i_cell d_name test_data idt 1 ic ic_s idt ic2 ic_e idt allocate upr_data ic1 ic2 lregister user data for domain idt call user_
401. nd Collapse Expand All Collapse All Figure 8 9 R Desk setup Defining the multi component phase in terms of species Left click on the Species node opens a new panel Figure 8 9 middle where a new species can be added to the mixture by clicking on the Add Child Species push button In this case the species tree will be refreshed and show just added species Right click on any Fluid Species node pops up a panel Figure 8 9 right which can be used to add a new species or delete the current species if the mixture contains more than two species Physical properties of each species are specified in the Species Properties panel which is displayed after clicking on the Properties node of the considered species see Figure 8 9 left In addition to the properties of single component phase this panel shown in Figure 8 10 contains sub panels to specify calculation options for the Mass Ricardo Software December 2009 144 8 MODELLING CONTINUA 8 6 SELECTING CONTINUUM AND ITS PROPERTIES Diffusion coefficient Thermal Mass Diffusion Heat Formation Temperature Formation and Turbulent Schmidt Number The heat enthalpy of formation and the temperature of formation are relevant to the reacting mixture flows Mass Diffusion Thermal Mass Difference Option Constant Values Option l Constant Values 4 Value 1 Value 0 Heat Formation Temperature Formation Option Constant Values 2
402. nd its compressibility in Fluid Phase panel 10 3 4 Inviscid and viscous regime Fluid flows in nature are always viscous and described by the Navier Stokes equations For high Reynolds number compressible flows inviscid flow model is sometimes used where the viscous and turbulent stress tensors in Equation 10 2 are neglected The resulting momentum equations are then called Euler equations The inviscid fluid can not stick to walls and therefore the slip wall boundary conditions are used The inviscid flow model is selected in the Turbulence Model panel after displaying with left click the Turbulence Modelling Approach list box and selecting the Inviscid Flow option Figure 10 4 Turbulence Model Turbulence Modelling Approach a E Inviscid Flow No Turbulence Models ba Turbulence Family Inviscid Flow Turbulence Model 4 gt 4 gt Figure 10 4 R Desk setup Selecting inviscid flow regime in the Turbulence Model panel 10 3 5 Potential flow model An additional condition can be imposed on the inviscid flow the irrotational velocity field V x U 0 This leads to the simplest flow model potential flow The velocity field can be then described Ricardo Software December 2009 172 10 MODELLING SINGLE PHASE FLOWS 10 3 MODELLING FLUID FLOW by a scalar potential function a velocity potential Y as r o U V or U 10 16 Ox The equation for the velocity potential is obtained
403. nd the area averaged wall temperature near wall temperature and heat transfer coefficient 13 8 1 Basic wall setup In the Solver Setup Tree left click on a boundary region to get the boundary setup panel Under Boundary Condition Type select the Wall option from the ListBox and the Wall Boundary panel containing wall boundary settings is displayed as in Figure 13 9 The Cartesian velocity components for wall motion can be specified in the Wall Velocity section where zero values for all velocity components indicate a non moving wall In case of modelling a turbulent flow where the wall roughness effects are significant the user should provide values for the Roughness Height and Roughness Constant These values are used in the wall function approach modified to take into account the roughness effects 13 8 2 Wall thermal conditions When solving the energy equation the user can specify one of the following thermal conditions prescribed wall temperature prescribed wall heat flux external convective heat transfer O external radiation heat transfer and O combination of external convective and radiation heat transfer Ricardo Software December 2009 223 13 BOUNDARY amp INTERFACE CONDITION TYPES 13 8 WALL Bnd_Reg_5 Wall Boundary a E Region Name Bnd_Reg_5 Region ID 5 Material ID 1 Heat Flux 0 y Boundary Report Temperature 293 15 Coupled Link Number
404. ndicator function phase fraction which is equivalent to the volume void fraction of the corresponding phase in case of steady state flows In general the phase fraction which has been introduced in the section dealing with properties of multiphase mixture is the probability that a certain phase exists at a certain point in space and time The averaging process introduces in addition to the turbulent Reynolds stresses or diffusion fluxes unknown extra terms into basic equations These terms describe inter phase mass momentum and energy transport and have to be modelled in terms of known Ricardo Software December 2009 197 12 MODELLING MULTIPHASE FLOWS 72 2 EULER EULER MODELLING EQUATIONS base variables The Reynolds stress and inter phase transport closure problem is the crux of multi fluid methodology since this closure depends on the nature of the flow i e it relies on empirical and problem dependent in put However the Euler Euler modelling can be applied to all multiphase flow regimes and its predictive capabilities are satisfactory in many application areas cf Lahey and Drew 2001 The remainder of this section deals only with the Euler Euler multi phase modelling 12 2 Euler Euler Modelling Equations The averaging of instant multiphase flow equations can be done in different ways The mass density weighted averaging gives a simple form of conservation equations For example the ensemble average velocity of the
405. nding distance 2 RD 1 option seton 92 Boundary refinement depth 2 blending distance 2 ooa ooo o 93 GH ODESSA Ree A A a a aaa e Se Gs 99 Control volume and MOON c m sa a a o EER Se ee eS 99 Various 2D and 3D cells CVs o ei a ta a a a a ee ew 101 Meshiguallity checKs ec idence on A ena whe Bh eed Be ae Bg og Be ands ee 104 Picture of 2 materials partitions with slight coincidence Green arrow shows how partition interface is moved from smaller 1 to larger partition 2 00 108 Example of colouring for 7 partitions o o 108 Picture of three meshes to be joined Gaps between meshes are exaggerated 109 xvii LIST OF FIGURES LIST OF FIGURES ZA T2 13 7 4 8 1 8 2 8 3 8 4 8 5 8 6 8 7 8 8 8 9 8 10 8 11 8 12 8 13 9 1 9 2 9 3 9 4 9 5 10 1 10 2 10 3 Simple multi domain simulation example o o e 125 R Desk setup Creating domain structure ee 127 R Desk setup Example of created multi domain structure left and previewed fluid domain right for conjugate heat transfer o o 128 R Desk setup Example of a single fluid domain tree with a porous sub domain left and its sub domain interface regions right o ooo o 128 R Desk setup Selecting a multiphase model o ooo e e 139 R Desk setup Adding a fluid phase lt o sena ad ke a a ee ae a
406. ne in the same way as for to the Mass Flow Rate Boundary type Ricardo Software December 2009 220 13 BOUNDARY amp INTERFACE CONDITION TYPES 13 5 STAGNATION INLET Bnd_Reg_6 Pressure Boundary a Ex Region Name Bnd_Reg_6 Region ID 6 Material ID 1 y Boundary Report Coupled Link Number 0 i Given Static Pressure Boundary Condition Type Pressure Total Boundary Condition Op Given Static Pressure wa Average Boundary Setting Uniform Values ES ES Inflow Outflow Inflow 211 Outflow Figure 13 5 Setting up a Pressure Boundary type in R Desk 13 5 Stagnation Inlet The stagnation inlet describes stagnation or total boundary conditions which correspond to the stagnant fluid with zero velocity upstream of the boundary It should be used for sub sonic compressible flows The implementation of this type of boundary is explained in section Subsonic Inlet Apart from the stagnation pressure Equation 14 66 and temperature Equation 14 67 the velocity direction has to be specified 13 5 1 Setting up a stagnation inlet boundary condition To set up a Stagnation Boundary left click on a boundary region in the Solver Setup Tree to display the boundary condition setup panel Under Boundary Condition Type select Stagnation option from the ListBox and the Stagnation Boundary panel with relevant settings is displayed as Figure 13 6 shows The velocity direction
407. nel was described in Section 8 6 2 The subpanel Initial Condition shown below lets the user specify initial flow velocity pressure temperature etc The Postprocessing Output Initial Condition a K X Velocity 0 Y Velocity 0 Z Velocity 0 Pressure 100000 Temperature 293 15 Turbulent Kinetic Energy 1 Turbulent Dissipation 1 Turbulent Viscosity 0 Volume Fraction 0 Figure 17 8 Initial Conditions Panel subpanel is for specifying which quantities are to be written to the postprocessing file post See Figure 17 9 17 8 Fluid Species The species panel see Figure 8 9 allows the user to add new species for given phase It also contains subpanels Fluid Species used to define species properties which contains the single component phase properties as well the species specific properties see Figure 8 10 Below the Fluid Species panel there is Output panel used to specify the quantities to be written to the postprocessing file post See Fig 17 10 17 9 Boundary Region Definition The Boundary Region panel was mostly explained in Section 13 1 The Boundary Report Check Box is used to indicate whether ASCII SDF reporting is required for this boundary See Sec tion 17 6 on alpha numeric reporting for more information Typical contents of an io report file are shown below Ricardo Software December 2009 284 Postprocessing Output o
408. ng boundaries conformal can complete its task successfully 4 Exact triangles defining interfaces Interfaces on all complementar meshes need to be defined by exactly the same triangles Also it is not possible to have the situation that triangle defined as interface triangle does not have any corresponding triangle on the complementar mesh Also all the corresponding triangles must be marked as interface no non interface interface couples 5 Mesher needs to be informed about interfaces When preparing meshes for multidomain the mesher needs to be aware which boundaries are interfaces int or interface command line option needs to be applied because these boundaries are treated differently Patches on the in terface are more simplified so as the later work to make the grids conformal is easier Also the min max extents of the model are not found on the geometry but it is taken from the positions of mesh lines This is necessary to ensure the same extents of boxes when generating cells on ORicardo Software December 2009 94 4 MESHING 4 9 WARNINGS AND ERRORS complementary meshes Also a different settings for deleting of small cells is applied here see the option smallint 6 Post meshing tool ensuring conformal faces When all meshes for multidomain calculation are generated it is necessary to use the tool of the mesher that ensures conformal faces on interfaces between each couple of meshes conform or conformrew command lin
409. nlet outlet 3 the flow outlet with the prescribed flow split can not coexist with pressure boundaries In case of compressible fluid flows the recommended combinations are given as a function of the flow type Mach number throughout a fluid domain O Subsonic flow Ma lt 1 within the fluid domain Inflow Outflow Stagnation Inlet Pressure Outlet specified pressure Mass Flow Inlet Pressure Outlet specified pressure Pressure Inlet Ma lt 0 3 Pressure Outlet specified pressure 213 13 BOUNDARY amp INTERFACE CONDITION TYPES 13 1 SETTING UP Transonic flow Ma lt 1 and Ma gt 1 within the fluid domain Subsonic Inflow Subsonic Outflow Stagnation Inlet Pressure Outlet specified pressure Mass Flow Inlet Pressure Outlet specified pressure Subsonic Inflow Supersonic Outflow Stagnation Inlet Pressure Outlet all variables extrapolated Supersonic Inflow Supersonic Outflow Stagnation Inlet all variables specified Pressure Outlet all variables extrapolated O Supersonic flow Ma gt 1 within the fluid domain Supersonic Inflow Supersonic Outflow Stagnation Inlet all variables specified Pressure Outlet all variables extrapolated Boundary regions associated with solid materials can have either Wall or Symmetry Plane condi tion type A boundary condition type along Interface Regions is relevant for fluid solid interfaces where wall boundary type is applicable to the solution of fluid flow equ
410. nlets and flow outlets In practice CFD problems are usually characterised by complex physics and complex geometry involving a number of fluid solid materials and associated material domains This means that we have to model both geometry replace real spatial domains with the material domains and 112 6 SOLVER FUNDAMENTALS 6 1 INTRODUCTION physics After modelling a numerical method is employed to solve the governing equations The numerical method converts the mathematical model into a system of algebraic equations through the discretisation process This process involves the discretisation of DO spatial domains i e space O time O and transport equations The space discretisation creates a numerical mesh which describes the solution domain in terms of a finite set of non overlapping control volumes or cells enclosed by faces and supported by vertices With the mesh support the governing equations are discretised over each cell In summary the numerical simulation comprises the following basic steps CFD PROBLEM DEFINITION lt Real world described by O Spatial domains geometry amp their bound aries interfaces O Conservation laws O Continuum properties MATHEMATICAL MODELS lt Modelling real world O Solution computational domain amp physical bound ary or interface conditions O Governing equations Constitutive and thermodynamic relationships thermo physical properties
411. no means of enforcing read only arrays when read via a pointer 18 7 Examples 18 7 1 Example of the user properties routine Description User programming routine to specify thermo physical properties of individual phase or species Ricardo Software December 2009 353 18 USER PROGRAMMING 18 7 EXAMPLES Filename upr_properties f90 subroutine upr_properties pro_name mat iph isp icell icel2 jbnd1l jbnd2 Jiul ee yp Ge Le fio 1_by 1 014211 This routine is called for a phase iph species isp present in fluid solid Modules used imported type definitions parameters scalars and arrays use upr implicit none character len intent in pro_name Iname of a property field integer iwp intent in mat amp index of material domain 1ph amp fluid solid phase index isp amp fluid species index icell icel2 amp start amp end cell indices yp jbnd1 jbnd2 amp start amp end bnd face indices Jjiul jiu2 amp start amp end high interfaces 3111 3112 Istart amp end low interfaces real wph intent inout fi_c icell icel2 amp property s cell values fi_b jbndl1l jbnd2 amp boundary values fi_ui jiul jiu2 amp values at upper interfaces 21114 Gill 3112 Ivalues at low interfaces optional ss fi by EU 61 11 Local Variables real wph pointer SS EL gt null amp cell temperature tb gt null amp bnd temperature ti gt null amp uper interface t
412. node This can be done by checking against the minimum and maximum values of over the cells that share each vertex of the central cell C bu max one min Gis ogag 14 21 Our experience has shown that the computationally less expensive procedure which involves only the neighbours that share faces of the central cell produces satisfactory results and this procedure is used in this work The TVD and CBC conditions are illustrated in Figure 14 4 in the form of a Normalised Variable Diagram NVD It follows that the cell face values should lie within the shaded areas in the monotonic range 0 lt dc lt 1 and on the line j dc outside the monotonic range Obviously the convective schemes with linear characteristics in the NVD diagram CDS LUDS QUICK may violate boundedness criteria and some form of a non linear or piecewise linear scheme need to be used It is useful for many reasons to express a general function f c in a form that involves the flux limiter Sweby 1984 0 lt py r lt 1 j c o 1 dc 14 22 The flux limiter itself is a function of the argument r which is defined as c _ Oc bu pice A 14 23 1 c 0D E c In terms of unnormalised variables Equation 14 22 becomes m 9 Op OI p op 14 24 J Ricardo Software December 2009 239 14 NUMERICAL SOLUTION 14 1 FINITE VOLUME DISCRETISATION Soi SS SMART AVLSMART 7 A Pe 0 25 05 0 75
413. normal is recalculated and will consequently point in the opposite direction Ricardo Software December 2009 24 3 GEOMETRY 3 7 TRIANGLE CREATION OPERATIONS Move Connected Marked Area Button a This function works in conjunction with connected faces or the faces defined by the mark com mand A face is selected by clicking with the left mouse button A panel will then appear to allow the specification of a transformation to apply to the selected triangles A scaling rotation or translation may be applied For a scaling only the scale factor needs to be supplied ij Geometry Movement Specification Choose Move Style Scale Rotate Translate Scale multiply by fi Apply Cancel Scaling Transformation Panel For a rotation the angle of rotation specified counter clockwise the axis vector to rotate about and a point that the axis vector passes through are needed The rotation axis can be defined by entering the x y z vector or by selecting 2 points 1 line a triangle normal or a normal to 3 points The point to rotate about can be entered or defined by selecting an existing point Ricardo Software December 2009 25 3 GEOMETRY 3 7 TRIANGLE CREATION OPERATIONS Rotation Transformation Panel For a translation the distance and direction are specified The direction vector for the transforma tion can be defined by entering the x y z vector or by selecting 2 points 1 line a triangle
414. nother cell in this column triangles may be moved from boundary to boundary The type of the boundary as far as the calculation is concerned is set in the fourth column Clicking the left mouse button in the cell cycles between the options Wall Zero Grad Inlet Outlet and Cyclic Boundaries may be given unique names by typing them into the sixth column Ricardo Software December 2009 55 3 GEOMETRY 3 15 BOUNDARY PROCESSING The fifth column indicates whether the boundary has refinement defined for it Selecting the table cell pops up the boundary refinement panel in which the boundary refinement is defined Then selecting Ok pops down the panel and the cell displays Yes to indicate that boundary refinement has been defined The sixth column is used to define a boundary name for each universal boundary number Each boundary name can be up to 80 characters long Comments can be defined for each boundary The seventh column indicates whether the boundary has a comment defined for it Selecting the table cell pops up the boundary comment panel in which the boundary comment can be set Then selecting Ok pops down the panel and the cell displays Yes to indicate that boundary comment exists The highlighting upon one row of the table indicates the currently selected boundary This may be changed by clicking the left mouse button in the identifier number column for the boundary to be selected This selection determines which bound
415. ns Stitch Tools Boundary Painting Mesh Set up Tools and Mesh View Tools buttons can be used to switch between the modes and pop up the respective panels Elle al Toggle Surface Display Button A The toggle surface display button switches the triangle surface display on off Toggle Line Display Button Al The toggle line display button switches the triangle edge display on off Cancel Command Button XI The cancel command button cancels any current command selected from the tool panels and restores the cursor Undo Button oj The undo button undoes the action of the previous command View Reset Button The view reset button resets the view orientation and re centres the model Ricardo Software December 2009 13 3 GEOMETRY 3 3 GRAPHICS INTERACTION N MX 3 3 Graphics Interaction Manipulation of the model image on the graphics canvas can be performed via the mouse or the keyboard Model translation The middle mouse button is used to alter the view of the model On its own the middle mouse button is used for translation holding the button down and moving the mouse makes the model move with the mouse Model rotation If the SHIFT key is held down at the same time as the middle mouse button it performs rotations With the mouse pointer near the centre of the window moving horizontally rotates about the y axis and moving vertically rotates about the x axis Near the edge of the window
416. nsists of many surfaces that are individually triangulated the triangles on adjacent surfaces do not match and thus require stitching together Some of this process may be completed automatically by the Phase 1 Auto Stitch command This tool may be invoked by selecting the Auto Stitch option from the Operations menu Edit View Toolbars Operations Help Auto Stitch Ctrl U Check Self Intersection Check for Unstitched Decimate Triangles Make Geometry Set Model Time Boundary Painting Mesh File Setup IJK Refinement Blocks Generate Mesh Slice Mesh View or from the Auto Stitch button f Ricardo Software December 2009 2 3 GEOMETRY 3 7 TRIANGLE CREATION OPERATIONS on the stitching tools panel An example of a section of geometry before and after stitching is shown in the Example Figure It can be seen how the stitching inserts extra triangles in order to close the gaps between the surfaces Stitching only operates on the active visible part of the model i e that part not removed by the Chop Area function My Example of the Results of Auto Stitching LS LAA Ne QU Occasions may arise in which the introduction of additional triangles may result in spurious changes to the geometry In such cases the automatic stitcher makes cautious decisions to avoid introducing errors into the surface Under these circumstances it is unlikely that a fully stitched model will r
417. nt boundaries for the CFD calculation 3 To set up multiple geometry positions for use in moving mesh calculations 4 To set up the global mesh and other control parameters for the mesher 5 To start up the mesh generator and view the resulting mesh 3 2 User Interface The Phase 1 GUI consists of the main menus the button bar the OpenGL canvas the information area and the tool panels 11 3 GEOMETRY 3 2 USER INTERFACE Ricardo Vectis Phasel None 2 5 x Fie Edt View Toolbars Operations Help BAMA S e E JO A N 2 E w Yl Options Stitch Mesh Lights TT Visualize Leak Paths Morel Fast Rotete Move I Show Surfaces Flat Shading FT Fip Surface Display TT Show Mesh Setup Number cels 3d Subdivide IJK Block Display Cor Solo G Al Outine Colours Off C Muki Mono Highight J Parts IV Holes Y Sharp Edges Number Nodes Mesh View e Show Faces Bounds Domans VECTIS 3 2 1 The Button Bar commands Open File Button S Model files and mesh files can be opened using the file open button Save Geometry Button TT Geometry can be saved using the Save Geometry button Ricardo Software December 2009 12 3 GEOMETRY 3 2 USER INTERFACE Save Mesh Set up Button El Mesh set up files can be saved via the Save Mesh Set up button Print button a The print button pops up the print panel The Toolset Buttons The View Optio
418. nt with surface triangles need to be found because only triangles linked to the shoeboxes intersected by the segment need to be taken into account 6 CONSISTENTLY ORIENT TRIANGLES Since the orientation of the input triangulated sur faces is random in the trifile it is necessary to orient them so as the normal vectors of the triangles always point inside into the flow domain 7 PREPARE IN OUT STATUSES Here in out statuses of the vertices of the global boxes which are defined by meshlines are found by ray method This method inheres in counting inter sections of the segment connecting a point with known in out status and the tested point with surface triangles If the number of intersections is odd the in out status of the tested point will be opposite when the number is even the in out status stays the same 8 GENERATION OF BOXES There is a special section below 4 6 which describes in detail how the boxes are generated 9 PREPARE VELOCITY LOCATIONS The common rectangular sides of the generated boxes are called velocity locations According to which direction they are perpendicular U V and W velocity locations are distinguished perpendicular to x y and z respectively The velocity locations serve for navigation through boxes and for generation of inner faces on them In this part of the algorithm the generated boxes are indexed first Then velocity locations are generated Ricardo Software December 2009 85 4 MESHING
419. nvergence Criterion 0 001 Figure 19 77 The Global Domain Panel be specified here Again here we will use the GRD file that was previously generated normally coolant GRD Convergence Criterion This is the criteria used to determine whether the simulation has converged In general as the complex flow within the coolant circuit is never completely steady and tends to fluctuate in localised regions the simulation will have higher residual levels than other types of calculations As such a more relaxed criteria can be used in order to prevent excessive computation after global levels such as mass flow through gasket holes have converged In this case a criterion of le 4 will be specified Timebase Click on the Timebase entry in the SolverSetupTree This is were the time base iteration settings are input These should have been read in from the VECTIS 3 input file The should be Steady State with 500 Iterations Output 2 8 xi Frequendes For Printing Data into Project Fles Output Fie Outer Iterations 1 Output Fie Time steps fio Post processing Fle Frequency iterattons 1000 Co smuation Post processing Fie Frequency Iteratons pooo Report Frequency Save Residuals To Asc Save Residuals To SOF Summary Frequency Report Level Report Boundary Fluxes For Mass And Energy Report Frequency 100 T Save Initial Conditions Restart Fie I Save Initial Conditions Post Fie I Linear Equations Solv
420. nvergence problems in case of high Mach number flows Since the potential flow field is just used as an initial field the user is en couraged to experiment with the smaller Max Number of Iterations default 20 sometimes two iterations can deliver the good initial velocity field However the Max Normalised Residuals should not be increased above 1073 Also the Solver Tolerance of the linear equation solver should be below 107 10 3 6 Laminar and turbulent regime Viscous flows are either laminar or transitional or turbulent Basic physical aspects of turbulent flows are introduced in the section devoted to turbulence modelling Laminar regime characterises the low Reynolds number flows which are well ordered and free of macroscopic random distur bances Transitional flow regime encompasses the transition from a laminar to turbulent flow The VECTIS MAX can not predict turbulence transition and the available turbulence models are ap plicable only to fully turbulent flows If it is a priory known that a flow regime is laminar this flow model is selected in the Turbulence Model panel in a similar way as the inviscid model Figure 10 6 Ricardo Software December 2009 173 10 MODELLING SINGLE PHASE FLOWS 10 3 MODELLING FLUID FLOW Equations Solver amp e Turbulent Viscosity Volume Fraction Potential y Passive Scalar Potential Equations Relaxation Factor 1 Convective Scheme UDS Blending
421. o apply E o e command gt f Figure 19 86 Using the mouse to create a slice To do this first create a plane through the gasket holes In this case the plane will be defined using mouse input Firstly orientate the model to view from one end so that a line can be drawn to define a plane through all the gasket holes see figure 19 86 Note It may be beneficial to turn off perspective viewing when defining the plane Then click on the slice tool and select With mouse Next left click once in the canvas to start the mouse definition Next define the plane by left clicking two points in the canvas A slice plot will appear in the plottree Note The mouse mappings set in the preferences for pick will determine the mouse and keyboard input needed for defining the slice If the slice plot is not already plotted in the 3d canvas drag the slice plot from the plottree to the canvas Remove the main model plot from the canvas by right clicking the corresponding canvas name under the main model plot in the plottree and selecting remove Ricardo Software December 2009 435 19 TUTORIALS 19 3 COOLANT FLOW roenan PEEEA ala WE le Edt View Optons Window Hep lex DAD b amp leaf y e ve rv jeale gt Iene man lx Br i negative _mass_fux _acoss pia y vortoty_magnitude vorboty_magnitude In Plane bsolute_pressure Elements mass_flux_acro jk Max 5349 ss_plane 250 625 150
422. o get the list of boundary regions then use the following integer iwp pointer 1_bc call get_reg 1l_reg_cond 1_bc There are nbcop entries in the array 1_bc and for the given region index ir this ar ray will return one of the following integer variables O ibmat1 0 index of adjacent parent in case of interfaces material domain 1_ bc ibmatl ir ibct 1 boundary condition type 1_bc ibct ir ibcop 2 boundary condition option related to the boundary type 1_bc ibcop ir ibreg 3 original boundary region index assigned in the pre processor 1_bc ibreg 1r ibinout 4 inflow or outflow condition 1_bc ibinout ir 1 defines the outflow region 1_bc ibinout ir 0 inflow and 1_bc ibinout ir lt 0 the wall or symmetry region ibeset 5 option to set up boundary conditions 1_bc ibceset ir Default uniform i e region wise boundary conditions will be provided if 1_bc ibcset ir 0 Other op tions are user defined boundary conditions 1_bc ibcset ir 1 use of tabulated data 1_ bc ibcset ir 2 and provision of time dependent boundary conditions 1_bc ibcset ir 3 Ricardo Software December 2009 315 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES oO ibcomp 6 region compressibility flag The values assigned to the 1_bc ibcomp ir
423. o specify whether the fluid contains a single phase or a homogeneous mixture of a number of different phases Initial Values Initial velocity pressure temperature passive scalar turbulence and species mass fraction for the solution domain at the start time as well as providing some other initial condi tions options Designed by User This option allows initial conditions to be specified using a user rou tine Potential Flow This option will calculate the initial flow field based on the boundary condition data specified The calculated velocity values can then be scaled using the Po tential Flow Scale Factor This is useful because the calculated potential flow may be much higher than the actually flow as turbulence and viscous effects are not taken into account Uniform Monitoring Monitoring points are specified within the model domain where values of the solution are output at the end of each time step Pressure Monitoring A cell ID number or X Y Z co ordinate can be specified for the reference pressure values Velocity Monitoring A cell ID number or X Y Z co ordinate can be specified for moni toring velocity values These values are written to the screen and to the output file Ricardo Software December 2009 378 19 TUTORIALS 19 1 BASIC TUTORIAL Figure 19 15 The Fluid Domain Panel Reference Data The reference values should have representative values for the sol
424. oach is to use wall functions for the k as designed in Kalitzin et al 2005 and Absi and Bennacer 2006 According to the latter reference k is a function of the wall distance Yt pU Y p and we can write ki pels yer 14 82 N where B 0 14 and A 8 are the constants With known k at the poor quality cell J the dissipa tion rate can be specified without an interpolation see Equation 9 47 Note that in case of laminar flow or viscous sub layer the above interpolative functions for the state of variables at intermediate nodes i are reduced to the linear ones For the poor quality cells not adjacent to the walls there is no interpolation plane and Equation 14 74 is used in conjunction with the qualifying neighbour and possible boundary values The pressure and mass conservation equation do not require any special treatment at poor quality cells The interpolative scheme is applied after solving the linearised discretized equations and after that interpolated variables are used in the same way as variables at normal cells Switching On Off poor quality cell treatment option From the Solver Setup Tree under Fluid Domain Solid Domain left click on Discretise This opens up the Discretise panel as seen in Figure 14 7 Under Bad Cell Treatment a number of options are available from the ListBox see Figure 14 7 The Off option simply switches off the cell treatment option To apply cell treatment option to all varaibles except pressu
425. odelling Consequently it can be applied to more complex geometries and higher Reynolds numbers While DNS is valuable research tools the solution of Reynolds Averaged Navier Stokes RANS equations has been the only viable approach in industrial CFD for over three decades LES might eventually emerge as the future industrial standard but it can be argued cf Hanjalic 2005 that for long RANS will play an important role especially in industrial applications With appearance of some hybrid RANS LES methods which combine advantages of RANS and LES for wall bounded flows it is expected to see more extensive use of advanced RANS methods Hanjalic 2005 VECTIS MAX employs the RANS approach and currently provides closure of the RANS equa tions via linear two equation k e models The reminder of this section contains O An overview of turbulence models for RANS O Description of linear two equation k e models implemented in VECTIS MAX Near wall turbulence treatment O How to select a model and wall treatment 9 2 Overview of Turbulence Models for RANS It has been shown that RANS equations contain unknown quantities statistically one point second moments p u j representing turbulent mass diffusion momentum and heat flux for equal to c ui and T respectively Consequently the turbulence closure problem arises that is these quantities have to be determined by a turbulence model The current turbulence m
426. odels can be basically classified into two groups namely Eddy Viscosity Diffusivity Models EVM These models are known also as the first order models They are based on the assumption that turbulent fluxes for species momentum and heat depend directly on the mean flow variables c U and T Differential Second Moment Closure DSM SCM Models These models are described by Reynolds Stress Flux transport equation i e a separate modelled differential equation is solved for each turbulent flux p u Ricardo Software December 2009 149 9 MODELLING TURBULENCE 9 2 OVERVIEW OF TURBULENCE MODELS FOR RANS As VECTIS MAX provides eddy viscosity k models the eddy viscosity approach will be outlined next 9 2 1 Eddy viscosity formulation Eddy viscosity models have a root in an in principle false analogy between molecular and turbu lent transport formulated by Bousinessq in 1877 Accordingly Reynolds stresses can be obtained from the constitutive equation similar to Stokes law Equation 8 2 1 2 Tij pul ul 24 s Sud 3P koi 9 4 where the mean flow strain tensor S is OU OU Sij 9 5 LE o F Ox Xi i and 4y is the turbulent or eddy viscosity The term 2pk0 3 represents the mean turbulence pres sure which is together with velocity divergence term Syx usually added to the unknown static pressure of the resolved motion ps by replacing it by the sum OU 2 3 1 ur Six P
427. of simulation process and then it is called before the beginning of each solution iteration Ricardo Software December 2009 296 18 USER PROGRAMMING Begin i Initialise upr_init upr_bnd_cond i Do timestep Do iteration End iteration End timestep Update properties upr_properties Y upr_generic Y Y upr_generic Y Y Update properties upr_properties Y Y upr_generic Y Y upr_generic Y End Ricardo Software December 2009 Segregated solver upr_generic upr_sources upr_sources upr_sources upr_sources upr_sources upr_sources upr_sources upr_sources Solve u v w velocities Solve continuity Update pressure u v W Solve volume fractions Solve turbulent kinetic energy Solve dissipation of turbulent kinetic energy Solve species Solve energy Figure 18 1 Solver flow chart and the corresponding UPR stages upr_ 18 2 FUNCTIONALITY AND CALLING SEQUENCE 297 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES 5 upr_sources This UPR can be used to specify additional sources for each of the transport equations solved by VECTIS MAX except passive scalars which contain no source This UPR is called for every solution outer itera
428. oined regions will appear as with boundary type Unjoined Boundary The full listing of all the jtype options are shown in Table 5 1 The options above 0 are in general less reliable robust and should in general be treated by the user as experimental 5 3 1 Sub domain joining Certain physical models such as porous media fan and radiator require creation of sub domains Since any fluid or solid material domain can contain any number of non overlapping sub domains the surface mesh along sub domain interfaces should be conformal In order to perform a conformal joining firstly vmesh is used to make interfaces conformal in a similar way described above Consider a catalyst test case consisting of a single domain air_no_cat that should be joined with 2 sub domains cat amp cat2 Furthermore interfaces 7 amp 8 of the domain should be made con formal to interfaces 2 amp 3 of the sub domains To make these interfaces conformal the following vmesh commands are issued vmesh int 7 8 air_no_cat mesh vmesh int 2 catl mesh vmesh int 3 cat2 mesh vmesh conformrew cat1l GRD 2 air_no_cat GRD 7 vmesh conformrew cat2 GRD 3 air_no_cat GRD 8 Then jtype option is used in vpre to perform the joining as follows vpre no material offset jtype 0 air_no_cat GRD subdom cat1 GRD subdom cat2 GRD o air_cat where subdom is used to indicate that the next mesh file is a sub domain of the preceding mesh parent no material offset ens
429. ojectname rep_runnumber It contains the data for all the additional output It can be viewed and plotted in R Desk using the SDF File Manager and X Y Plot Manager Summary output Summary data can also be written to the screen and output file Co Simulation Post processing Frequency Controls the frequency at which post processing data is written when WAVE requests postprocessing output during a coupled simulation Save Initial Conditions The Save Initial Conditions Restart File and Save Initial Conditions Post File toggles allow the user to select whether the initial data used in the calculations is saved to a file The initial conditions are saved after any flow potential calculations are performed but before the first timestep iteration Linear Equations Solver Residuals Information This toggles whether residual information is writ ten for each of the equations at every timestep This will write to the screen and output files in the format shown below This pipe flow example uses the following main options Steady state 1000 iterations Convergence criterion le 5 During the calculation restart files are written at a user defined frequency throughout the cal culation specified in the Restart File Frequency field Two restart files are generated and are written alternately at the restart frequency time step interval Restarting the calculation To restart the calculation from a saved restart file the Re
430. ollowing three values O ak 1 the fluid k occupies completely the cell o ak 0 no fluid k in the cell O 0 lt a lt 1 gt the interface between fluid k and other fluids is present Predictions of the volume fraction should be accurate enough to get the sharp resolution of the interfaces This requires the use of High Resolution Interface Capturing HRIC schemes and finer numerical grids The governing equations can be derived by summing the phase Equations 12 2 12 4 12 8 12 10 and 12 11 and having in mind that the concept of slip velocity is usually applied to the momen tum and volume fraction equations 12 3 1 Mixture equations o Mass conservation for the mixture op d m m z EAG u 0 12 13 where superscript m designates the mixture The mixture density and velocities are Nph N akakyyk _ 0 p U k Ak m k 1 i p Op U i 2 p 12 14 O Momentum conservation for the mixture d m m m m MY __ dp m er 3 P U oe U Un fi e a Yo Kikak 12 15 ax WT Ox cd A Xj p Ricardo Software December 2009 200 12 MODELLING MULTIPHASE FLOWS 3 MULTI PHASE MIXTURE AND VOF MODELS where the viscous and turbulent stress tensors Tj and T are given respectively as ij t m m 1 m tm m m 1 m 2 m m 2u Sij 3504 gt Tij 24 Sij z Snn ij 3P k ij 12 16 and the mixture mean strain tensor is a det a 12 17 yo 2 Ox OX i l Both laminar and
431. omain 5 amp Body Force El Options Boussinesq Approach Fixed Disabled Gravity e Enabled Standard Approach Force COMA a COM Figure 10 8 R Desk setup Setting body force buoyance fluid model in Fluid Domain panel In addition it is important to specify correctly the reference density p e f and in case of the Boussi nesq model the reference temperature T ef and the coefficient of volumetric expansion p Setting of these reference values is dealt with in the section which explains the fluid domain panel Fig ure 17 6 Ricardo Software December 2009 176 10 MODELLING SINGLE PHASE FLOWS 10 4 MODELLING HEAT TRANSFER 10 4 Modelling Heat Transfer Three modes of heat transfer conduction convection and thermal radiation can occur within fluid and or solid continuum This section considers modelling of conductive and convective heat trans fer in O fluid domains O solid domains and O conjugate heat transfer In each case a variation of the general energy equation will be solved 10 4 1 Heat transfer in fluids All convection modes forced natural and mixed can be simulated by solving momentum and energy equations in fluids Natural convection is associated with the buoyancy driven flows The energy equation introduced earlier represents the conservation of total enthalpy WECTIS MAX solves the total enthalpy equation for the compressible flows gases while the total energy equation i
432. on The automatic mesh generation capability of VECTIS MAX makes the CFD analyses of engine coolant jackets particularly quick since the complex computational coolant jacket mesh is cre ated very quickly from the closed CAD STL geometry This allows investigations such as gasket hole position and size optimisation to be integrated into Engineering Development projects It is also a common technique to perform coolant analysis and then extract the surface Heat Transfer Coefficient results and apply these to a Finite Element analysis as thermal boundary conditions CFD analyses of coolant jackets require boundary conditions for the inlets and outlets and the wall surfaces The inlet and outlet boundary conditions are not usually an issue since the coolant mass flow rate and or inlet and outlet pressure is usually known However the coolant jacket surface temperature distributions are not usually known There are several options to overcome this issue 1 Run the coolant jacket analysis as a fluid only analysis and as an iso thermal analysis such that both the coolant fluid temperature and the wall temperatures are the same The Heat Transfer Coefficients HTCs are then mapped on to a Finite Element FE analysis mesh and used as thermal boundary conditions for the FE analysis The predicted wall surface temperatures from the FE analysis can then be extracted from the FE analysis results and mapped on to the VECTIS computational mesh The VECTIS CFD analy
433. on File groupBox is used to specify the name of the input file radfile to be subsequently used by the radiation modules The Start box indicates the time step iteration number when the radiation solving should begin The Frequency box denotes how often time step iteration the radiation solver should be called Global Surface to surface Parameters PAT VFM amp CON filenames These are auxillary filenames which will normally adopt the base name of the mesh file in the current input file The user can however freely override this The PAT file is generated when radprep is run and contains superpatch information The CON is the corresponding connectivity file The VFM file is created by radvfm and contains the superpatch view factor matrix Reflectivity This reflectivity option allows for the inclusion of secondary radiation i e radiation that is emitted from a surface that is also receiving radiation heat flux By default this is set to YES so that radiation can be reflected from a surface Transient Problem The transient option is used to define if the radiation analysis will be for a transient or steady state analysis Default Ambient Temperature This value is going to be used whenever a super patch is not fully surrounded by patches i e its view factor summation is less than 1 In such case the radiation from the background needs to be included The ambient temperature is needed because in such a case where the
434. on are displayed The user can either interrogate a boundary region or a region of triangles which have been marked with the marking commands Triangle8587 BoundaryHumber4 HormalVYector x 0 004768 y 0 581374 2 0 813623 Interrogate Boundary Marked Region display 3 13 Hints for Manual Stitching There is usually more than one possible way to use Delete Triangle Create Triangle and Join Triangles operations in a given situation The user will find that the correct form of the surface is very often clear The way to approach an area that needs stitching is usually to delete triangles that are wrong and then to fill in the gap created with new triangles Another feature of the surface which should be checked once stitching is complete is that the model does not contain overlapping or tightly folded surfaces To allow the identification of such regions any edges in the model which lie on two triangles whose normals have an angle of more than 170 degrees between them are highlighted in green at all times It is good practice to look for these regions by turning off the display of surfaces and outlines There may be very sharp edges in the geometry which are quite correct but there may also be areas where triangles lie on top of each other or where the surface is folded back on itself These may become a source of ambiguity in the surface for the mesher and they should be eliminated with the use of the manual stitching operations R
435. on each partition and the concatenating onto partition 1 Partition 1 has the sole responsibility of writing the file y vs t subroutine upr_generic id icall_pos This general model independent routine is called from several positions in Ithe solution cycle use upr implicit none Global public declarations types parameters scalars arrays integer iwp intent in id amp global solution domain id 1 icall_pos calling position flag in the solution cycle integer iwp pointer iscd integer pointers iecd linteger pointers real wp pointer 22 xec amp cell centre coords tfield amp cell temperature ufield cell velocity vector integer iwp ic amp lcell counter idt Isolution variable index iupl 21 amp file unit idom amp Idomain counter n_dom amp Inumber of fluid solid domains i_parts amp this partitions number Nn_part inumber of paritions mn counter for compressed arrays integer iwp allocatable save dsize number of stored values on each partition integer iwp save gsize total number of stored values real wph allocatable save Xval tval compressed arrays for y coordinate and temperature character len 80 filename amp loutput filename proj_name amp project name proj_num project number Iwrite temperature profile at x 0 5 For parallel cases the values are calculated on each partitions and then a Iconcatonated list
436. on has global_max and global_min and has support for character integer real and double precision format The first section deal with scalar variables whereas the second section deal with vector variables a Get global maximum minimum for scalars Ricardo Software December 2009 343 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES subroutine global_sum svaluel svalue2 Arguments Description svaluel gt integer global sum svalue2 gt integer optional global sum svaluel real global sum svalue2 gt real optional global sum svaluel double precision global sum svalue2 gt double precision optional global sum subroutine global_sum values nvaluesl nvalues2 Arguments Description values gt global sum values can either be 1D or 2D values integer sum values real sum values double precision sum values integer sum values real sum values double precision sum nvalues1 gt integer number of elements in dimension 1 values 1 nvalues1 nvalues2 gt integer optional number of elements in dimension 2 val ues 1 nvalues1 1 nvalues2 Used if values is a 2D array Table 18 32 Subroutine global_sum to get global sum over partitions real wph mach_max amp estimated maximum Mach number calculate mach_max for each partition Get global maximum value over all partitions call g
437. on value and i is the value from previous iteration In case of the velocity vector scaled approach the velocity direction must be specified The new boundary velocity is then evaluated as ui gt 13 6 13 3 1 Setting up a mass flow rate boundary condition A mass flow boundary condition can be specified by left clicking on the boundary region from the Solver Setup Tree and selecting the Boundary Condition Type from the ListBox to be Mass Flow Rate Boundary The Mass Flow Rate Boundary setup panel is then displayed as in Figure 13 4 The mass flow rate can be given the appropriate value in Mass flow Rate InputBox and can have either Out or In attribute If the In attribute is selected it means the flow enters a fluid domain inflow boundary otherwise it leaves a fluid domain outflow boundary Bnd_Reg_2 Mass Flow Rate Boundary a E Region Name Bnd_Reg_2 Region ID 2 Material ID 1 Boundary Report Coupled Link Number lo Normal Velocity Scaled Boundary Condition Type Mass Flow Rate E Velocity Vector Scaled Boundary Condition Op Normal Velocity Scaled Boundary Setting Uniform Values 2 Out l Massflow Rate 0 Out gt i Velocity Direction Velocity Direction x o y fo z o Ba x o pr o z o Figure 13 4 Setting up a Mass Flow Rate Boundary type in R Desk There are two Boundary Condition Options ava
438. ondensation reported in literature are loaded automatically for each of the cavitation model The user may either increase or decrease the evaporation and condensation by entering the appropriate values for Evaporation and Condensation Setting up Saturation Pressure is mandatory for all three models Surface Tension is only required in the Singhal et al model Ricardo Software December 2009 212 13 BOUNDARY amp INTERFACE CONDITION TYPES Two types of boundaries characterise fluids walls and flow boundaries Walls are natural boundaries usually impermeable to fluid flow and can be stationary or moving Flow boundaries are introduced to model fluid domains i e to reduce the size of fluid domains by cutting through the flow A number of flow boundaries and in turn boundary condition types can be defined de pending on the type of flow In VECTIS MAX a boundary region can be associated with the one of the following boundary condition types O Velocity Inlet O Mass Flow O Pressure Inlet or Outlet Stagnation Inlet O Flow Outlet O Symmetry Plane O Wall Not all combinations of the above flow boundary condition types are physically compatible With regards to the flow outlets within the given fluid domain 1 the flow outlet boundaries with the specified flow split and fixed mass flow rate can not coexist 2 the flow outlet with specified mass flow rate must be used with at least one pressure i
439. onic at the inlet all flow variables p T pp ip the density is calculated from the equation of state must be specified Subsonic Outlet When the outlet is subsonic the pressure is fixed and all other variables are extrapolated The boundary velocity is obtained from Equation 14 53 applied to a boundary face This velocity and the corresponding velocity correction are given as Vp A p p A p Pp Vpp d 14 70 Pap Ap dy SN r gt Vp Ab 3 ee se 100 E E 14 71 ia b Pp where the boundary pressure correction is zero p 0 The mass flux correction at the pressure boundary then reads see Equations 14 58 14 59 gt gt Q rit pUJ Ap pr Cpr max rit 0 p p 14 72 b and serves to derive the pressure correction coefficient a The above treatment of pressure bound aries is applicable to both compressible and incompressible flows Supersonic Outlet At supersonic outlets all variables including the pressure are extrapolated from the upstream However the computations usually start with the initial fixed pressure until establishing the stable supersonic conditions in the outlet region 14 2 5 Setting up pressure correction algorithm For each Fluid Domain node in the Solver Setup Tree panel left click on Algorithm node and the Algorithm panel is displayed to the right as in Figure 14 9 Now the Solution Algorithm can be selected to either be Simple or Simplec Number of Pressure Correction
440. only used for incompressible flows For compressible flows the conservation of total enthalpy 22 a O S H E h 6 13 A 5 6 13 is the preferred form The total enthalpy equation reads E phav f aA 0 U 0 ak 6 14 d Nsp a aT a AA SJ pav o 4 as el g dA AG U av 1 PA PA 0 Us Py 7 Y hn sat rs PFU A 6 15 6 2 2 4 Space conservation law For control volumes with moving surfaces U 0 the space geometric conservation law SCL should be satisfied d gt a dV qU dA 0 6 16 ah E oe If the above law is not enforced numerically the conservation of mass and other quantities is not guaranteed as artificial mass sources or sinks can be generated 6 3 Closure Problem and Averaging In order to solve the above instantaneous equations they need to be closed by O Basic constitutive relations specifying Species diffusion fluxes i k 1 Nsp Viscous stress tensor 7 Heat flux vector q O Equation of state linking density and internal energy or enthalpy with basic thermodynamic variables pressure p and temperature T O Thermo physical properties associated with the constitutive relations and equation of state for example dynamic viscosity thermal conductivity mass diffusion coefficient density and specific heat Ricardo Software December 2009 117 6 SOLVER FUNDAMENTALS 6 3 CLOSURE PROBLEM AND AVERAGING In addition well posed boundary
441. ons is a matter of user preference and will depend on the particular application If a boundary has a refinement specification in both the triangle file and the mesh file the specification in the mesh file will take priority Boundary refinement specifications are written to the triangle or mesh input files by phasel For reference the format of the information in these files is as follows O Triangle file specification SDF 2D integer array name VEC BOUNDARY_REFINEMENT 3 NumberOfRefinementSpecs where the three elements per specification are 1 Boundary number NB 2 Refinement depth DEEP 3 Blending distance BLEND signed to indicate the blend to type as above O Mesh input file specification Any number of lines of the form BOUNDARY _ REFINEMENT NB DEEP BLEND where the variable meanings are as above Ricardo Software December 2009 59 3 GEOMETRY 3 15 BOUNDARY PROCESSING 3 15 5 Boundary Refinement examples The effect of different parameters available for the surface refinement are shown below x Specification for boundary 2 K gt Delete Refinement depth at boundary C Refinement blending distance ji Refinement Blending Blend to boundary depth Blend to boundary depth 1 Refinement Specification Destination Save specification in triangle file Save specification in mesh file sp to Surface refinement set
442. originated from from the dispersed phase evaporation of liquid droplets or any user defined source 166 10 MODELLING SINGLE PHASE FLOWS 10 1 GOVERNING EQUATIONS Oo Momentum conservation o o E L 2 pu 2 pu 0 00 tont ur 102 Ot Ox Xj J In case of linear k models the turbulent stress tensors 7 is given by Equation 9 4 The sum of laminar and turbulent stresses can be then expressed as 1 2 Tij Tij heff s 55m 3Pkoij Heff U Mb 10 3 where the resolved mean strain tensor Sj is given by Equation 9 5 The turbulent viscosity u is predicted according to Equation 9 14 Molecular viscosity calculation options are discussed in relation to the setting of fluid phase properties see also Table 8 1 Note that the pressure p in the above momentum equation represents the modified static pressure defined by Equa tion 9 6 The body force per mass unit mass can include the effects of gravity buoyancy gi system rotation fi rot porous media fi por and user defined forces fi usr fi git firot fi por fiusr 10 4 Only the gravity body force is currently supported O Species mass fraction conservation z Pci T pci U Uz a Jej 56 3 Sci 10 5 In the context of the k modelling the turbulent diffusion flux of species i de is given by Equation 9 8 Thus the sum of molecular and turbulent mass diffusion fluxes de dc Jaj Jaj Diett Ta Gi eff
443. ort file The ASCII data files are used by the Live Update utility in R Desk The SDF binary format is written to a single report file named projectname rep_runnumber It contains the data for all the additional output It can be viewed and plotted in R Desk using the SDF File Manager and X Y Plot Manager Summary output Summary data can also be written to the screen and output file at the chosen fre quency and detail level Save Initial Conditions The Save Initial Conditions Restart File and Save Initial Conditions Post File toggles allow the user to select whether the initial data used in the calculations is saved to a file The initial conditions are saved after any flow potential calculations are performed but before the first timestep iteration Linear Equations Solver Residuals Information This toggles whether residual information is writ ten for each of the equations at every timestep This will write to the screen and output files in the format shown below Fluid Domain In this case the calculation will consist of one single phase fluid domain Domain Name A name can be assigned to the fluid domain to allow for easier reference later in the calculation Material Name A name can be assigned to each material in the fluid domain again to allow for easier reference later in the calculation Each fluid domain contains only one material Material ID This refers to the material ID number referenced in the GRD file Each
444. ote that transport properties discussed in the previous section are also true thermodynamic properties With the help of exact differentials T ds equations given as d d de Tds pdv Tds p5 dh Tds vdp Tds E 8 12 s is the specific entropy and v 1 p is the specific volume various thermodynamic relations and properties can be defined 8 2 1 Coefficients of expansion and compressibility Considering the specific volume as a function v v T p two thermodynamic properties related to the total differential dv are defined 1 dv 1 Se 2 1 2 e 21 g 8 13 p v sr p 5 j v p r pxop s i which are the coefficient of volumetric expansion p and the coefficient of isothermal compress ibility Considering an isentropic compression the coefficient of isentropic compression 1 2 EA al 8 14 65 8 14 appears in the definition of a speed of sound a B Op _ 1 8 2 2 Specific heat Under certain conditions the specific heat is the amount of energy added by heat transfer to 1 kg of material to increase the temperature by 1 K Depending on the process in which the energy is added there is a distinction between the specific heat at constant pressure dh p and the specific heat at constant volume de v Ep 17 E en Their ratio K 8 18 Ricardo Software December 2009 133 8 MODELLING CONTINUA 8 3 THERMALLY PERFECT FLUIDS 1s called the specific heat ratio Employing the equat
445. ovements over the previous mesh generator taking advantage of a clean sheet design and the unstructured grid capability of the new solver For example separated volumes that are found in one box are now stored as separate cells Boundary faces are now tied exactly to each other Automatic scaling of the input geometry now prevents problems with very large and very small geometries Faster meshing times are now achieved because of advanced techniques for storing and sorting data with the additional benefit of reduced memory consumption O Conformal meshing for multi domain modelling The mesher is able to produce separate meshes for each domain that are conformal at the domain interfaces These meshes are then coalesced into a single multi domain grid file using the vpre utility 2 INTRODUCTION 2 1 MAIN FEATURES AND CAPABILITIES Cell splitting This feature deals with highly non convex cells It is automatically activated when running vmesh Each non convex cell is split into two cells with better less concave shape O Geometry surface import Import of ASCII and binary VECTIS MAX tri triangle files VDA files and Stereo lithography stl files The import of geometry is done by using phase module 2 1 2 Solver vsolve The vsolve program is a state of the art solver using advanced CFD algorithms Design Concepts O Modular code There are two major parts CFD kernel and Application modules O Data s
446. oyancy Ricardo Software December 2009 281 17 USING SOLVER 17 5 MONITORING POINTS Reference Pressure Location Use Cellid Cell ID For Reference Pressure 0 Use Coordinate E gi ja Figure 17 5 GroupBox for Pressure Monitoring Reference Pressure 100000 Reference Temperature 293 15 Reference Gas Constant 287 Reference Area 1 Reference Length 1 Reference Velocity 1 Reference Density 1 189 Reference Viscosity 0 001824 Reference Spec Heat 1004 Reference Altitude xlo Y fo iz o Body Force Options Boussinesq Approach Fixed Disabled Gravity Enabled Standard Approach Force x o HE HE Figure 17 6 Reference values LineEdit s driven flows Figure 10 8 17 5 Monitoring Points This panel is used to monitor various flow quantities velocity pressure etc at specific locations Either the cell Id or xyz location can be used Monitoring values are written to the mon files The first monitoring point is also written to the screen and main out file ORicardo Software December 2009 282 17 USING SOLVER 17 6 FILE OUTPUT Monitoring Points a amp Use Cell ld Cell ld Xx Ls E Y 1 0 0 0 Delete 0 1 1 1 Delete Add Figure 17 7 Monitoring Points Panel 17 6 File Output Here the frequency of data reporting output is chosen The monitoring and convergence data written
447. p aa aa maojal_ _ AN ol E Y Yl Optons Stitch mesn J Te 7 Triangles pan E A eat Figure 19 52 Using the cap hole tool to fill the open edge Ricardo Software December 2009 409 19 TUTORIALS 19 2 STEADY STATE PORT FLOW ioixi Fie Edt View Toobars Operatons Help daa a m aE INAS Y yl Optons Stitch mesh Tnangles lolo ela gt 10 1 gt gt gt 2 Gr cda gS thE Jere eg S jm E pl a aa 2 z Figure 19 53 Chop out the required surface lol cala maca 10 Noel Y Yl Options Stich mesn 4 Triangles Figure 19 54 Final Arbitrary Surface Ricardo Software December 2009 410 19 TUTORIALS 19 2 STEADY STATE PORT FLOW The arbitrary surface does not have to be planar it can be any form Once the surface has been defined then the report region in the solver setup tree can be setup Select Report Regions in the Solver Setup Tree Then add a report region using the add child report regions button Report Regions xi Add Child Report Regions Figure 19 55 Add report region The geometry filename needs to be specified along with a tolerance Arbitary Surface xj Domain ID 1 Filename arbitary_surf1 tri Tolerance 0 01 Figure 19 56 Specify triangle filen
448. p modelling approach options left bottom and k e model family variants right bottom Le the Reynolds Averaged Navier Stokes RANS equations will be solved All approaches can be listed by left clicking on the Turbulent Modelling Approach box and they are shown in Figure 9 5 left bottom Apart from RANS equations the solution of laminar or inviscid flows both do not require turbulence modelling can be selected The Turbulence Family box contains only k e model family Within each family all available members are listed after left click on the Turbulence Model ComboBox The current models belonging to the k family are depicted in Figure 9 5 right bottom These are the Standard k model RNG based and Standard model with the realisable time scale bound The default k model is the standard one It remains to select the Near Wall Modelling method which can be either Low Re number or Wall Functions approach Three variants of wall functions are available namely Scalable Wall Functions Standard Wall Functions and Unified Wall Functions The Standard wall functions are the default choice The low Re modelling option can be selected only for the Standard and Standard realisable i e TSB models and it is always employed in conjunction with the Unified Wall Conditions Ricardo Software December 2009 165 10 MODELLING SINGLE PHASE FLOWS This chapter describes currently available physical models for
449. passive scalar index ieq integer equation id iget integer describes object type takes in iget_dom iget_mat iget_phase iget_ specs iget_ps Table 18 25 Function eq_idt to return identifier idt for derived data types 18 3 2 18 get_field This UAR is used to get field values either for a scalar or a vector for domain type idt see Table 18 27 If idt is not known then use eq_idt UAR to get idt A complete list of names that can be passed to get__field subroutine through var_name are given in Table 18 26 together Ricardo Software December 2009 332 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES with availability of variables at grid objects There are 2 different options for get_field as specified in 2 sections in Table 18 27 a The first method is used to obtain field values at one grid object Example Obtain velocity values at cell centre firstly obtain domain type idt for momentum and then obtain velocity values real wph pointer vel_cell cell velocity values integer 23 idt dom dom is domain id dom 1 domain 1 used as an example Get velocity domain index idt eq_idt ifmom dom iget_dom get cell velocity values call get_field idt iget_cell velocity vel_cell Note the use of a 2D array for velocity vel_cell For a scalar a 1D array must be used Argument iget_uif corresponds to upper interface values These are fi values at cell faces along material
450. pe 1 The choice of which mesh to extrude is by default the mesh containing the larger on average cells along the AGI interface This can be overridden via the AGI_LEXTRUDE environment variable Any new boundary faces should inherit the boundary numbers from those faces removed as a result of extrusion Same as jtype 2 except mesh is extruded both sides of the AGI interface 4 Similar to jtype 0 except any unconnected faces are then dealt with using the jtype 2 extrusion re mesh approach This method is appropriate when the interface regions are known to be nearly conformal Oo Table 5 1 Table of all the jtype options This would generate 8 grid files in the directories P001 P002 etc along with the correspond ing 8 restart files POO1test rst0_010 etc Vpre expects the parallel corresponding to np 4 grid and restart files from the previous run to be present in the parallel directories P0O1 etc If no filename argument is passed to rest then the latest restart files corresponding to the grid file name e g POO1flow rst1_010 etc in this case are used Ricardo Software December 2009 11 SOLVER FUNDAMENTALS 6 1 Introduction The solver is a software module which performs the numerical simulation of a CFD Continuum Mechanics problem In other words it produces the numerical results which simulate transport of mass momentum energy and other associated physical phenomena in the considered continuum Continuum is
451. pecified tolerance the calcu lation will terminate 2 Unsteady Time Scheme This can be set to either 1st Order Euler or 2nd Order Implicit Time Base Here the time base is selected 2 stroke Seconds 4 stroke Non Dimensional Time Dependant Boundary Conditions The boundary conditions can be steady or unsteady transient which is determined by selecting the time dependent boundary conditions toggle Timesteps The number of timesteps is chosen for the calculation the end time being the the chosen time step multiplied by the number of timesteps specified Additionally a Maximum Number of Outer Iterations is chosen This is the maximum number of iterations performed per time step Ricardo Software December 2009 3735 19 TUTORIALS 19 1 BASIC TUTORIAL Typically the number of iterations required per step will reduce after the initial period of the calculations However this will be dependent on the accuracy of the applied initial conditions Output 18 xj Frequendes For Printing Data into Project Files Output File Outer Iterations 1 Post processing File Frequency Iterations 200 Co simulation Post processing File Frequency Iterations 0 Report Frequency asii OO wha V Save Residuals To Asc Save Residuals To SOF Summary Frequency Report Level oF amp Report Frequency 100 SCSCS lt lt 7 7 7 lt J save Initial Conditions Restart Fle T Save In
452. pected however the name of the input meshfile was not specified ERROR 1513 Identical triangles have been detected in the trifile filename triangles T and T2 are identical ERROR 1514 A flaw in triangle topology was detected in trifile filename ERROR 1515 Triangle T has an unstitched edge in trifile filename Repair the triangulated surface in the preprocessor first ERROR 1516 Neighbouring triangles have opposite orientation TZ and 72 The orientation of tri angles should be harmonized in the preprocessor first ERROR 1519 Processing of command line viewijk needs to be followed by six integer values A fatal problem occured when command line option viewijk was processed The help section 4 4 should be consulted ERROR 1520 Invalid mesh size and or placement No boxes were generated The placement of meshlines should be checked in the preprocessor ERROR 1521 Cannot write changed grid to file X It is not possible to write down the gridfile changed by routine making boundaries of two complementar grids conformal ERROR 1522 Velocity location X neighbours to a box with incorrect cell index X This is an internal logic error the system of generated boxes seems to be corrupted Ricardo Software December 2009 07 4 MESHING 4 9 WARNINGS AND ERRORS ERROR 1523 The modified surface cannot be written to file X It is not possible to write down the file X ERROR 1524 The common face X has less than three vertices X ver
453. ple where boundary region with id ir 1 The following example illustrates the use of local_force subroutine Ricardo Software December 2009 339 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES integer ir j ic idt dom integer iwp pointer Jsbc jsbc integer scal_field 1 scalar rank integer vect_field 1 1 vector rank real wph pointer pb boundary pressure real wph pointer vis_b Iboundary effective viscosity real wph pointer vol_frac_b boundary volume fraction real wph pointer vel_cell amp cell velocity vel_bnd amp boundary velocity yvel_ui amp upper interface velocity vel_li lower interface velocity real wph pointer bfs_vec boundary face surface vector real wph pointer dnb real wph 22 Epree l 3 amp pressure force fvis 1 3 amp viscous force are boundary face area scal_field 1 Iscalar rank vect_field reshape 1 1 1 1 vector rank dom 1 Idomain 1 used as an example Get pressure domain index idt eq_idt ifmas idom iget_dom Get boundary pressure field Call get_field idt iget iget_bnd var_name pressure fi_out pb Get velocity domain index idt eq_idt ifmom dom iget_dom Get velocity at cell boundary lower and upper interface call get_field idt vec_field var_name velocity fi_c vel_cell amp i_b vel_bnd fi_ui vel_ui fi_li vel_1li Get boundary effective viscosity
454. plit into 2 sections Each section has different arguments The first section deal with scalar variables whereas the second section deal with vector variables and vector variables can have 1 or 2 dimensions a Get global sum for scalars Suppose that we want to get the region area calculated over all partitions Then using section 1 of the table ylobal_sum svaluel svalue2 we get real wph sum_area total region area Icalculate sum_area for each partition get global sum over partitions call global_sum sum_area Variable sum_area now contains the total region area Note that svalue2 is optional and is in tended for use with 2 scalars each of them returning its global sum over partitions b Get global sum for vectors To calculate mean region pressure use section 2 of the table global_sum values nvalues1 nvalues2 as real wp p_reg mean region pressure integer nreg get number of regions nreg get_number n_bnd_regs calculate p_reg for each partition Get sums over all partitions call global_sum p_reg nreg Here p_reg 1 nreg passed to global_sum is a 1D array A 2D array can also be used when required The examples presented above use a real type argument for sum_area and p_reg but integer and double precision are also available 18 3 2 24 global_max global_min This UAR can be used to obtain maximum minimum over partitions Table 18 33 is split into 2 sections Each secti
455. post uprdata upr_data idt d_type d_name end if if icall_pos icp_end_iter then Set upr_data to something upr_data end if end subroutine upr_generic User data contained in array upr_data is saved to the post processing file Data type d_ type is used to establish whether data written is cell based d_type i_cel1 boundary based etc see Table 18 4 for further details The data in the post file is identified under the data base name d_name Domain type id is used for output levels If data base name is USER_PHASE and data is written for 2 phases then USER_ PHASE is identified under USER_PHASE_ 1 corresponding to phase 1 and USER_PHASE_2 cor responding to phase 2 in the post file Note that upr_data 1s a 1D array It can also be used as a 2D array depending on the variable subroutine user_post uprdata idt dtype dname Arguments Description uprdata gt real pointer store pointer of the user data uprdata real rank one array uprdata real rank two array idt gt integer store domain type id dtype gt integer store data type e g cell based dname gt character store data base name Table 18 31 Subroutine user_post to store user data Ricardo Software December 2009 342 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES 18 3 2 23 global_sum This UAR can be used to get global sum over partitions used in parallel runs Table 18 32 is s
456. pplied then the transient term is discretised as d _ 3 pVo p 4 pVO p pV g evo 2At 14 31 14 1 5 2 Convection term The convective flux C Equation 14 4 can be discretised as Cj mjs yo mjlo Ge or 14 32 where Y is the blending factor between Upwind Differencing Scheme UDS and higher order bounded schemes and represents the flux limiter for the selected bounded scheme The first term in the above equation is treated implicitly while the second underlined term is calculated by using values from the previous iteration step and treated as an additional source term deferred correction approach Khosla and Rubin 1974 14 1 5 3 Diffusion term With the cell face gradient defined by Equation 14 29 the diffusion flux D in Equation 14 4 becomes A S a De T y bp bp 41 Vo 4 24 14 33 J wag j Oj Yj J And J Dj n normal diffusion D jc cross diffusion The non orthogonal cross diffusion term which vanishes on orthogonal grids is treated in a deferred correction manner Limiting cross diffusion As Equation 14 33 shows the diffusion flux consists of the normal D jn and cross diffusion D c terms The latter is important as it maintains the second order ac curacy of the diffusion discretisation The amount of the cross diffusion depends on the face skewness angle 0 which is defined as an angle between the face area vector A j and the distance vector d
457. py of formation h J kg is the energy released in an exothermic reaction h lt 0 or absorbed in an endothermic reaction h gt 0 when the compound is formed from its elements all being at T ef and Pref At a state p T different than the standard state the specific energy and enthalpy for thermally perfect fluids are defined as T e h e p T elPref Tref h cydT h 6 T T T Tres Tref 8 40 Tref T he h h p T h Pref Tref f cpdT h 2 T T p Tref Tref 8 41 ref The above specific internal energy and enthalpy thus include the enthalpy of formation and they are often called thermo chemical internal energy and enthalpy The enthalpy of formation is required for each species i in order to define the mixture internal energy and enthalpy Assuming that the standard state is the same for all species involved Tye ri Tref we have ec X cieci cih ci CiT CviTref i Y cihi OT Tyef 8 42 i i i i he Y cihei Y cih Y ci Cp iT Tp Tres Y cih T TpT e 8 43 i i i i In the above equations cy and Cp denote the mixture specific heats see Equation 8 39 For non reacting mixtures it is sufficient to work with thermal internal energy and enthalpy in which case Equations 8 42 and 8 43 do not include enthalpy of formation terms and they reduce to Equations 8 24 and 8 25 Reacting mixture flows are not supported yet 8 5 Properties of Multiphase Mixture
458. quation of State Models ocres ko Ba e Pe d 2S be Ped e eae des 168 10 2 1 Reference and solver working pressure 2 ac aaa ee ee eee 169 10 3 Modelling Fluid FlOW io ae ae hs a a eh a ee ae OE A a ee 170 10 3 1 Two and Three Dimensional Flows 000000 eee 170 10 3 2 Single and multi component phase o e e e 171 10 3 3 Incompressible and compressible flow o o 171 10 3 4 Inviscid and viscous regime o s ss su at A a a e E 172 10 3 5 Potential How model cuicos har a ad a a 172 ORicardo Software December 2009 y 10 3 6 Laminar and turbulent regime o s chocado Re a a a 173 10 3 7 Steady and unsteady OWS s oce a era ie aaah eared EM ee ge ad de 174 10 3 8 Gravity buoyancy driven flows o e ee eee eee 175 10 4 Modelling Heat Transfert seso a ews OO A a A ee ae ed 177 10 4 1 Heat transier in DIOS c ceca 24 haha da ede dake de ae acbe ess 177 104 2 Heat transfer in solids slides a A ee a E 179 104 3 Conjugate Heat transfers 2 4 04 fe rra ea a a a a ee 180 10 5 Modelling Mass Transie s sd 40048 2 4428 24 Be ees hoe eke eek ve ded we 180 10 3 1 Passivescalat cocos Be be bE eS EW a we Ba we Re ee ee Ge 181 11 MODELLING POROUS MEDIA 182 11 1 Theoretical Background coe toee ae a a SS aw ROE ee RE a ee 183 11 1 1 Volume averaging procedure e 183 11 1 2 Double decomposition concept s s e pe ecese 0 2 0 0
459. r bound n_cells and other upper bounds in Table 18 17 then use get_number function Alternatively use Fortran 95 2003 utilities such as ubound array dim and lbound array dim for upper and lower array bounds max_cells ubound cell_ volumes 1 where max_cells n_cells b To get the x y z co ordinates of vertices then do the following real wph pointer vertices call get_grid_geom xyz_vert vertices Starting and ending interface indices are needed for each material domain to be able to use interf_dist_v and interf_normal_d Example integer iwp pointer Ad L2G real wph pointer i_dist_vec call get_mat ise_high_interface 11 12 call get_grid_geom interf_dist_v i_dist_vec Now i_dist_vec 1 3 i1 1 i2 n_mat_doms represents all distance vector from a cell to interface for all material domains If distance vector from a cell to interface for material domain are required then use i_dist_vec 1 2 i1 1 i2 1 A similar approach is used for option interf_normal_d 18 3 2 10 get_grid_connect This UAR is used to retrieve information related to grid connectivity This UPR can be called in 2 different ways as represented in Table 18 18 and explained here a b a Consider the example given below integer iwp pointer f_verts integer iwp pointer internal_face_vert integer iwp pointer isa_fv call get_grid_connect n_face_verts f_verts ca
460. r details can be found in the GEOMETRY chapter of the Documentation The boundary refinement specification can be saved either in the triangle model file tri or in the global mesh file mesh Often it is more convenient to save the refinement in the mesh file as it can be more easily transferred to different models olx al al al aj E A el Blend to boundary depth 1 r Refinement Specification Destination le Paint Face Auto Paint Paint Line Motion zmin 0 036690 zmax 0 326520 Refinement Motion Info Paint Angle 25 AN E w w Options Stitch Mesh W Tangles veme A A fal A Parts Boundaries a snow Tri Type __ Refine A ay a 1 Yes 44924 Wall No bou Slicing Inlet Outlei No S bo Inlet OutleiNo A Al o E E G o VE xj Sn fh 50 Specification forboundary2 lt gt Delete Reet Delete Boundary Marking gt Refinement depth at boundary 3 geua Hiat Ex E e r Refinement blending distance 2 Toggle Compress wi E 5 EE Refinement Blending Reduce Paint All ts C Blend to boundary depth Figure 19 27 Setting Boundary Refinement Save the mesh file as port mesh and this then completes the mesh preparation Ricardo Software December 2009 392 19 TUTORIALS 19 2 STEADY STATE PORT F
461. r from cell to 2 rval 1 3 1 n_internal_face bndf_dist_v distance vector from a cell to boundary face rval 1 3 1 n_bnd_faces interf_dist_v distance vector from a cell to interface rval 1 3 ise_high_interface i 1 12 n_mat_doms interf_normal_d normal distance from the near wall cell centre to the interface between fluid and solid rval 1 2 ise_high_interface i 1 12 n_mat_doms Table 18 17 Subroutine get_grid_geom to get grid geometry variables defined in access group jacc_grid_geom integer iwp pointer b_verts integer iwp pointer bnd_face_vert integer iwp pointer isa_bv call get_grid_connect n_bndf_verts b_verts call get_grid_connect 1l_bndf_verts bnd_face_verts call get_grid_connect isa_bnd_verts isa_bv For the face jb with k 1 b_verts jb vertices the particular vertex is returned as kv bnd_face_verts isa_bv jb k where array isa_bv provides starting address to be used in the array Lbver which provides the list of boundary face vertices integer iwp pointer nc_faces integer iwp pointer cell_faces integer iwp pointer isa_cf call get_grid_connect n_cell_faces nc_faces call get_grid_connect l_cell_faces cell_faces call get_grid_connect isa_cell_faces isa_cf b To get the number of faces that enclose each cell then Ricardo Software December 2009 324 18 USER PROGRAMMING 18 3 ACCESSING S
462. r head wall No mes 1264 Wall No outlet valve No A 6 85 Inlet Outle No port inlet No an 7 240 Wall No boundary_ No A e J 384 Vall No boundary_8 No Add Boundary Delete Boundary Show All Hide All Toggle Compress Reduce Paint All Paint Face Auto Paint Paint Line Motion Refinement Motion Info Auto Paint Angle 45 Before a calculation may be run on a model it is necessary to define the boundaries of the model For example when calculating the flow in the inlet port and cylinder of an engine the user may define the piston crown as one boundary the cylinder liner as another the valve surface as another the port inlet plane another etc A panel of tools is provided to set up this information The table as shown to the left displays information related to each boundary The first column simply holds a unique identifier number for that boundary The second column is filled with the colour that triangles on that boundary are filled with By clicking the left mouse button on a cell in that column the text in the cell will alternate between Active and Inactive An inactive boundary is not displayed on the screen and so boundaries may be quickly removed to gain an unobstructed view of hidden parts of the model The third column shows the number of triangles currently assigned to that boundary By left clicking and holding the mouse button over a cell in this column and dragging the mouse pointer to a
463. r interface index 12 1 n_domains ending lower interface index 12 1 n_domains Table 18 8 Subroutine get_domain to get values for variables defined in the access group iacc_do main describing domain structure mainly starting and ending indices of domain objects Ricardo Software December 2009 306 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES 18 3 2 3 get_mat This UAR is used to retrieve information related to materials Information regarding material reference properties starting and ending indices for different variables related to materials can be obtained though this UAR for all materials From Table 18 9 1t can be seen that this UAR can be called in 2 different ways a b a The first way of calling only needs 2 arguments and the information obtained has to do with material reference properties For example to get the reference pressure for all materials the fol lowing code can be used real wph pointer p_ref call get _mat r_pressure p_ref where p_ref now contains all reference values for all materials b The second way of calling needs 3 arguments where argument 2 and 3 need to be declared as integers and the information obtained is related to starting amp ending indices Consider a simulation with 2 materials where material 1 contains 2 fluid phases and material 2 contains 1 fluid phase In dices of starting amp ending phases over all materials can be obtained us
464. r is shown below by Eq 15 3 pS V kVT 15 3 Finite Volume Method The finite volume method is based on flux balance It means that sum of fluxes which enters the system must be equal the sum of fluxes which leaves the system Using Green s theorem equation 15 3 can be transformed to ar aT f odas car 15 4 In this transformation the volume integral was replaced by the integration over the boundary There are several types of boundary conditions that the heat transfer equation should allow so that the different types of heat transfer problems can be modelled The above equation only allows for a heat flux boundary condition Another boundary condition type is a fixed temperature boundary condition This is done by simply replacing the unknown function T in equation 15 4 with its known value and adding it to the right hand side as a source This approach is described in more details in the Discretization section Boundary Conditions oT T T k k q onl 15 5 on r T T Tro Pad h qc onl 15 6 on r AL tT 4 on T3 15 7 n where I boundary on which heat flux is specified I2 boundary on which convective heat loss is specified I3 boundary on which radiation is specified Discretization oT foge adr f adr f grdV 15 8 r ot T T gt T Equation 15 8 can not be solved directly therefore it needs to be discretized into discrete points Because the Finite Volume Method is used in
465. r of examples to show how UARs can be called to exchange information with the solver Note the the exclamation mark indicates a comment in Fortran 95 2003 no Variables within quotation marks given in this section describe options available and should be defined as integers before use Ricardo Software December 2009 303 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES 18 3 2 1 get_number This function is used to retrieve main variables and it is seen as a starting point in user program ming A number of variables related to the total number of domains materials phases species boundary amp interface regions geometric quantities such as number of cells vertices etc are given in Table 18 7 The following code integer iwp n_doms n_doms get_number n_domains declares n_doms to be of type integer and invokes the get_number n_domains func tion to return the number of domains in the simulation A similar approach is used to get other main numbers defined in Table 18 7 function get_number var_name Input Output var_name get_number return type integer n_domains Total number of fluid and solid domains n_fluid_doms Number of fluid domains n_solid_doms Number of solid domains n_mat_doms Total number of materials n_fluid_mats Number of fluid materials n_solid_mats Number of solid materials n_bnd_regs Number of boundary regions n_interf_regs Number of interface region
466. racing out a polygon and all faces within the polygon become selected To start the polygon Shift Left Click anywhere in the 3D canvas if the canvas is not already selected then it may take two clicks to begin the polygon The first point of the polygon is shown as a square subsequent Shift Left Clicks add extra nodes to the polygon To finish the polygon requires a Shift Double Left Click RubberBand Rubber band picking is similar in concept to Polygon picking except that a rectangular box is traced out with the mouse instead of a polygon R Desk 3d1 p1 tube GRD 5 xj IFR Fie Edt View Options Window Hep 18 x FT gt metas z e 200 2 4 Pes Aa i ly erm 30 x gt 2 gt 11 02 28 INFORMATION a fi e training_tutorials V4 mesh_import tube GRD Figure 19 95 Use Flat Picking to select the required face In this case Flat Pick will be used to select the other end of the tube to boundary 2 Rotate the model to see the other end of the model Shift and left click on a face in the required region to paint the entire end of the tube Figure 19 95 Once the new face sets have been defined then these new regions need to be saved to the GRD file Right click on one of the face sets and select save all sets Figure 19 96 Ricardo Software December 2009 44 19 TUTORIALS 19 4 IMPORTING THIRD PARTY MESHES La R Desk 344 p1 tube GRD 2 5 x IR File Edt Vi
467. radient l od ror O Yci V DicmidA By 11 24 where Y denotes the tortuosity diffusion tensor i jk In the context of the k e modelling the turbulent diffusion flux of species i is given by Equa tion 9 8 Its volume averaged counterpart is given as It pelul 2 ge 11 25 J where the turbulent mass diffusion coefficient is defined by Equation 9 10 The concentra tion dispersion term p GU j can be also expressed as a function of the intrinsic concentration gradient ae 1 ais 9 Yci p cU j D Pl l j y i jk OX 11 26 Ricardo Software December 2009 186 11 MODELLING POROUS MEDIA 11 1 THEORETICAL BACKGROUND by introducing the dispersive diffusion tensor Dei Finally the last term in Equation 11 22 represents the mass flux at fluid solid interface hucha is Zero The above mass diffusion coefficients can be lumped together defining an effective mass diffu sion tensor DE p Dit Dig Sje 2i Br 11 27 With this definition Equation 11 22 becomes d d 0 e 9 yc ay ypc dx ypciU ax aea YSci 11 28 O Energy conservation Volume averaging of the simplified fluid phase energy Equation 10 7 fluxes due to species diffusion omitted gives a O pp 21 4 2 Lea ro poa Hot 5 ar E A yp fiUi 11 29 vda t ax gt YP dv YPSiVi The intrinsic laminar heat flux is defined as pope 1 O yT 1 x w ap gt A a 7 f Tn 11 30
468. raged turbulent kinetic energy is defined as pS A ne 6 38 2p gt 6 38 It is useful to relate the definition of k with the Favre fluctuating turbulent kinetic energy k uu 2 k and notice that pk 0 Similarly the Reynolds acne and Reynolds fluctuating turbuleni kinetic energy are defined as k ulu 2 and k ulu 2 k respectively It can be shown that k k K u u 2 In relation to the turbulent kinetic energy Equation 6 40 the two terms in Equation 6 31 aut Ui pku correspond to viscous molecular diffusion and turbulent transport of k respectively One can derive the turbulent kinetic energy equation by averaging the instantaneous momentum Equation 6 8 multiplied by u see for example Wilcox 1998 pk o TIF aU ETA o y i cd DA Pr pe D _ Ot Op du 7 J u TNT As z a Oe pa A E A Pp Compressibility terms The terms on the right hand side of the above equation can be identified as P turbulent energy production pe dissipation rate of k O Dx contains viscous diffusion turbulent transport and diffusion associated with pressure velocity interactions O Compressibility part includes the viscous stress and pressure work terms as well as the pressure dilatation term These terms vanish in case of incompressible flow Ricardo Software December 2009 12 6 SOLVER FUNDAMENTALS 6 4 REYNOLDS AVERAGED EQUATIONS O Pp turbul
469. rbulent kinetic energy at inlet kp as u V 2 3 2 bos ee IU 13 1 us Us b z b le Subscript denotes the free stream conditions Using Kolmogorov relation the inlet values for the dissipation rate p can be estimated by 4 3 2 Ci ky 13 2 Ep X b L where Le e 1 represents a fraction of the characteristic inlet dimension one tenth of the shear layer width or the domain size Furthermore may be defined by using the ratio of the turbulent and molecular viscosity at the inlet as follows 2 1 aye LEPE 13 3 u Vu The values of to 10 for u can be often used at inlet boundaries representing the free stream boundaries 13 2 1 Setting up a velocity inlet boundary condition The setup panel for the Velocity Inlet boundary condition is given in Figure 13 2 In order to enter variable boundary values for each phase the Boundary Phase setup panel shown in Figure 13 3 top right should be open by selecting the particular Boundary Phase from the Solver Setup Tree Figure 13 3 top left In a similar way Boundary Species and Boundary Passive Scalar panels are displayed as illustrated in Figure 13 3 bottom left and bottom right respectively Phase Values The inlet velocity values are entered in terms of Cartesian velocity components la belled with X Velocity Y Velocity and Z Velocity Next Temperature and estimated absolute Pressure are specified A boundary value of Volume Fraction i
470. rces are normalized as follows 2F Cr _ Pref Ur Aref 17 3 see Section 13 8 for more information The Coupled Link Number is used to specify the WAVE link number for this boundary see Sec tion 17 16 17 10 Report Regions The Report Regions panel shown below allows the user to define extra regions other than bound ary regions defined by the geometry a x Report Regions Add Child Report Regions Arbitary Surface Domain ID 1 Filename xxx1 tri aA Tolerance 0 01 Figure 17 11 Report Regions Arbitrary Surface Panel The Filename must be a VECTIS geometry triangle file Mapping is done from the defined triangles to the solver mesh Each triangle is successively divided along its longest edge until one of the following criteria are reached 1 All nodes of a triangle are contained within the same cell 2 All nodes of triangle are greater than or less that the mesh extents ie triangle is outside mesh 3 Longest edge is less than the user specified edge Tolerance 17 11 Solid Domain The Solid Domain panel is very similar to its Fluid Domain counterpart Only the differences are as follows For the Postprocessing Output panel there are fewer available quantities that can be saved see below Ricardo Software December 2009 287 17 USING SOLVER 17 11 SOLID DOMAIN Postprocessing Output x Y Temperature y Density Y T
471. re S Pres_AVE versus iteration number Iter_No The update interval default 5000ms for the XY plot can be adjusted if necessary 17 14 Solver Control This section describes how to control the simulation The Solver Control is invoked via the View pull down menu The associated working directory can be changed if necessary The Run GroupBox displays the current jobs running in this working directory If any jobs have recently started it may be necessary to click on the Refresh button The Solver Options GroupBox list all the available options that can be modified In order to apply these changes the user must then click on the Send button In the Figure 17 17 run 005 is selected and the maximum iterations Ricardo Software December 2009 289 17 USING SOLVER 17 14 SOLVER CONTROL Hu B tele gp ell xBO0B AA oA Live Update ok Pr Pressure Browse for tun Pressure Files H Domain E WO O 1 10 3 Run 1 Run 2 Wall XY Canvas Refresh Available Files Figure 17 15 Live update amp XY Canvas pressure versus iteration number Value s la Iter_No Area_TOT Vel_AVE Temp_AVE Mflow_SUM Density_AVE T Pres_AVE Tke_AVE Ted AVE Update Interval ms 5000 Figure 17 16 Live Update Value GroupBox is changed to 100 In order to stop the simulation suddenly and cleanly with restart postprocessing write out it is probably simplest to
472. re based on the linear relationship between the Reynolds stress tensor and local mean strain rate tensor Equation 9 4 where the proportion ality coefficient is a scalar quantity the turbulent viscosity u While this simple framework is computationally very efficient it has some purely physical shortcomings The relation 9 4 can be expressed through the anisotropy stress tensor b t 2pk6 3 Y 1 bij Sij 385405 9 58 j 2pk a mys i a In reality the stress anisotropy is far from what is predicted by linear models This and other deficiencies can be summarised as follows cf Hanjalic 1994 Linear Reynolds stress strain relationship A consequence is the isotropic eddy viscosity Poor predictions for flows with turbulent stress transport for example flows involving strong separation and buoyancy Insensitivity to stress anisotropy Poor predictions wherever normal stresses are important e g stress driven secondary flows in the straight non circular ducts cannot be predicted at all Insensitivity to extra strain No physical mechanism to deal with complex flows three dimensional characterised by streamline curvature swirl and rotation However the k models have shown remarkably good agreement with experimental data for simple flows two dimensional attached boundary layers channels with pressure gradients and even for some recirculating flows dominated by pressure gradients or flows with infirm st
473. re correction then the user should select All Variables Except Pressure Correction The user can also choose to apply bad cell treatment to pressure correction as well but not the Ist pressure correction by selecting All Variables Except 1st Pressure Correction 14 5 Parallelisation The whole solution process is parallelised including above pre conditioners by employing domain decomposition strategy and MPI based message passing Ricardo Software December 2009 256 15 MODELLING RADIATION 15 1 Introduction And Overview Thermal radiation between different surfaces in a flow domain model can be predicted using the VECTIS radiation module VECTIS calculations can therefore include the heat transfer by con vection and conduction and radiation This help manual describes the radiation module and its individual programs in terms of what they do and how to use them This document assumes that the user has a basic knowledge of using of VECTIS The VECTIS radiation module is currently based on diffuse radiation theory so that the emission and reflection of radiation is independent of direction The distribution of the radiated heat to the surrounding surfaces is determined by the view factors of each patch Only wall to wall radiation 1s currently modelled and the fluid between the surfaces is assumed to have no interaction 15 2 Mesh File Setup The radiation modules require that the input mesh is gap free between faces To ensure this
474. re given in Figure 18 1 labelled with upr_generic This generic UPR is used to implement general solver functions such as initialisation of variables executed after the default initialisation adjustment manipulation of variables called every time step or iteration post processing of results at the end of every time step iteration etc 2 upr_init This UPR is used to specify initial cell values for VECTIS MAX solution It is called for each phase species present in fluid solid material domains Called once at the begin ning of each simulation after reading the input file and before beginning the solution process Any values set at this stage will overwrite values read from the input file 3 upr_bnd_cond This user programmable routine is used to specify boundary conditions for the VECTIS MAX solution variables This is called from the same access point as upr_init Boundary values set through the input file are overwritten at this stage of UPR Other boundary profiles may also be defined This UPR is called for each boundary region and each phase species present in the corresponding fluid solid material domains 4 upr_properties Thermo physical properties of individual phase or species can be specified through the use of this user programmable routine For example if a property variable of a phase species can be modified then this UPR can be called to adjust the values of that variable This UPR is called once at the start
475. re of species The molecular mass transfer of species by diffusion including the definition of the species concentration or mass fraction c is outlined in Section Mass transport and mass diffusion coefficients All mechanisms of mass transfer molecular diffusion convection and turbulent transport are described by the species modelling equation Thermo physical properties featuring in the mass momentum and energy equations are properties of mixture of species They depend on the species mass fractions and their prediction is explained in Section Properties of Multicomponent Phase Ricardo Software December 2009 180 10 MODELLING SINGLE PHASE FLOWS 10 5 MODELLING MASS TRANSFER Note that the transport of energy due to species diffusion in Equations 10 7 10 26 Nsp ax be Ver a j is currently neglected This is justified when the Lewis number Le Ler 10 29 7 PCp for each species k is much greater than unity The mass transfer is an important ingredient of reacting flows including combusting flows The reacting flows are also described by the species modelling equation but are not yet supported Modelling mass transfer implies setting of the multi component phase see also Figure 8 3 If either multi component or WAVE based phase is selected this will enable the definition of con stituent species the WAVE based phase has the pre defined set of species and their thermo physical properties see Figure 8
476. reamline curvature Considering the conventional wall functions they have been derived for simple boundary layer near equilibrium flows It is very unlikely that wall functions will hold in complex flows The same conclusion can be made for the scalable wall functions They are expected to improve con vergence properties of the solution method The unified wall functions can potentially reduce the sensitivity of predicted results i e remove the grid dependence of the results Again it is unlikely that these functions used with the current high Reynolds number models will significantly reduce the grid dependence of the results For this the near wall models that can be integrated up the wall and properly account for the viscous and kinematic blocking wall effects should be used see Popovac and Hanjalic 2007 The current low Reynolds number model variants provide integra tion up to the wall The characteristic flow areas such as the stagnation region the boundary layers the shear layers and the wake shown in Figure 9 4 for the flow around a circular cylinder are commonly found in Ricardo Software December 2009 163 9 MODELLING TURBULENCE 9 6 INHERENT LIMITATIONS OF K MODELS typical engineering flow configurations For such flows the k models and most of advanced b Boundary Shear layer Stagnation Unsteady wake Figure 9 4 Characteristic flow regions for a circular cylinder a
477. refers to define turbulence as the general solution of the Navier Stokes equations it is entirely true and it adds nothing to what was known already i e that turbulent flows are three dimensional and unsteady irregular rotational diffusive and dissipative Turbulence appears when the local viscous force can t suppress random disturbances and flow instabilities amplified by the inertial buoyancy and other external forces In practice the minimum critical values of non dimensional numbers Reynolds number Re pUL u Rayleigh Ra BAT gL3p7cp UA O or Grashof number Gr Ra Pr Pr uc A Prandtl number are used as criteria for turbulence generation The above numbers are based on the characteristic flow variables velocity U dimension L and the temperature difference AT For values of non dimensional numbers below the critical ones a laminar flow exists Direct Numerical Simulation DNS of Navier Stokes equations is intractable for most turbulent flows as all length and time scales must be resolved numerically The smallest of these scales are associated with dissipative eddies which are responsible for the dissipation of turbulence mechanical energy into heat due to viscosity They are known as Kolmogorov length and time scales amp and T 1 4 v v 1 2 u l ie 9 1 where represents the viscous dissipation rate per unit mass and v is the kinematic viscosity The large en
478. region was proposed by Wolfshtein 1969 with y 0 263 To get a correct value in the wall limit the parameter y 0 2 needed to be used Rung et al 2000 9 4 3 Scalable wall functions The scalable wall functions Esch and Menter 2003 attempt to remove inconsistencies with the standard ones caused by the non uniform near wall grid resolution This is achieved by limiting the normalised wall distance y from below by y 11 63 The above standard wall functions are therefore used in which y max yp yz replaces yp The y limiting artificially moves near wall cells from the viscous sub layer into the turbulent layer Ricardo Software December 2009 159 9 MODELLING TURBULENCE 9 4 NEAR WALL MODELLING 9 4 4 Enhanced unified wall functions Enhanced wall functions described in Przulj 2009 have been designed in terms of non dimensional wall distances y star units in order to provide a smooth distribution of flow vari ables across viscous sub layer buffer layer and logarithmic law region In the viscous sub layer they satisfy corresponding wall limiting expressions and in the fully turbulent region they are identical to the standard wall functions O Velocity distribution Figure 9 2 shows the scaling of the channel and boundary layer DNS data Moser et al 1999 Spalart 1988 in a form U F y All data collapse well for y lt 15 20 Bnd layer Re 300 o Bnd layer Re 670 15H
479. represents the intersection point between the wall limiting and fully turbulent expressions Introducing the function fe y Elog Equation 9 54 can be expressed as E feElog fe 1 G E 9 39 Figure 9 3 right compares the above g blending function labelled as Present a against DNS data Unified expressions proposed by Wolfshtein 1969 Rung et al 2000 Rodi et al 1993 and Popovac and Hanjalic 2007 are also evaluated Large differences between the considered expressions and between each individual expression and DNS data are evident Another expres sion denoted as Present b and having a similar form as Equation 9 55 is in good agreement with DNS data This expression is formulated as fe 1 exp 0 031y E esp 0 077 9 56 and currently used to calculate near wall dissipation rate according to Equation 9 55 9 5 Low Reynolds number modelling The low Reynolds k e models have been devised as a synergy between the low Reynolds num ber model of Yang and Shih Yang and Shih 1993 and imposed physical bounds on the turbulence time scale pioneered by Durbin Durbin 1991 1996 2009 Yang and Shih defined the dumping function f in Equation 9 14 in terms of the mean strain rate tensor magnitude S see Equa tion 9 17 fa yl exp a R aR a3R R a 9 57 where the a coefficients read as aj 3 x 1074 a 6 x 1075 and az 2 x 1076 This make
480. ress for boundary face vertices ival 1 n_bnd_faces L_bndf_verts list of vertices for a boundary face ival 1 nabve subroutine get_grid_connect var_name ival Input Output var_name cha ival integer pointer l_face_cells list of 2 cells adjacent to an internal face ival 1 2 1 n_internal_face n_cell_faces number of faces which enclose a cell ival 1 2 1 n_cells isa_cell_faces starting address for cell faces ival 1 2 1 n_cells l_cell_faces list of faces that enclose a cell ival 1 2 1 n_internal_face Table 18 18 Subroutine get_grid_connect to get grid connectivity variables defined in access group jacc_grid_connect Ricardo Software December 2009 325 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES 18 3 2 11 get_turb This UAR is used to retrieve information related to turbulence model identifiers and constants Table 18 19 is split in 3 sections and represented as a c here a Turbulence modelling control parameters can be obtained as integer iwp pointer turb_par Call get_turb turb_ctrl_vars turb_par Array turb_par 1 nturb_par 1 n_domains now contains a list of turbulence control variables Number of turbulence control parameters is nturb_par 4 and are identified by pointers as O iturb_meth 1 physical method DNS RANS LES DES Physical model identified by iturb_meth returns one of the following iturb_inv 1 inviscid flow
481. rictly necessary to define the boundary types in Phasel as they are later defined in the solver input file However it is useful to define regions that are inlets and outlets as the mesher contains advanced options to allow different treatment depending on the boundary type Phasel contains 4 possible boundary types that can be applied to the boundaries Wall Zero Gradi ent Inlet Outlet and Cyclic Symmetry Only the Wall and Inlet Outlet types have any significance in VECTIS MAX and as stated above the solver input file allows these settings to be overridden Once the four boundaries have been defined the model should be saved as port tri File gt Save As Boundary Description Boundary Number Wall Wall back of valve Inlet Outlet 1 2 3 4 Table 19 1 Boundaries for Port Geometry ORicardo Software December 2009 389 19 TUTORIALS 19 2 STEADY STATE PORT FLOW Figure 19 24 Back of valve defined as separate wall boundary to allow specification of boundary refinement 19 2 6 Mesh Specification Having loaded a triangle file into phasel select Toolbars gt Mesh Setup from the menu The purpose of this section of Phase 1 is to set up the global mesh lines which vmesh uses as the basis for mesh generation A full description of the functionality of the Mesh Menu is described in the GEOMETRY chapter of the Documentation For this example focus the mesh in the valve region
482. rinting Data into Project Files Output File Outer Iterations 1 Output File Time Steps 10 Post processing File Frequency lterations 400 Co simulation Post processing File Frequency lterations lO Report Frequency Ascii 0 SDF 1 Save Residuals ToAscii y Save Residuals To SDF Summary Frequency Report Frequency 20 Report Level Report Boundary Fluxes For MassAnd Energy Save Initial Conditions Restan File Save Initial Conditions Post File Linear Equations Solver Residuals Information Figure 17 1 GroupBox for Frequencies For Printing Data into Project Files Referring to figure 17 1 Output File Outer Iterations controls the outer iteration output frequency to screen and out file Output File Time Steps sets the output frequency for unsteady runs Post processing File Frequency sets the frequency to write to the post files Depending on the Time Base chosen this frequency changes its meaning for Non dimenensional it refers to the frequency per cycle for 2 Stroke 4 Stroke or Seconds it is an interval in degrees or seconds respectively Report Frequency controls the frequency to write to the rep SDF file and the equivalent ASCII files By default the SDF frequency is set to 1 and the ASCII frequency is set to zero off The ASCII output files are explained in Section 17 6 Summary Frequency controls the frequency of writing run summaries to screen and ou
483. ro_const ipro_user O itempf 10 formation standard state reference emperan for reacting flow n ce cos involving solution of the energy equation Calculation options ipro_const ipro_user O itsch 11 turbulent Schmidt number which is used to calculate turbulent mass diffusion ce ces flux Calculation options ipro_const ipro_user Ricardo Software December 2009 319 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES To get the list of species properties calculation options then do the following integer iwp pointer species_pro_opts call get_property 1l_species_opts species_pro_opts The solution of the energy equation for solid materials requires the same properties as for fluid phases the density thermal conductivity and specific heat The calculation options for these prop erties are restricted to the constant or user defined values In summary considering solution of phase equations including solid phases nproph 6 properties may be needed density viscosity thermal conductivity specific heat molecular weight ther mal expansion coefficient id s from idens to itexc 6 Multicomponent fluid phases involve nprosp 11 physical properties density viscosity thermal conductivity specific heat molecular weight thermal expansion coefficient mass diffusion coefficient thermal diffusion coefficient formation enthalpy formation reference
484. ro_suth 6 see Equation 8 3 The next two options labelled as the user defined property ipro_user 7 and user supplied ex pression ipro_expr 8 are applicable to any property The last option is not implemented yet The options ipro_invis 9 and ipro_nonwt 10 are related to the treatment of molecular viscosity for an inviscid and non Newtonian fluid respectively The option ipro_nonwt invokes calculation of generalised non Newtonian viscosity and it is not yet implemented The last calculation option ipro_bouss 11 enables approximation of density by Boussinesq s Ricardo Software December 2009 321 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES formula Equation 8 27 To apply this formula the constant density at reference temperature Pre e Tre 1 and the corresponding thermal expansion coefficient B are required It is important to realise that an option is applicable to the certain properties only see the specifi cation of properties above b The constant or reference property values of phases species and passive scalars can be obtained as see Table 18 16 real wph pointer ph_pro_ref real wph pointer sp_pro_ref real wph pointer ps_pro_ref Call get_property r_phase_values ph_pro_ref call get_property r_species_values sp_pro_ref Call get_property r_ps_values ps_pro_ref For example ph_pro_ref ivis iph represents the referen
485. rol Timebase Output E Fluid_1 Fluid Domain Algorithm Turbulence Model Discretise E Phase_1 Fluid Phase Initial Condition Postprocessing Output ld Grid Extract Dialog Materias Remiro A Figure 19 38 Importing the mesh into R Desk Solver Setup ORicardo Software December 2009 398 19 TUTORIALS 19 2 STEADY STATE PORT FLOW Each material in the grid file needs to be allocated as a fluid domain or as a material in a solid domain In this case the computational grid contains a single fluid domain As such the default setup for fluid solid domains in the mesh import dialog is correct for this case bid R Desk 3d1 AutoPreview port GRD lol xj we Fie Edt View Options Window Help 18 x DAB S been y 30 8 w B QQ gt xBDE8bB ADA ax Solver Setup Tree See E Global_domain_1 Global Domain Restart Control Timebase Output E Fiuid_1 Fluid Domain Monitoring Points ase_1 Boun E Bnd_Reg_4 Inlet Given Velocity Boundary Re Phase_1 Boundary Phase Y Plots Solver Setup Tree Soler Sep AX Extract Materials And Boundaries From Mesh File Filename l port GRD Os Figure 19 39 Grid file imported into R Desk By importing the grid file the Solver Setup Tree is automatically populated with the correct number of domains and boundary regions If show mesh preview is
486. roperties m are calculated in a similar way as the density Equation 8 45 Npn Om Y 00 8 46 k 1 where denotes the property value of the constituent phase Note that for the multicomponent phase the property should be calculated according to Equation 8 39 8 6 Selecting Continuum and Its Properties It follows from the previous sections that the physical properties can be associated with species phases and with the mixture of phases The association depends on the type of fluid flow Single phase single component fluid The phase is the same entity as the species i e pure substance and the phase properties appearing in the transport equations are required Single phase multicomponent fluid Some species properties are required to solve mass frac tion equations and others to calculate the properties of the multicomponent fluid phase see Equation 8 39 The latter are needed for the fluid phase transport equations Multiphase multicomponent fluid The properties required depend on the multiphase flow modelling Full Euler Euler model Phase properties are required For the multicomponent phases the properties can be derived from constituent species properties Equation 8 39 Mixture and Volume of Fluid VOF model Properties of the multiphase mixture must be defined see Equation 8 46 from properties of constituent phases For the multicomponent phases the properties can be calculated from consti
487. routine to specify initial cell values for VECTIS 4 solution subroutine upr_init var_name idt icell icel2 This routine is called for each phase species present in fluid solid Imaterial domains after the default initialisation is done by the solver Modules used imported type definitions parameters scalars and arrays use upr implicit none integer iwp intent in Pe Ade amp domain type index yicell amp material domain start cell index icel2 Imaterial domain end cell index character len intent in var_name name of a variable field Local Variables real wph pointer fi gt null pointer array for a scalar field real wph pointer fi_vec gt null pointer array for a vector field character len len_var_name usr_obj_name amp user selected name for either phase or species object Usr_var_name user selected variable name integer iwp He iget amp flag to get phase species PS index usr_obj_id amp index of phase species or PS object corresponding to upr_obj_name nic lcell index real wph pointer xcell gt null lcell centre coordinates usr_obj_name aluminium usr_var_name temperature if usr_obj_name or usr_var_name return iget iget_phase if var_name spec_mass_frac iget iget_specs if var_name mass_frac_ps iget iget_ps call get_id usr_obj_name usr_obj_id iget if abs idt usr_obj_id and var_name
488. ry Species Boundary Passive Scalar Turbulent Length 0 001 Passive_Scalar_1 Boundary Passive Scalar pa Passive_Scalar_2 Boundary Passive Scalar Volume Fraction 0 Species_1 Boundary Species a E Passive_Scalar_1 Boundary Passive 3 X Species Concentration 0 Passive Scalar Concentration 0 Species Flux 1 Passive Scalar Flux 1 Figure 13 3 Setting up an Inlet Boundary type outline of phases species and passive scalars in R Desk 13 3 Mass Flow The mass flow rate is usually specified at the inlet to the fluid domain In case of compressible flows either the magnitude of the normal velocity is scaled or the velocity vector is scaled For the incompressible flows the mass flow inlet type is equivalent to the velocity inlet type For incompressible and weakly compressible flows the mass flow rate can be prescribed at an outflow boundary In this case at least one pressure inlet or outlet boundary should exist The mass flow rate at a boundary face is given as mp Pp tip Ap 13 4 where mp const There are two approaches to maintain the mass flow rate at the boundary constant normal velocity scaled and velocity vector scaled In the first normal velocity scaled approach the boundary velocity is evaluated as de el me 2 As Uy Up 2 25 a 13 5 b p A Ricardo Software December 2009 218 13 BOUNDARY amp INTERFACE CONDITION TYPES 13 3 MASS FLOW where i is the new iterati
489. ry or interface regions The import of a grid file is done by reading in the appropriate grid file project_name GRD The VECTIS MAX grid file format GRD is generated using VECTIS MAX mesher vmesh The following file formats are supported in VECTIS MAX solver VECTIS MAX grid file GRD Universal grid file unv O Vectis 3 DAT O Nastran nas Cedre ccm O Spider flma CGNS cgns When the vsolve command is run without specifing a grid file name a message appears advising the user of accepted grid file types 1 linxx vsolve USAGE vsolve np 2 grid type debug level proj_name where type O GRD Default 1 unv Universal 2 DAT Vectis 3 3 nas Nastran 4 ccm Star CCM 5 flma Spider 6 cgns CGNS o 2 linxx Ricardo Software December 2009 103 4 MESHING 4 12 MESH QUALITY CHECKS Wall normal distance Wall distance ratio Non orthogonality Ps Wrap angle Volume change ratio 2 Ao37 Figure 4 15 Mesh quality checks 4 12 Mesh Quality Checks The finite volume discretisation deals with cells of any shape such as hexahedra prisms tetrahedra pyramids and Vectis cut Cartesian cells However in terms of the numerical accuracy and stability the body fitted hexahedral or prismatic grids near boundaries are the best choice The quality of the mesh used in a CFD simulation is very important Ther
490. s PU dbuik Re PEC Cbulk 12 51 ui i c paa 12 52 Ai The quantity U in Equation 12 51 is the mixture velocity magnitude Ricardo Software December 2009 206 12 MODELLING MULTIPHASE FLOWS 12 4 PHASE CHANGE MODELLING O Boiling at a wall surface The formulation of the mechanistic model takes into account the single phase heat flux qsp the nucleate boiling heat flux qnuc and the quenching heat flux qyue to give the total heat flux q spl nuc Qque 12 53 The nucleate boiling heat flux nuc is given by the following relation Anuc V fNa Pghfg 12 54 where V is the bubble volume at departure from the wall and is evaluated as v Pi 12 55 A number of models have been proposed to calculate the wall bubble diameter Dpw in rela tion 12 55 A modified version of the Tolubinsky and Konstanchuk 1970 model is employed as Dpy min 1 2 xp ae mm 1 4 mn 12 56 The bubble release frequency in relation 12 54 is given by the following expression _ 48 Pi Pe f 3p Dbw 12 57 whereas the nucleation site density Na is given according to as Na 210 Twan Trat 12 58 The quenching heat flux in relation 12 53 is evaluated by using the following expression que Q Twal T 12 59 where 2 is the effective wall area covered with bubbles and is calculated as Q min one 0 25Dbw tNan 12 60 The dimensionless number 7 is given as n 4 8e 80 The Jacob number Ja is defined ac
491. s The wall shear stress can be used to calculate the friction or viscous shear force which acts on the wall boundary region Thus the total time dependent force exerted on the given wall boundary Ricardo Software December 2009 224 13 BOUNDARY amp INTERFACE CONDITION TYPES 13 8 WALL consists of the pressure force F and friction force Fy F pidA tjidA F F 13 8 wall wall Here p is pressure T j is the wall stress tensor dA and denote differentially small wall surface and its unit vector respectively The total force is usually decomposed into the force parallel to the oncoming stream the drag force Fp and into the force normal to the stream the lift force F The forces or force components F are usually normalised with the help of reference density Pref velocity U ef and area Aye so that the corresponding force coefficients are defined as 2F 13 9 Pref Ur Aref Cri The provision of the above reference quantities is explained in Fluid Domain Inputs Figure 17 6 Both total and viscous force coefficients defined for each Cartesian coordinate axis are written by default into the SDF report file When the Boundary Report box Figure 13 9 is checked on the force coefficients will be written to the ASCII wall file projectname wall_runnumber see File Output section Apart from the force coefficients and other variables the above report files contain the total energy transported through the wall a
492. s none Data Faces F Color Scales Opacty 0 5 Plot Geometry Number of Polyhedron Elements read 181027 Read 181027 elements read 181027 elements 634271 faces 272398 nodes 3 Y a extents 0 15387 lt x lt 0 193 0 020499 lt y lt 0 2705 0 03669 lt z lt 0 32652 PlotProperties Solver Setup J Li R Desk 3d1 p2 port GRD 10 xf we File Edit View Options Window Help lex DAB SI rd 3 160 8 a EH ERE A ae ll xBODB then Ba gt Plots Solver Setup Tree J Properties Plot ax Lines imesh Faces TT Color 1 el Y Opacity 0 5 T I Auto apply Apply Number of Polyhedron Elements read 181027 Read 181027 elements read 181027 elements 634271 faces 272398 nodes El Y x extents 0 15387 lt x lt 0 193 0 020499 lt y lt 0 2705 0 03669 lt 2 lt 0 32652 Sets __Plot Properties Solver Setup commana gt A Es Figure 19 34 Removing a plot from a canvas Ricardo Software December 2009 396 19 TUTORIALS hie R Desk 341 me File Edit View Options Window Help 19 2 STEADY STATE PORT FLOW BASEE protect MR Viewer 1 y Plots lis creato new E Type E Viewer Plot Properties Lines none faces y Opacity 0 0 ud F Auto apply Apply Figure 19 35 Open a new VECTIS project The new project will open with the default layout as shown in Figure 19 36 The
493. s n_regions Number of boundary and interface regions n_phases Total number of phases n_fluid_phases Number of fluid phases n_solid_phases Number of solid phases n_species Number of species n_ps Number of passive scalars n_cells Number of internal and halo buffer cells n_bnd_faces Number of boundary faces n_internal_face Number of internal faces n_mat_interface Number of internal faces at material interfaces n_vertices Number of vertices n_partitions Number of partitions processors n_current_part Current partition processor number n_halo_cells Total number of halo buffer cells Table 18 7 Function get_number to get values for main variables numbers defined in the access group lacc_numbers Ricardo Software December 2009 304 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES 18 3 2 2 get_domain This UAR is used to retrieve information related to domains Thus a list of information is returned for all domains Information such as number of materials indices of starting and ending phases species boundary amp interface regions etc see Table 18 8 for a complete list are provided for all domains For example if a simulation case has 2 domains and domain 1 has 100 cells domain 2 has 200 cells then the following code integer iwp pointer il i2 call get _domain ise _cell 11 12 will give the following information 1 i2 1 100 for domain 1 101 12 2 300 for domain 2
494. s vapour dl dispersed liquid dv dispersed vapour gt From Lahey and Drew 2001 O Discrete Phase Modelling DPM or Eulerian Lagrangian approach Discrete phase mod elling assumes the dispersed topology of the multi phase flow Individual dispersed elements droplets bubbles particles are tracked in time through the flow domain by solving momentum equations given in the Lagrangian form The dispersed elements exchange mass momentum and energy with a continuous fluid phase The transport equations of the continuous phase are those used in Eulerian Eulerian approach and they are usually coupled with dispersed element equa tions This modelling approach is usually applied to flows with the low volume fraction of dispersed phase elements less 10 Spray modelling atomisers spray breakups droplet collisions dynamic drag accounting for changes in the droplet shape wall films turbulent interactions etc belong to this class of multiphase flow Discrete phase modelling in VECTIS MAX will be considered in future O Multi Fluid Modelling or Eulerian Eulerian approach The Euler Euler or multi fluid ap proach describes each phase by Eulerian conservation equations for mass momentum and energy The phases are treated as inter penetrating continua with their own field variables such as velocities pressure and temperature The space and time ensemble averaging of lo cal instantaneous phasic equations requires a phase i
495. s Iname for each VECTIS MAX object 330 18 23Subroutine get_id to get VECTIS MAX objectsid o o 331 18 24Subroutine get_parent to get parent identifier o oo o 332 18 25 Function eq_idt to return identifier idt for derived datatypes 332 18 26L1st of accessible variables 2 ine id Bk ee ee A eS 335 18 27 Subroutine get_field to get field values either for a scalaroravector 336 18 28Subroutine get_grad to get gradient of scalar or vector variables 338 18 29Function deln_star to calculate non dimensional wall distance 000 339 18 30Subroutine local_force to calculate viscous and pressure forces at individual boundary faces Se IMIEPIACES ate tase ih Gio a ne ek a dt eae al eae ete amp ad ley a ioe oath dian ee IA 341 18 31 Subroutine user_post to store user data 2 ee ee 342 ORicardo Software December 2009 xv LIST OF TABLES LIST OF TABLES 18 32Subroutine global_sum to get global sum over partitions o oo 344 18 33Subroutine global_max Subroutine global_min to get max min over partitions 345 18 34Subroutine concat_array to take local array from each partition and concatenate into global artayion partilod le waa uoic eee a Mele daa pa ea a a a ha daw ad Mee bee 347 18 35Subroutine upr_properties to modify properties 2 02 0 004 348 18 36Changeable
496. s and errors are identified with a four digit number These messages are written to the program information area The first digit identifies the type of message O Warning 1 Array bounds error 2 Other error Those messages which require explanation are listed and explained below Others are self explanatory WARNING 0102 VECTIS PHASE 1 THREE TRIANGLES FOUND ON LINE m n TRIANGLE NUMBERS ijk This warning can be produced when loading triangle data into the triangle processing part of Phase 1 Three triangles have been found with an edge between the two nodes m and n This is not permitted in the definition of a closed surface which is that each edge should lie on exactly two triangles so the situation is treated as two separate edges Places where this occurs will appear as red lines lines which need stitching A large number of these warnings may indicate overlapping or duplicate surfaces WARNING 0103 VECTIS PHASE INCORRECT IDENTIFIER IN FILE filename IT DOES NOT APPEAR TO BE A VECTIS TRIANGLE FILE A valid triangle file must start with VECTIS_TRIANGLES A file not containing this identifier is ignored WARNING 0201 VECTIS PHASE 1 POOR CONVERGENCE IN BISECTION SOLUTION EXITING ROUTINE This warning may occur during the tracking of VDA trim curves on trimmed surfaces It is caused by round off errors in the numerical scheme used to locate the intersection between a trim curve and a line of constant s or t on the
497. s are called linear EVM models Non linear and algebraic Reynolds stress models have the Reynolds stress tensor defined as a general poly nomial function of the mean velocity gradient cf Gatski and Rumsey 2002 For example models with the quadratic terms can be cast in the following form 2 1 k 1 Tij gt 3Pk 2 by s Sud us A sus Fsm Smi k k 1 14 Be SW S Wri Mr Bs it 5 Wann 9 12 where fj are closure coefficients and W represents the mean rotation tensor _ 1 dU dU ll ES g i LB In case of non linear models the closure coefficients are calibrated with the help of experimental or numer ical data and physical constraints For the algebraic stress models the coefficients are derived in a consistent way from the full DSM models Depending on the number of differential equations which are used to predict turbulence scales in Equa tion 9 7 the EVM models can be classified as Zero equation algebraic models Both turbulent length scale 4 and time scale T are specified alge braically employing the empirical Prandtl mixing length theory Examples are Smagorinsky 1963 and Baldwin and Lomax 1978 models One equation models In these models one of turbulence scales is obtained from its own transport equa tion The turbulent kinetic energy is usually used to define the velocity scale v k 2 while the distribu tion of the length scale 4 is prescribed
498. s can also be specified for each solution algorithm The default value is 2 Algorithm a A Solution Algorithm e Simple Simplec Number Of Pressure Corrections 2 Figure 14 9 R Desk setup for pressure correction algorithm 14 2 6 Linear solvers The bandwidth of the sparse matrix is reduced by applying the Reverse Cuthill McKee re ordering algorithm to cells George and Liu 1981 The system of linearised algebraic equations 14 48 Ricardo Software December 2009 252 14 NUMERICAL SOLUTION 14 3 IMPLEMENTATION OF BOUNDARY CONDITIONS is under relaxed implicitly Patankar 1980 It is solved by a family of preconditioned conjugate gradient solvers Van Der Vorst 1992 Sleijpen and Fokkema 1993 Setting up linear solver gt From R Desk under each Fluid Domain node in the Solver Setup Tree panel left click on Equations amp Solver node and the Equations Solver panel is displayed to the right The Solver Type can be specified for each transport equation to either be the Bi cg Stabilised method or Symmetric Conjugate Gradient method Figure 14 5 The default solver is Bi cg Stabilised A number of pre conditioners are also available such as zero and first order incomplete lower upper ILU factorisation pre conditioners Meijerink and Van Der Vorst 1981 Benzi 2002 Jacobi including a new unstructured grid adaptation of the well known Stone s SIP solver as an efficient pre conditioner The list of
499. s fraction Ricardo Software December 2009 13 8 MODELLING CONTINUA 8 2 EQUATION OF STATE of the i th species c is defined as Ns Pp Ns p ci mi m m mi Y ci 1 8 7 El where m is the mass of i th species m is the total mass of the fluid phase occupying the control volume V and Nsp is the number of species in the mixture Since various species move at different velocities Us the mass flux vector Je relative to the local mass averaged velocity U re ciU is defined as I pci 0 0 8 8 where p m V is the mass density of the mixture In a laminar flow the mass flux can be often modelled as a VT Je P PiV ci Dr 5 8 9 where Y denotes the mass diffusion coefficient for i th species and Yr is the thermal diffusion or Soret coefficient The first term represents the ordinary mass diffusion described by Fick s first law of diffusion whereas the second term is the thermal diffusion In general the ordinary mass diffusion is a complicated function of concentration gradients of all the mixture species and it is not always governed by Fick s law Modelling such fully multicomponent diffusion is important for diffusion dominated laminar flows e g chemical vapour deposition and it is not currently sup ported The driving forces for the ordinary diffusion are concentration gradients while the thermal diffusion is driven by a temperature gradient examples are devices designed for separation of mixtures
500. s outflow boundaries for which fully developed flow conditions can be assumed In order to comply with the fully developed flow conditions a zero gradient condition and negligi ble boundary diffusion flux the outlet boundaries ought to be placed downstream far away from the regions with significant flow changes Two sub types are available either prescribed mass flow rate split for single multiple outlets or prescribed mass flow rate for the single outlet The flow outlet should not be used for highly compressible flows Ricardo Software December 2009 222 13 BOUNDARY amp INTERFACE CONDITION TYPES 13 7 SYMMETRY PLANE 13 6 1 Setting up a flow outlet boundary condition To set up an Outlet Boundary condition left click on the relevant boundary region in the Solver Setup Tree The boundary setup panel is displayed Under Boundary Condition Type se lect Outlet from the ListBox and the Outlet Boundary setup panel is displayed A number of settings becomes available as shown in Figure 13 7 Bnd_Reg_3 Outlet Boundary a amp Region Name Bnd_Reg_3 Spilt Factor 1 Region ID 3 Massflow Rate 0 Out Material ID 1 A Boundary Report Out Given Mass flow Rate In Boundary Condition Type Outlet Boundary Condition Op Given Flow Split Spilt Factor 1 Boundary Setting Uniform Values a Mas fiow Rate 0 Out a a Spilt Factor 1 Mass flow Rate 0 Out
501. s port tr Plot Properties commana gt A Figure 19 30 The Launch Mesher dialog box Browse to the mesh file that was saved in phasel Click on Launch to start the mesher A terminal window should open displaying the output from the mesher Once the meshing is complete the terminal window will close The screen output is also written to a file in the same directory as the mesh file The file should be called port OUT In this case the mesh will have around 200 000 cells eX C WINDOWS system32 cmd exe lt z done gt done gt done gt done gt successfully generated STATISTICAL DATA OF GENERATED MESH Number of generated cells 182754 boundary 83882 internal 99672 Patching method d by Marching Cu bes 75372 89 10 gt d by Exact Fit 447 5 29 gt 1 73_ of boundary cells of all attempts correct number of volumes e volumes with too low quality so the undo was applied ndary cells 81356 96 18 gt tive volumes they are deleted now 4 84 7 ivated 3236 3 83 gt 2 81999e 804 u3 81396 boundary ce MESH FILE Figure 19 31 Screen output from the mesher The mesher will write the generated computational mesh to a grid file In this case port GRD This can be opened in R Desk for visualisation purposes It is also possible to modify the boundary regions contained in the grid file within R Desk however this is outside the scope of this tutorial Open the grid fi
502. s printed onto screen menu Boundary feature resolution size It is possible to set fr sizes for boundaries This is currently available only in the Wrapper parame ters file using keyword BOUNDARY_FEATURE_RESOLUTION_SIZE Boundary fr sizes can be smaller than the global fr size Base mesh will be refined to the level not exceeding the specified sizes This can enhance resolution of particular features for example close surfaces Maximum refinement Instead of boundary fr sizes it is possible to set certain boundaries refined to the maximum selected fr size This also can be done only in the Wrapper parameters file using the following key word MAXIMUM_REFINEMENT_ BOUNDARY Leak detection In case of base mesh leaks inside the geometry will be wrapped from inside and outside In this case usually memory and CPU usage are significantly increased One of the ways to prevent this is to cap large holes in the original geometry The Wrapper can be run in leak detection mode Ricardo Software December 2009 47 3 GEOMETRY 3 14 GEOMETRY WRAPPING generating WrapperLeakPaths stl for visualisation of leak paths This mode runs faster and does not require some parameters Leak paths connect external points with user specified internal ones The maximum number of internal points is 24 If an internal node is too close to the geometry a warning will be printed and the closest internal point will be selected
503. s required when modelling multi phase flow i e when the volume fraction equation is solved for If a turbulent flow is simulated then boundary conditions for turbulent flow include Turbulent Intensity and Turbulent Length Species Values In case of a multi component phase made of two or more species boundary values of Species Concentration mass fractions for each species should be specified The Species Flux value is not used at inlets Passive Scalar Values The Passive Scalar Concentration value is required when solving the passive scalar transport equation The Passive Scalar Flux value is not used at inlets Note that the boundary values of volume fractions and species passive scalar concentrations mass fractions must be between zero and one and that their sum over all phases or species passive scalars should be one Ricardo Software December 2009 217 13 BOUNDARY amp INTERFACE CONDITION TYPES 13 3 MASS FLOW Phase_1 Boundary Phase K Fluid_1 Fluid Domain Monitoring Points Algorithm X Velocity 1 Multiphase Turbulence Model Y Velocity 1 Equations amp Solver Discretise Z Velocity 1 Phase_1 Fluid Phase Boundary Regions Pressure 100000 Bnd_Reg_1 Inlet Given Velocity Boundary gt Phase_1 Boundary Phase __ _ _ _ _ _ _ _ _ _ _ _ __a Temperature 300 Boundary Species Species_1 Boundary Species Turbulent Intensity 0 02 Species_2 Bounda
504. s solved for the incompressible flows As the total energy can be expressed as UU p UW P E k h H 10 2 er gt A gt k p 10 25 its conservation equation can be obtained from the total enthalpy equation as 0 ra Nsp at PE Jy PE U U o hy J cj tS Ck am n S el tens TA Convection Due to Species Diffusion o l Ok t qj j 37 tUat pu U 10 26 Ox q di Oj Ox ij ij T Pav pfiUi Heat Fluxes Work of Stresses Pressure Work Source Body Force Work Turb Energy Based on the selection of the fluid compressibility the total energy equation will be solved for incompressible flows and total enthalpy equation for all compressible flows An important issue when solving either the total enthalpy or energy equation is the design of a conservative discretisation scheme As Equation 10 26 shows the convective and unsteady terms are expressed in terms of total enthalpy or energy while the diffusion term is naturally defined in terms of the temperature gradient It is beneficial especially in case of conjugate heat transfer modelling to discretise and solve the energy equation in terms of temperature as a primary vari able For this an efficient numerical technique is devised and described in the section presenting Ricardo Software December 2009 177 10 MODELLING SINGLE PHASE FLOWS 10 4 MODELLING HEAT TRANSFER the discretisation of the energy equation If the energy equatio
505. s the model sensitive to local strain rate and removes ambiguity associated with dumping functions using wall normal distances Away from the wall the model reduces to the standard one as it employs the standard model constants Another of the model s feature is the limiting of the turbulence time scale T from below by the Kolmogorov time scale T Eq 9 1 i e T k Th As reported by Watterson et al 1999 Yao et al 2002 a truncated version of the model where the additional source term S in Equation 9 20 is neglected has produced satisfactory results for separating flows including vortex shedding in complex geometries Thus the same truncated version of the model is also employed here The low Reynolds variants of either Standard k e model SKE or Standard k e model with realisable time scale bound TSB are available In the case of the SKE model the turbulent time Ricardo Software December 2009 162 9 MODELLING TURBULENCE 9 6 INHERENT LIMITATIONS OF K MODELS scale is bounded from below by the Kolmogorov time scale Eq 9 14 For the TSB model in addition to the lower bound Durbin s Durbin 1996 realisability condition is imposed as the upper bound on the time scale see Eq 9 23 9 6 Inherent Limitations of k e Models and Wall Functions The linear k models available in VECTIS MAX and other two equation EVM models rep resent the simplest and complete turbulence closu
506. s with the case when a 1D array is used whereas the second part if a 2D array is used Each part of the table deals with 3 data types integer real and double The following example illustrates the use of a real array t_val in concat_array subroutine Ricardo Software December 2009 345 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES number of stored values on each partition integer iwp allocatable save dsize ltotal number of stored values integer iwp save gsize compressed array of temperature real wph allocatable save t_val integer iwp i_part amp this partitions number n_part amp number of paritions n counter for compressed arrays n_part get_number n_partitions i_part get_number n_current_part allocate dsize n_part dsize 0 dsize i_part n number per partition get global sum over partitions call global_sum dsize n_part gsize sum dsize total number allocate memory for compressed array of temperature lon partition 1 globally sized arrays are created if i_part 1 then global arrays allocate t_val gsize else Ilocate arrays allocate t_val dsize i_part endif concatonate local array to global array on partition 1 call concat_array dsize t_val Ricardo Software December 2009 346 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES subroutine concat_array dsize array garray Arguments Description
507. scid laminar or turbulent flows can be performed The potential flow model is also available as a part of the flow initialisation Body forces Modelling of the gravity force effects is provided for all flow regimes Heat transfer All convection modes forced natural and mixed can be simulated by solving momentum and energy equations in fluids In solids the heat conduction is modelled CHT modelling describing coupled heat transfer through adjacent fluid and solid material domains 1s also available Mass transfer In case of multi component phase which is defined as a mixture of species modelling of mass transfer for non reacting flows is supported Passive scalars Similar to standard species transport of passive scalars can be modelled Turbulence Modelling The RANS Reynolds Averaged Navier Stokes approach is currently employed and closure of the RANS equations is provided via linear two equation k e models standard standard with realisable time scale option and RNG In the near wall regions viscous and wall blocking effects are accounted for by using either standard or scalable or enhanced wall functions Multi Phase Physical Models The current version of VECTIS MAX solver includes two models O Homogeneous mixture model This model neglects the relative motion between phases In addition other relevant fields such as temperature and turbulence quantities are shared by all phases O
508. searched corresponding vertices triangles on the boundaries bB1 bB2 bBn from the B geometry The summary of the comparison is printed cmpbw TrifileA bA1 bA2 bAn TrifileB bB1 bB2 bBn This switch is similar to cmpb but in the case that there are some topological or geometrical differences two trifiles TrifileA_tested and TrifileB_tested are generated The triangles are separated into several temporary boundaries corresp_tri problem_of_topology and problem_of_geometry If there is a possibility to repair the differences than also the trifile TrifileB_corrected is generated The corrected triangles are marked as a boundary named corrected_tri 4 5 Basic Scheme of VMESH When normal meshing task is run none of the parameters test rep sep info locate is used the scheme of the work looks like this 1 READ INPUT FILE FOR MESHING TASK The input ascii file meshfile is read Ricardo Software December 2009 84 4 MESHING 4 5 BASIC SCHEME OF VMESH 2 READ SURFACE TRIANGLES The input SDF file trifile is read For each triangle the geo metrical extents in x y z are found Then the node to triangles connectivity is gathered Also the connectivity between the triangles is found Then normal vectors of triangles are prepared and area of each triangle is calculated 3 PREPARE DUAL LEVELS In order to avoid collisions of sides of boxes with triangles perpen dicular to x y or z axes meshlines are shifted to some d
509. sec 1 kg sec momentum velocity source gt kg sec m sec N total enthalpy energy source gt kg sec J kg W turbulent kinetic energy source gt kg sec x m sec 2 W turbulent energy dissipation source gt kg sec m 2 sec 3 W s As the cell volume occupied by phase vol_ph is available the user will have to evaluate first the source terms per unit volume or per unit mass times density and multiply it by phase volume vol_ph In case of single phase fluid or multi phase mixture model the phase cell volume is the actual cell volume 18 4 5 User generic routine subroutine upr_generic id icall_pos Arguments Description id integer global id icall_pos integer calling position Table 18 40 Subroutine upr_generic to modify variables in arbitrary fashion This routine is used to modify variables in a general way It is called at various positions during the simulation see Table 18 41 See Section 18 7 5 for example 18 5 Writing and Compiling UPR Dynamic Shared Objects Currently UPRs must be written in Fortran 95 2003 The subroutine names and argument list must adhere to those listed in Section 18 4 In addition the Fortran 95 2003 module upr mod must Ricardo Software December 2009 352 18 USER PROGRAMMING 18 6 UPR CHECK REPORT MESSAGES CALL ID PROGRAM POSITION icp_beg_run start of simulation after initialisation icp_beg_time beginning of ea
510. ser Programmable Routines UPR are explained in depth in section 18 4 The methodology for writing a UPR compiling and using a dynamically shared object is the subject of section 18 5 Section 18 6 lists the messages that can be generated from the use of UPRs These include errors warnings and information messages Section 18 7 describes some examples of user programmable routines and their capability The examples are written in the Fortran 95 2003 programming language 293 18 USER PROGRAMMING 18 2 FUNCTIONALITY AND CALLING SEQUENCE 18 2 Functionality And Calling Sequence The UPRs enable access to the internal parameters of a simulation The five different user pro grammable routines are 1 upr_generic 2 upr_init 3 upr_bnd_cond 4 upr_properties 5 upr_sources The first UPR upr_generic is the generic UPR which is model independent Most of the solver parameters can be accessed using this UPR The other UPRs are more specific These user pro grammable routines are called at different points in the simulation cycle The purpose of each one 1s explained below 1 upr_generic This is the generic user programming routine It is called at five different stages of a simulation mainly at the beginning of a simulation before the start of each time step after the end of each outer iteration after the end of each time step and the end of the simula tion These five stages at which upr_generic may be utilised a
511. set the maximum number of iterations time steps to anything less than the current iteration time step The equivalent command line utility run_control can be used outside R Desk This program must run from the relevant working directory and is invoked without any arguments If any jobs are currently running the user will be prompted to change a parameter via a text menu Once the parameter modification has finished the user must enter 0 in order to apply these changes See Figure 17 18 ORicardo Software December 2009 290 17 USING SOLVER 17 14 SOLVER CONTROL Solver Control Working Directory N Jusr2 scratch BOIL_CASES ANNULAR_nightly Run Job Project Run Number 1 zeitoun 001 2 zeitoun 002 Refresh Solver Options Value Property Solver Control Number Of Time Steps Maximum Number Of Iterati 100 Post Processing Frequency Restart Frequency Post Processing Dump Number Of Iterations Equation Status Under Relaxation Factor Send Figure 17 17 Solver Control panel pbranch Changeable run time parameters q to quit Figure 17 18 Solver Control via command line run_control Ricardo Software December 2009 29 17 USING SOLVER 17 15 TOOLCHAIN LAUNCHER 17 15 Toolchain Launcher The mesher vmesh solver preprocessor vpre and solver vsolve can all be launched from within R Desk see Figure 17 19 Ha B gt Figure 17 19 Launcher buttons from l
512. sh line Add a horizontal red control mesh line Delete a red mesh control line by selecting this icon and then the red mesh line Set the number of verticaldivisions between red control lines Set the number of horizontal divisions between red control lines Measure the distance between two parallel mesh control lines HPS SoS Sle jj u Figure 19 5 Options to manipulate mesh lines Open R Desk either by typing rdesk at the command prompt or using the start menu Create a new VECTIS sesion through the menu using Menu gt New The default panel layout for a VECTIS project will be shown however panels can be re arranged by dragging and docking them into the desired format Additional panels can be opened by right clicking the project menu bar Fig 19 6 Once a desired layout is found it can be saved and also be made as the default using File gt Layout gt Save LA o Setup E Global_domain_1 Global Domain Restart Control Bnd_Reg_1 Wall Boundary Interface Regions Sub domans Report Regions Sold Domains Figure 19 6 R Desk Layout ORicardo Software December 2009 371 19 TUTORIALS 19 1 BASIC TUTORIAL 19 1 6 Generating the Grid File Once the global mesh has been defined the mesh generator vmesh can be started using Launch Mesher button in R Desk Fig 19 7 Once the panel has opened browse to the tube mesh file saved from Phasel Then click launch Alternati
513. should be close to zero Figure 19 59 i lol we File Edit View Options Window Help la x hi BD Pre VECTIS 1 X DIS wile xRDBseenhR BE gt 2 gt Live Update ax xyi Browse for run 0 015 4 c Tutorials Port A 2 0 01 4 f J E El Domain 0 005 4 0 E Arbitrary Surface 5 Arbitrary Surface 1 Fl Wall Y 0 q gt m Refresh Available Files 5 Y 20 005 4 Value El ta J X Ang_Mom j J Y Ang_Mom 0 01 4H Z Ang_Mom f 311 Tke_AVE a al Y Ted AVE oppa Update Interval ms 5000 r m WIP a 1 0 100 200 300 400 500 600 700 Live Update Solver setup Tree J Iteration XY Plot Manager E X Messages px XLabel fteration Xx Min o X Max 700 W Auto Y Label mentum YMin 0 015 YMax 0 015 IV Auto I Showlegend IV Show X Grid IV Show Y Grid Name x Y xY 4 Tter_No Z Ang_Mom 4 po la Solver Setup___ XY Plot Manager Jar _ rrr Figure 19 59 Live Update 19 2 14 Post processing Once the simulation converges it will stop and write restart and post processing files The pro jectname post_0 file contains simulation data for the entire calculation domain at specified post processing times Open the post file in R Desk using the File gt Open gt File menu option It should appear in the plot tree By default the plot will be added to the active 3D canvas Note this beh
514. sible that maximum minimum sub domain fluid temperature goes above below the reference temperature Tef for the heater cooler respectively The solver takes care about these non physical model situation by adjusting heat exchanger reference temperature T ef depending on the exchanger type The exchanger type cooler or heater is not currently explicitly specified It can be however specified through values of the minimum temperature difference AT nin as follows HO AT min gt 0 heater Tref gt Tf max AT min O ATmin lt 0 cooler Trey lt Ty min AT min AT min 0 allows both heating and cooling Ricardo Software December 2009 195 12 MODELLING MULTIPHASE FLOWS This chapter introduces basic physical and modelling aspects of multiphase flow and describes two modelling approaches multi fluid or Euler Euler approach and single fluid Mixture and VOF approach together with corresponding modelling equations After that setting of available models is presented 12 1 Introduction Multi phase flow can be defined as the simultaneous flow of several phases where the phase is the one of the three thermodynamic states of matter either solid or liquid or gas In many industrial flow environments multiphase flows are very common they are generally the rule rather than the exception Typical examples in the automotive industry are engine injection systems diesel fuel injectors spray combustion coolant systems
515. simulations are separated fluid streams and conjugate heat trans fer where temperatures at the fluid solid interfaces have to be coupled Figure 7 1 can help to understand better VECTIS MAX multi domain structure The material domain is the basic con stituent in this case there are two fluid materials namely hot liquid material 1 and cold air material 2 Corresponding fluid domains coincide with these fluid materials The fluid domains are separated by one solid domain which in turn consists of two solid material domains material 3 and 4 Thus the whole computational domain has four material domains grouped into two fluid domains and one solid domain The first fluid domain hot liquid has three boundary regions wall at the left inlet at the top and outlet at the bottom Similarly the cold air fluid domain has three boundary regions wall at the right inlet at the bottom and outlet at the top Finally for each solid material domain there are two wall boundary regions at the top and bottom Two fluid solid interface regions and one solid solid region can be identified If the energy equation is solved for all fluid and solid domains there will be one global domain which coincides with the Ricardo Software December 2009 124 7 MODELLING SPATIAL DOMAINS 7 2 BOUNDARY REGIONS Hot Liquid Reg 1 Inlet 7 Wall 9 Wall Reg 5 Outlet l nn O T PPPE NEEE ERER O E A ERRE Reg 4 Inle Cold Air Figure
516. sis is then run again but this time has realistic wall temperatures For general gasket development and coolant jacket development it is common to simply perform the iso thermal analysis and use these results without considering the wall temperature effects 2 Run the coolant jacket analysis as a fluid only analysis and as an iso thermal analysis until a steady state converged solution is achieved Then use this fluid only analysis result to start a VECTIS conjugate analysis where both the fluid and the solid domains are solved in VECTIS as well as the heat flux between them and the surface wall temperatures This is more accurate than option 1 and generally easier as all the thermal analysis work is performed in VECTIS Both the surface and internal thermal results can then be mapped onto an FE mesh if required 3 The calculation can be extended to include the solid components around the coolant water jacket using a conjugate heat transfer analysis In this case the fluid will be solved non isothermally This will calculate the temperature distribution in the solid components and fluid taking into account the variation in heat transfer and wall temperature between the fluid and solid Ricardo Software December 2009 416 19 TUTORIALS 19 3 COOLANT FLOW 19 3 2 Aims This tutorial assumes that the user has the required knowledge to use VECTIS MAX to a level that 1s achieved after completing the basic tutorial After completing this tutorial the
517. solution Schemes Using Flux Limiters for Hyperbolic Conservation Laws SIAM Journal Numer Anal vol 21 pp 995 1011 14 1 3 1 Ricardo Software December 2009 449 19 TUTORIALS 19 4 IMPORTING THIRD PARTY MESHES Tennekes H and Lumley J 1986 A First Course in Turbulence MIt Press Cambridge MA 6 3 1 Tolubinsky V and Konstanchuk D 1970 Vapor bubbles growth rate and heat transfer intensity at subcooled water boiling Heat Transfer vol 5 12 4 1 2 Ubbink O 1997 Numerical Prediction of Two Fluid Systems with Sharp Interfaces ph D Thesis Imperial College University of London 12 3 2 Van Der Vorst H A 1992 Bi CGSTAB A Fast and Smoothly Converging Variant of Bi CG for the Solution of Nonsymmetric Linear Systems SIAM J Scientific Computing vol 13 pp 631 644 14 2 6 Vandromme D 1993 Turbulence Modeling for Compressible Flows and Implementation in Navier Stokes Solvers in Introduction to the Modeling of Turbulence Lecture Series von Karman Institute for Fluid Dynam ics 6 4 Veynante V and Vervisch L 2002 Turbulent combustion modeling Progress in Energy and Combustion Science vol 28 pp 193 266 6 4 Wark K 1983 Thermodynamics 4th ed McGraw Hill New York 8 3 2 Watterson J K Dawes W N Savill A M and J W A 1999 Predicting Turbulent Flow in a Staggered Tube Bundle Int J Heat amp Fluid Flow vol 20 pp 581 591 9 5 Whitaker S 1999 The Method of Volume
518. ssed in an integral form of the general conservation law Its semi discrete form is obtained after the approximation of surface and volume integrals by using the values of integrand that prevail at the face and cell centroids respectively 14 1 1 Generic transport equation Equations presented in section 10 1 can be written in the generic integral form nf DAD a mid e rece Lda fs avb f sAdAy 14 1 dt V JA lt A J J A Ems Transient Y V Convection Dif fusion Sources Each of the terms in Equation 14 1 T SG and S are given appropriate values according to the type of the transport equation being solver for 234 14 NUMERICAL SOLUTION 14 1 FINITE VOLUME DISCRETISATION Similarly the continuity equation is written in the following integral form d y a ev de p Ux Ung dAy 0 14 2 V In order to arrive at a set of closed set of modelling equations the space conservation law which is needed for problems with moving boundaries In the integral form this equation is written as d of ag f Und 6 143 L n bed Ax 14 3 An analytical solution of the governing equations casted in the integral form does not exist ex cept in the most simplistic form Most real world problems require the solution of the governing equations presented A numerical method is therefore used to convert the system of governing equations into a system of algebraic equations that can be solved for This process of conversion is known as
519. ssors 4 Metis Partitioning Method K way _ Repartition Restart Files X Arbitrary Grid Interface Arbitrary Grid Interface Output Filename DIATION_VECTIS4 STAR_TEST COALESCED GRD O Number of Additional Mesh Files 2 Additional Mesh Files gt 4 STAR_TEST solid1 GRL A CTIS4 STAR_TEST fluid2 o Mesh Join Type v Refrain From Merging Stitched Triangles Launch Cancel Figure 17 21 Vpre launcher dialog box JOIN TYPE COMMAND JTYPE NO Conformal 0 No extrusion with meshing 1 Extrusion with meshing 2 Table 17 7 Mesh join types and corresponding vpre jtype option numbers The vsolve gt launch dialog Fig 17 22 allows the user to set the input file number of processor to run on and the hosts set to localhost if local 17 16 WAVE VECTIS Co simulation An important capability of VECTIS is the ability to couple at a time step level with the Ricardo one dimensional engine performance simulation code WAVE WAVE VECTIS coupling allows one or more parts of a WAVE network to be replaced by a VECTIS model This can be useful for system ORicardo Software December 2009 293 17 USING SOLVER 17 16 WAVE VECTIS CO SIMULATION R Launch Solver Dialog i o x Number of Processors 4 Run Hosts comp comp2 Figure 17 22 Vsolve launcher dialog box components
520. stant volume porosity y the local thermal equilibrium T T and heat flux continuity across the fluid solid interface area i e 1 OT 1 OT 1 A A da dA h Aj T T 0 11 38 am 4 dx gt V Ja dx V T T hi is the interfacial heat transfer coefficient the energy conservation for both fluid and solid medium can be derived as follows d _dyp 0 sa Or e by Ok t ypH 1 Y PsHs F Ox ypHU Ot dx iy ag ERA U Tij 1 OK dx y 2 UJU YP fiUi 109 1 Y Pss v 11 39 where the effective porous conductivity Aie gt is assembled as ap a4 ta he 8 TAGS 4 ator 11 40 jk y s t Ojk jk por jk and the porous tortuosity conductivity Aor je is introduced to model the tortuosity molecular heat flux tor YT 1 a y J AAs Fnjda por jk Oxy 11 41 o Macroscopic turbulent kinetic energy If volume averaging is applied to the ensemble Favre averaged turbulent kinetic energy Equa tion 9 19 the intrinsic or volume average of k is given as 1 1 ay limra Pawar k k ul km WN km uy ur 11 42 2 2 l l 2 1 1 2 1 1 The modelled transport equation for k can be then derived as see Nakayama and Kuwahara 1999 Pedras and de Lemos 2001 by applying the volume average operator to Equa tion 9 19 Ok He 2 S 1143 a a a 57 YPK 5 YPKU Y Pe Po PE 5 y u Te oe Xj Xj Ok Ricardo Software
521. start With Pre vious Results toggle must be selected By default this will use the most recent restart file written for the current project Additionally a filename can be explicitly specified to restart the calculation from an alternate rst file Once the Set Filename Toggle is selected a restart file name can be entered into the Filename box Timed restarts can also be specified to allow the writing of restart files at specific times with specific names A time value in terms of iteration or timestep depending on whether the calculation is steady or unsteady and a filename to define a restart specification This pipe flow example uses the default options 19 1 8 2 Fluid Domain In this case the calculation will consist of one single phase fluid domain Domain Name A name can be assigned to the fluid domain to allow for easier reference later in the calculation Ricardo Software December 2009 377 19 TUTORIALS 19 1 BASIC TUTORIAL Restart Reading TF Restart With Previous Results Restart Writing Restart File Frequency 400 Figure 19 14 The Restart Panel Material Name A name can be assigned to each material in the fluid domain again to allow for easier reference later in the calculation Each fluid domain contains only one material Material ID This refers to the material ID number referenced in the GRD file Each separate material will have a unique ID number Multiphase Modelling This allows the user t
522. steady separation b unsteady separation vortex shedding EVM models used in conjunction with the wall functions cannot simulate the transition to tur bulence in the laminar boundary layers and or predict accurately the points of flow separation and re attachment There is limited scope to improve the predictions in the stagnation regions by using either the RNG or Standard k model with the realisable turbulence time scale bound TSB Further improvements can be expected if the low Reynolds number variants of the standard and TSB model are employed on the refined grids Compared to the standard k model the RNG based model improves predictions of separated flows in particular the length of recirculating regions However it might spoil predictions of accelerating flows It generally performs better in the flow regions where the normal stresses govern the production of turbulence In these regions usually associated with the stagnation point flow the standard k model grossly over predicts the turbulent kinetic energy and thus the level of turbulent viscosity Transported downstream this excessive level of k can seriously affect the accuracy of model predictions Typical examples include the flow around a circular or square cylinder and impinging jet flow The k model with the realisable turbulence time scale bound TSB prevents excessive levels of the turbulent kinetic energy in the stagnation flow regions by
523. super patches view factor summation is less than 1 the patch is not only losing energy by radiating to the ambient but Ricardo Software December 2009 258 15 MODELLING RADIATION 15 3 RADIATION SETUP 7 Radiation Modelling Surface to Surface y Radiation File Filename heat_shield rad 7 Start 0 Frequency 1 Surface to Surface PAT File shield pat VFM File shield v m CON File shield con Reflectivity Yes X Transient Problem No 4 Ambient Temperature 300 00000000 Max Iterations for Linear Solver 100 Max Iterations for Non Linear Solver 40 JOR Relative Tolerance for Linear Solver 0 0001 Relative Tolerance for Non Linear Solver 1e 08 Figure 15 2 R Desk radiation setup panel it also receives some energy back from the background This default ambient temperature is used for all radiating boundaries unless a boundary has its own ambient temperature specified in the Radiation Boundary panel Solver control parameters The 4 remaining parameters control the radiation solver radsolv and are fairly self explanatory The default values chosen should normally be fine Boundary Specific Surface to surface Parameters The selection of the participating boundaries is done from within the boundary panels as shown in Figs 15 3 15 4 E Bn
524. surface There are usually no adverse effects resulting from this but a very large number of these warnings may indicate a poorly defined CONS curve in the VDA Ricardo Software December 2009 74 3 GEOMETRY 3 18 WARNING AND ERROR MESSAGES file WARNING 0202 VECTIS PHASE1 NO TRIANGULATION FOUND FOR POLYGON ON FACE facename The program has not been able to triangulate an individual polygon on a trimmed surface This may result in holes in the triangulated surface It may indicate a self intersecting trim curve or other poorly defined trim curve in the VDA file WARNING 0203 VECTIS PHASE1 NO VDA HEADER BLOCK IN FILE filename EXITING ROUTINE The file is not a valid VDA file as it does not contain a correct VDA header block WARNING 0204 VECTIS PHASE1 INCOMPLETE CONS LOOP CLOSED ON FACE facename The VDA standard requires that a set of CONS curves form a closed loop in order to define a FACE a trimmed surface The program has detected a non closed loop but has determined that the loop can be closed by joining its end points This warning indicates a file which does not conform to the VDA standard WARNINGS 0401 0413 Various messages These warnings indicate errors in the mesh coordinate section of the mesh input file The action resulting from these errors is to produce a uniformly spaced mesh with the specified numbers of mesh lines in the i j and k directions WARNING 0414 VECTIS PHASE1 NUMBER OF
525. t file via the Report Frequency LineEdit Report Level controls amount of information contained therein Co simulation Post processing LineEdit is used to control the output frequency of variables when VECTIS is coupled with WAVE see Sec 17 16 Save Initial Conditions CheckBoxes are used save the initial state of the system to the Restart Ricardo Software December 2009 279 17 USING SOLVER 17 3 RESTART CONTROL 0 Off 1 Report CPU timings 2 Report boundary fluxes for mass and energy Table 17 2 Report levels for Summary Frequency GroupBox and or Post file The initial conditions are saved after any flow potential calculations are per formed but before the first time step iteration The Linear Equations Solver Residuals Information provides extra residual information for each equation and each inner iteration of the solver This will write to the screen and output files in the format shown below Domain 1 FI U RESO 2 466729E 00 Domain 1 FI U ITER 1 RES 3 730890E 03 Domain 1 FI V RESO 5 618043E 01 Domain 1 FI V ITER 1 RES 5 258057E 03 Domain 1 FI W RESO 6 936036E 02 Domain 1 FI W ITER 1 RES 2 324654E 03 Domain 1 FI P RESO 8 562884E 03 Domain 1 FI P ITER 1 RES 8 590540E 01 Domain 1 FI P TER 2 RES 8 876634E 01 Domain 1 FI P TER 3 RES 9 081014E 01 17 3 Restart Control During the calculation restart files are written at a user defined frequenc
526. t gt real optional scalar boundary values fi_ui real optional scalar upper interface values fi_li real optional scalar lower interface values Table 18 35 Subroutine upr_properties to modify properties The list of changeable properties pro_name are shown below The routine upr_properties is called for each material phase and species The availability of a given property can be tested for using the Fortran 95 2003 present command E g Ricardo Software December 2009 348 18 USER PROGRAMMING 18 4 USER PROGRAMMABLE ROUTINES Cell Boundary Interface Solid Species Phase density x x x x x Xx conductivity x x x x x specific_heat x x x x xX gas_constant x x Xx lam_viscosity x x x lam_mass_diffusion x x x Table 18 36 Changeable properties and types available if prop_name was conductivity then present fi_b would return false Property values are typically modified by looping over the extents passed down e g in Fortran 95 2003 do ic icell icel2 fi_c ic lt some value gt end do As this routine modifies properties of pure substances which in general depend on the temper ature and pressure the user should use the get_field routine to acquire the temperature and or pressure field If the argument isp is non zero then the property for that species can be changed otherwise only the phase property can be modified See Section 18 7 1 for example
527. t Control Timebase Discretse E Phase_1 Fluid Phase Tritial Condition Postprocessing Output Equatons amp Solver Fan Model El Boundary Regions Bnd_Reg_1 Wall Boundary EJ Bnd_Reg_2 Inlet Given Velocity Boundary Phase_1 Boundary Phase J Bnd_Reg_3 Inlet Given Veloaty Boundary Phase_1 Boundary Phase Figure 19 10 The solver setup input tree 19 1 8 1 Solution Control The Solution control panel allows general simulation parameters to be set Domain Name A name can be given to each global domain in the calculation For example Coolant or Cylinder Head Ricardo Software December 2009 374 19 TUTORIALS 19 1 BASIC TUTORIAL Global_domain_1 Global Domain 2 a x Domain Name Global_domain_1 Input Mesh Filename Set Filename tube GRD w Convergence Criterion 1e 05 Figure 19 11 The Global Domain Panel Input Mesh Filename The solver will default to using a computational grid file with the name pro jectname GRD A different grid file can be specified here otherwise the projectname is taken as the inp file prefix Timebase 215 r Time Mode C sy C Unsteady Time Scheme Steady y Time Base Steady Number Of Iterations 10 Figure 19 12 The Timebase Panel Time Mode Here the type of calculation time mode is selected 1 Steady A maximum number of iterations is chosen for the calculation If all the residuals for all the equations are less than the s
528. t and mass transfer than to momentum transfer In some situations a simple empirical approach where k and are defined algebraically is more appropriate instead of solving turbulence transport equations This approach requires knowledge of porous media turbulence intensity por and length scale lpor 3 4 3 2 Cu k Cage Lt k i Tor 11 54 N l por Appropriate values for por and lpor are problem dependent typical values might be 0 01 and 0 1 x characteristic pore dimension Ricardo Software December 2009 190 11 MODELLING POROUS MEDIA 11 3 POROUS MODEL TYPES AND USER INPUTS 11 3 Porous Model Types and User Inputs The generalised flow resistance Darcy Forchheimer s model given by Equation 11 21 is ap plicable to non isotropic porous media and includes the viscous Rij and inertial tensor Te These resistance tensors depend on the type of porous structure and have to be specified by the user Porous model types commonly used in automotive industry and required user s inputs are described next A porous media region has to be associated with the fluid sub domain an example is shown earlier in Figure 7 4 The Figure 11 1 shows R Desk Sub domain panel where Subdomain Name Subdomain Type and Turbulence Model can be edited or selected Either Standard or Heat Exchanger or Porous type can be associated with the given fluid sub domain The standard sub domain contains the same continuum as the parent flui
529. t calculations this can be set to either 1st Order Euler or 2nd Order Implicit Time Base For transient calculations the following timebases are available 1 2 Stroke 2 Seconds 3 4 Stroke 4 Non Dimensional Time Dependant Boundary Conditions The boundary conditions can be steady or unsteady transient which is determined by selecting the time dependent boundary conditions toggle Timesteps The number of timesteps is chosen for the calculation the end time being the chosen time step multiplied by the number of timesteps specified Additionally a Maximum Number of Outer Iterations is chosen This is the maximum number of iterations performed per time step Typically the number of iterations required per step will reduce after the initial period of the calculations However this will be dependent on the accuracy of the applied initial conditions r Frequencies For Printing Data into Project Files Output File Outer Iterations 1 Output File Time Steps 10 Post processing File Frequency iterations S00 Co simulation Post processing File Frequency Iterations o Report Frequency Asai 1 SDF 1 IV Save Residuals To Ascii Save Residuals To SDF r Summary Frequency Report Frequency 100 Report Level Report Boundary Fluxes For Mass And Energy T Save Initial Conditions Restart File I Save Initial Conditions Post File TF Linear Equations Solver Residuals Information
530. t_reg to get integer variables defined in the access group iacc_region 18 3 2 8 get_property This UAR is used to retrieve information related to phase species and passive scalar material properties and reference values Table 18 16 is split in two sections and explained here a b Ricardo Software December 2009 318 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES a To get the list of phase properties calculation options then do the following integer iwp pointer ph_pro_opts Call get_property 1l_phase_opts ph_pro_opts Array ph_pro_opts 1 nproph 1 n_phases now contains all calculation option for phase properties For example ph_pro_opts 1 2 represents a density option for phase 2 The next list specifies parameters used to define fluid properties which are required when solving transport equations of either individual phases or the mixture of phases 66 ces cc ccs cc cos O idens 1 ren Calculation options ipro_const ipro_mix ipro_igas ipro_poly ce cos ipro_user ipro_bouss cc cor cc cor O ivis 2 laminar viscosity Calculation ge ipro_const ipro_mix ipro_poly ce cos ce ces ce cos ipro_power ipro_suth ipro_user ipro_invis ipro_nonwt O icon 3 thermal apo if the energy equation is solved Calculation options cc cos cc cos ce ces ipro_const ipro_mix ipro_pol
531. tal problem with IN OUT test There is a problem with in out test The input geom etry is probably not clean enough ERROR 1003 Fatal problem with IN OUT test There is a problem with in out test The input geom etry is probably not clean enough ERROR 1004 A flaw in triangle topology detected check triangle n Unexpected topological situa tion near the triangle n was detected Check the geometry around the triangle Ricardo Software December 2009 95 4 MESHING 4 9 WARNINGS AND ERRORS ERROR 1005 The triangulated surface cannot be harmonized There must be a flaw causing Mo bius strip effect near the triangle n Unexpected topological situation near the triangle n was detected Check the geometry around the triangle ERROR 1200 Exact fit routine not enough memory ERROR 1201 Marching Cubes routine zero or even number of intersections Marching Cubes rou tine encountered the situation when there is zero or even number of intersections between two points with different in out statuses From the logic of the matter the number of intersections here must be always even There is probably a serious error in the input geometry ERROR 1202 The file filename cannot be opened ERROR 1203 Problem with syntax found when processing the keyword keyword The keyword is known however its additional parameters cannot be properly read The syntax of the keyword needs to be checked ERROR 1204 Unknown keyword keyword ERROR 1205 Rout
532. tant thermal_exp_coeff lam_mass_diffusion thermal_diff_coeff enthalpy_form_heat en thalpy_form_temp turb_schmidt_numb where phase can have the first 6 property names In case of passive scalars then the following list is obtained density lam_mass_diffusion and turb_schmidt_numb 18 3 2 14 get_name_list This UAR can be used to obtain a list of names for each VECTIS MAX object see Table 18 22 The list of material names is obtained as character len pointer mat_names call get_name_list mat_names iget_mat where character array mat_names now contains a list of all material names A similar way is used for all other iget options in Table 18 22 Ricardo Software December 2009 329 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES subroutine get_name idt cname iget Input Output idt range integer iget integer cname cha 1 n_domains iget_dom domain name 1 n_mat_doms id iget_mat material name n_phases 1 iget_phase phase name 1 n_species iget_specs species name l n_ps iget_ps passive scalar name 1 n_bnd_regs iget_breg boundary region name 1 n_interf_regs iget_ireg interface region name 1 nte iget_eq equation name 1 nbtype iget_bctype name of boundary condition type 1 nproph iget_ph_pro name of phase property 1 nprosp iget_sp_pro name of species property 1 nprops iget_pS_pro name of passive scalar property Table 18 21 Subroutine get_name to get VECTIS
533. te Boundary Cut Chop Boundary Unchop Boundary Boundary pop up menu The Boundary Pop up menu allows the user to add and delete boundaries cut and paste boundaries into other parts and chop and unchop boundaries Ricardo Software December 2009 53 3 GEOMETRY 3 15 BOUNDARY PROCESSING Add Delete Boundary The Add Boundary function adds a boundary definition without assigning triangles to it and assigns it a default boundary name in the form boundary__ If the Delete Boundary function 1s selected the user is prompted to delete both the boundary and the triangles on the boundary or to delete just the triangles on the selected boundary Cut Boundary Selecting Cut selects a boundary to cut which can subsequently be pasted into another part using the part pop up menu Chop Unchop Boundary Selecting Chop Boundary and Unchop Boundary makes the boundary Active or Inactive respec tively and is the same as switching a boundary on or off by using the Show column in the bound ary panel An inactive boundary is not displayed on the screen and so boundaries may be quickly removed to gain an unobstructed view of hidden parts of the model Ricardo Software December 2009 54 3 GEOMETRY 3 15 BOUNDARY PROCESSING 3 15 2 Boundary Processing Parts Boundaries Show Tri Type Refine Name Comment 1 Yes 2 2149 Wall No piston bowl No E 3 4024 Vall No inlet valve No a 4 14218 Wall No cylinde
534. th the continuity and momentum equations On non staggered grids a special interpolation practice is required for the mass conserving face velocity U in Equation 14 6 In what follows we consider the pressure correction method which provides velocity pressure and density coupling for flows at all speeds 14 2 1 Pressure correction scheme The SIMPLE algorithm effectively couples the velocity and pressure fields by converting a discrete form of the continuity equation 14 7 into an equation for the pressure correction The pressure corrections are then used to update the pressure and velocity fields so that the velocity components obtained from the solution of momentum equations satisfy the continuity equation 14 2 2 Face velocity for mass conservation On non staggered grids the use of linear interpolation to obtain the pressure and velocities at the cell faces usually leads to de coupling of two fields In order to overcome this problem a special interpolation practiceRhie and Chow 1983 Ferziger and Peric 1997 is required for the mass conserving face velocity in Equation 14 6 The discretized momentum equations can be re written in a vector form as a a aU AS Up hp eae hp pe eth ad 14 49 ap ap where the source g does not include the pressure term A similar equation for the CV around the cell face centre j reads gt gt V U hj VP 14 50 ap J If the variation of the veloci
535. that can be produced by vmesh and any other mesher In case of vmesh the low quality cells are typ ically produced when the insufficient mesh refinement can not resolve fine geometry features Ricardo Software December 2009 a 2 INTRODUCTION 2 1 MAIN FEATURES AND CAPABILITIES Conjugate Heat Transfer CHT Novel and efficient techniques are introduced for the implicit solution of the energy equation The CHT simulations require conformal grids at fluid solid or solid solid interfaces Parallelisation Scalable parallelisation of VECTIS MAX is based on the domain decomposition strategy and MPI based message passing Domain decomposition is performed using METIS to decompose the mesh into partitions taking into account load balancing for each material for maximum efficiency The number of partitions used for the solution can be changed during a calculation by utilising the re partitioning of restart files feature of the vpre module Boundary and Interface Condition Types O Wall and flow boundaries In addition to walls and symmetry planes various flow boundary condition types can be specified as fixed static average or total pressure at inlet or outlet inlet velocity and mass flow rate Inlet stagnation conditions are also available along with an outlet boundary condition O Wall thermal conditions For the solution of the energy equation various thermal conditions can be specified the wall temperature or wall heat flux
536. the Joolbars gt Mesh Setup option in the Phasel menu The tools are displayed and explained in Fig 19 4 A default outline mesh is displayed which can be manipulated to provide the required control mesh Here a global mesh size of approximately 2 5 mm is used It is important to remember that the mesh needs to be defined in all three directions so at least two of the x y x z y z views need to be used Once the global mesh is defined save the mesh File gt Save mesh as as tube mesh Y Ricardo VECTIS Phase 1 tube tri Pa File Edit View Toolbars Operations Help a a a e Paseo NE al El Y Options Stitch Mes Rg Setup View View Style OOD Mesh Lines el a al El EE Information El IJK Refinement m njn Add Delete j Refinement Parameters DEEP o FORCE io pal a 14 40 24 abet dimensions xmin 0 040000 xmax 0 010000 ymin 0 025000 ymax 0 005000 _ zmin 0 000000 zmax 0 100000 gt Figure 19 4 Global Mesh Definition 19 1 5 R Desk R Desk GUI Ricardo Software December 2009 370 19 TUTORIALS 19 1 BASIC TUTORIAL Set the mesh view to the X Y plane Set the mesh view to the X Z plane Set the mesh view to the Y Z plane Set the mesh view to be 3D so both mesh and geometry cabn be rotated Move ared mesh control line by selecting this icon and then the red mesh line Add a verticalred control me
537. the Mark Face command Example of Line and Face Marking Mark Area Button ORicardo Software December 2009 32 3 GEOMETRY 3 10 TRIANGLE MARKING OPERATIONS This operation is used to select a part of the model to which further commands will be applied Triangles are selected by describing the perimeter of a polygon which is drawn about the model by clicking the left mouse button Two clicks in the same place will end the drawing process and close the polygon The triangles inside the polygon will then be marked to which further commands can be applied Mark All Button de This operation is used to mark all of the model Delete Marked Area Button This operation works on the faces defined using the marking functions described above When the left button is clicked over a triangle that triangle and all other triangles upon the same face are deleted Keep Marked Area Button El This function is the inverse of the previous one in that only triangles on the selected face remain all other triangles are deleted Swap marked and Unmarked Button This swaps over the faces that have been marked All unmarked triangles become marked and the previously marked triangles are cleared Clear Marked Area Button Ed 1 Pressing this button on the toolbar clears any marked faces from the model The triangles return to their previous colour Ricardo Software December 2009 33 3 GEOMETRY 3 11 TRIANGLE CHOP
538. the boundary that the selected triangle belongs to Chop Boundary by Number Button El Selecting this command pops up the Chop Boundary by Number panel Ku Chop Boundary Ea Chop Boundary Enter Boundary Number fi Apply Cancel ly Lie Specifying the boundary number and selecting OK makes all the triangles invisible on the specified numbered boundary Unchop Selected Boundary Button ES This makes all the triangles visible on the boundary that the selected triangle belongs to Unchop Boundary by Number Button aj Selecting this command pops up the Unchop Boundary by Number panel Ricardo Software December 2009 35 3 GEOMETRY 3 11 TRIANGLE CHOPPING OPERATIONS Yy Unchop Boundary X Unchop Boundary Enter Boundary Number 1 Apply Cancel p Lie Specifying the boundary number and selecting OK makes all the triangles visible on the specified numbered boundary Grow Active Set Button EX This operation allows the user to grow the chopped model by a specified number of triangles A panel is popped up in which the user enters the number of triangles to grow the set by and inactive triangles adjacent to the active ones are made active up to the specified depth Only triangles on active boundaries are made visible Triangles on inactive boundaries remain invisible Grow Inactive Set Button fed This operation allows the user to grow the chopped model by a specified number of triangles
539. the section about setting fluid phase properties The effect of enthalpy transport due to species turbulent diffusion represented by the correlation de hy in Equation 6 31 can be modelled in a similar way as the effect due to species laminar diffusion In Equation 10 7 these effects are taken into account for each species k The total enthalpy H takes the usual form N U U aE H h k h Y chy 10 9 2 k where hz denotes specific thermal enthalpy of the k th species and A is the thermal enthalpy of the fluid phase For thermally perfect fluids the thermal enthalpy and internal energy are given by Equations 8 25 and 8 24 respectively Their evaluation requires specific heat for which various calculation options have been presented in the section setting fluid phase properties Depending on the selected physical model the R Desk Equations Solver panel will show which transport equations are to be solved Figure 10 1 The momentum and mass equations are always Equations Solver a E Momentum y Mass y Energy v Species y Turbulent Energy y Turbulent Dissipation Turbulent Viscosity Volume Fraction Potential Passive Scalar Figure 10 1 R Desk setup List of equations solved for in VECTIS MAX activated as well as the energy equation In case of incompressible fluid with constant physical properties the energy equation is decoupled from others and the user might decide not to solve it Apart from the
540. the single phase single or multi component O Fluid flow and associated O Heat transfer and O Mass transfer All the models are described by corresponding governing equations of resolved flow which ought to be closed by an Equation of state model equation The flow resolved equations are presented first followed by the description of physical models and their selection by the user 10 1 Governing Equations of Resolved Flow Instantaneous equations for the mass momentum and energy conservation describe both lami nar and turbulent flows The main strategies to deal with the problem of turbulence namely DNS RANS LES and hybrid LES RANS have been outlined in the introduction to turbulence modelling section It can be noted that the above strategies result with an identical form of the governing equa tions However the meaning of flow variables is different the variables represent instantaneous quantities in DNS they are time or ensemble averaged quantities in RANS and filtered quantities in LES As a common feature of these strategies is to resolve unsteady turbulent motion in time and space the resulting conservation equations are known as the mean or resolved flow equations The governing single phase resolved flow equations have the same form as the Reynolds averaged equations They read as follows O Mass conservation J J p gt p U Ug sm 10 1 The source term Sm can be for example the mass source
541. the solver setup tree with the correct number of materials and boundaries A GRD file needs to be entered into the Filename box Clicking on the browse button will open a dialog box and allow the relevant grid file to be selected Once a file has been selected The name will appear in the Filename box Next click the extract button A dialog box pops up and displays the materials found in the imported grid file Then each material in the grid file needs to be allocated into a fluid or a material in a solid domain R Desk 10lxi Fie Edit View Options Window Help DEBES PENEITIA NEH EA gt a 9 gt e a erre vc E Solver Setup S Global_domain_1 Global Domain Extract Materials And Boundaries From Mesh File Restart Control Timebase Output Radation tube GRD Browse Fluid_1 Fluid Domain Monitoring Points Algorithm Turbulence Model Discretse Phase_1 Fluid Phase Tritial Condition Postprocessing Output Equatons amp Solver Fan Model Boundary Regions Bnd_Reg_1 W Filename ning ne D RPe traning_tutorials V4 bast e GRD Closing file D RPe training_tutorials V4 bash e GRD Opening fle D RPe traning_tutoriais V4 basic_tutonial tub J e GRD xl Figure 19 8 Importing the mesh into R Desk Solver Setup By importing the grid file the relevant regions of the grid are highlighted when they are selected in the Solver setup tree For example in 19 9 the first boundary region is s
542. the star units 1 4 U _ pUlk y pUw _ pci k 2y U U 3 Uk Tw u p 9 26 Numerous experimental and DNS data as well as dimensional analysis suggest that the inner part of the near wall region comprises three layers Viscous sub layer the flow is dominated by the viscous force represented by viscosity Buffer layer Both molecular and turbulence effects are important Fully turbulent or logarithmic law layer Turbulence inertial force plays a dominant role As Figure 9 1 shows the distribution of the non dimensional velocity parallel to the wall Ut f y is governed by distinctive laws throughout the viscous and fully turbulent layers In the viscous sub layer the velocity obeys the linear law while the so called universal velocity profile or log law describes the fully turbulent layer ite l ys viscous sub layer y lt 5 ylin Ey log law y gt 30 50 9 27 where k 0 41 is the von Karman constant while E is another empirical constant whose value depends on the wall roughness for smooth walls E 9 Unified law 24 4 Tr TT TTITT T TT TTTITI T T TTT T T TTT C DNS data H o Channel Re 180 20 a E a Channel Re 395 fe o Channel Re 590 16 T A Boundary layer Re 300 C v Boundary layer Re 670 F 5 7 12 E Linear ae y gt t Log law U In Ey 8 4 0 0 1 1 10 100 y Figure 9 1 Velocity distribution in the inner wall region
543. the superficial volume average and intrinsic volume average operators can be introduced as Ont F Art Y A t aV 11 1 9 est v 80 0 Ar 0d 11 2 where an averaging volume V Vf V consists of the fluid Vy and solid V parts The position vector Ax of any point inside the volume is defined relative to the centre of the volume while the vector x denotes the position vector of the volume centre A fluid phase distribution function yp defined as _ Jf 1 5Ax EV n 0 ifAx eV ia identifies the fluid volume The superficial and intrinsic averages are related through the volume porosity Y s s_ Vf C2 10 a 11 4 Similarly to the Reynolds and Favre decomposition Equations 6 17 and 6 22 respectively the variable can be split into its intrinsic volume average and the spatial deviation of from the intrinsic average p with 0 11 5 The usual rules for a sum and a product of two variables apply Oty 9 W OY 9 W OH 0 9 0 11 6 Ricardo Software December 2009 183 11 MODELLING POROUS MEDIA 11 1 THEORETICAL BACKGROUND In order to derive the volume averaged RANS equations it is necessary to recall two important the orems which relate volume averages of derivatives to derivatives of averages The local averaging theorem see for example Whitaker 1999 is given as dP s IV do ayo om OX f oma ot de e Ox 5 ona oe where A represents th
544. ther species of the considered phase iph By visiting all fluid materials material 1 n_fluid_mats Le all fluid phases iph 1 n_fluid_ phases there will be nspreg addresses This number is the sum of species regions over all fluid domains where the number of species regions within each domain is the product of the number of boundary regions and the number of domain species To obtain starting and ending addresses for boundary region wise values of species variables do the following integer iwp pointer il 12 call get_species ise_specs_bnd_reg i1 i2 subroutine get_species var_name il i2 Input Output var_name cha i1 integer pointer optional i2 integer pointer optional ise_specs_bnd_reg index of starting boundary region index of ending boundary region 11 1 n_species 12 1 n_species isa_specs_bnd_reg starting allocation addresses for species at boundary regions 11 1 n_species ise_specs_interf_ index of starting interface region index of ending interface region reg 11 1 n_species 12 1 n_species isa_specs_interf_ starting allocation addresses for reg species at interface regions 11 1 n_species Table 18 12 Subroutine get_species to get values for variables defined in the access group iacc_ species mainly start end boundary interface regions for species Ricardo Software December 2009 313 18 USER PROGRAMMING 18 3 2 6 g
545. tices Check the triangulated surface in the global box I J K ERROR 1525 The boundary face X has less than three vertices X vertices Check the triangulated surface in the global box I J K ERROR 1526 The IP face X has less than three vertices X vertices Check the triangulated surface in the global box 1 J K ERROR 1527 The cell X has less than four faces X faces Check the triangulated surface in the global box I J K ERROR 1528 Incorrect patches were generated in box X Check input geometry near triangles X and X Check the triangulated surface in the global box I J K WARNING 1050 IN OUT test seems to be uncertain the confidence is percent The in out test has been evaluated however not all released testing rays had shown the same result WARNING 1051 IN OUT test of tested side seems to be uncertain the confidence is percent The in out test has been evaluated however not all released testing rays had shown the same result WARNING 1052 IN OUT test seems to be uncertain The in out test has been evaluated however not all released testing rays had shown the same result WARNING 1053 IN OUT test of tested side seems to be uncertain the confidence is percent The in out test has been evaluated however not all released testing rays had shown the same result WARNING 1800 N of overlapped triangles were detected This warning occurs when the rapid test of overlapping triangles finds any overlap For details see sec
546. time step sizes more time steps are required but with fewer iterations per time step Obviously there is a set of the optimal values for the time step size under elaxation factors and for a number of iteration per time step The user is advised to start with a small number of iteration per time step 5 10 10 3 8 Gravity buoyancy driven flows Various body forces per mass unit are given in Equation 10 4 and these forces appear as volume sources in the momentum equation Equation 10 2 Currently VECTIS MAX provides modelling of the gravity force effects The gravity force per unit volume F Pg 10 19 always plays an important role in free surface and natural convection flows The latter are flows driven by thermal or concentration buoyancy where gravity acts on the variable density field In mixed convection flows the gravity force should not be neglected if the ratio of Grashof Gr and Reynolds Re numbers Gr 8 BAT ve pLre f Re 7 Use f 10 20 is close to or above one In the above equation f represents the coefficient of volumetric expansion Ricardo Software December 2009 175 10 MODELLING SINGLE PHASE FLOWS 10 3 MODELLING FLUID FLOW and AT ef Lref and U er are reference temperature difference length and velocity respectively Another dimensionless number Rayleigh number Ra El BAT eLo P cp LA indicates the strength of the buoyancy induced flow in pure natural convection The flow is consid
547. ting and converting external mesh files Solver Setup A controlling inp input file needs to be defined for the CFD solver vsolve This is created in R Desk using the solver setup panels The main parameters specified by the input file include Selected time base to be steady or unsteady O Time step and convergence criteria Specify monitoring and reference locations O Frequency of restart and post processing file dumps O Restart options such as whether to start from an existing restart file or not O Fluid solid property data Equations to be solved Numerical schemes used to solve the equations O Boundary conditions Walls Inlet Outlets O Initial conditions CFD Solver Vsolve is the solving stage of VECTIS this is the phase that actually performs the computational fluid dynamics calculation The CFD solver uses the generated computational grid file and the input file Post processing The results produced by the CFD computation and written to the post file can be viewed in the R Desk post processor This allows the simulation data to be visualised manipulated and extracted for further use Ricardo Software December 2009 10 GEOMETRY 3 1 Introduction Phase 1 has a number of main functions 1 To read in geometry from an external source and convert it to a form acceptable to the VECTIS mesh generator 2 To identify regions of the geometry as differe
548. tio sets the ratio SMALL_CELL_VOLUME BOX_VOLUME which defines the maximum volume of cell which should be considered as small for all cells that contains patches from boundary bouinx Option smallio has higher priority in the case when it can also be used for the cell small ratio sets the ratio SMALL_CELL_VOLUME BOX_VOLUME which defines the maxi mum volume of cell which should be considered as small for all other cells where smallio or smallb is not defined the default value is 1 0 x 1073 SAFE POSITIONS OF MESHLINES noduallevels switches off the technique of dual levels when meshlines lie on different discreete levels than vertices of triangles this technique ensures that there is no collision between sides of boxes and triangles when dual levels are switched off moving of meshlines to safe positions is automatically switched on mvfrac fraction sets fraction which defines the tolerance for moving of meshlines out from danger ous positions the tolerance is calculated as dist x fraction where dist represents the distance of the tested meshline from the closer of the two neighbouring meshlines this option can be used only with noduallevels POLYGON SIMPLIFICATION PARAMETERS Ricardo Software December 2009 8l 4 MESHING 4 4 COMMAND LINE OPTIONS ps_sfang degangle sets the minimal angle from which the edge between two polygons is considered as sharp feature in polygon simplification Default value is 35 0 degrees Wh
549. tion x_mass rgmax 10 a 1 rga 12 45 The ratio density ry is given by Pg rg 12 46 j Pl Ricardo Software December 2009 205 12 MODELLING MULTIPHASE FLOWS 12 4 PHASE CHANGE MODELLING 12 4 1 2 RPI boiling model The RPI model is mainly used in Euler Euler multi phase flows but was adopted for use in homoge nous multi phase mixture flows In the RPI model the vapour bubble diameter and quenching heat flux are taken into account The mass flow from the wall is calculated directly from the nucleate boiling heat flux O Volume of Fraction Equation The mass conservation equation 12 2 is used in its exact form O Evaporation and Condensation in the bulk The evaporation condensation rate in the bulk flow is given as _ hiAi T Tsar pr 12 47 hfg where A is the interfacial area density and is calculated as lsg 1 0 1 A 6 12 48 1 isg dpuk where lsg min 0 25 according to Kurul and Podowski 1990 and d 1x is the bulk bubble diameter and is calculated as a function of local sub cooling as follows Kurul and Podowski 1990 1 51074 AT gt 13 5K der lt 1 51073 10 ATap 0 lt AT up lt 13 5K 1 5107 ATO The heat transfer coefficient h in relation 12 47 is expressed as _ N uA A bulk hi 12 49 The Nusselt number Nu is calculated according to Ranz and Marshall 1952 correlation as Nu 2 0 6Re2Pri 12 50 The Reynolds and Prandtl numbers are evaluated a
550. tion from within the SIMPLE segregated solver and it is called for each transport equation which is being solved by VECTIS MAX 18 3 Accessing Solver Variables The following sections contain a brief description of available variables that a user can access The user may access through UPRs User Accessible Variables UAV and User Accessible Routines UAR 18 3 1 User accessible variables User accessible variables UAV are those variables that a user can access directly from the solver by specifying its identifier name Most of these variables are used to identify certain property types of the VECTIS MAX solver For example the UAV with the name iget_cell is used to identify cell values Each of these identifiers is assigned an integer value to make the logic simpler The user does not need to know its integer values The purpose of each UAV will be much clearer through UPR examples provided A complete and consistent list of all UAV is provided in this section in the form of tables Each UAV is described in terms of its identifier name description and the type integer real character cha A similar approach is used in UARs but in addition to the type of argument a keyword pointer is used where appropriate to indicate a pointer variable Ricardo Software December 2009 298 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES Precision format Table 18 1 defines the numeric precision used in VECTIS MAX Both integer and r
551. tions 4 5 and 4 7 WARNING 1801 Not all faulty triangles have been repaired The routine for automatic reparation of triangles was not able to repair all problems WARNING 1802 Boundary refinement information is set in trifile and also in meshfile It is not clear which type of definition of boundary refinement should be read because both types were used The boundary refinement which was set in the trifile will be ignored in this case The user should check what was his intent For more information about setting up the refinement see section 4 6 WARNING 1850 There are no patches on the first geometry that can be matched to non conformal patches of the second geometry please check the list of complementary boundaries It seems that lists of boundaries that form the interface between two complementary boundaries are not correct WARNING 1851 There are no patches on the second geometry that can be matched to non conformal patches of the second geometry please check the list of complementary boundaries It seems that lists of boundaries that form the interface between two complementary bound aries are not correct Ricardo Software December 2009 98 4 MESHING 4 10 GRID DATA STRUCTURE WARNING 1852 There are patches which have no possible patch to make a pair X in GEO 1 and X in GEO 2 There are patches that cannot be matched with any other patch The lists of boundaries that form the interface between two complementary bound
552. tly more time The output data is appended into Devia tion rp file providing distance vs percentage of area of triangles further from the geometry than this distance aa is not available in ld mode Statistics Geometry Wrapper returns the wrapping time number of output triangles and wrapped surface area In case of aa mode minimum maximum and average distances from triangles to the original geometry is also reported The latter is not available in st and ld modes Output file The output information about progress and results is reported into Phasel menu and appended to PHASE1 OUT 3 14 3 Geometry Wrapper Input File A file browser in Geometry Wrapper Panel allows the user to specify the use of wrapper param eters input file of the format as shown by the example below The definitions in this file will be used instead of the GUI panel input values Refer to Appendix D for more information regarding the wrapper parameter file IVECTIS_WRAPPER_INPUT VERSION 3 9 FEATURE_RESOLUTION_SIZE Ricardo Software December 2009 49 3 GEOMETRY 3 14 GEOMETRY WRAPPING 145 2 0 001 3 0 0005 NTERNAL_POINT 0 115518 0 007180 0 000471 NTERNAL_POINT 0 033556 0 038563 0 002235
553. to 0 With this setup there is no surface refinement applied x Specification for boundary 2 ja gt Delete Refinement depth at boundary 2 Refinement blending distance jy Refinement Blending Blend to boundary depth Blend to boundary depth 1 m Refinement Specification Destination Save specification in triangle file Save specification in mesh file mu ts hi If the Refine to Boundary depth 1 option is selected only the region attached to the surface is refined Surface refinement set to 2 with a blending distance of 1 and blend to boundary depth 1 Ricardo Software December 2009 60 3 GEOMETRY 3 15 BOUNDARY PROCESSING Now the surface refinement level is set to 2 and the cells next to the boundary are split into 4 smaller cells Only the region of the global cell that is attached to the boundary are sub divided since the blend to boundary depth 1 option is used x Specification for boundary 2 Ej 2 Delete Refinement depth at boundary j2 Refinement blending distance jl m Refinement Blending Blend to boundary depth Blend to boundary depth 1 m Refinement Specification Destination Save specification in triangle file Save specification in mesh file woo Corea Wi Back to Mesh refinement stepped global down every 2 cells until cell size global cell size reached applied 4 If the Blend
554. to ensure that the generated cells fulfil certain quality criteria VMESH has no problems to gen erate cells automatically when the input surface is correctly defined However in real geometries imported from CAD systems there usually are flaws For the mesher it is essential that the user resolves serious problems such as gaps and intersections of triangles However even when there are no unstitched lines and test of intersected triangles reports no problems small errors might re main in the surface In Phasel it is possible to recognize these problems according green edges of triangles VMESH works quite robustly so it can overcome majority of problems with the surface imperfection However it might happen that due to problems on the surface the generated grid has some problematic cells In this section the possibilities of dealing with problems of the input triangles are summarized A reasonable procedure of mesh generation might look like this Ricardo Software December 2009 92 4 MESHING 4 7 PROBLEMS WITH QUALITY OF INPUT SURFACE x Specification for boundary 2 E B Delete Refinement depth at boundary as Refinement blending distance IA m Refinement Blending 6 fiendis boundary d Blend to boundary depth 1 m Refinement Specification Destination Save specification in triangle file Save specification in mesh file m9 cu Figure 4 11 Boundary refinement depth 2 blen
555. to find and correct degenerate parts of the geometry such as overlapped or intersected triangles The resulting trifile is then stored with postfix _wt both corrected and uncorrected triangles are painted as a boundary most relevant to their position The mesher does not proceed to generate the mesh sep separate this switch activates a thorough search of intersected and overlapped triangles these are repainted as a new boundary and the whole geometry is then stored in a new trifile with postfix _wt The user should check this trifile with Ricardo graphic tools such as Phasel and correct the problematic parts The mesher does not proceed to generate the mesh rep sep When these two switches are used together the mesher first attempts to heal problematic triangles then repaints the remaining faulty triangles and stores the result in a new trifile with postfix _wt not no problematic triangle test switches off the default rapid test of overlapped triangles in this case problematic triangles are not detected at all PREPARE GRIDFILES FOR MULTIDOMAIN int b1 b2 bn interface b1 b2 bn This command informes the mesher that boundaries b1 b2 bn form the in terface between domains Patches on the interface will be treated differently more simplified so as the later work to make the grids conformal is easier Also the min max extents of the model are not found on the geometry but it is taken from the positions of mesh lines
556. tolr tolerance If this option is used the tolerance used for linking corresponding vertices is directly set and not calculated from the diagonal length of the bounding box see description of the option cb_tolrnlevs viewnonconform minx maxx miny maxy minz maxz vne minx maxx miny maxy minz maxz When one of these synonyms is used together with conform or conformrew the situation is visualised only instead of trying to make the comple mentar boundaries conformal The conformal patches are drawn as blue polygons nonconfor mal parts are grey Patches that are recognized to be alone they have no possible corresponding patch to be tied on are drawn as red polygons If red polygons appear the input specification should be checked If the six values representing limits of the area are added only part of the geometry will be shown In order to learn hot keys for manipulation with the visualization see basic visualization tools above TEST TRIFILES FOR MULTIDOMAIN MESHING operates on two trifiles cmpb TrifileA bA1 bA2 bAn TrifileB bB1 bB2 bBn This feature can serve for detection of problems in definition of interfaces for multidomain meshing The common interface must be exactly the same in both input trifiles and this tool can check whether it really is When this switch is used the two trifiles TrifileA and TrifileB are loaded and for all vertices triangles belonging to boundaries bA1 bA2 bAn on the A geometry are
557. ton located on the button bar Lights Pressing the Lights button will bring up a new panel that allows lights to be turned on and off The lights are placed in default positions around the model Light Positions m Lights ff amp Model ma EH E The position of the tick box with respect to the word Model corresponds to the location of the light with respect to the model on the screen Fast Rotate Move Selecting Fast Rotate Move switches on the fast rendering of the model when mouse rotate and move operations are being performed In this mode a fraction of the model triangles are drawn allowing the screen to refresh at a much faster rate Show Surfaces Selecting the Show Surfaces option will cause the triangulated surface of the model to be shown By default the surface is shaded in the colour assigned to the boundary that the triangles belong to and as if lit by the lights defined above Flat Shading If the Flat Shading option is chosen the lighting is ignored and the triangles are simply filled with the appropriate colour Ricardo Software December 2009 18 3 GEOMETRY 3 5 GENERAL VIEW OPTIONS Flip Surface Display The Flip Surface Display option causes the triangles to appear as if they were lit from the other side This may help to show the detail of cut interior surfaces or to light the model properly if it has been defined with the surface normal facing in the opposite direction to that wh
558. transport equations for mass momentum species energy and turbulence quantities VECTIS MAX provides a solution of equations for the volume fraction in case of multi phase modelling velocity potential in case of potential flow initialisation and passive scalar 10 2 Equation of State Models The resolved flow equations need to be closed with an equation of state in order to determine density Two models for the equation of state are available Ricardo Software December 2009 168 10 MODELLING SINGLE PHASE FLOWS 10 2 EQUATION OF STATE MODELS O Incompressible substance model Density is either constant or depends on the temperature The incompressible model is used for liquids gases and solids O Ideal gas model Density is a function of pressure and temperature and it is calculated from Equation 8 28 _ Pabs 10 10 with paps being the absolute pressure The absolute pressure is defined as Pabs Ps Pref 10 11 where ps is the relative static pressure and pref denotes the reference pressure The model is used only for gases The density derivative with respect to the pressure under constant temper ature 1 e compressibility factor as defined by Equation 8 13 is also required when solving the pressure correction continuity equation see Equation 14 58 The equation of state model selection depends on the type of fluid phase gas or liquid and its compressibility options see Figure 10 3 The incompressible mo
559. tructure The data structure and software design of VECTIS MAX kernel supports multi domain and multi physics modelling of a general multi phase and or multi component continuum Unstructured grid solver The ability to perform CFD simulations on an unstructured grid either generated with the native mesher vmesh or imported from a 3rd party format CGNS Ideas universal unv etc All mesh types are accommodated by general polyhedral control volume formulation Continua Properties Thermo physical and thermodynamics property specification is designed to be as general as pos sible with properties specified for each species and phase separately using a range of options These options include fixed values ideal gas temperature dependent polynomials or exponentials Properties can also be specified using user functions or from WAVE property files N umerical Method O Discretization Advanced finite volume based discretization applicable to arbitrary polyhedral cells Linear equations solvers Both bi conjugate gradient and symmetric conjugate gradient linear solvers are available with a number of pre conditioning options including incomplete Lower Upper Modified Stone s SIP and Jacobi Pressure velocity density coupling Segregated implicit SIMPLE like approach for all speed incompressible and compressible flows Poor quality mesh cells A novel cell interpolative scheme deals with low quality cells
560. tube inp input file The mesh file name needs to be specified in the solution control panel Also select Save Initial Conditions Post File in this panel Next switch on the potential flow option for the initial values in the Fluid Domain data panel Note When potential flow solver is used to initialise the calculation it is important that the refer ence values and initial condition values specified are representative as they are used by the potential flow solver Save the file as tube_pot inp file gt save as Run the simulation again type vsolve tube_ pot inp at the command prompt Using live update the convergence of the two initialisation methods can be compared Using the potential flow initialisation has made the simulation converge much more rapidly when compared to the uniform constant initial conditions This can be seen by the fact that the curves for the boundary mass flows when using the potential flow initialisation are much shorter see Figure 19 21 The initial flow field predicted by the potential solver can be examined by opening the tube_ pot post file The step panel allows the iteration for which the data is plotted to be selected for each plot In this case a slice is made though the plot The slice is then copied right click then copy The first copy is placed in one canvas and iteration zero is chosen this shows the flow field after initialisation The second copy is dragged into
561. tuent species properties Equation 8 39 Ricardo Software December 2009 138 8 MODELLING CONTINUA 8 6 SELECTING CONTINUUM AND ITS PROPERTIES 8 6 1 Setting a multiphase mixture After importing the mesh and creating the domain structure the single phase single component fluid will be assigned to fluid domains In order to define the multiphase mixture left click on a Fluid Domain node in the Solver Setup Tree This will open the Fluid Domain Setup whose top part is shown in Figure 8 1 Here the default Domain and Material names can be changed The Material ID value is currently not used Left click on the Multiphase Modelling button and select Mixture Domain Name Fluid_1 Material Name air Material ID 1 Multiphase Modeling Mixture gt lam Figure 8 1 R Desk setup Selecting a multiphase model The Solver Setup Tree will be refreshed and the user is now able to create a multiphase mixture by adding the required number of fluid phases To add a phase right click on the Fluid Phase node and select Add Fluid Phases Figure 8 2 Initial Condition Output Delete air Fluid Phase Species Species_1 Fluid Species Expand Properties Output Collapse Species_2 Fluid Species Expand All aa Collapse All Figure 8 2 R Desk setup Adding a fluid phase 8 6 2 Setting a phase and its properties To define a phase model the Fluid Phase Setup panel needs to be open
562. tum equations for U p and O see Equations 14 49 14 50 provide the link between the velocity and pressure corrections h Vp 14 56 j With the cell face gradient V pi given by Equations 14 29 the face velocity correction is obtained as SS Vp A Vp E 22 a E I d Nae Jj jpd S Na The mass flux corrections are then given as We 7 4 14 57 Ajd mj p pi 0 Aj m m nil p 0 Aj piU Aj piU Aj ORicardo Software December 2009 249 14 NUMERICAL SOLUTION 14 2 SIMPLE BASED SOLUTION PROCEDURE Obviously the last term in the above expression is smaller than others Next the density correc tions are approximated by ps 2 p Cpp 14 58 dp r where the coefficient Cp can be found from the equation of state For an ideal gas we have Cp 1 R T The cell face density correction p is approximated here by a fraction of the upwind scheme Pj p 2 Bp 14 59 mj where we use Pp Oy p being the under relaxation factor for pressure It should be noticed that such a choice has no influence on the final solution since the pressure corrections tend to be close to zero at convergence P Cpr max rir 0 Pp Cpp max m 0 p We recognise that the velocity corrections from neighbouring cells represented by hp i are not known at this stage In the SIMPLE algorithm they are neglected which is a crude oversimpli fication It is also common pract
563. turbulent mixture viscosities are phase weighted Nph k 2 eee Y a ui p Cu 12 18 The last term in Equation 12 15 represents momentum diffusion due to the relative motion of phases and contains the drift or diffusion velocities defined as yet uk U The diffusion velocity can be determined by employing the algebraic relations for the slip velocity If the phases move with the same mixture velocity terms containing the drift velocity omitted the mixture model reduces to a homogeneous multiphase model O The k model for the mixture 0 k e 0 Li N ok lo k pr P P p e p 12 1 AG gg eu a PE 3 ur o ae o mm 0 mrm Cei P Cap e Cez P 3 P e Ox p e Uj I m 0 m Ce4p AT 2 E R 12 20 dx Ox O Energy conservation for the mixture d d Op P a p H U7 ar tP a fi 0 u u oT u ok di m 12 21 ON A Pr Cp Ox oi Ox me ai 17 In the above equation the total enthalpy of the mixture H takes the usual form H CT U U 2 k while the specific heat and molecular thermal conductivity of the mixture A u C Pr are defined as m LE o pc m eE op HEL an Eta 12 22 O Volume fractions for the mixture dakpk a krym d f k krdk k s a akptym z aie r 12 23 Ricardo Software December 2009 201 12 MODELLING MULTIPHASE FLOWS 3 MULTI PHASE MIXTURE AND
564. ty field and the pressure gradient between nodes P and Pj is linear as would be the case on very fine grids then vector h j denoted for this case as RO would be given ORicardo Software December 2009 248 14 NUMERICAL SOLUTION 14 2 SIMPLE BASED SOLUTION PROCEDURE as TaN V o a Vn 14 51 ap J By replacing the value of h j in Equation 14 50 by an the following formula for the face velocity vector is obtained E V a j VP Vp 14 52 After replacing the pressure gradient Vp by the face gradient definition according to Equation 14 29 the face velocity vector reads gt V A vu J eg Y a PEPENE d 14 53 i i a5 ef ea i Z _ E 2 14 54 ap 2 ap ap where the preferred interpolation practice for the face pressure gradient is the arithmetic averaging Vp 0 5S Vpp Vp 14 2 3 Pressure corrections The discretized momentum equations are solved for U p by using the existing pressure and density fields p and p The mass fluxes computed by using the cell face velocity from Equation 14 53 do not generally satisfy the continuity equation 14 7 A mass source would thus result defined by d piv Y ri 14 55 j The basis of the SIMPLE algorithm is to drive this mass source to a negligible small value This is achieved by introducing the corrections Up U3 Up pp p pp and pp p pp The discretized momen
565. ua tion 11 15 is commonly provided by Darcy Forchheimer s model f 1 E pCelO UF 11 19 where K m7 and Cr m are the permeability and Forchheimer s parameters respectively which in general depend on the volume porosity y The Forchheimer s parameter is often ex pressed as Cp Cr I VK where Cp is the Forchheimer s non dimensional constant Introducing the permeability and Forchheimer s tensors K and C respectively the total porous resis tance tensor can be defined in terms of the viscous linear Ar and inertial quadratic Ri j parts as follows Rij Bi Bi V with Bi uK and 2i pC 11 20 This leads to the generalised Darcy Forchheimer s model applicable to non isotropic porous media p y uK pC 10s Us y 2 20 Us 1121 The viscous and inertial tensors accommodate various flow resistance models whose parameters are user supplied Their dimensions are kg m gt s and kg m respectively It can be noticed that the superficial volume averaged velocity U Uj is used in the Darcy Forchheimer s model O Species mass fraction conservation a a o a 3 Pei 5 vPciU an KORRAS P amp U Wet 5 Te dA 11 22 Xj where the intrinsic molecular diffusion flux Je j is defined as 110 1f oz Jag dan y 2 ax y i BicinjdA 11 23 The last term in the above equation can be modelled in terms of the intrinsic concentration g
566. uation have been neglected having in mind their current state of knowledge Leschziner et al 2000 Barre et al 2002 The dissipation rate of turbulence kinetic energy T 0u 0x p appears as the sink term in the k equation While an exact transport equation for e can be derived and interpreted in the physical terms cf Speziale 1996 the model equation relies much more on the physical and empirical reasoning than on the rigorous modelling of the exact equation An important issue regarding the modelled and exact equation is that the dissipation rate used in the defini tion of the eddy viscosity Equation 9 14 actually represents the energy transfer rates from large energy containing eddies This quantity is generally different than given by the exact equation describing the dissipative process in the small eddies 9 3 1 Standard k e model The following model equations for the turbulent kinetic energy and its dissipation rate describe the basic standard SKE k model cf Hanjalic 1994 Speziale 1996 ae Be uy k 57 PK 5 kU Unj R tP p u E 9 19 Ce1Pk Ce2p Ce3Pp T o o J p ax pe U Ug j Ceap ESkk ue de x Se 9 20 xj Ricardo Software December 2009 153 9 MODELLING TURBULENCE 9 4 NEAR WALL MODELLING 9 3 2 RNGk e model The Renormalisation Group Theory RNG applied to turbulence modelling by Yakhot and Orszag
567. uations c T and cp T are temperature averaged specific heats Their use sim plifies the calculation of internal energy and enthalpy of thermally perfect fluids A general poly nomial function of temperature T a a T a3T7 a T 8 26 Ricardo Software December 2009 134 8 MODELLING CONTINUA 8 3 THERMALLY PERFECT FLUIDS n is the number of polynomial coefficients aj i 1 n is often useful to express specific heat and other property dependence on the temperature Some fluid properties are based on the two commonly used thermally perfect fluid models incompressible substance model and ideal gas model 8 3 1 Incompressible substance model For liquids and weakly compressible gases at moderate pressures the density depends mainly on the temperature The following expression also known as Boussinesq density approximation is appropriate for small temperature differences for liquids temperature range less than 4K and for gases less than 15K p Pref Pref P T Tef gt 8 27 where P ef is the density at a reference temperature T ef The above expression is not valid for water around 4 C For liquids and solids it is often appropriate to assume constant density Such an incompressible substance is characterised by B 0 0 cp cy K 1 de c T dT or e CyT and dh c T dT dp p or h CpT p p The speed of sound becomes infinite since the coefficient of isentropic compressibility 6 tends to zero
568. ub sonic E IT Solve All Species I Define Passive Scalar Density Option Ideal Gas tt sti Viscosity Option Constant Values Value 0 008240 r Conductivity oon Constant Values SSS Value 0 0257 r Specific Heat Option Constant Values y Value 1004 Molecular Weight Option Constant Values y Value 29 3 r Thermal Expansion Coefficient Option Constant Values y Value 0 Figure 19 46 Fluid Phase Options Phase Type and compressibility The phase type can be selected to be either a gas or a liquid In the case of liquid simulation the compressibility is set to be incompressible For gases there is a choice of compressibility options Incompressible Fluid Weakly Compressible Fully Compressible Subsonic Fully Compressible Supersonic For the tutorial we specify a fully compressible subsonic gas More details of the different options can be found in Setting a phase and its properties Solve all species This determines whether the mass fraction equations for all species are solved or whether the last species is taken as the remaining mass fraction after the other species are solved Property Specification In general there are a number of ways of specifying the properties constant values or relationships Initial Conditions The initial conditions for the fluid can be specified in the initial conditions panel This allows the
569. ulated as hix ay 09U 43 U a4 U 45 U a6 U 11 58 After defining the heat transfer coefficient the Heat Exchanger Temperature T ef and the Minimum Temperature Difference ATjni have to be provided in the HeatExchangerData sub panel The minimum temperature difference is defined as the difference between the minimum sub domain fluid temperature Tf min and the heat exchanger radiator reference temperature Tye i e ATinin Tf min Tref gt O if the sub domain fluid is cooled If the sub domain fluid is heated heater AT nin Tref Tf max gt O where Tf max denotes maximum sub domain fluid temperature As a radiator is a heat exchanger more details about the heat exchange modelling is given in the next section 11 3 5 Heat exchangers For a general heat exchanger the flow resistance can be specified as for the general porous model In terms of heat exchange heat exchangers can be classified as O Heaters Sub domain fluid is heated reference heat exchanger temperature T should be greater than the maximum fluid temperature Ty max for at least ATinin where T ef and ATmin are specified by the user inside Heat Exchanger Data group box Figure 11 3 O Coolers Sub domain fluid is cooled i e Tref lt Tf min AT nin For both heat exchange modes either Heat Load or Heat Transfer Coefficient the Heat Transfer Coefficients hy are provided as for the radiators For the given hy F U it is pos
570. ult to all phase properties If the user specifies a different option than Mixture it means that the given phase property will not depend on the composition of the mixture i e Equation 8 39 will not be used Ricardo Software December 2009 143 8 MODELLING CONTINUA 8 6 SELECTING CONTINUUM AND ITS PROPERTIES Option Expression No of Coeffs Intended Usage Constant values All properties Boussinesq Density see Eq 8 27 Exponential f T c3 c4exp cy c2T 4 All properties Inviscid Viscosity Piecewise Linear f T As above All properties Polynomial f T As above All properties Ideal Gas f T ci aT cpT 7 5 Wark 1983 Specific heat for common gases density Eq 8 28 Power Law f T Ci z S 3 Viscosity Eq 8 4 Sutherland Law Cj Z l Ja 3 Viscosity Conductivity Eqs 8 3 8 6 Mixture Multi component phase User Subroutine All properties Table 8 1 Options for calculation of single component phase and species properties Conductivity Option Sutherland Law f T Number of Coefficients 2 53 Coeff 1 11 Coeff 2 11 Coeff3 1 Figure 8 8 R Desk setup Property calculation options Sutherland formula O Species Solver Setup Input a amp Species_1 Fluid Species Properties Output Species_2 Fluid Species Properties Output Add Child Species Add Species Delete Species_1 Fluid Species Expa
571. umbers Pr and Pr with corre sponding Schmidt numbers Sc and Sc respectively Numerically the wall fluxes shear stress heat flux and diffusion flux of species can be defined in terms of the wall effective viscosity Uw thermal conductivity Ay and mass diffusion coefficient Pi w respectively AU Ty T epee Tw io dw q ea Jiw a PD Ci Cir 9 39 from which and Equations 9 32 9 34 9 37 the effective wall diffusion coefficients follow as E Din Di 9 40 As before u A and 9 are molecular diffusion coefficients for momentum heat and mass transfer respectively In the viscous sub layer linear laws apply Uvis NN YP Tiis Pryp gt a Scyp 9 41 so that the wall effective coefficients reduce to the molecular ones The standard wall function approach excludes the buffer layer by extending the viscous and tur bulent layers up to their point of intersection y 11 63 see Figure 9 1 The following formulae describe the distribution of velocity temperature mass fraction of species and turbulence variables O Velocity and wall shear stress 7 e viscous sub layer yp lt y a i on 9 42 n Ey p log law yp gt ye The wall shear stress is calculated from Equation 9 39 with the wall viscosity obtained from Equation 9 40 Ricardo Software December 2009 158 9 MODELLING TURBULENCE 9 4 NEAR WALL MODELLING O Temperature and wall heat flux For the entire wall region a unifie
572. unnin she Solver lea a AA A a A a ew 432 LERNER 432 19 39 Postprocessing or 2s 2 ha k a a a e a ha hd ad 434 1939 1 Dt uExXtACUON sae eagen 6k ee A a Ral a ee a ee 434 19 4 Importins Third Party Meshes c 2444024 24 cao ba dede de aedse ba ets 438 19 4 1 Introductoni is 2 a wl we a A a ee a eee E 438 1942 Simple Example o osans a a o do A we ee A ee ee 438 19 4 2 1 Creating Boundary Regions e 439 19 422 Editing Sets crio A A AA A dd 440 19 4 2 3 Importing grid file data into solver Setup 443 Bibliography ae ORicardo Software December 2009 xiii List of Tables Sal 8 1 9 1 17 1 17 2 17 3 17 4 17 5 17 6 17 7 18 1 18 2 18 3 18 4 18 5 18 6 Table oball the Ype OPlONS ci o E da e ia ek Bol ee ee ew 111 Options for calculation of single component phase and species properties 144 Model coefficients for the K models os cesas aoe eh ae wh a A eee ee ed 154 Description of engine components in the RTH file and in the solver x cylinder number y port valve number HF heat flux EHT external heat transfer 2 228 Grid types SUPPOTIEd sos a oad a o oo ob Gok Ge Bok HOR we at we Ee ee ee 278 Report levels for Summary Frequency GroupBOX gt s ece o eme neren deea 280 Naming convention for different output file types o ooo o 283 Table of all the variables contained in io files o
573. up the homogeneous mixture model The VECTIS MAX solver provides the homogeneous mixture model which neglects the relative motion between phases i e the drift velocity term in the mixture momentum equation 12 15 is omitted In addition other relevant fields such as temperature and turbulence quantities are shared by all phases The mixture modelling equations are solved together with the volume fraction equations Equation 12 23 where the interphase mass transfer is neglected I 0 Multiphase modelling approach i e the mixture model can be enabled within each fluid do main The actual model selection and definition of phases have been described in Section Setting a multiphase mixture Figure 8 1 When the multi phase model is defined the Volume Fraction equation will be activated in the Equations Solver panel with the default control variable values as Figure 12 2 illustrates ies ia Nph The computed values of volume fractions have to satisfy the compatibility condition L ak 1 As the phase volume fraction should be bounded 0 lt ak lt 1 it could be difficult to satisfy numerically the compatibility condition and ensure the bounded solution Regarding this the compatibility condition can be enforced by either solving volume fractions for all phases or solving for Npn 1 phase and calculate the volume fraction of the last phase as Non 1 rola Y o 12 25 k 1 The Diseretise panel shown in Figure 10 10 prov
574. updated showing new fluid solid domains and materials as well as associated sub domains and boundary and interface regions as Figure 7 3 and Figure 7 4 illustrate At this stage the default names have been assigned to the fluid solid domains sub domains fluid phases solid materials and to all boundary and interface regions If the relevant domain component from the tree fluid Ricardo Software December 2009 127 7 MODELLING SPATIAL DOMAINS 7 5 CREATING A DOMAIN STRUCTURE Monitoring Points Algorithm Turbulence Model Discretise Phase_1 Fluid Phase Initial Condition Postprocessing Output Equations amp Solver Fan Model E Boundary Regions outlet Pressure Boundary Phase_1 Boundary Phase inlet Mass Flow Rate Boundary Phase_1 Boundary Phase head Unjoined Boundary gasket_head Wall Boundary block Unjoined Boundary gasket_block Wall Boundary gasket_holes Unjoined Boundary inlet_pipe Wall Boundary E Interface Regions Interface_1 Interface Region Interface_2 Interface Region Interface_3 Interface Region Sub domains Report Regions Solid_2 Solid Domain Monitoring Points Equations amp Solver Discretise Postprocessing Output Solid Materials Solid_Material_2 Solid Material Initial Condition El Boundary Regions Figure 7 3 R Desk setup Example of created multi domain structure left and previewed fluid do main right for conjugate heat transfer
575. ure 10 7 R Desk setup Setting a steady or unsteady simulation in the Global Domain panel To perform unsteady simulations real time information are required such as the time step size a number of time steps and in some situations the time dependent boundary conditions The discreti sation of the transient term Figure 14 2 describes setting of steady and unsteady simulations If a steady state calculation converges it is reasonable to conclude that the simulated flow is steady The same conclusion can be made after performing the unsteady simulation for which the time histories of monitoring flow variables indicate the steady state solution The best way to confirm a steady state solution is to work with the optimal values of the time step size and under relaxation factors so that the solution converges within each time step after just one outer iteration per step Many flows are inherently unsteady for example vortex shedding flows or buoyancy driven flows and they often show periodic behaviour If such unsteady flows are run in steady mode their solu tion will not converge However if the simulated flow is physically steady the transient behaviour indicated by the non convergent solution in steady mode can be falsified by numerical effects In such situations it is advisable to perform an unsteady simulation In case of unsteady simulations the physical time step size is governed by the physics of the flow being simulated For smaller
576. ures that offsetting of material id s with the same id is not al lowed when reading multiple grids The generated grid file air_cat GRD now contains both sub domain grids 5 4 Restarting On A Different Number Of Partitions In order to stop restart the solver vsolve on a different number of processors it is first necessary to run vpre with the np and rest argument in order to repartition the grid and restart files concomitantly For example in order to switch from 4 to 8 processors vpre rest test rst0_010 np 8 flow GRD Ricardo Software December 2009 110 5 READING amp MANIPULATING MESHES 5 4 RESTARTING PARTITIONS JTYPE USAGE 0 Conformal join This method is the simplest and should be used when the input meshes are known to be conformal i e boundary faces are the same along AGI boundaries Any non conformal faces are left unconnected and remain as boundary faces 1 This method should be used when there is an initial gap between 2 input meshes This gap is closed with a tetrahedral conformal mesh Typically the 2 input meshes would be 2 fluid domains and the gap would represent the solid domain Any new boundary faces generated are all painted with the same boundary number 2 This method is appropriate when the boundaries to be joined on each input mesh are nearly coincident but not necessarily conformal and thus require mesh extrusion on one side in order to create a gap to be re meshed tetrahedrally as in jty
577. urrent project run number proj_nums Table 18 20 Subroutine get_run_ctrl to get run control variables defined in access group iacc_ run_ctrl Ricardo Software December 2009 328 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES 18 3 2 13 get_name This UAR is used to retrieve the solver object name see Table 18 21 The list of variables under iget in Table 18 21 are predefined in VECTIS MAX and are readily available To get the name of domain 1 idt 1 then do the following character len dom_name call get_name idt dom_name iget_dom Variable dom_name contains the name of the domain To get the name of species 2 idt 2 then do the following character len specs_name call get_name idt specs_name iget_specs To get the name of an equation with idt 4 then Character len x eqn_name call get_name idt eqn_name iget_eq Variable eqn_name may contain one of the following transport equation names idt given in brackets for illustration momentum ifmom mass_pressure ifmas energy iene spec_mass_ fracs ifcs turb_energy ifte dissipation ifed volume_fracs ifvf and passive_scalar ifps Phase species and passive scalar property names that can be obtained with option iget_ph_pro iget_sp_pro and iget_ps_pro were explained in get_property subroutine Phase and species list the following density lam_viscosity conductivity specific_heat gas_ cons
578. using four or five major mesh lines However do not use too many cells This is a training example and should run quickly Use Figure 19 25 as a rough guide it shows a global cell size of around 3mm in the valve region and 6mm in areas away from the valve In addition to this it is possible to use the IJK blocks to increase the detail level around the valve and inlet port Ricardo VECTIS Phase 1 port tri 10 x File Edit View Toolbars Operations Help daa a E e al NOE ul Y zmin 0 036690 zmax 0 326520 14 40 18 Finished Printin 14 40 43 14 42 51 Mesh input file real A LL Figure 19 25 Example of Global Mesh ORicardo Software December 2009 390 19 TUTORIALS 19 2 STEADY STATE PORT FLOW 19 2 7 IJK Refinement IJK refinement blocks are used to apply mesh refinement in specific regions of the geometry The computational cells in the refined region are sub divided according to the level of refinement applied In this case a single IJK block should be added to the geometry The block should cover the valve region The IJK regions need to be defined in the three dimensions However as it would be difficult to define IJK blocks within the 3 D section of the mesh set up menu they must be defined two dimensions at a time In a 2 D Mesh View mode the IJK Refinement Blocks panel as shown in Figure 19 26 w
579. ut the surface thickening mode To reduce artificial mesh disturbances near concave or poorly resolved features nodes located rel atively far from the original geometry can be smoothed using Distant Node Smoothing mode This option can be recommended for the majority of geometries The quality of wrapper approximation can be assessed using distances from triangles to the original geometry Using this mode would take slightly more time The output data is appended into a file called Deviation rp which can be opened in the Ricardo plotting program RPLOT and displays a distance vs percentage of area of triangles further from the geometry than this distance The approximation assessment is not available in when using the leak detection mode through the wrapper input file The geometry wrapper returns the wrapping time number of output triangles and wrapped surface area When using the approximation assessment mode the minimum maximum and average distances from triangles to the original geometry is also reported The latter is not available when using the surface thickening and The output information is displayed on the screen and is written into file PHASE1 OUT Note that the new information is appended Ricardo Software December 2009 44 3 GEOMETRY 3 14 GEOMETRY WRAPPING Y vecris a A T A A Rough capping of Large Holes Features ORicardo Software December 2009 45 3 GEOMETRY 3 14 GEOMETRY WRAPPING
580. ution Body Force Options to allow the simulation of body force Here the pipe flow example uses water as the fluid the following reference values are used Density 1012 Ricardo Software December 2009 379 19 TUTORIALS 19 1 BASIC TUTORIAL Viscosity 0 000719 Specific heat 3620 For the moment the the initial values will be set as uniform For all the other data entries on the panel the default values will be used Additionally for this example the remaining fluid panels algorithm turbulence model equations amp solver and discretise will be left at their default values Solver Setup Tree DN Solver Setup Global_domain_1 Global Domain Restart Control Timebase Output Fluid_1 Fluid Domain Algorithm Turbhulancea Mera x Monitoring Points Use Cell ld Cell id x Y zZ Y 1 0 1 1 1 Delete Add Solid Domains Figure 19 16 The Solver Setup Tree Next the fluid phase needs to be set up 19 1 8 3 Fluid Phase The fluid phase panel contains data for each fluid phase Each phase may be made up of a single or multiple species Phase Name Each phase can be given a name to aid reference Mixture of Species Option Each phase can be made up of a number of different compo nents species In this tutorial a single component phase is considered More details of the different options can be found in Setting a multiphase mixture Phase Type and compressibility The phase type can be
581. value ibttot 1r O ibrwh 5 wall roughness height 1_reg_value ibrwh ir O ibrwc 6 turbulence constant for rough walls 1_reg_value ibrwc ir O ibsplit 7 flow splitting factor at the outlets 1_reg_value ibsplit ir ibturin 8 turbulence intensity 1_reg_value ibturin ir O ibturls 9 turbulence length scale 1_reg_value ibturls ir c To get the list of region boundary conditions then do the following real wph pointer rval call get_reg 1_reg_value rval To get region boundary values for phases then do the following real wph pointer ph_val call get_reg phase_value ph_val Index pointers in Table 18 11 are needed for array ph_val see subroutine get_phase and is used as follows integer iwp pointer is_ph_bndreg ie_ph_bndreg integer iwp pointer a_ph_bndreg get index of starting amp ending boundary region for phases call get_phase ise_phase_bnd_reg is_ph_bndreg ie_ph_bndreg get the starting allocation addresses for phase variables lat boundary regions call get_phase isa_phase_bnd_reg a_ph_bndreg do ir is_ph_bndreg iph ie_ph_bndreg iph ph_val ivar irt ta_ph_bndreg iph end do Ricardo Software December 2009 317 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES where ivar is the variable identifier ir is the region index and iph is the phase index
582. ve equation was used by Murthy and Mathur 1998 for both unsteady and convection terms 14 1 6 1 Diffusion flux at interfaces for conjugate heat transfer The total enthalpy or energy equation is solved in terms of temperature in a fully implicit manner over the entire solution domain involving all participating fluid and solid domains At material interfaces a conformal mesh is required so that these interfaces are topologically the same entities as internal fluid or solid cell faces Considering the interface j between two different materials Figure 14 8 left both temperature and heat flux continuity must be satisfied ensuring the overall conservativeness of the discretisation This means that heat transfer from the cell P to the interface J A T E Dj e hp T Tp Djc php p Djc p VIP Ted hpdp i 14 42 dp and from the interface j to the cell N AS S Dn hy TT Dje whw Tn Djelv VI CwAj hvdi 1443 jej ORicardo Software December 2009 246 14 NUMERICAL SOLUTION 14 1 FINITE VOLUME DISCRETISATION must be in balance Dj Dj p Dj n 14 44 Note that I in the above equations represents the thermal conductivity for fluid materials it is the effective wall conductivity given by Equation 9 40 Eliminating 7 from the last identity we can derive the interface heat flux and temperature as y lt Aj Dj hj Iy Tp A Fj Dj fi Dje w hj Tj ss 14 45 Try ic dj
583. vely the mesher can be started from a command window Type vmesh tube mesh at the prompt MEA x aj henna fare nen ial iann cms Figure 19 7 Launching the mesh generator in R Desk This will write output to the command window orton 50 done 0 20 60 done ae AAA 70 done EI 80 done A 90 done a we 100 done all common faces successfully generated STATISTICAL DATA OF GENERATED MESH Number of generated cells 14461 boundary 5281 internal 9180 Patching method Number of cells processed by Marching Cubes 4881 92 43 Number of cells processed by Exact Fit 299 5 66 Number of volumes broken by Cell Splitting 0 0 00 of boundary cells Cell quality Number of correct boundary cells 5180 98 09 NO PROBLEMS with negative volumes NO PROBLEMS with gaps There were 130 small volumes they are deleted now 2 46 Number of cells which had to be deactivated 101 1 91 Total mesh volume 7 12103e 005 u3 6 39150e 005 u3 9180 inner cells 7 29531e 006 u3 5180 boundary cells WRITING THE MESH FILE successfully finished Total time elapsed 4 seconds SUCCESSFULLY DONE A computational mesh is produced tube GRD ORicardo Software December 2009 372 19 TUTORIALS 19 1 BASIC TUTORIAL 19 1 7 Importing the Grid File The materials and properties defined in a grid file can be imported into the solver setup This will populate
584. via the use of vpre The use of the acronym AGI Arbitrary Grid Interface below denotes boundary regions to be joined Prior joining the meshes it is important to generate them in the correct way i e the AGI bound aries should be specified when vmesh is run see Sec 4 8 for a detailed explanation This is to ensure the best possible join There are currently five ways of joining separate grids via the jtype option The following command vpre jtype 0 fluidl GRD solidl GRD would create a 2 domain grid fluid amp solid from the 2 single domain grids It is important that the fluid grid files are listed FIRST on the command line The preferred method of joining is jtype O where the input meshes have near or total conformal boundaries An example of creating a 3 material grid file beginning from the mesh creation stage is shown below The middle mesh solid1 in Fig 5 3 represents a solid domain the two surrounding meshes are fluid domains The coincident boundaries to be joined are fluid1 boundary 1 to solid1 boundary 2 and fluid2 boundary 1 to solid1 boundary 3 The three meshes are initially created using the int option to specify the boundaries to joined and subsequently converted to interface regions ORicardo Software December 2009 108 5 READING amp MANIPULATING MESHES 5 3 MESH JOINING Figure 5 3 Picture of three meshes to be joined Gaps between meshes are exaggerated vmesh int 1 fluidl mesh vmesh
585. w gt nf gt dA Y Af 0 4 2 where ny is the number of cell faces The cell face area is then calculated as M Af y A 3 Vn Aay Any AL Aj 4 3 t 1 t The coordinates of the face centre Ff x djs Z j are computed as the average values of the centre coordinates of decomposed triangles 7 t 1 M weighted by their areas A lA M M i Ear Ea 4 4 The face centre coordinates of a triangle are simply average values of its vertex coordinates Ricardo Software December 2009 101 4 MESHING 4 11 MESH IMPORT The cell volume formula is based on the Gauss theorem 1 2 4 j M f 7 Via frd vp 5 7 A1 3 Y Y ade 4 5 7 A f 1 f 1 1 1 The geometric centre of a control volume around Pj is defined as 5 E p z rdV 4 6 Vp Jv With the help of Gauss theorem and by using the property of a planar face 7 1 const the following equation for the position vector at the cell centroid can be derived E 1 MD eee 3 Aj Fp 37 7 A Jr 47 P We E Er s 7r 4Vp 2 2 ir Air To id The interpolation factor defined in terms of distances between the node P the centre of face f and Pj 1 see Figure 14 1 is often used when the cell face values are obtained by linear interpolation FP Fs 7 Ef 7 T Am F Wf 4 8 The distance vectors for internal d f boundary faces d f and faces along material interfaces d are also required They are defined as
586. we ER ew 428 19 76Importing the computational grid into the solver setup oo o 428 19 77 The Global Domain Panel 4 sex 440 412 e Rs la A aon ae 429 19 78 Output Options lt s s cos hae ABA ba a eh a a eee dds 429 19 79 The Fluid Domain Panel cid eel a ee E ed A ee Bad Dee aa 430 19 S0Fluid Phase Opuonsa 4 4 0 24 28 Seda Bawiea ke Ha oe Pe d OS be Ped se ware des 431 19 51 Boundary Tesion Settings o c steg a A wh ea A ew 432 TO S2 LNE UPAS ee dca ee ge ee ete BONE ce as AE a eA ae ee OR ee we we et A 433 19 83The residual Values for the coolant simulation o 433 19 84Massflow at the outlet pressure boundary 2 o o e e 434 19 85Pressure variation at the model surface s cono w s sor oo o e ee a 434 19 86 Using the mouse to createa slice n cito uoa ek ee Se Re ee i E 435 19 87Visualising a slice through the gasket holes secre sose oa adaa aa aa e a a a 436 19 88 Extract Flux Options 2 4 4 sos aia i 4 a alae ace Oa a a a e a a a we ee wee a 436 19 89Using a polygon to define the region to extract flux data from a aooo aaa 437 19 90Flux data written to information panel c scsi sacs esaa n ia a a 437 19 91 The GRD file when loaded into R Desk s ceea nocna ta a eoa a wo s 439 19 92 Creatine a new face Set s onokia ia ia a a OR 439 19 93 Editing the tace Set n ie soree A ea ee Gh we Se A ee O E E Aa 440 19 94Pick options available for set editing lt lt o
587. weakly compressible fluid is defined for Ma lt 0 3 isubsonic 2 subsonic gas flow 0 3 lt Ma lt 1 O isupsonic 3 supersonic gas flow Ma gt 1 The calculation options for mixture of species for fluid phases are obtained as integer iwp pointer ph_mix_opts call get_phase mixture_opts ph_mix_opts Array ph_mix_opts may have one of the following identifiers O iphmix_none 0 single component phase Ricardo Software December 2009 310 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES O iphmix_vec 1 vectis calculation method OD iphmix_wave 2 combination of wave density Cp and vectis for other properties O iphmix_wavec 3 wave properties with combustion The get the starting amp ending species indices for all phases then integer iwp pointer il i2 call get_phase ise_specs il i2 If for example phase 1 has 2 species and phase 2 has 3 species then arrays il and i2 will contain the following data 2 for phase 1 for phase 2 oll WwW Be p N N ol ul Ricardo Software December 2009 311 18 USER PROGRAMMING 18 3 ACCESSING SOLVER VARIABLES subroutine get_phase var_name il 12 Input Output var_name cha il integer pointer optional i2 integer pointer optional phase_type type liquid gas solid 11 1 n_phases phase_compress compressibility for each phase 11 1 n
588. which can not be adequately represented by a one dimensional WAVE model or for VECTIS calculations where a coupled solution with WAVE would provide more accurate boundary conditions For further background and theory consult the co simulation manual available within the WAVE help manual In a WAVE VECTIS coupled calculation each VECTIS model is started as a separate child process by WAVE Current restrictions for the VECTIS model are as follows 1 The domain containing coupling must be a single phase fluid domain 2 Participating boundaries must be defined as Mass Flow 3 All participating boundaries must belong to the same fluid domain 4 The calculation must be unsteady It is recommended that the WAVE and VECTIS files be stored in separate directories One con venient way to do this is to have WAVE and VECTIS subdirectories below the main project directory Typically a run_VECTIS script is used to execute VECTIS as a child process An example script which changes to the VECTIS directory and starts vsolve could be as follows cd VECTIS vsolve np 2 project The VECTIS input file requires the corresponding link WAVE junction references These can be set via the Boundary Regions panel See Fig 13 2 Ricardo Software December 2009 294 18 USER PROGRAMMING 18 1 Introduction And Overview In general a user may not be able to do everything that he wants with a CFD solver For example the user may w
589. which contains any postprocessing data as selected by the user in R Desk The restart files rst are explained in Section 17 3 The input file read in by VECTIS MAX has an extension inp without the runnumber part However each run automatically generates a copy of the inp with a runnumber attached 17 13 Monitoring The Solution This section describes the Live Update panel Fig 17 15 used to monitor the solution The user is able to change the run directory to monitor it will initially default to the current working directory Clicking on the Refresh Available Files is necessary if a run has just started in the current run directory The Files panel displays the averaged data for boundaries I O wall and domains In addition within each boundary domain number there is a list of runs to choose from As shown in Fig 17 16 this panel is split according to the type of averaged data Once the values have been chosen for a particular run it is then necessary to open an XY Canvas Dragging the first value e g Iter_No from Value will make X axis the next drag and drop of value will make up the Y axis This can be repeated for subsequent graphs on the same or different canvas where the value pairs are taken to be the X Y axis There is no limit to the number of graphs that can be plotted on the same XY canvas The XY Canvas can be renamed by right clicking anywhere within the canvas Figure 17 15 displays the static pressu
590. with k equation Based on the work of Davydov 1961 Hanjalic 1970 see also Hanjalic and Launder 1972 proposed a simple modelled e equation for high Re number flows He also established model coefficients that differ a little from Ricardo Software December 2009 151 9 MODELLING TURBULENCE 9 3 LINEAR TWO EQUATION K MODELS values proposed by Jones and Launder 1972 Launder and Spalding 1974 The model pub lished by Launder and Spalding 1974 is often referred to as the standard k model Some weaknesses of the standard model can be removed by using the RNG k cf Yakhot and Orszag 1986 Yakhot et al 1992 and realisable model of Shih et al 1995 The term realisable de scribes satisfaction of certain mathematical constraints so that generation of physically unrealistic variable values e g negative values of k and its components normal turbulent Reynolds stresses is prevented Neither standard nor RNG models are realisable Shear Stress Transport SST model This model developed by Menter 1994 is formulated as the k model but effectively it is a combination of the k and k models Close to the wall it reduces to the k model and away from the wall to the k model It was claimed to perform better than the Wilcox k and k e models especially for separated and adverse pressure gradient flows Advanced models This group of models provides so
591. wo global cells The figure below is a snapshot of a VECTIS control mesh that shows this setup Refinement block overlap regions The different IJK blocks have different refinement settings so with the forced level of refinement increasing by 1 for each block from the outer block to the inner block Cell shape It is best to maintain square cells throughout the computational domain as much as possible since this improves the numerical accuracy This is particularly important with a mesh that includes IJK refinement regions The figures below show the nonideal setup and the ideal setup ORicardo Software December 2009 72 3 GEOMETRY 3 17 MESH SET UP VIEW OPTIONS Neighbouring cells with significantly different aspect rations Neighbouring cells with suitable aspect ratios If it is necessary to have a variation in global cell size at the edge of an IJK refinement block then overlapping the refinement block will minimise the numerical error as much as possible The figure below shows this concept Refinement region overlaps different global cell size regions reducing neighbouring cell size ratios Ricardo Software December 2009 73 3 GEOMETRY 3 18 WARNING AND ERROR MESSAGES 3 18 Warning and Error Messages Note that these pages are part of the ongoing improvements to the VECTIS documentation please contact RS_Support Ricardo com for specific questions about certain warning or error messages In Phase 1 warning
592. y ipro_suth ipro_user O icph 4 temperature averaged specific heat either at constant pressure the total enthalpy equation is solved or constant volume the total energy equation solved for Calculation ce cos ce cos ce cos options ipro_const ipro_mix ipro_igas ipro_poly ipro_user O imolw 5 molecular nee which is used to calculate the gas constant Calculation cc cos cc cos options ipro_const ipro_mix ipro_user O itexc 6 volumetric thermal expansion coefficient used to calculate the density for small ce cos temperature ranges Calculation options ipro_const ipro_user In case of the multicomponent fluid phase i e when species mass fraction equations are solved additional fluid species properties have to be defined oO idifm 7 mass diffusion coefficient which governs the laminar mass diffusion flux Cal culation options ipro_const ce cos ipro_poly ipro_user O idift 8 thermal diffusion coefficient which governs the mass diffusion flux caused by the temperature gradient Soret effect Calculation options ipro_const i user ce cor ipro_poly ipro_ O iheatf 9 formation standard state or heat of formation enthalpy for reacting flow aA 66 cor lems involving solution of the energy equation Calculation options ip
593. y join points that are very close together in order to avoid distortion of the model An example of the use of this function is given in the Figure Joining Triangles to mend the surface Ricardo Software December 2009 23 3 GEOMETRY 3 7 TRIANGLE CREATION OPERATIONS Split Line Button A This function will create a new node at the location selected upon an edge The triangles sharing this edge will be re triangulated to use this node MO ADS ESTU Edge Splitting Cap Hole Button A On selecting this operation the cursor will change to a cross hair If a line is then selected by clicking upon it and the line forms part of a closed loop of a hole edge drawn in red in the surface a number of triangles are created within the loop Triangles are created between the lines describing the smallest angles first and stops once the hole is filled Auto Stitching Button A 7 4 at This is a direct way to do Auto stitching it is an alternative to using the menu command Operations gt Auto stitch this button can be used to do the same operation The user is prompted for a confir mation before the operation continues and the operation can be cancelled at this point if it is not required Flip Connected Marked Triangles Button Ol This function reverses the normals of a connected set of triangles or a set defined with the mark triangles command The triangle topology of the selected triangles is reversed and the
594. y of the mesh generation written at the end of mesh generation In this case it can be seen that two warnings are produced These are non fatal and the GRD computation mesh file is written The summary output is shown below Ricardo Software December 2009 423 19 TUTORIALS 19 3 COOLANT FLOW STATISTICAL DATA OF GENERATED MESH Number of generated cells 64181 boundary 48024 internal 16157 Patching method Number of cells processed by Marching Cubes 36802 72 48 Number of cells processed by Exact Fit 9165 18 05 Number of volumes broken by Cell Splitting 2722 5 36 of boundary cells 651 attempts to split a volume failed 19 30 of all attempts 274 attempts gave incorrect number of volumes 377 attempts gave volumes with too low quality so the undo was applied Cell quality Number of correct boundary cells 48714 95 94 5 NO PROBLEMS with negative volumes Number of cells with problems of gaps 2 0 00 There were 2348 small volumes they are deleted now 4 62 Number of cells which had to be deactivated 2057 4 05 Total mesh volume 1 75590e 003 u3 1 09865e 003 u3 16157 inner cells 6 57243e 004 u3 48883 boundary cells WRITING THE MESH FILE successfully finished Total time elapsed 54 seconds DURING THE MESHING 2 OF WARNING S OR NON FATAL ERRORS APPEARED 1X ERROR 1506 Tying patches routine cannot move the required node in box 37712 in these IJK positions 42 31
595. y then the following holds rr lt i 12 26 liquid vapour The saturation temperature Tsat depends on absolute pressure The absolute pressure is defined in section Reference and solver working pressure Paps P Pref App 12 27 where p is the working pressure preg is the reference pressure and Ap is the hydrostatic pressure defined in relation 10 15 The set up of body force reference altitude and reference density needed for calculation of hydrostatic pressure together with reference pressure is given in section Fluid Domain Figure 17 6 12 4 1 1 VECTIS3 boiling model This model was first proposed in Bo 2004 and is based on the assumption that boiling flow is modelled in the nucleate regime The detailed analysis of the forces acting on bubbles such as Ricardo Software December 2009 203 12 MODELLING MULTIPHASE FLOWS 12 4 PHASE CHANGE MODELLING buoyancy drag surface tension inertia etc are not included The drift velocity between the two phases is also not considered O Volume of Fraction Equation For a multi fluid homogeneous mixture model the mass conservation Equation 12 2 is re written in the following form aak Et a atut r 12 28 where I is the evaporation condensation mass source O Evaporation and Condensation in the bulk The source term in Equation 12 28 is expressed as ri a na 12 29 where 2 a a F oai max 0 12 30 tevap re min o Can
596. y throughout the calcula tion Restart files are written with an interval specified in the Restart File Frequency field Two restart files are generated projectname rstO_runnumber and projectname rstl_runnumber are written alternately every restart frequency time steps The Restart Control panel controls the restart file reading and writing Shown below is the restart reading GroupBox Restart Reading y Restart With Previous Results SetFilename Filename _UNSET__ Figure 17 2 GroupBox for Restart Reading Restart reading is activated by clicking on the Restart With Previous Results This will instruct the solver to restart from the newest rst files with the same project base name To override this behaviour the user can activate the Set Filename CheckBox and enter a specific filename in the Filename LineEdit The ideal restart frequency is a compromise between how far back in the calculation you will have to go to restart in the case of the calculation crashing or system failure and how much time the calculation spends writing the restart data to disk If a very short frequency Ricardo Software December 2009 280 17 USING SOLVER 17 4 FLUID DOMAIN 1s chosen then it can reduce the overall speed of the calculation The Restart File Frequency LineEdit in the Restart Writing GroupBox specifies how often to write to the restart rst files In addition the user can add specific restart files to be written at cert
597. yed Recall that porous media conservation equations are written in terms of intrinsic volume averaged variables Bearing in mind uncertainties in modelling various dispersion terms tortuosity fluxes and other additional terms arising from volume averaging procedure these terms have been neglected Com pared to the non porous equations the simplified porous equations contain volume porosity y Note that the momentum equation is fully implemented with a general flow resistance model applicable to non isotropic porous media The simplified volume averaged equations read O Mass conservation ayp 9 J I pu YSm 11 48 Ricardo Software December 2009 189 11 MODELLING POROUS MEDIA 11 2 SIMPLIFIED MODELLING EQUATIONS Oo Momentum conservation a a 9 PUUN yp o g a od where fp represents the resistance force Equation 11 21 O Species mass fraction conservation ao d Yci TETT pciU Ox 2 2 Oxy YSc 11 50 O Energy conservation d d yp 2 OT eo Ly Ok 5 PH a pHU gt t F ra A a U tij Tij a YRijUjU pfiU 11 51 O Turbulent kinetic energy and its dissipation rate d d e d Lt dk 3 Wek an PkU y Pk P p T r u H z 11 52 o o 5 Ce1Pk Ce2P Ce3Pp z PE 5 PEU Y T j 2 Lt dE YCe4P Et os r n 11 53 As porous structure tends to suppress turbulence the turbulence modelling is more relevant to predictions of hea
598. ys labsolute call get_field idt iget_bnd var_name fib lupr gt start do jb jbndl jbnd2 end do lupr gt end fib gt null case turb_energy In case of pressure total and mass flow boundary types the user lhas to supply turbulence intensity instead of actual turbulence lenergy The turbulence intensity is also saved in the array fib Initialise velocity and turbulence at inlet from tabulated lexperimental data for the boundary layer profile k z associated with the z axis xbnd 3 Get access to fields call get_field idt iget_bnd turb_energy teb turbulent energy Get equation index idm eq_idt ifmom usr_obj_id iget_phase call get_field idm iget_bnd velocity ub Ivelocity Get bnd face centre coordinates call get_grid_geom xyz_bndf_c xbnd IGet region wise uniform bnd values call get_reg phase_value reg_bval zinl zero uz zero Ricardo Software December 2009 18 7 EXAMPLES 360 18 USER PROGRAMMING tez zero ifun 0 call get_funit ifun call open_file ifu ifun fln wing inl Imodule file_utilities st old acc sequential fm formatted read ifun read ifun ndata do k 1 ndata read ifun zinl k uz k u2z v2z uvz tez k 0 5 u2z v2z 2 v2z u2z xreg_bval 1 ir 2 end do do jb jbndl jbnd2 k ndata if xbnd 3 3b gt zinl k then ub 1 3 jb uz k reg_bval 1 3 ir teb jb tez
599. z direction is zero and it is not solved for The selection of two or three dimensional simulation is done by checking one of Dimensionality RadioButtons in the Discretise panel Figure 10 2 Ricardo Software December 2009 170 10 MODELLING SINGLE PHASE FLOWS 10 3 MODELLING FLUID FLOW Discretise o amp Dimensionality e Three dimensional Flow Two dimensional Flow b 4 gt Figure 10 2 R Desk setup Selecting dimensionality of the simulation in Discretise panel This panel is open from the Solver Setup Tree under any Fluid Domain or Solid Domain node by left clicking on the Discretise sub node Before selecting the two dimensional simulation the user should ensure that the geometry is nominally two dimensional i e that O two parallel and plane symmetry boundaries are placed in the z direction other boundary types are not allowed in this direction and O there is only a one layer of cells between symmetry planes in the z direction This might be difficult to achieve with VECTIS MAX mesher using various levels of refinement A refine ment level depth of 0 does not do any refinement and with this option the mesh should have a one layer of cells in the z direction 10 3 2 Single and multi component phase For either single phase or multi phase fluid flow the fluid phase can be selected as a single component or multi component see Modelling Continua and Their Properties The selection has been expl

Download Pdf Manuals

Related Search

Related Contents

Manual de Usuario - Sede electrónica del Ministerio  TracVision R5/R4 Installation Guide  CX501E - Clarion  平 成 2 4 年 度 南部浄水場 自家発電設備更新工事 設 計 書  InduKey TKH-MAUS-IP68-GREY-OPT-USB  申請に対する受理等の通知書 - 一般財団法人環境優良車普及機構  Mode d`emploi Appareils de sauvetage P660SG SUPER SILENT    User Manual AMiS-60  communique - Ville de Chassieu  

Copyright © All rights reserved. Failed to retrieve file