Home

(DPLR) Code Version 4.01.0 User Manual

1. 2 6 09 Using DPLR iover ioint XXXXX ispace dxmin slength nxtot C ntx nty ntz iconr isim ifree initi ibadpt iflx iord omgi ilim idiss epsi jf1lx jord omg jlim jdiss epsj kf1lx kord omgk klim kdiss epsk iextst nrlx ildir ibcu iblag ilt ibdir cflm Boundary condition type ibc imin imax jmin jmax kmin kmax irm density M Re V cx cy cz Tin Trin Tvin Tein turbi tkref subp0 subT0 pback cs Species order N2 02 NO N O DPLR Code Version 4 01 0 User Manual 4 4 2 6 09 Using DPLR Step 3 Save the input deck file Action At the command line type save yourdplrinputfilename inp Result The input deck for your problem is saved Step 4 Run DPLR Action At the command line prompt type mpirun np X machinefile machine inp Spath dplr2d or dplr3d lt yourdplrinputfilename inp Result DPLR performs the simulation on X number of processors for the number of iterations specified in the input deck to achieve a solution During the run diagnostic output called the standard out STDOUT is echoed to the screen to provide feedback on the action s being performed including any warning messages If a fatal error is encountered a descriptive message will be displayed in the STDOUT and the run will terminate When a specified convergence level is reached or you halt the run an output solution or restart file ps1x is created along with an on screen run summary See Section 4 3 fo
2. Tech Tips 1 You can adjust the CFL number during a DPLR run by using a runtime control ctrl file See Section 6 4 2 You can use exact timesteps At instead of CFL numbers by entering negative numbers into the CFL Number Listing area of the DPLR Input Deck however you cannot use both CFL numbers and At values in a single DPLR simulation Note that the term 1 in the CFL number list tells DPLR to stop reading the CFL number list and refer to the istop flag for the final iteration number DPLR Code Version 4 01 0 User Manual 4 52 2 6 09 Using DPLR 4 3 Neptune Sample Case The sample case used throughout the DPLR Code User Manual to illustrate how the Code Package works describes a Neptune entry type probe with an ellipsoidal body as shown in Figure 4 1 This case is an example of aerocapture where drag from the atmosphere is used to decelerate the vehicle and bring it into orbit Figure 4 1 Neptune Probe 4 3 1 Neptune DPLR Input Deck The DPLR input deck below shows the problem specific entries made before running the initial DPLR run was made DPLR Code Version 4 01 0 User Manual 4 53 2 6 09 INPUT DECK FOR DPLR2D DPLR3D CODE gname fname bname rname cname dname neptune 8PE neptune none none none sdel fs twhite SF dpcodeV4 00 0 cfdinput neptune5sp leibowitz76 chem nblk itrmod 0 istop 500 igdum 1 xscale 1 0d0 itime 0 ifstat iaero
3. 64 ny 64 nz The summary information states that the problem has been decomposed or hardwired for execution on seven processors but there are only two physical master blocks that must be considered during the problem setup Using this strategy once a DPLR or POSTFLOW input deck has been created for a given problem the same input deck can be used regardless of actual the number of processors employed DPLR Code Version 4 01 0 User Manual 8 18 2 6 09 8 3 4 Appendices in the solution This means that you are not required to visualize or work with the parallel decomposition of the problem except when running FCONVERT If you select a parallel archival output format for the decomposed file ouform 1 11 21 a single output file will be created This type of file actually contains only as many master blocks as specified in the original input grid file but additional information is written to the file header to tell DPLR how to perform the appropriate decomposition at run time This virtual decomposition information is written only to the grid file header Therefore parallel archival restart files never need to be decomposed or recomposed Once a parallel archival grid file has been created it is considered to be hardwired for a given number of processors This will be reflected in the output messages produced when FCONVERT is run If you want to run or restart the problem on a different number of processors t
4. It is important to understand that DPLR2D and DPLR3D are separate codes and even though many common subroutines are shared each code requires properly dimensioned input A common misconception is that DPLR2D reads a three dimensional grid file with the third dimension set to 1 and all z coordinates set to zero This is not the case When you prepare a plot3d grid for solution by DPLR2D your grid must be in 2D format If a three dimensional grid is read as input to FCONVERT with idim 2 the results will be unpredictable and probably not what you intended DPLR Code Version 4 01 0 User Manual 8 11 2 6 09 8 3 8 3 1 Appendices Parallel Decomposition This Appendix offers a detailed discussion with several examples of the parallel decomposition process performed by FCONVERT on computational grids submitted to the DPLR Code Package for processing DPLR is a distributed memory parallel code so all blocks in a computational grid are computed simultaneously rather than sequentially Multi block information transfer is handled through MPI data constructs so simulations must be run on at least as many processors as there are master blocks in the original computational grid Because running on more processors than master grid blocks is often advantageous in terms of solution speed large blocks can be split decomposed into smaller pieces to increase computational efficiency and decrease turnaround time This decomposition if required
5. The output will be an ASCII format file with the suffix los DPLR Code Version 4 01 0 User Manual 5 40 2 6 09 5 4 8 Using POSTFLOW When using this option you must specify single body normal line of sight either with the iexbc flag e g iexbc 14 will extract the stagnation line of an axisymmetric body or by specifying a 1D line for extraction with the Tecplot specifier flags Because RADEQUIL requires a certain set of species mole fractions in a certain order POSTFLOW will compare the input mole fractions with the expected set in RADEQUIL and reorder as necessary Species expected by RADEQUIL that are not in the current CFD dataset will be filled in with zeros as required The resulting file should then be ready for direct processing in RADEQUIL NaN s Not A Number POSTFLOW can extract the locations of any NaN s in the restart file to help you determine where the simulation begins to diverge Although rarely used in practice this option can be a handy tool to use in locating the occasional evil bug You can accomplish this by setting ouform 18 The output data generated by this operation consists of a list of ijk locations of all NaN s in the volume listed block by block However the results are only written to the screen in the standard out STOUT not to an output datafile Tech Tip Note that once a NaN is generated by DPLR it will quickly be convected throughout the solution domain so if you want
6. gname bname DPLR Code Version 4 01 0 User Manual Specifies the name of the restart file to process Required Specifies the name of the output file to create Required Specifies the name of the grid file to process Optional Only needed if ingrid is NOT set to 0 Specifies the name of the boundary condition file to process Optional Only needed if ingrid is NOT set to 0 5 25 2 6 09 5 3 5 3 1 Using POSTFLOW Neptune Sample Case The sample case used throughout the DPLR Code User Manual to illustrate how the Code Package works describes a Neptune entry type probe with an ellipsoidal body as shown in Figure 1 This case is an example of aerocapture where drag from the atmosphere is used to decelerate the vehicle and bring it into orbit a eS SS SS SOS SS S N WNS A Meee Figure 1 Neptune Probe Neptune Input Deck The input deck below shows two problem specific entries to make for POSTFLOW to process a restart file generated during the DPLR simulation of the Neptune case One input deck focuses on data generated at the surface of the probe postsurf inp and one examines data on the pitchplane postpitch DPLR Code Version 4 01 0 User Manual 5 26 2 6 09 Using POSTFLOW Input file for postflow imemmode itruev 2 1 inrest ingrid 11 0 nzones 10 interp 19 ivarp dataset istat 0 inbcf ouform iwrtd 0 26 0 isep istyp iunits 0 1 1 zmc imrx imry
7. ifree DPLR Code Version 4 01 0 User Manual Specifies the total number of computational cells in the i jJ k directions for a block Should be set to the number of interior cells not including dummy or ghost cells Note that ntz is only used for 3D flow simulations Specifies how to initialize this block for simulation when then global modeling flag iinit 10 Allowable values are 0 Start by initializing to the values in the freestream specification identified in ifree Restart from saved file 2 Start with a stagnant interior at low pressure Start with artificial boundary layer in place Includes or excludes a master block from the simulation Allowable values are 0 Do not include this block in the simulation 1 Include this block in the simulation Tech Tip Because excluding blocks does not save on the computational intensity of the simulation this setting is only used when you want to freeze problem blocks while the remainder of the solution is allowed to converge Identifies the number of the freestream specification to use for this block 4 37 2 6 09 initi ibadpt ijk f 1x ijk ord DPLR Code Version 4 01 0 User Manual Using DPLR Identifies the number of the freestream specification to use to initialize the interior of this block Indicates whether grid adaption will be applied to a specific block Allowable values are 0 Do not perform adaption on this block Not Working in DPLR
8. pgrx grid file dplr inp file gt convergence file DPLR log file standard out adapt grid to capture shock pslx restart solution file inp file upsequence restart postflo He POSTFLOW dat plt Tecplot input file standard out Flow Simulation Graphic DPLR Code Version 4 01 0 User Manual 7 3 2 6 09 7 1 1 DPLR Workflow Initial Simulation Run Geometry of Interest GridGen plot3d grid file This first time you run a flow simulation problem for a geometry of interest most likely a vehicle of some sort using the DPLR Code Version 4 01 0 Package you will need to generate a new or adapt an existing structured computational grid that you believe will capture the shock wave the object will encounter at hypersonic speeds within a specified flow environment This first guess can be created from specifications in a CAD design or from scratch In either case you will need to use a grid generation program such as GridGen or GridPro to develop the grid file for use with DPLR In most cases the preferred form of the structured grid file you create will be plot3d plot3d grid file feonvert inp FCONVERT perx grid file standard out Once you have created your structured grid file you will need to create an input file for FCONVERT that specifies among other things how many processors will be used to run your simulation and how your grid file sh
9. 2 Second order upwind biased 3 Third order upwind biased recommended value Specifies the value of w as defined by Yee to employ in the MUSCL extrapolation scheme Recommended value 3 Tech Tip DPLR will automatically reset this value to 2 for second order simulations Specifies the type of flux limiter to use in the Euler flux extrapolation Allowable values are 1 Minmod recommended value 2 Superbee 3 Van Albada Tech Tip The Superbee and Van Albada flux limiters while somewhat less dissipative are also less stable and should only be used when low dissipation schemes are actually necessary to obtain highly accurate solutions such as reactive mixing layer flows Specifies the type of eigenvalue limiter to use in the Euler flux extrapolation Allowable values are 0 No added dissipation recommended value for body normal direction 1 Standard eigenvalue limiting recommended value for radial and circumferential directions 2 Standard eigenvalue limiting on linear fields only Standard eigenvalue limiting on non linear fields only Tech Tip Normally eigenvalue limiters should be used in the radial and circumferential flow directions but should be avoided in the body normal direction when possible to avoid adding dissipation in the boundary layer However in those very rare instances when a normal direction limiter can be helpful set ijk diss 3 to apply the limiter only to fluxes with non linear eigenva
10. Specifies whether multiple output datasets are to be written to a single or multiple files Allowable values are 0 all active output datasets are written to a single file 5 8 2 6 09 istyp iunits lref aref xmc ymc zmc DPLR Code Version 4 01 0 User Manual Using POSTFLOW 1 each active output dataset is written to its own file Specifies how to extract boundary condition data with iexbe Allowable values are Not working in DPLR 4 01 0 1 extract entire volume of data for each boundary condition 1 extract a single plane of data for each boundary condition 2 extract 2 planes of data for each boundary condition Specifies whether a POSTFLOW will include SI units associated with data in output files formatted for Tecplot ouform 5 6 25 26 Allowable values are 0 do not include units in the output file l include SI units in the output file Tech Tip If outformis set to create output files NOT specifically formatted for Tecplot this flag will be ignored Specifies the reference length in SI units meters used for the normalization of moment coefficients Tech Tip The extraction of moments and moment coefficients can either be performed directly in POSTFLOW or with an included utility program Moment If Moment is used the value you enter into Iref will be passed to the ultility for use in computation Specifies the reference area in SI units square meters used for the normalization of for
11. a grid file that was previously decomposed by setting iact ion 3 nborig number of blocks in the recomposed file and init 1 Tech Tip Although FCONVERT will recompose an input grid file it does not recreate the zonal interface file for the recomposed file Therefore it is important to save the original interface file to avoid having to recreate it after the recompose is completed Mesh Sequencing Computational grids composed of a large number of data points typically take longer to solve than grids with fewer points As a result grids used for initial solutions of CFD problems are sometimes coarsened or sequenced to reduce the number of points while maintaining the topology of the mesh After an acceptable first guess is acquired the grid is restored in a step wise fashion to its original number of points for final solution and post solution data reporting Also there may be a problem specific advantage to obtaining a solution on a coarser mesh as in the case of wake flow problems or for performing grid convergence studies For both these reasons an option is included in FCONVERT to sequence coarsen a grid radiation boundary condition or restart file and create a new output file that maintains point matching fidelity Sequencing an Input Grid To sequence an input grid set imseq 1 then enter a sequencing factor in iseq jseq and kseq for each master block in the grid A sequencing factor of n implies tha
12. if not entered Specifies the output file name This is the DPLR readable file that will be created by FCONVERT to be solved by DPLR The filename should be surrounded by single or double quotes and can be specified with either a relative or an absolute path as shown in the example below XDRParallelgridfilename pgrx 3 12 2 6 09 nsin nerin nevin necin DPLR Code Version 4 01 0 User Manual Using FCONVERT Tech Tip The suffix used in the file name is optional FCONVERT will assume the default suffix for the file type specified in ouform if not manually entered See Appendix A for a list of file types and associated default suffixes Note If the output filename with suffix is the same as the input filename the input file will be overwritten not a typically desired result Also if an output interface file is requested by setting ouint 1 the suffix inter will be appended to the prefix specified by oname Specifies the number of chemical species to be considered in the CFD solution This is only read if you are trying to create a restart file from an input file other than DPLR s pslx format so that FCONVERT can correctly determine the location of the velocity components in the file Specifies the number of unique rotational temperatures energy conservation equations to be considered in the CFD solution This is only read if you are trying to create a restart file from an input file other t
13. opt intel fce 9 1 037 lib opt ompil 1 2 lib gt LFLAGS home atipa hpl libgoto opteron 64 r0 99 3 so home atipa hpl xerbla o home lib working tecio64 a usr lib gcc x86_64 redhat linux 3 4 5 libstdctt a gt CPPFLAGS cpp D_i686linuxipf gt FFLAGS r8 extend_source 03 pad ip W0 cm gt F77 opt ompil 1 2 bin mpif90 gt FXDRLIB home lib libfxdr a DPLR Code Version 4 01 0 User Manual 4 68 2 6 09 Using DPLR Summary of enabled CPP compiler directives gt AMBIPOLAR 1 gt PARKTEXP 0 50 gt NOHTC WARNING CPP macro AMBIPOLAR 1 uses a simplistic model for ambipolar diffusion INFORM Compiled for 32 bit compatible execution Overset Logic is disabled Dual time stepping is disabled Neptune Mechanism 5 species 5 reactions Liebowitz 1973 amp 1976 Model gt Species List H2 H H He e gt Reaction rates from neptune5sp_ leibowitz76 chem gt Reaction Status 11111 gt Keq Fit Used 00000 gt Park 1990 fits for Keq n 10 16 gt Assume molecules created destroyed at mixture Tve Catalytic wall BC enabled gt Constant accomadation coeff gamma 1 000 gt Fully catalytic to ion recombination Radiative equilibrium BC enabled gt Constant wall emissivity epsilon 0 85 gt Maximum wall temperature 3000 00 K Rotational Equilibrium Fully Excited Vibrational Equilibrium SHO Electronic Energy Neglec
14. 3 16 2 6 09 Using FCONVERT Input Flag Setting Explanation cont cont cont iseq jseq kseq 1 1 1 Values ignored imseq 0 1 1 1 iname neptune The name of the input grid file is neptune xname neptune The name of the input zonal interface file is neptune cname none When iaction 1 FCONVERT breaks master blocks according to the values in ibrk jbrk kbrk and ignores information in from a CFD input deck file oname neptune 8PE The name of the output DPLR readable grid file is neptune 8PE file convention that notes how many processors were used to run the problem in this case 8 nsin 5 Ignored value not trying to create a restart file from a non DPLR pslx input file nerin 0 Ignored value not trying to create a restart file from a non DPLR pslx input file nevin 1 Ignored value not trying to create a restart file from a non DPLR pslx input file necin 0 Ignored value not trying to create a restart file from a non DPLR pslx input file ntbin 0 Ignored value not trying to create a restart file from a non DPLR pslx input file 3 3 3 Neptune Output Summary When you run FCONVERT See Section 3 1 Step 4 the program provides an on screen summary of the actions performed along with some supplemental information as shown below DPLR Code Version 4 01 0 User Manual 3 17 2 6 09 Using FCONVERT fconvert
15. 98 turbulent Schmidt number Sc_t 99 turbulent Prandtl number Pr_t Mixture Flow Properties Note that stagnation quantities density pressure and temperature are computed assuming isentropic relations and thus are not valid for a chemically reacting flowfield 100 mixture density rho 101 mixture number density N_ tot 102 stagnation mixture density r_0 110 pressure p 111 dynamic pressure Q 112 stagnation pressure p_o 113 Pitot pressure p_pitot 114 pressure coefficient C_p 120 translational temperature T 121 bulk temperature T_b 122 stagnation temperature T_o 124 rotational temperature Tr 125 vibrational temperature Tv 126 electronic temperature Te 127 free electron temperature Tel 132 total enthalpy per unit mass h 5 14 2 6 09 DPLR Code Version 4 01 0 User Manual 133 134 135 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 175 180 Using POSTFLOW static enthalpy per unit mass h_s total enthalpy per unit volume rh static enthalpy per unit volume rh_s total energy per unit mass e total translational energy per unit mass et total rotational energy per unit mass er total vibrational energy per unit mass ev total electronic energy per unit mass ee total free electron energy per unit mass eel total chemical formation energy per unit mass eh total kine
16. Catalytic radiative equilibrium wall with slip Not Working in DPLR 4 01 0 Non catalytic radiative equilibrium wall with slip Not Working in DPLR 4 01 0 Viscous adiabatic wall with slip 4 44 2 6 09 60 61 62 Using DPLR Input primitive variables ps u v w Tv T Input primitive variables p cs 2 ns u v w Tv T Input conserved variables ps pu pv pw Ev E Tech Tip Entering one of these numbers will tell DPLR to look in the pcba file to find values for the indicated variables If such a file does not exist a runtime error may occur 70 71 75 76 85 DPLR Code Version 4 01 0 User Manual Input species mass flow rate amp T extrapolate for p with thermal equilibrium assumed Input c mass flow rate amp T extrapolate for p with thermal equilibrium assumed Activate surface kinetic mechanism isothermal icatmd 1001 Activate surface kinetic mechanism radiative equilibrium icatmd 1001 Subsonic reservoir inlet same as 6 Subsonic inlet specify mass flow rate density M ReV normal velocity amp T extrapolate p Subsonic exit specify static pressure pback extrapolate others 4 45 2 6 09 DPLR Code Version 4 01 0 User Manual Using DPLR 86 Subsonic inlet specify subsonic temperature subTO and subsonic pressure subp0 Developmental in DPLR 4 01 0 87 Subsonic exit specify static pressure pback extrapolate others Uses method of characteris
17. Controls the output of debugging information Allowable values are 0 Do not output debugging information 1 Output debugging information Tech Tip This information is intended for software developers Specifies the DPLR Code version of the output file Allowable values are 1 Do not attempt to change file version 2 Upgrade file to the current version of DPLR Code 3 Convert file to the DPLR Code version specified by nvers 3 5 2 6 09 nvers inform inint DPLR Code Version 4 01 0 User Manual Using FCONVERT Tech Tip All parallel grid restart boundary condition and radiation files are DPLR Code Version specific With the exception of boundary condition files ivers allows the format conversion of all supported files between all release versions of the DPLR Code Package Specifies the DPLR Version to convert the output file to when ivers 3 Allowable values are the real numbers of the major and minor releases of the DPLR Code Package from 2 31 through 4 01 0 Specifies the format of input file See Appendix A for more information about supported I O formats Allowable values are Unformatted parallel file Unformatted plot3d grid or q file 3 Unformatted plot3d grid or function file 11 XDR parallel file 21 ASCII parallel file used for debugging 22 ASCII plot3d grid or q file used for debugging 23 ASCII plot3d grid or function file used for debugging Tech Tip Because DPLR is a double
18. T To 5 Two temperature model using LeRc curve fits Ty Te STe T T NOT WORKING in DPLR 4 01 0 11 Te Independent VOT WORKING in DPLR 4 01 0 irad Specifies the model used for shock layer radiation modeling Allowable values are 0 No radiation model for weakly radiating flow fields DPLR Code Version 4 01 0 User Manual 4 19 2 6 09 DPLR Code Version 4 01 0 User Manual Using DPLR 1 Read pointwise A Q froma file use only if rname is defined 2 Optically thin emission Carbon Nitrogen Violet 3 Optically thin emission Carbon Nitrogen Red 4 Optically thin emission Carbon Nitrogen Violet Red 102 198 Same as 2 98 but with input surface heating information read from the rname file Tech Tip 1 DPLR does not compute shock layer radiation directly but several hooks are provided for coupling either loosely to external radiation transport codes or tightly for optically thin emission Typically for weakly radiating flowfields shock layer radiation is either neglected or computed in an uncoupled manner For these cases irad 0 2 If the radiation field is known to be optically thin DPLR supports tight coupling by computing the A gt Q source term at each volume cell using curve fits generated by comparison to more exact computations Currently DPLR supports this option for CN radiation only irad 2 4 In this case DPLR reads the curve fit coefficients from the file emission rad from th
19. Transition at specified constant x value 2 Transition at specified constant y value 3 Transition at specified constant z value Tech Tip Setting i trans to a positive value implies transition proceeds with increasing ordinate while setting it to a negative value implies that transition proceeds with a decreasing ordinate Specifies the transition onset location This flag is used whenever a turbulent flowfield ivis 2 12 with an input transition model itrmod 1 2 3 is chosen trloc is a real dimensional number tied to the value of itrans For example if itrans 1 and trloc 2 5 DPLR will initiate transition at a value of x 2 5m with turbulent flow for larger values of x and laminar flow for smaller values Specifies the extent of transition to turbulent flow This flag is used whenever a turbulent flowfield ivis 2 12 with an input transition model itrmod 1 2 3 is chosen trext is areal dimensional number equal to the width of the TANH function from 0 01 0 99 The other transition models do not permit user modification of the transition length so the trext flag is not used in those cases Specifies the model used for shock unsteadiness modeling when itrmod 100 Allowable values are 0 No shock unsteadiness model 1 Model of Sinha et al AIAAJ V43N3 Mar 2005 NOT WORKING in DPLR 4 0 Specifies the number of iterations to run before stopping the simulation This is a relative value For example if a simulation is start
20. be formed in the direction normal to the last body surface detected Specifies how often to update the implicit boundary conditions during the relaxation process for DPLR or FMDP iextst 1 or 2 Recommended value 1 Tech Tip Updating implicit boundary conditions improves parallel efficiency on machines for which message passing is very inefficient ibcu 1 forces the implicit boundary conditions to be updated during each relaxation step Specifies whether to lag the implicit boundary conditions when using DPLR or FMDP iextst 1 or 2 Allowable values are 4 4 2 6 09 ilt ibdir DPLR Code Version 4 01 0 User Manual Using DPLR lor0 Do not lag implicit boundary conditions recommended value 1 Lag implicit boundary conditions Tech Tip Although it is generally desirable to lag the implicit boundary condition update in order to better mask the message passing latency and improve parallel performance of the time advancement method there are certain instances when the block topology employed makes lagged boundary conditions dangerous In the future DPLR may automatically determine whether latency can be masked for a given application and this flag will be automatically set by the code Specifies whether to employ global or local time stepping for implicit simulations Allowable values are 1 Global time stepping 2 Global time stepping with maximum CFL limit Local time stepping 2 Local time stepping
21. igalign 0 imax igrid 11 ikt 1 ireqmd 101 ikeq 3 itrans 0 nplot 100 kb1 0 ngiter 500 DPLR Code Version 4 01 0 User Manual irest ll ikv 11 twall 3 0d3 ivib 2 trloc 1 0d0 iplot 1 kdg 0 Le Sc 5d0 rvr 1 3d0 ibcf 0 ivmod 3 2 trext 0 1d0 iaxi 1 istate 0 LeT ScT 1 00d0 resmin 1 0d 20 tfinal 9 0d99 nalign 1 4 54 Using DPLR iinit 1 iradf nfree islip iblow 0 1 vwall 0 irad 0 2 iresv 1 prtl 0 72d0 prtlT 0 90d0 2 6 09 Using DPLR imedge imradial ngeom ismooth 1 2 2 3 fs_scale ds_mult gmargin 0 95 2 5 0 0 dsl cellRe ds 1mx ds2fr 0 0 1 0 1 0d 4 0 3 iover ioint XXXXX ispace dxmin slength nxtot 0 1 0d 5 1 0d0 1000 BLOCK 1 ntx nty ntz iconr isim ifree initi ibadpt 32 16 64 1 1 1 1 1 iflx iord omgi ilim idiss epsi 4 3 2 0d0 1 1 0 3 jflx jord omg j jlim jdiss epsj 4 3 2 0d0 1 1 0 3 kflx kord omgk klim kdiss epsk 4 3 2 0d0 1 0 0 03 iextst nrlx ildir ibcu iblag ilt ibdir cflm i 4 0 1 1 lt 1 1 1 0d20 Boundary condition type ibc imin imax jmin jmax kmin kmax 20 20 20 19 26 1 BLOCK 2 ntx nty ntz iconr isim ifree initi ibadpt 48 64 64 ip 1 1 1 1 iflx iord omgi ilim idiss epsi 4 Si 2 0d0 1 1 0 3 jflx jord omgj jlim jdiss epsj 4 3 2 0d0 1 1 0 3 DPLR Code Version 4 01 0 User Manual 4 55 2 6 09 kf
22. is a region where two grid blocks abut sharing the same grid points Information about these areas of abutment in the form of an ASCII zonal interface file must be provided as input for DPLR to ensure that data are mapped correctly across grid blocks during the CFD computation Zonal interface files can be prepared in three ways e manually through direct observation of the serial plot3D input grid init 1 e automatically by FCONVERT init 2 4 e automatically by using the TEMPLATE utility See Section 8 1 5 When the plot3D input grid describes a relatively simple set of master blocks and resulting zonal boundaries developing the information for a zonal interface file by hand may be a straightforward way to proceed However when a multi block input grid contains a large number of blocks or describes complex geometries as is often the case in aerospace problems using FCONVERT or TEMPLATE to develop the detailed data required to accurately describe each zonal interface is likely to be the more productive approach Creating Zonal Interface Files by Hand Step 1 Examine the plot3D input grid to determine the location of all zonal interfaces making note of how the points in each master block abut those in another block Step 2 Open the text editor program for your system and create an ASCII file with the format shown below ZONAL BOUNDARY INFORMATION Cell Matching No dummy cells zvers izdum nblk ninta nintc Zonal Boun
23. l Parallel archival file native unformatted 11 Parallel archival file XDR format Recommended 21 Parallel archival file ASCII DPLR Code Version 4 01 0 User Manual 4 7 2 6 09 irest ibcf iradf nfree iinit DPLR Code Version 4 01 0 User Manual Using DPLR Specifies the format of the restart file fname Allowable values are 1 11 21 Parallel archival file native unformatted Parallel archival file XDR format Recommended Parallel archival file ASCII Specifies the format of the boundary condition BC file bname if any Allowable values are 0 1 11 21 Do not read a BC file Parallel archival file native unformatted Parallel archival file XDR format Parallel archival file ASCII Specifies the format of the input radiation file rname if any Allowable values are 0 1 11 21 Do not read a radiation file Parallel archival file native unformatted Parallel archival file XDR format Parallel archival file ASCII Indicates the number of freestream specifications i e areas of the freestream with preconfigured flow conditions that are characterized in the DPLR input deck Specifies how to initialize the simulation Allowable values are l or 0 10 11 Start all blocks by initializing to freestream values in the specification identified in the block Restart from saved file Start with a stagnant interior at low pressure Start with artificial boundary la
24. recombination pathways Enhanced Mitcheltree CO2 model additional recombination pathways in this model may make it more appropriate for higher flow enthalpies 4 13 2 6 09 ireqmd DPLR Code Version 4 01 0 User Manual Using DPLR 300 Generalized CO catalysis Bose Wright used in sensitivity analyses not design work Tech Tip Catalysis refers to how the wall surface facilitates chemical reactions that can deposit energy on the vehicle surface during flight Much work is currently being done to enhance empirical knowledge of material responses under hypersonic flight conditions in a variety of atmospheric flows As the materials knowledge base increases DPLR can be used to simulate how the overall flow environment contributes to and in turn is affected by chemical reactions on flight vehicle surfaces Specifies the model to be used for surface radiative equilibrium Allowable values are 0 Disable radiative equilibrium 1 Constant emissivity model value for set with the espr flag 3 98 Material specific emission rates for different materials are experimentally obtained and given in the emission rad file in the cfdinput directory 99 Input material map allows DPLR to access an emission radiation map specifving pointwise material properties in a boundary condition file 101 199 Material specific with a maximum wall temperature used to model physical material temperature limits duri
25. 0949864923525E 01 cfl 1 0E 00 nit 198 rmsres 1 5724959884754E 02 cfl 5 0E 01 nit 199 rmsres 1 5236815979400E 02 cfl 5 0E 01 nit 200 rmsres 1 4743944713869E 02 cfl 5 0E 01 writing restart file neptune pslx solution written at Thurs Feb 5 07 35 27 2009 nit 201 rmsres 1 4303718730322E 02 cfl 1 0E 02 nit 202 rmsres 1 7744774664322E 02 cfl 1 0E 02 nit 203 rmsres 1 7218162636161E 02 cfl 1 0E 02 nit 298 rmsres 2 3267700178866E 05 cfl 1 0E 03 nit 299 rmsres 2 1545250679943E 05 cfl 1 0E 03 nit 300 rmsres 2 0022263684723E 05 cfl 1 0E 03 writing restart file neptune pslx solution written at Thurs Feb 5 07 49 42 2009 nit 301 rmsres 1 8676626863417E 05 cfl 2 0E 03 nit 302 rmsres 1 9613965114431E 05 cfl 2 0E 03 nit 303 rmsres 2 1038915489521E 05 cfl 2 0E 03 DPLR Code Version 4 01 0 User Manual 4 70 2 6 09 4 3 4 4 4 Using DPLR nit 398 rmsres 1 4656581391990E 08 cfl 5 0E 03 nit 399 rmsres 1 3612594864547E 08 cfl 5 0E 03 nit 400 rmsres 1 2656218641573E 08 cfl 5 0E 03 writing restart file neptune pslx solution written at Thurs Feb 5 08 03 56 2009 nit 401 rmsres 1 1780771213662E 08 cfl 5 0E 03 nit 402 rmsres 1 0979147316707E 08 cfl 5 0E 03 nit 403 rmsres 1 0245683275129E 08 cfl 5 0E 03 nit 498 rmsres 2 1745095792356E 10 cfl 5 0E 03 nit 499 rmsres 2 0982611245646E 10 cfl 5 0E 03 nit 500
26. 16 1 16 64 64 64 The ending point in the first extent direction for one block in the first zonal interface is 16 and for the abutting block it is 1 The ending point in the first extent direction for one block in the second zonal interface is 16 and for the abutting block it is 64 The ending point in the first extent direction for one block in the third zonal interface is 64 and for the abutting block it is also 64 DPLR Code Version 4 01 0 User Manual 6 7 2 6 09 DPLR Codes Package Input Output Files Input Value Setting Explanation cont cont cont ndr2 3 3 3 3 1 2 The second extent direction for the first interface zone is k The second extent direction for the second interface zone is also k The second extent direction for the third interface zone is i to j nst2 1 1 1 1 1 17 The starting point in the second extent direction for one block in the first zonal interface is 1 and for the abutting block it is also 1 The starting point in the second extent direction for one block in the second zonal interface is 1 and for the abutting block it is also 1 The starting point in the second extent direction for one block in the third zonal interface is 1 and for the abutting block it is 17 nen2 64 64 64 64 The ending point in the second extent direction for one block 32 48 in the first zonal interface is 64 and for the abutting block it is also 64 The ending point in the second extent direction
27. 4 0 itrmod Specifies the model used for turbulence transition modeling This flag is used whenever a turbulent flowfield is specified by ivis 2 12 Allowable values are 0 Neglect transition flow is fully turbulent 1 TANH transition function 2 Dhawan and Narashima model 3 Sigmoid function 100 Input transition map allows simulation of local turbulent regions in a laminar flow Tech Tip Although some turbulence models in DPLR are capable of predicting transition the code also allows you to use several models to impose transition in locations of your choosing When itrmod 1 2 3 the location and extent of the transition regions are defined by the values you put into itrans trloc and trext When itrmod 100 you have the option of creating a transition map consisting of a turbulence intensity value ranging from 0 fully laminar to I fully turbulent at each surface point and placing the information into a boundary condition file bname See Section 6 3 for more information on boundary condition files Remember however choosing any of these models will impose and not predict turbulent transition s in your simulation itrans Specifies the ordinate of the transition onset location This flag is used whenever a turbulent flowfield ivis 2 12 DPLR Code Version 4 01 0 User Manual 4 21 2 6 09 trloc trext itshk istop Using DPLR with an input transition model itrmod 1 2 3 is chosen Allowable values are
28. 4 01 0 User Manual 8 5 2 6 09 Appendices layer of blocks For Shuttle Orbiter applications including local grids around damage and repair the process has been fully automated with the help of ancillary input files Using Template To generate all or most of the two DPLR control files using Template perform the following steps in the working directory containing your grid in PLOT3D multiblock form formatted or unformatted Caution Existing dp1r inputs dplr inputs 2 or dplr interfaces file will be overwritten Step 1 Step 2 Step 3 Step 4 Optional Copy the generic inp file in the cfdinput directory as sample inputs See below for ancillary input file details Optional Copy the template inp 2 file from the utilities directory See details below Run TEMPLATE You will be prompted for the name of the PLOT3D grid fileand a tolerance to use in its detection of matching faces Check the outputs listed below Note that some boundary conditions may need changing while the free stream flow conditions CFL schedule and flow solver iteration limit typically need editing Your working directory now contains five new files dplr inputs file containing the block specific mid section of your DPLR input deck boundary conditions etc or possibly all of the input deck depending on Step 1 above dplr inputs 2 variant of dplr inputs intended for possible grid sequencing dplr int
29. 620 621 622 623 624 625 626 650 651 652 653 660 661 662 663 670 671 672 673 700 701 702 703 710 711 712 713 DPLR Code Version 4 01 0 User Manual Using POSTFLOW viscous force on a face in all directions viscous force on a face in x direction Fx_V viscous force on a face in y direction Fy_V viscous force on a face in z direction Fz_V viscous force on a face in x direction per unit area Fx_Va viscous force on a face in y direction per unit area Fy_Va viscous force on a face in z direction per unit area Fz_Va total force coefficient on a face in all directions total force coefficient on a face in x direction Cx total force coefficient on a face in y direction Cy total force coefficient on a face in z direction Cz pressure force coefficient on a face in all direction pressure force coefficient on a face in x direction Cx_P pressure force coefficient on a face in y direction Cy_P pressure force coefficient on a face in z direction Cz_P viscous force coefficient on a face in all direction viscous force coefficient on a face in x direction Cx_V viscous force coefficient on a face in y direction Cy_V viscous force coefficient on a face in z direction Cz_V total moment on a face in all directions total moment on a face in x direction Mx total moment on a face in y direction My total moment on a face in z direction Mz pressu
30. Action At the command line prompt type make Result Assuming there were no problems all executable files in the package are created Links to executables dplr2d DPLR Code Version 4 01 0 User Manual 2 4 2 6 09 Installation Guide dplr3d fconvert postflow are located in the bin directory 2 4 Directory File Contents The directories and files resulting from untarring the DPLR Code Package contain the following components bin cfdinput cfdlib config defs dplib dplr2d dplr3d fconvert include makefile makefile comm post utilities links to compiled binaries physical modeling data files used by DPLR during execution subroutines common to DPLR2D and DPLR3D a configuration script used to set up the makefile for the specific machine architecture makefile templates for supported machines subroutines common to the entire package subroutines unique to the DPLR2D code subroutines unique to the DPLR3D code subroutines unique to the FCONVERT code modules common blocks and other include files that are incorporated into the various executables makefile for the package compiler architecture and library options for building DPLR may be created by config script subroutines unique to the POSTFLOW code utility codes and scripts distributed with the package Tech Tip The contents of each directory distributed with the DPLR Code Package are requir
31. Allows you to impose a constant blowing vwal1 gt 0 or sucking vwal1 lt 0 when used with the iblow flag If iblow 1 expressed as velocity m s If iblow 2 expressed in mass flux per unit area kg m2 s Specifies the model employed for chemical reactions in the gas phase Allowable values are 0 Frozen chemistry no chemical reactions occur in the flowfield 1 Finite rate chemistry Arrhenius style reaction kinetics used to model the chemistry are given in the chem file specified with the dname flag and found in the cfdinput directory Tech Tip DPLR does not currently support equilibrium chemistry 4 15 2 6 09 ikeq DPLR Code Version 4 01 0 User Manual Using DPLR Specifies the model used for computing equilibrium constants This is required when ichem 1 Allowable values are 1 11 19 21 29 31 39 No reverse reactions debugging tool that turns off reverse reactions completely should not be used for actual simulations Park 1985 fits not recommended for use due to potential instabilities in some simulations Mitcheltree 1994 fits not recommended for use due to potential instabilities in some simulations Park 1990 fits n 10 em not recommended for use due to potential instabilities in some simulations Park 1990 fits n 10 cm not recommended for use due to potential instabilities in some simulations Computed from NASA Lewis thermodynamic fits 1994 preferred
32. Tech Tips 1 Formats for the output files will usually be plot3d ouform 3 23 33 standard CFD output formats that can be read by most commercial post processing tools and Tecplot ouform 5 6 25 26 4a file format used only by Amtec s Tecplot post processing visualization software 2 The plot3d q file output option ouform 2 is included primarily for historical purposes Although still technically available writing data to a q file requires the user to know and include the specific variable set specified by that file format With no error checking performed by DPLR data file validity becomes the 5 6 2 6 09 iwrtd interp DPLR Code Version 4 01 0 User Manual Using POSTFLOW entire responsibility of the user 3 To generate Tecplot binary files ouform 5 6 the Amtec provided tecio a or tecio64 a runtime library must be installed on your system Specifies whether a subdirectory called INPUTDECKS containing reconstructions of the DPLR input decks including the physical property data decks used to run the simulation will be created in your working directory Allowable values are 0 do not create a subdirectory containing reconstructed input decks 1 create a subdirectory containing reconstruct input decks recommended Tech Tip One of the more powerful features of POSTFLOW is the ability to recreate usable DPLR input and physical data decks directly from the restart file Because of t
33. Using DPLR Block 1 Flags Setting Explanation ntx 32 There are 32 computational cells in the x direction in Block 1 nty 16 There are 16 computational cells in the y direction in Block 1 ntz 64 There are 64 computational cells in the z direction in Block 1 iconr 1 This value is ignored because iinit 1 isim 1 This block will be included in the simulation ifree 1 Use freestream specification 1 for this master block initi 1 Use freestream specification 1 to initialize the interior of this master block ibadpt 1 Grid adaption will be performed on this block iflx 4 The Euler flux extrapolation method to use in the i direction is MUSCL Steger Warming with Ap iord 3 The Euler flux extrapolation order of accuracy is third order upwind biased omgi 2 0d0 The value of w to employ in the MUSCL scheme is 2 ilim 1 The MinMod flux limiter is used in the Euler flux extrapolation idiss 1 A standard eigenvalue limiter is used in the flux extrapolation epsi 0 3 The magnitude of the eigenvalue limiter is 0 3 in the flow direction jflx 4 The Euler flux extrapolation method to use in the j direction is MUSCL Steger Warming with Ap jord 3 The Euler flux extrapolation order of accuracy is third order upwind biased omgj 2 0d0 The value of w to employ in the MUSCL scheme is 2 jlim 1 The MinMod flux limiter is used in the Euler flux extrapolation jdiss 1 A standard e
34. are several other possible decompositions that would achieve the same result The sample output for this case would be identical to that shown in the previous example The choice of using iaction 1 or 2 is really dependent on the situation For example iaction 1 can be used prior to generation of the DPLR input deck In addition iaction 1 gives you more direct control over the decomposition performed Because it is preferable for the sake of efficiency to decompose the grid so that the generation of additional zonal interfaces is minimized using iaction 1 and manually specifying the decomposition strategy can help you meet this condition For example in the previous test problem a decomposition strategy of Decomposition information for each master block ibrk jbrk kbrk 1 1 1 6 1 1 would result in the following output Decomposing block 1 into 1 ibrk 1 jbrk 1 kbrk 1 Decomposing block 2 into 6 ibrk 6 jbrk 1 kbrk 1 creating 7 total blocks 7 Blocks Total load imbalance 6 49 Output Block 1 size il 16 jl 32 kl 128 65536 cells Output Block 2 size il 11 jl 64 kl 128 90112 cells Output Block 3 size il 11 jl 64 kl 128 90112 cells DPLR Code Version 4 01 0 User Manual 8 16 2 6 09 Appendices Output Block 4 size il 11 jl 64 kl 128 Output Block 5 size il 11 jl 64 kl 128 64 kl 128 64 kl 128 90112 cells 90112 cells 81920 cells 81920 cells Output Block 6
35. areas this way the total given in nfree and use them to initialize master blocks and set freestream boundary conditions on a block by block basis using the initi and ifree flags respectively In DPLR freestream values are always expressed in standard international SI units irm Specifies whether a velocity Mach number or unit Reynolds number will be given as input Allowable values are 1 Mach number assumed to be the equilibrium as opposed to frozen value 2 Reynolds number per meter Velocity recommended value Tech Tip The most common and least ambiguous input for most free flight simulations is velocity because each of the other entries requires the velocity to be derived from the thermodynamic and transport models employed in the given simulation DPLR Code Version 4 01 0 User Manual 4 48 2 6 09 density M Re V c xyz Tin Tvin Trin Tein turbi DPLR Code Version 4 01 0 User Manual Using DPLR Specifies the input freestream mass density Specifies the input Mach number unit Reynolds number or velocity depending upon the value of irm Tech Tip Whichever choice is made in irm DPLR will determine the remaining two values using the input thermodynamic and transport models distributed with the code Specifies the input velocity vector direction cosine Allowable values are 0 0 lt value lt 1 0 Tech Tip These values must be nondimensionalized Specifies the input translation
36. as a laminar flow simulation and telling DPLR to ignore turbulence and transition related flags trloc 1 0d0 This value is ignored because ivis 1 defining the problem as a laminar flow simulation and telling DPLR to ignore turbulence and transition related flags trext 0 1d0 This value is ignored because ivis 1 defining the problem as a laminar flow simulation and telling DPLR to ignore turbulence and transition related flags itshk 0 Not used by DPLR at this time istop 500 DPLR will run 500 iterations before stopping the simulation nplot 100 DPLR will write a restart file every 100 iterations iplot 1 DPLR will save only the most recently written restart file iaxi 1 This is a non axisymmetric simulation ires 2 Screen output for this simulation will include the iteration number the global residual and the CFL number being used and will include a comparison with these values from the first iteration of the simulation igdum 1 DPLR will compute grid dummy cell coordinates kbl 0 DPLR will ignore this flag kdg 0 DPLR will ignore this flag istate 0 DPLR will use the equation of state for a perfect gas iresv 1 DPLR will track the total density of all species in the simulation xscale 1 0d0 DPLR will not scale the input grid at runtime ils 2 Input numbers governing laminar diffusion coefficients will be interpreted as Schmidt Numbers Le Sc 0 5d0 The Schmidt number to be used in the simulation is 0 5 LeT ScT 1 00d0 This value is ignored
37. at any viscous walls will be reinitialized on restart which will lead to a significant perturbation to the flowfield and L2norm residual Open an FCONVERT input deck file See Section 3 1 and set iaction 10 ifile 2 inform 3 or 23 nsin of chemical species nerin nevin necin of independent temperatures in each mode ntbin of turbulence variables and the rest of the flags to problem specific values Save the input deck file Run FCONVERT DPLR Code Version 4 01 0 User Manual 6 22 2 6 09 6 6 6 7 DPLR Codes Package Input Output Files Tech Tip During the conversion process FCONVERT will generate all necessary header elements and format the file properly for DPLR However the resulting restart file does not contain all of the CFD modeling flags and thus cannot be post processed with POSTFLOW until it has been run at least 1 iteration and re saved in DPLR Chemistry Files The DPLR Code Package contains a large number of chemistry model files that are automatically placed in the cfdinput directory when the software is installed along with other physical property databases See Section 2 4 for more information on the contents of directories installed with the DPLR Code Package Version 4 01 0 Chemistry files contain the input information DPLR needs to define species lists chemical kinetic reactions and reaction rates for a simulation A chemistry file is required input for all simulations and should be specified
38. avoid the need to recreate it after the recompose is completed DPLR Code Version 4 01 0 User Manual 8 21 2 6 09 8 4 8 4 1 Appendices POSTFLOW Output Variables A complete listing of all POSTFLOW output variables is provided in Section 5 2 of this Users Manual This appendix provides additional detailed information about some of these variables The output variables in POSTFLOW are selected via the ivarp integer array where each variable is assigned a unique integer quantity These integers are a superset of those defined in the Plot3d and GASP programs and are expressed either as non dimensional quantities or in SI units Note DPLR does not support English units Grid Related Variables Path Length 11 path length along grid lines in i direction si 12 path length along grid lines in j direction sj 13 path length along grid lines in k direction sk Path length is determined by computing the distance from grid point to grid point in the mesh along the selected coordinate direction For example if ivarp 11 POSTFLOW will compute the path length for each constant i line in the output datasets The path length is assumed to begin at zero for ijk 1 and increases for increasing index Body Normal Distance 21 body normal distance dn The body normal distance at a surface is defined as the distance from the cell center of the first interior cell to the face center on the surface This is the distance used in the firs
39. block faces to see each other and thus describing how energy behaves within a concave surface geometry Not used by DPLR at this time Not used by DPLR at this time Not used by DPLR at this time Specifies the total number of computational cells in the 7 j k directions respectively for a master block Tech Tips 1 Although unlabeled each line of values corresponds to one master block 2 Values also found in the ntx nty and ntz flags in the block specific areas of the DPLR input deck 6 16 2 6 09 DPLR Codes Package Input Output Files ibc numbers for each block Specifies the boundary condition values entered for each face of a master block in the DPLR input deck imin imax jmin jmax kmin kmax respectively Tech Tip Although unlabeled one line of values corresponds to one master block Optional Lines Catalytic material map specifiers if nmc 1 Surface radiation map specifiers if nme 1 Transition map specifiers if nmt 1 View Factor map specifiers if nmv 1 Profile Data for Block Face Listing of variable values specified by DPLR Code Version 4 01 0 User Manual the ibc number for a particular master block face For example if Block 1 Face 6 has a value of 60 DPLR will look at this file in this place for numeric values for ps u v w Tv T Tech Tips 1 For boundary conditions to be accurately simulated the data in this area must appear in the exact order the va
40. data in the restart file being read by POSTFLOW Allowable values are 0 get format from restart file Recommended 1 parallel archival file native unformatted 11 parallel archival file XDR format 21 parallel archival file ASCII Tech Tip POSTFLOW will automatically look at the restart file to determine if any boundary condition file is required and what the format is If no BC file was used during the simulation the value of inbcf is ignored 5 5 2 6 09 ouform DPLR Code Version 4 01 0 User Manual Using POSTFLOW Specifies the desired format of the output data Allowable values are 2 plot3d grid or q file native unformatted 3 plot3d grid or function file native unformatted 5 Tecplot block ordered data binary 6 Tecplot point ordered data binary 7 compute max min values for variables and output to STDOUT 8 integrate variables over given surface s and output to STDOUT 9 RESERVED 10 print selected freestream quantities to STDOUT 11 output datasets for Moment calculations 17 compute max min amp maxloc minloc and output to STDOUT 18 print a list of NaN locations to STDOUT 22 plot3d grid or q file ASCII 23 plot3d grid or function file ASCII 25 Tecplot block ordered data ASCII 26 Tecplot point ordered data ASCII 28 RADEQUIL LOS file ASCII 32 gzipped plot3d grid or q file ASCII 33 gzipped plot3d grid or function file ASCID 110 print freestream quantities to STDOUT in tabular format
41. flag ibc must be an array of comma or space separated entries Entering a value of 1 disables this feature Tech Tips 1 To extract data from the intersection of two surfaces enter a reference boundary value then a forward slash then the boundary condition that is desired to intersect with the reference boundary This expression becomes one value and can then be added to the array of numbers on this input line For example iexbc 26 18 3 would extract data from the intersection between a catalytic radiative equilibrium wall ibc 26 and the y symmetry plane ibc 18 in addition to extracting data from the supersonic exit plane ibc 3 Note that the order of the two numbers is important as the wall must be defined before an intersecting plane of symmetry can be specified 2 If the simulation contains multiple instances of a boundary condition specified in iexbc the resulting data extraction will be saved as separate blocks for a plot3d output file or zones for a Tecplot output file Both designations refer to the same regions in the 5 11 2 6 09 ivarp DPLR Code Version 4 01 0 User Manual Using POSTFLOW simulation and will be named by the iexbc setting For example if iexbc 19 POSTFLOW will display the words zone t BC19 for each block in which that boundary condition is extracted 3 Entering appropriate values into iexbc is a quick and easy way to extract defined surface data from a complex multiblock gr
42. formed for the DPLR time advancement method iextst 1 Allowable values are 0 Autodetect direction recommended value 1 i direction 2 j direction 3 k direction 4 40 2 6 09 ibcu iblag DPLR Code Version 4 01 0 User Manual Using DPLR 4 Alternate directions changes orientation of lines with each iteration alternating between i j and k direction solves Provided for use with separated flows although no significant advantage has been shown with this method Tech Tips 1 The DPLR time advancement method is based on the Gauss Seidel Line Relaxation GSLR method and the lines should be formed in the body normal direction for maximum performance Setting ildir 1 2 or 3 will cause the code to orient the solver in that block so that lines are formed in the i j or k directions respectively If DPLR detects that a line is formed in a non body normal orientation a warning message will be printed It is not a fatal error to run DPLR with the lines in non body normal directions but the convergence rate and stability of the method will be degraded 2 When ildir 0 DPLR will automatically determine the best direction to form the lines for each block at runtime by examining the block boundary conditions For those blocks for which no body surface boundary condition is detected the lines will be formed in the i direction For those blocks with a body surface boundary condition at more than one face the lines will
43. ignored because ilt 1 not 2 or 2 imin 20 Use a zonal interface boundary condition at this computational cell face imax 20 Use a zonal interface boundary condition at this computational cell face jmin 20 Use a zonal interface boundary condition at this computational cell face jmax 19 w w is the plane of symmetry at this computational cell face kmin 26 The wall at this computational cell face is set to catalytic radiative equilibrium kmax 1 The boundary conditions at this computational cell face are fixed at freestream conditions DPLR Code Version 4 01 0 User Manual 4 63 2 6 09 Using DPLR Block 2 Flags Setting Explanation ntx 48 There are 48 computational cells in the x direction in Block 2 nty 64 There are 64 computational cells in the y direction in Block 2 ntz 64 There are 64 computational cells in the z direction in Block 2 iconr 1 This value is ignored because iinit 1 isim 1 This block will be included in the simulation ifree 1 Use freestream specification 1 for this master block initi 1 Use freestream specification 1 to initialize the interior of this master block ibadpt 1 Grid adaption will be performed on this block iflx 4 The Euler flux extrapolation method to use in the i direction is MUSCL Steger Warming with Ap iord 3 The Euler flux extrapolation order of accuracy is third order upwind biased omgi 2 0d0 The val
44. imradial ngeom ismooth fs_scale ds_mult gmargin dsl cellRe eR KR KR KR KR KR KR KH KR HK ds 1mx ds2fr HEAHEAHAHEEAEAEEE AEE AAEEEE EERE AEEEAREEEAEEAE EAE EEEE AEE AAEEEE EEE Miscellaneous controls HEHEHE EAAHAEEE AEE EAEEEEEEEEAAEEAEEEEAEEAE EAE EEEE AEE AAEEAA EEE cfl DPLR Code Version 4 01 0 User Manual 6 19 2 6 09 DPLR Codes Package Input Output Files 6 4 2 Input Flags for Runtime Control Files As of the DPLR Code Version 4 01 0 the following input flags can be dynamically changed in a runtime control file see Section 4 2 for more information on each flag Global Flags istop nplot iplot nruntime freq Tech Tip nruntime_fregq is not found in the standard DPLR input deck It is used to specify how often the control file is to be read by DPLR i e after n iterations Default value 100 Grid Adaption Flags igalign ngiter nalign imedge imradial ngeom ismooth fs_ scale ds_mult gmargin dsl cellRe dslmx ds2fr Timestepping Flags cf1 Tech Tip Note that changes to the cf 1 number list should be made via an iteration specific command as illustrated below 6 4 3 Syntax for Runtime Control Files Unlike other DPLR file formats you do not have to use any of the control file input flags in any particular order Also the syntax for this type of input file is as follows e For comments type a sign first and everything after that on that line will not be seen by the cod
45. imrz 0 0 0 0 0 ymc 0 0 czs 0 0 lt list of BC numbers to extract from dataset lt list of variable numbers to extract from 0 110 120 154 150 151 152 Tecplot plot3d zone iwrt ifac imin imax jmin jmax kmin kmax 1 1 A 2 0 0 L r 1 information bkmin bkmax zonetitle 1 f 1 1 stag2d F 1 1 1 body2d sle T 1 flow2d y vl T 1 terminator fname pname gname bname neptune postpitch Figure 2 POSTFLOW Input Deck for Pitchplane Analysis of Neptune DPLR Code Version 4 01 0 User Manual Probe 5 27 2 6 09 Using POSTFLOW nput file for postflow imemmode itruev istat 2 1 0 inrest ingrid inbcf ouform iwrtd 11 0 0 26 0 interp nzones isep istyp iunits 10 1 aref imrx imry imrz 0 0 0 iwind 0 iexbc lt list of BC numbers to extract from dataset 25 26 ivarp lt list of variable numbers to extract from dataset 0 110 120 507 521 Tecplot plot3d zone information iwrt ifac imin imax jmin jmax kmin kmax bkmin bkmax zonetitle 0 1 1 1 1 stag2d 0 1 L 1 1 body2d 0 r 1 1 1 flow2d 1 A 1 1 1 terminator fname pname gname bname neptune postsurf Figure 3 POSTFLOW Input Deck for Surface Analysis of Neptune Probe DPLR Code Version 4 01 0 User Manual 5 28 2 6 09 5 3 2 DPLR Code Version 4 01 0 User Manual Neptune Input Deck Settings Using POSTFLOW The following table explains the meaning of t
46. in this DPLR input deck ivis 1 DPLR will perform a laminar full Navier Stokes simulation ikt 1 Translational thermal conductivity is modeled in a manner consistent with the baseline model used to compute mixture viscosity and thermal conductivity specified in the ivmod flag ikv 11 Vibrational thermal conductivity is modeled with standard expression with T gradients ivmod 3 The baseline model used to compute mixture viscosity and thermal conductivity is the Yos approximate mixing rules which is preferred for all reacting gas flows idmod 1 The species diffusion coefficients for this simulation are computed with a constant Lewis Schmidt number itmod 0 This value is ignored because ivis 1 defining the problem as a laminar flow simulation and telling DPLR to ignore turbulence and transition related flags islip 0 Slip wall calculations will be disabled iblow 1 Blowing wall calculations will be disabled DPLR Code Version 4 01 0 User Manual 4 57 2 6 09 Using DPLR Global Flags Setting Explanation cont cont cont icatmd 2 Wall catalysis is calculated with the constant y fully catalytic to ions but supports homogeneous surface reactions such as N N N2 amp 0 0 Odz ireqmd 101 The radiative equilibrium wall is modeled with constant wall emissivity set by the value in epsr and a maximum wall temperature set by value in twa11
47. is performed using FCONVERT Although the ideal number of processors to use for a given job is sometimes a matter of personal preference it is often a function of the total number of processors that are available and the number that are necessary to achieve a reasonable load balance Once the desired number of processors to use during the run has been selected the input grid file must be decomposed into one block per processor This is accomplished by setting iaction 1 or 2 in the FCONVERT input deck Load Imbalance One of the primary metrics by which the quality of a parallel decomposition is judged is the amount of load imbalance that results In FCONVERT this load imbalance is computed as a measure of the average amount of wasted CPU time assuming that the total CPU time is directly proportional to the number of grid points on a given processor The total load imbalance Z is then given by nb Lot X N max N N max nb n 1 where nb is the total number of parallel blocks N is the size of block n and Nmax iS the size of the largest block In practice things are more complex than this The type of boundary condition on each face the number of zonal interfaces and the relative speed of each processor all contribute to the amount of time spent on a given decomposed block in DPLR However the load imbalance metric is sufficient to provide a first order estimate The estimated total load imbalance is always reported by FCO
48. itruev 0 then echoes a warning to the screen Specifies the type of processing POSTFLOW performs on flow variables Allowable values are 0 process instantaneous flow variables default 1 process instantaneous statistical flow variables Not working in DPLR 4 01 0 compute mean Not working in DPLR 4 01 0 3 compute RMS Not working in DPLR 4 01 0 5 4 2 6 09 4 01 0 inrest ingrid inbcf DPLR Code Version 4 01 0 User Manual Using POSTFLOW 4 compute standard deviation Not working in DPLR Specifies the format of the restart file to be read by POSTFLOW Allowable values are 1 parallel archival file native unformatted 11 parallel archival file XDR format 21 Parallel archival file ASCII Specifies the format of the grid file to be used when post processing the simulation data Allowable values are 0 get format from restart file Recommended 1 parallel archival file native unformatted 11 parallel archival file XDR format 21 parallel archival file ASCII Tech Tip Setting igrid 0 ensures that the grid file being used to post process the data is the same as the one used to generate the data in the first place However if the name of the grid file or its location relative to the restart file is ever changed you must use one of the other settings in igrid to point POSTFLOW to the original grid file Specifies the format of the boundary condition file if any that was used to generate the
49. jl 64 kl Largest block is ll N nb 2 original block il 64 jl 64 kl 128 Read input interface file neptune inter Found 3 valid zonal interface blocks in 2 block grid file Decomposing block 1 into 1 ibrk 1 jbrk 1 kbrk 1 Decomposing block 2 into 6 ibrk 3 jbrk 2 kbrk 1 creating 7 total blocks 7 Blocks Total load imbalance 6 49 Output Block size il 16 jl 32 kl 128 size il 22 jl 32 kl 128 90112 cells size il 21 jl 32 kl 128 86016 cells 65536 cells size il 21 jl 32 kl 128 86016 cells Output Block Output Block Output Block Output Block size il 22 jl 32 kl 128 90112 cells size il 21 jl 32 kl 128 86016 cells size il 21 jl 32 kl 128 86016 cells Output Block y DN U AeA W N Be Output Block The load imbalance for this case is slightly larger 6 49 vs 4 32 but the increased performance of the implicit algorithm would far outweigh the increase in load imbalance Alternatively this outcome can be accomplished by setting DPLR Code Version 4 01 0 User Manual 8 15 2 6 09 Appendices iaction 1 and using the block decomposition flags to specify the desired decomposition For this example the following decomposition would give output identical to that obtained by using iaction 2 Decomposition information for each master block ibrk jbrk kbrk 1 1 1 3 2 1 Note that this solution is not unique there
50. method of computing equilibrium constants using Gibb s free energy method where species enthalpy and entropy are computed using curve fit expressions given by Gordon and McBride with the final equilibrium constant determined via the van Hoff t equation Same as 1 9 with ramped limiter slowly approaches more aggressive limiter value for advanced users only Same as 1 9 with aggressive limiter begins with larger limiter as default as used in simulations of the Fire II entry vehicle at early trajectory points for advanced users only Same as 1 9 with conservative limiter scales down the 1 9 model selected by 75 Tech Tip Equilibrium constants used to compute the reaction kinetics for chemical reactions in the gas phase are computed by default in DPLR via Arrhenius expressions The curve fit model offered in ikeq 9 has been found to yield the most stable equilibrium constants for most simulations However in those situations where the equilibrium constants are either very small or very large e g ionized wake flow simulations and low density ionized flow simulations with non or weakly catalytic cold 4 16 2 6 09 ivib DPLR Code Version 4 01 0 User Manual Using DPLR walls DPLR offers ways to minimize the potential solution instability that may occur with these extreme values By setting ikeq 11 19 DPLR slowly increases the value of the constant limiter as the solution progresses By setting ikeq 21 29 DPLR begins w
51. o new inter sage dplr gasp g grid g where old inter infile the interface file you are converting new inter outfile the output file for the process The script automatically detects the format of the input interface file and converts it to one of the supported formats specified by the sage dplr or gasp flags DPLR Code Version 4 01 0 User Manual 8 2 2 6 09 8 1 2 8 1 3 Appendices Tech Tip If the output format is sage SAGe you must also specify the associated ASCII plot3d grid file using the g flag as shown above This is because SAGe requires knowledge of the grid size in the input deck and this information is not available in the interface files for either DPLR or GASP dpconvert dpconvert is a Perl script that can be used to change the format of DPLR input decks for use with different release versions of the software Although it is used primarily to enable rapid conversion of older DPLR input decks to a format that is compatible with the current release it can also be used to convert a newer deck to a format that works with older versions of DPLR The script is run from the command line dpconvert i old inp o new inp V where old inp infile the original file you are converting new inp outfile the modified file vV DPLR Release Version for which file is being modified At runtime the script will automatically determine the version of the provided DPLR input deck
52. parallel line relaxation nrix 4 Four implicit data parallel line relaxation steps will be used in simulating this master block ildir 0 The lines will be formed automatically in an appropriate direction when simulating this master block ibcu 1 Implicit boundary conditions will be updated during each line relaxation step iblag 1 Implicit boundary conditions will not be lagged when simulating this master block ilt 1 Global timestepping will be employed when simulating this master block ibdir 1 This value is ignored because nb1k 2 cflm 1 0d20 This value is ignored because ilt imin 20 Use a zonal interface boundary condition at this computational cell face imax 3 Use a first order extrapolation supersonic exit boundary condition at this computational face jmin 19 w w is the plane of symmetry at this computational cell face jmax 19 w w is the plane of symmetry at this computational cell face kmin 26 The wall at this computational cell face is set to catalytic radiative equilibrium kmax 1 The boundary conditions at this computational cell face are fixed at freestream conditions DPLR Code Version 4 01 0 User Manual 4 65 2 6 09 Using DPLR Freestream Setting Explanation Specifications Flags irm 3 Velocity will be used as input for this area of the freestream density 1 6313d 5 The density of this area of the freestream is 000016313kg m M Re V 3 1045d4 The v
53. plane 5 29 2 6 09 Using POSTFLOW Input Flag Setting Explanation cont cont cont imry 0 Do not enforce symmetry about the xz plane imrz 0 Do not enforce symmetry about the xy plane iwind 0 Do not alter the raw output data Ignored CXS 1 Cosine of the global wind axis in the x direction 1 Ignored cys 0 Ignored CZS 0 Ignored iexbc 19 Extract the boundary conditions from the z symmetry plane pitch plane 25 Extract data from the intersection between a catalytic 26 isothermal wall ibc 25 and a catalytic radiative equilibrium wall ibc 26 i e the probe surface ivarp 0 Extract all grid coordinates 110 Extract pressure data 120 Extract translational temperature data 150 Extract velocity in the x direction u 151 Extract velocity in the y direction v 152 Extract velocity in the z direction w 154 Extract the frozen Mach number 507 Extract the total wall shear stress tau 521 Extract total wall heat transfer iwert 0 Do not extract data in these zones fname neptune The restart file to be post processed by POSTFLOW is named neptune pname postsurf The output files created by POSTFLOW for use by Tecplot postpitch are named postsurf and postpitch 5 3 3 Neptune Output Summary Upon execution POSTFLOW will create an on screen summary of the problem for each input deck run as shown below DPLR Code Version 4 01 0 U
54. precision code be sure that all grid files being imported were also created as double precision Specifies whether an input zonal interface file is to be read by FCONVERT or whether zonal interfaces are to be computed automatically Allowable values are 0 Do not read input interface file 1 Read input interface file 2 Auto detect full face interfaces ONLY 3 Auto detect all interfaces fast method 4 Auto detect all interfaces accurate method 3 6 2 6 09 Using FCONVERT Tech Tip Input zonal interface information is only required when the input file is a serial plot3D file The zonal interface file required in this instance can either be created by hand inint 1 or can be created automatically by FCONVERT which will then embed it into the DPLR readable grid file it creates inint 2 4 Because the automatic creation of zonal interface files may require a substantial amount of time and computing resources you can tell FCONVERT to also write the information it generates to a separate file by setting the ouint flag gt 0 thus eliminating the need to regenerate the information every time the problem is run See Section 3 4 for more information on zonal interface files idummy Specifies whether or not the input grid file contains dummy a k a ghost cells Allowable values are 0 Input file does not contain dummy cells 1 Input file contains dummy cells Tech Tip Because DPLR automatically generates grid dummy cells
55. residual and CFL number 13 Output nit block residual and CPU time 14 Output nit block residual and flow time 15 Output nit block residual CFL number and aero data 22 Output nit global residual and min max CFL 32 Output nit block residual and min max CFL Tech Tips 1 nit iteration number global residual summed residual over all computational blocks At timestep for the iteration number CFL number CFL number for the iteration number CPU time elapsed CPU time at the iteration number flow time elapsed flow time at the iteration number only useful for time accurate simulations aero data data written toa aero file for each iteration when viscous fluxes are included in the simulation block residual residuals computed for each master block in the 4 24 2 6 09 igdum DPLR Code Version 4 01 0 User Manual Using DPLR simulation written only to the convergence file 2 To compare the output residual at each iteration with the computed residual in the first iteration enter the ires value as a negative number Although this technique is usually preferred by DPLR users for viewing progress toward solution conversion you should be aware that certain problems can have very small or zero residuals in the first iteration which would result in seemingly large or inappropriate residuals at later iterations 3 Although DPLR will automatically capture specific information in the convergence file co
56. resource If you are new to DPLR begin by reading Sections 3 4 and 5 This will give you the information you need to understand the basic functions and elements of the Code Package Next study Section 7 to gain insight on how to use what you have learned to run an actual simulation Finally use the Sample Cases described in Section 8 to practice what you have learned As your command of the software grows consult Sections 6 and 9 to deepen your understanding of the files types utilities and detailed capabilities of the DPLR Code Package Version 4 01 0 If you are already using an earlier version of DPLR you may want to begin by comparing the input decks in Sections 3 4 and 5 with the versions you are currently using then learning about the new or changed elements in the 4 01 0 version of the Code Package DPLR Code Version 4 01 0 User Manual 1 4 2 6 09 Chapter 2 Installation Guide Contents 2 0 Introduction 2 1 System Requirements 2 2 Software 2 3 Installing the DPLR Code Package 2 4 Directory File Contents DPLR Code Version 4 01 0 User Manual 2 1 2 6 09 Installation Guide 2 Introduction The DPLR Code has been designed to achieve optimal performance on distributed memory parallel machines making the code widely portable to a variety of architectures from laptops networked desktop workstations and simple LINUX clusters to dedicated supercomputers 2 1 System Requirements The DPLR Code has been suc
57. rmsres 2 0246851605138E 10 cfl 5 0E 03 writing restart file neptune pslx solution written at Thurs Feb 5 08 18 10 2009 Loop time 4227 63 seconds on 8 processors Neptune Output Summary Information In addition to verifying the values entered into the DPLR input deck the DPLR output summary displays information about computing resources required for the run values calculated by code and an initial snapshot of the set of iterations that are being performed as the problem converges to a solution In this sample case DPLR estimates that 187 megabytes of stack memory will be required for each of the 8 processors calculates the Reynolds and Mach numbers used in the simulation shows that the early iterations of the run began with very large residuals and and very small CFL timesteps writes restart files every 100 iterations and ends the run with a very small residual at iteration 500 when the neptune pslx solution file is written Note that this estimate may not include additional memory requirements for turbulence models Monitoring the DPLR Run When DPLR begins its execution of a simulation run the screen will display the standard out as discussed above in Section 4 3 4 DPLR Code Version 4 01 0 User Manual 4 71 2 6 09 Using DPLR In addition to the standard out you can actively monitor the simulation run by setting up a POSTFLOW input deck to read restart files as they are being saved during the DPLR run As the
58. run See Section 7 1 and 7 2 for more information on DPLR Workflow and Workflow Shortcuts DPLR Code Version 4 01 0 User Manual 3 23 2 6 09 Chapter 4 Using DPLR Contents 4 0 4 1 4 2 4 3 4 4 Introduction Running DPLR Input Flags for DPLR Neptune Sample Case 4 3 1 Neptune Input Deck Data field settings for the Neptune sample case 4 3 2 Neptune Input Deck Settings Explanation of the Neptune data field settings 4 3 3 Neptune Output Summary Output summary generated by DPLR for the Neptune sample case 4 3 4 Neptune Output Summary Information Explanation of the Neptune Output Summary Monitoring the DPLR Run DPLR Code Version 4 01 0 User Manual 4 1 2 6 09 Using DPLR Introduction DPLR2D and DPLR3D are the main CFD solver applications provided in the DPLR Code Distribution Package The two programs are closely related sharing a common input deck format and most of the physics and numeric subroutines and libraries However two dimensional or axisymmetric problems must be solved and run much faster with the DPLR2D executable whereas DPLR3D must be used for solving three dimensional problems For this manual the term DPLR will be used to refer to whichever solver application DPLR2D or DPLR3D is chosen for the problem under consideration 4 1 Running DPLR Step 1 Step 2 Open the text editor program for your system Action At the command line prompt open path to yo
59. simulation progresses you can extract the data you want to examine and launch Tecplot or some other graphics visualization program to read the POSTFLOW output files and create a graphic representation of the state of your solution at specific iterations or timesteps See Chapter 5 for more information on Using POSTFLOW This workflow set up can help you monitor the progress of your simulation run early enough to see if you are accurately capturing flow conditions along the shock wave If not you may be able to use a runtime control file to implement a grid adaption process during the run to improve the quality of the simulation See Section 6 4 for more information on runtime control files Ideally your simulation will achieve convergence when the residual from the latest iteration in your solution approaches zero and the resulting data visualization accurately represents the flow conditions you are simulating as shown in the standard out for the Neptune simulation at the 500 iteration However each case will have its own set of unique convergence parameters to tell you when you have achieved an acceptable solution In practice if your solution progresses far enough for the residual to stop dropping by orders of magnitude over time and appears to level out you may have achieved an acceptable result DPLR Code Version 4 01 0 User Manual 4 72 2 6 09 Using DPLR y z x M qw a 2 6E 07 2 2E 07 a 1 8E 07 1 4E 07 i 1E 07
60. size il 10 jl Output Block 7 size il 10 jl As you can see this decomposition strategy results in the same load imbalance but offers potentially improved performance because fewer additional zonal boundaries are created Finally when iaction 10 FCONVERT will generate an output file with the same number of blocks as the input file i e no further decomposition will be performed The same result could be achieved either by 1 setting iaction 1 and all ibrk jbrk kbrk flags 1 or 2 setting iaction 2 and nbreak equal to nborig FCONVERT will automatically compute all additional face edge and corner zonal interfaces created by the specified parallel decomposition In addition if the input grid contains one or more zonal interfaces these will be automatically decomposed along with the grid file This information will be written to the output grid file header if one of the parallel formats is requested You can request that the resulting zonal interface file be output for informational purposes by setting ouint 1 11 or 12 Decomposing a file in multiple directions can create a large number of output zonal interfaces particularly when edge and corner interfaces are considered Because each zonal interface represents a message that must be constructed and sent via MPI send and receive calls each iteration during the CFD solution it is generally a good idea to keep decompositions as simple as possible For example if you
61. specific heat at constant volume cv frozen specific heat at constant pressure cp translational specific heat at constant volume cvt rotational specific heat at constant volume cvr vibrational specific heat at constant volume cvv electronic specific heat at constant volume cve mixture gas constant R mixture molecular weight Mw Turbulence Quantities 70 71 12 73 75 turbulent kinetic energy TKE turbulent omega omega _t RESERVED RESERVED Spalart Almaras conserved variable mu_SA Laminar Transport Properties 80 81 82 83 84 85 86 87 laminar viscosity mu_1 laminar kinematic viscosity nu_l laminar thermal conductivity kap_1 laminar rotational thermal conductivity kapr_1 laminar vibrational thermal conductivity kapv_1 laminar free electron thermal conductivity kape_1 laminar binary diffusion coefficient D_1 laminar Lewis number Le 5 13 2 6 09 DPLR Code Version 4 01 0 User Manual Using POSTFLOW 88 laminar Schmidt number Sc 89 laminar Prandtl number Pr Turbulent Transport Properties 90 turbulent eddy viscosity mu_t 91 turbulent kinematic eddy viscosity nu_t 92 turbulent thermal conductivity kap_t 93 turbulent rotational thermal conductivity kapr _t 94 turbulent vibrational thermal conductivity kapv_t 95 turbulent free electron thermal conductivity kape_t 96 turbulent binary diffusion coefficient D_t 97 turbulent Lewis number Le _t
62. terminator line tells POSTFLOW to stop reading zone specification information Now assume further that the exit outflow plane of the problem can be completely defined as the imax surface of block 5 You can tell POSTFLOW to extract data from this surface by creating the following zone specification lines iwrt ifac imin imax jmin jmax kmin kmax bkmin bkmax zonetitle I r 1 l Lp Siy 1 l 5 5 outflow 1 0 dee cols Lo sey 1 lt 1 1 1 terminator These lines tell POSTFLOW to read the outflow line iwrt 1 extract the i face ifac 1 and extract the i surface imin 1 imax 1 from block 5 bkmin 5 bkmax 5 Then with iwrt 1 the terminator line tells POSTFLOW to stop reading zone specification information iexbc Flag Instead of using zone specification lines to extract surface data from a restart file a process that can be cumbersome to set up and which requires you to pre determine the locations of all surface sub zones in the simulation you can use the iexbc flag to accomplish the same result To extract data from all six surfaces of each master block in the simulation simply set iexbc to one or more values of the boundary condition settings allowed for the ibc flag in the DPLR Input deck See Section 4 2 For example if you want to extract ivarp specified data from all possible symmetry planes and outflow boundaries of a multiblock grid the iexbc setting would be as follows DPLR Code Versio
63. the number of blocks The four integers in the next line is the number of points in the first block in the i j k directions followed by a 1 for one variable then repeated for the points in the second block on the next line This is followed by the real block radiation data supplied by the offline radiation transport code Tech Tip Although optional if an rname is specified in the DPLR input deck a pdrx file must exist in the cfdinput directory to avoid a runtime error Convergence Files When ires gt 0 in the DPLR input deck DPLR automatically creates a convergence file when a simulation is run and places it in your working directory The convergence file contains information on the iteration number CFL number or timestep and L2norm of the flow variable specified by the iresv flag of the DPLR input deck An example of a convergence file where ires 2 and iresv 1 is given below Summary of enabled CPP compiler directives gt AMBIPOLAR 1 PARKTEXP 0 50 SCEI 1 00 NOHTC computing L2norm residual of density nit global resid cfl 1 1 000000000000000E 00 1 0E 03 2 9 999989632126156E 01 1 0E 03 3 9 999980913475810E 01 1 0E 03 4 9 999975461771219E 01 1 0E 03 5 9 999973229450715E 01 1 0E 03 sassaseneses etc 98 8 653047974124419E 01 2 5E 00 99 8 637292011393368E 01 2 5E 00 100 8 621722044117890E 01 2 5E 00 Loop time 8 75 seconds on 8 processors DPLR Code Version 4 01 0 User Manual 6 24
64. thermal conductivity An appropriate setting is required for all viscous simulations with vibrational nonequilibrium ivib 1 3 4 Allowable values are 1 Standard expression with e gradients 2 Hard sphere approximation with e gradients 11 Standard expression with 7 gradients preferred setting for all practical applications 12 Hard sphere approximation with T gradients Tech Tips 1 Hard sphere approximations are provided only for DPLR Code Version 4 01 0 User Manual 4 9 2 6 09 Using DPLR comparison to legacy codes and should not be used 2 The choice between e and T gradients is somewhat arbitrary and scales the resulting vibrational thermal conductivity k by the vibrational specific heat C ign OT oT e e q k k k on de on on 1 K K C For most simulations there is little difference between ikv 1 and ikv 11 However for cases where the flow is nearly completely dissociated using energy gradients becomes slightly unstable because there is little energy in this mode ivmod Specifies the baseline model used to compute mixture viscosity and thermal conductivity An appropriate setting for ivmod is required for all viscous simulations Allowable values are 1 Blottner Wilke model with an Eucken relation inaccurate at elevated temperatures 2 Sutherlands Law and constant Prandtl number available only for perfect gas flows but a reasonable estimate at low to moderate temperatu
65. to DPLR by the dname variable in the DPLR input deck along with an absolute pathname To help you choose the chemistry file that is most appropriate for your simulation file names typically contain a descriptive indication of the flow environment being modeled the number of chemical species included in the model the personal or institutional source of the model and the year the model was published followed by a chem suffix For example the file name air7sp park93 chem tells you that it is a model of earth air containing seven chemical species that it was developed by Park and that it was published in 1993 Tech Tip For a more complete description of the model contents reference publication and author s see the legend at the end of each chem file Radiation Files The radiation coupling file is an optional input file used to input pointwise V Qp information obtained from an offline radiation transport code This plot3D formatted file name typically has a pdrx suffix and is specified by the rname variable in the DPLR input deck An example of a radiation file for a 2 block grid where one block size is 32x64x64 points and a second block size is 64x32x64 points is given below DPLR Code Version 4 01 0 User Manual 6 23 2 6 09 6 8 DPLR Codes Package Input Output Files 2 32 64 64 1 64 32 64 1 block radiation data in the i j and k directions The first integer in the first line is
66. to view the location where the NaN first occurred you need to stop the simulation and write a restart file at the conclusion of the iteration in which the NaN was first generated typically the iteration PRIOR to when the residual itself becomes NaN DPLR Code Version 4 01 0 User Manual 5 41 2 6 09 6 0 6 1 6 2 6 3 6 4 6 5 6 6 6 7 6 8 6 9 6 10 6 11 Chapter 6 DPLR Code Package Input Output Files Contents Introduction Grid Files Zonal Interface Files Overview of zonal interface files 6 2 1 Creating Zonal Interface Files by Hand 6 2 2 Input Flags for Zonal Interface Files 6 2 3 Neptune Zonal Interface File 6 2 4 Input Values in the Neptune Zonal Interface File 6 2 5 Creating Zonal Interface Files Automatically Boundary Condition Files 6 3 1 Creating a Pointwise Boundary Condition File 6 3 2 Input Flags for Pointwise Boundary Condition Files Runtime Control Files 6 4 1 Creating Runtime Control Files 6 4 2 Input Flags for Runtime Control Files 6 4 3 Syntax for Runtime Control Files Restart Files 6 5 1 Converting Function Files to Restart Files Chemistry Files Radiation Files Convergence Files Aerodynamic Files Log Files Tecplot Files DPLR Code Version 4 01 0 User Manual 6 1 2 6 09 6 1 DPLR Codes Package Input Output Files Introduction This chapter of the DPLR Code User Manual discusses the 11 types of input and output files created and or used by the DPLR Code Package Vers
67. twall 3 0d3 Maximum temperature at the vehicle surface 3 000 degrees Kelvin epsr 0 85d0 The surface material is 85 efficient in emitting energy away from the vehicle gamcat 1 0d0 The value of y for the constant y homogeneous catalysis model is 1 XXXXX 1 0d5 This flag is currently ignored in DPLR vwall 0 This value is ignored because iblow 1 telling DPLR to disable blowing wall calculations ichem 1 Finite rate chemistry is employed for the chemical reactions in the gas phase ikeq 3 The equilibrium constants are computed from the Park 1990 model n 10 cm ivib 2 Vibrational energy is computed with vibrational equilibrium using statistical mechanics irot 2 Rotational energy is computed with rotational equilibrium using statistical mechanics ieex 0 Electronic energy of the gas is not modeled iel 1 Free electron energy of the gas is computed using the coupled free electron and translational energy model irad 0 No model is used to compute shock layer radiation ipen 0 Not used by DPLR at this time itrmod 0 This value is ignored because ivis 1 defining the problem as a laminar flow simulation and telling DPLR to ignore turbulence and transition related flags DPLR Code Version 4 01 0 User Manual 4 58 2 6 09 Using DPLR Global Flags Setting Explanation cont cont cont itrans 0 This value is ignored because ivis 1 defining the problem
68. zone specification array Allowable values are 0 do not extract the data defined by this zone specification 1 extract the data defined by this zone specification 1 terminator line Tech Tip You can enter any number of zone specification lines in the POSTFLOW input deck However only those that are turned on by iwrt 1 will actually be extracted at runtime This way you can set up a default input deck with multiple zone specification lines for all possible desired output Then each time POSTFLOW is run only 5 23 2 6 09 Using POSTFLOW the data that are actually required can be turned on while the rest are left inactive ifac Specifies the ijk orientation of the surface being extracted Allowable values are 0 No face selected 1 i face 2 j face 3 k face Tech Tips 1 ifac is only needed when surface oriented variables are specified in ivarp i e those marked with an asterisk such as skin friction or heat transfer 2 If ifac 0 in one or more turned on zone specifications and one or more surface oriented variables are specified in ivarp the variables will be removed from the output dataset and a warning message will be echoed to the screen imin imax jmin jmax kmin kmax Specifies the extent of the desired extraction in the ijk directions Numbering depends on the value of interp or the shorthand value as explained below interp 1 interpolate grid points to cell centers i j kmin cel
69. 0 6822392 The fraction of the freestream mass contributed by H2 is 0 6822392 DPLR Code Version 4 01 0 User Manual 4 66 2 6 09 Using DPLR Freestream Setting Explanation cont Specifications cont Flags cont cs H 0 The fraction of the freestream mass contributed by H is 0 cs H 0 The fraction of the freestream mass contributed by H is 0 cs He 0 3177608 The fraction of the freestream mass contributed by He is 0 3177608 cs e 0 The fraction of the freestream mass contributed by electrons is 0 CFL numbers Setting Explanation or timesteps for ramping 00001 Perform 20 iterations of the simulation at a CFL setting of 00001 0001 Perform 20 iterations of the simulation at a CFL setting of 0001 001 Perform 20 iterations of the simulation at a CFL setting of 001 01 Perform 20 iterations of the simulation at a CFL setting of 01 1 Perform 20 iterations of the simulation at a CFL setting of 1 1 Perform 20 iterations of the simulation at a CFL setting of 1 5 Perform 20 iterations of the simulation at a CFL setting of 5 10 Perform 20 iterations of the simulation at a CFL setting of 10 20 Perform 20 iterations of the simulation at a CFL setting of 20 50 Perform 20 iterations of the simulation at a CFL setting of 50 100 Perform 20 iterations of the simulation at a CFL setting of 100 200 Perform 20 iterations of the simulation at a CFL se
70. 09 irot ieex DPLR Code Version 4 01 0 User Manual Using DPLR Specifies the model used to compute the rotational energy component of the gas Allowable values are 1 Rotational nonequilibrium single T mainly used for radiation studies of high altitude flows 2 Rotational equilibrium using statistical mechanics recommended setting 3 Complete thermal equilibrium using NASA LeRC curve fits based on data from NASA s Lewis Research Center s now Glen Research Center computer program CEA NASA Reference Publication 1311 4 Two temperature model using LeRc curve fits T Ta T T Te not recommended 5 Two temperature model using LeRC curve fits T T T Te Ta NOT WORKING in DPLR 4 01 0 11 Rotational nonequilibrium multiple T VOT WORKING in DPLR 4 01 0 Tech Tip Unlike vibration the rotational mode of the gas is assumed to be fully excited and thus cannot be neglected for polyatomic species You must decide whether to model the rotational mode in equilibrium with the translational mode irot 2 4 or in nonequilibrium irot 1 In practice it is rarely necessary to solve for a nonequilibrium rotational energy so this feature is provided mainly for detailed radiation studies of high altitude flows Specifies the model used to compute the electronic energy component of the gas Allowable values are 0 Neglect electronic energy 1 Statistical mechanics T T recommended setting 2 St
71. 10 6E 06 2E 06 6 2 Figure 4 2 Graphic Representation of Mach Contours and Convective Heating at the Wall in the Neptune Simulation after 500 Iterations DPLR Code Version 4 01 0 User Manual 4 73 2 6 09 Chapter 5 Using POSTFLOW Contents 5 0 Introduction 5 1 Running POSTFLOW 5 2 Input Flags for POSTFLOW 5 3 Neptune Sample Case 5 3 1 Neptune Input Deck 5 3 2 Neptune Input Deck Settings 5 3 3 Neptune Output Summary 5 3 4 Neptune Output Summary Information 5 4 Extracting Datasets 5 4 1 Volume Data 5 4 2 Surface Data 5 4 3 Line Data at the Intersection of Two Boundaries 5 4 4 Zone Minima or Maxima 5 4 5 Integrated Surface Data 5 4 6 Freestream Data 5 4 7 Extracting Data for External Codes 5 4 8 NaN s Not A Number DPLR Code Version 4 01 0 User Manual 5 1 2 6 09 5 1 Using POSTFLOW Introduction Restart files generated by DPLR contain all the input deck settings and physical modeling parameters that were used in the simulation These data exist independently of the original input files used to run the simulation Using POSTFLOW you can identify and extract specific data from a DPLR restart file to use in a presentation or further process with graphics software such as Tecplot to create appropriate visualizations of the results of your CFD simulation POSTFLOW always runs in serial mode on a single processor regardless of the number of processors used to run the simulation that generated the restart fil
72. 1x 4 iextst Ae kord 3 nrlx 4 Boundary condition type ibc imin imax jmin jmax kmin kmax 20 3 19 19 Using DPLR 3 Tin 140 3 turbi 0 001d0 subp0 2 650d2 cs 0 6822392 0 0 0 3177608 density 1 6313d 5 Trin 140 3 tkref 0 00d0 subT0 2 650d2 omgk klim kdiss 2 0d0 1 0 ildir ibcu iblag 0 1 1 26 1 M Re V cx 3 1045d4 0 8090160044 Tvin Tein 140 3 140 3 pback 1 05d5 H He Species order H2 DPLR Code Version 4 01 0 User Manual 4 56 epsk 0 03 ilt ibdir cflm 1 1 1 0d20 cy cz 0 5877852523 0 0d0 2 6 09 Using DPLR 4 3 2 Neptune DPLR Input Deck Settings This is a three dimensional problem with one plane of geometric symmetry The original grid consists of two master blocks The following table explains the meaning of the DPLR input deck settings in this sample case Global Flags Setting Explanation nblk 2 There are 2 master grid blocks in this simulation igrid 11 The input grid file is a parallel archival XDR file irest 11 The restart file to be created from this simulation will be a parallel archival XDR file ibcf 0 Do not read a boundary condition file iradf 0 Do not read a radiation file nfree 1 There is one region of the freestream a k a freestream specification characterized in this DPLR input deck iinit 1 Start all blocks by initializing to the values in the freestream specification characterized
73. 2 6 09 6 9 6 10 DPLR Codes Package Input Output Files Convergence files named with the same prefix as the restart file and the suffix con are usually retained for archival purposes and can be used to plot the rate of convergence of a specified variable in a given simulation Tech Tip If the current job began with a restart file DPLR appends the new convergence data to the existing file if any If the current job is a fresh start a new file is created and any previous file with the same name is automatically overwritten Aerodynamic Files When ires 5 or 15 in the DPLR input deck DPLR automatically creates an aerodynamic datafile when a simulation is run and places it in your working directory The aerodynamic file contains information on the iteration number and the three force and moment coefficients computed as dimensional quantities Moments are computed about the origin 0 0 0 and vehicle symmetries are not incorporated An example of an aerodynamic file where ires 5 is given below nit FX Fy Fz Mx My Mz 2001 2 5777E 05 4 706E 04 8 5400E 02 1 0967E 01 6 2569E 02 2 7165E 05 2002 2 5777E 05 4 706E 04 8 5075E 02 1 0972E 01 5 8149E 02 2 7165E 05 2003 2 5777E 05 4 706E 04 8 4872E 02 1 0975E 01 5 4368E 02 2 7165E 05 Aerodynamic files named with the same prefix as the restart file and the suffix aero Typically used as an additional means of monitoring the progress of a simulation run toward conv
74. 4 01 0 1 Perform adaption on this block Specifies the method to use to extrapolate the Euler fluxes in the i j or k directions Note that the method for flux extrapolation can be set separately in each computational direction Allowable values are No flux evaluation Upwind modified Steger Warming with Ap 2 MUSCL Steger Warming with Ap p cs T recommended value 3 MUSCL Steger Warming with Ap p t T MUSCL Steger Warming with Ap p cs u T 5 Pure 2 order central difference should not be used for problems which contain shock waves could be unstable even for subsonic flows 11 Upwind modified Steger Warming without Ap 12 MUSCL Steger Warming without Ap p cs T 13 MUSCL Steger Warming without Ap p T 14 MUSCL Steger Warming without Ap p cs u T Tech Tip The settings ijk flux 2 4 and 12 14 use a MUSCL based adaptive stencil to attain higher order accuracy via a more sophisticated approach The difference between the selections is in the set of variables that are extrapolated to attain high order accuracy and whether a pressure gradient based switch is employed to smoothly transition from high order to first order in the vicinity of strong shock waves Specifies the nominal order of accuracy of the Euler flux extrapolation Allowable values are 1 First order upwind 4 38 2 6 09 omg ijk ijk lim ijk diss DPLR Code Version 4 01 0 User Manual Using DPLR
75. 6 09 7 1 2 DPLR Workflow pslx file postflow inp y POSTFLOW flow plt standard out When DPLR completes the specified number of iterations to achieve a solution and write a restart file you will need to create an input file for POSTFLOW that specifies the data you want extract from the solution and the format of the file you want POSTFLOW to write something that will depend upon the third party data reporting or visualization program such as Tecplot you are using When both files are available you will run the DPLR data extraction executable POSTFLOW to create an input file for your flow simulation graphics program If you are using Tecplot the form of this file will be p1t In addition to the new input file POSTFLOW will create a screen report called a standard out of the actions taken to create the file which again can be saved for archival purposes flow plt gt Tecplot gt graphic representation of simulation solution When the post process data file is available from POSTFLOW you can launch your data visualization program to read in the information and create a graphic representation of your simulation run Subsequent Simulation Runs Although you began your solution run with a structured grid that represented your best guess for capturing the shock wave in a hypersonic flow simulation problem it is common to find that some adjustment of this grid is needed for your solution to adequ
76. 999 pointwise residual res Species Data The following variables are species specific data In each case the user can choose to extract data for either a subset of the species by entering just the desired variable numbers 5 19 2 6 09 DPLR Code Version 4 01 0 User Manual Using POSTFLOW or data for all species by entering the appropriate macro value See Tech Tip 2 1000 1000 n 1200 1200 n 1400 1400 n 1600 1600 n 1800 1800 n 3400 3400 n 3600 3600 n 4000 4000 n 4200 4200 n 4400 4400 n 4600 4600 n all species densities density of species n n all species number densities number density of species n N_n all species mass fractions mass fraction of species n C_n all species mole fractions mole fraction of species n X_n all species densities normalized by Px normalized density of species n RnD_n all species rotational temperatures rotational temperature of species n Tr_n all species vibrational temperatures vibrational temperature of species n Tv_n all species total internal energies per unit mass total internal energy per unit mass of species n e_n all species translational internal energies per unit mass trans internal energy per unit mass of species n et_n all species rotational internal energies per unit mass rotational internal energy per unit mass of species n er_n all species vibrational energies per unit mass vib
77. AGe you must first convert the function file from that program in these cases a plot3d file into a pslx restart file a task that can be accomplished using FCONVERT Step 1 Step 2 Step 3 Step 4 Make sure 1 the input function file contains the following variables in the following order Pv sv w T Te Ty T turb where r are the species densities u v and w are the velocity components T is the translational temperature T is the rotational temperature T is the vibrational temperature Ta is the free electron temperature and turb are the turbulence variables 2 the input function file has dimensions of the number of internal cells in each grid block if idummy 0 or the number of internal cells 2 to account for a single row of dummy cells if idummy 1 Tech Tip The second layer of dummy cells used for high order flux extrapolations should never be included in the input function file Either style can be used to create restart files If dummy cell information is not provided in the function file values in the dummy cells will be extrapolated from the interior and then overwritten by the corrected values when DPLR is run If the dummy cells are included in the file the values contained in the dummy cells should be face centered values at all solid surfaces This allows for an exact specification of the viscous wall boundary condition If dummy cells are not included the boundary condition
78. Data Parallel Line Relaxation DPLR Code Version 4 01 0 User Manual 2 06 09 NOTICE This SOFTWARE falls under the purview of the U S Munitions List USML as defined in the International Traffic in Arms Regulations ITAR 22 CFR 120 130 and is export controlled It shall not be transferred to foreign nationals in the U S or abroad without specific approval of a knowledgeable NASA export control official and or unless an export license license exemption is obtained available from the United States Department of State Violations of these regulations are punishable by fine imprisonment or both The ITAR notice provided on the SOFTWARE shall not be removed by RECIPIENT and the ITAR notice shall remain on any modified versions of the SOFTWARE 2009 U S Government as represented by the Administrator of the National Aeronautics and Space Administration All Rights Reserved Acknowledgements The following individuals have made major contributions to the development functionality and documentation of the DPLR Code Package Dr Michael J Wright ARC Todd White ELORET ARC Dr Matt MacLean CUBRC Dr Dave Saunders ELORET ARC Dr James L Brown ARC Ryan McDaniel ARC Dr Grant Palmer ELORET ARC Dr Chun Tang ARC Nancy Mangini ELORET ARC DPLR Code Version 4 01 0 User Manual 1 Creator DPLR Project Manager Automatic Interface Detection Subsonic Inlet Boundary Conditions Spalart All
79. Face Zonal Interface DPLR Code Version 4 01 0 User Manual 6 10 2 6 09 DPLR Codes Package Input Output Files Setting inint 3 results in moderately rapid detection of both full face and sub face zonal interfaces and is the recommended option for most DPLR cases Using a modern 2007 era computer and working at a speed of approximately one minute per million grid cells analyzed this option works by examining all edge cells of all block faces for interfaces Figure 6 3 shows a typical block where edge cells that are checked are highlighted in red The only type of interface that will not be detectable using the inint 3 option is a case where an interior sub face of a 3D block face touches another interior sub face of the same block as shown in Figure 6 4 Z edge cells checked by inint 3 l i interior face cells checked only by inint 4 Figure 6 4 Edge Cells Checked with inint 3 shown in red Setting inint 4 results in accurate detection of all zonal interfaces including the interior to interior zonal boundaries that a setting of inint 3 would miss as shown in Figure 6 4 As might be expected this setting employs a slow search algorithm where every single exposed face cell is compared with every other exposed face cell requiring on the order of 15 minutes of 2007 era computer time for every million grid cells analyzed Cases requiring the use of this option are likely to be infrequent although it is a useful tool to have when n
80. LOW run if you present them in a space or comma separated list For example the following iexbc entry will extract intersections between a radiative equilibrium catalytic surface 26 and all 180 symmetry planes iexbe 26 17 26 18 26 19 You can extract boundary intersection data in conjunction with extracting standard boundary conditions and volume data For example iexbe 26 18 3 DPLR Code Version 4 01 0 User Manual 5 36 2 6 09 Using POSTFLOW would extract the intersection between boundary condition 26 and 18 as well as data for the exit plane Tech Tip Intersection extraction does not currently work properly with pointwise specified boundary conditions However you can extract the intersection with all pointwise specified boundary conditions by using ibc 0 in an intersection specifier 5 4 4 Zone Minima or Maxima POSTFLOW can extract the minimum or maximum values of selected output variables in each output dataset and if desired the ijk location of these values You can accomplish this by setting the value of ouform in the POSTFLOW input deck either to 7 or to 17 In both cases however the results of this operation are only written to the screen in the standard out STOUT not to an output datafile For example if you set ouform 7 in the Neptune Sample Case described in Section 5 3 the on screen output summary would be show block 1 nx 32 ny 16 nz 64 zone t BC19 i 34 j 1 k 66 Zone Maximum and Min
81. NASA Ames Version 4 01 0 Maintained by Mike Wright last modified 02 05 09 kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk Reading plot3d asciifile neptune g writing parallel XDR formatted file neptune 8PE pgrx Input file does not include dummy cells Output file includes dummy cells Input file is 3D Input Block 1 size il 32 jl 16 kl Input Block 2 size il 48 jl 64 kl 64 32768 cells 64 196608 cells Largest block is nb 2 original block 2 il 48 jl 64 kl 64 Read input interface file neptune inter Found 3 valid zonal interface blocks in 2 block grid file Decomposing block 1 into 1 ibrk 1 jbrk 1 kbrk 1 Decomposing block 2 into 7 ibrk 7 jbrk 1 kbrk 1 creating 8 total blocks 8 Blocks Total load imbalance 12 50 Output Block 1 size il 32 jl 16 kl 64 32768 cells Output Block 2 size il 7 jl 64 kl 64 28672 cells Output Block 3 size il 7 jl 64 kl 64 28672 cells Output Block 4 size il 7 jl 64 kl 64 28672 cells Output Block 5 size il 7 jl 64 kl 64 28672 cells Output Block 6 size il 7 jl 64 kl 64 28672 cells Output Block 7 size il 7 jl 64 kl 64 28672 cells Output Block 8 size il 6 jl 64 kl 64 24576 cells Largest block is nb 1 original block 1 il 32 jl 16 kl 64 Summary grid dimensions for CFD input deck Hardwired to run on 8 processors Block l1 nx 32 ny 16 nz 64 Block 2 nx 48 n
82. NVERT whenever a grid file is processed DPLR Code Version 4 01 0 User Manual 8 12 2 6 09 8 3 2 Appendices Decomposition Strategies When iaction 1 you manually specify how each block in the input file is to be decomposed using the ibrk jbrk and kbrk decomposition factors One set of decomposition factors is required for each master block in the input file A decomposition factor of n implies that the block should be broken n times in that direction For example a decomposition record of Decomposition information for each master block ibrk jbrk kbrk 2 3 1 indicates that the original block should be split into six by breaking it into two equal pieces in the i direction and into three equal pieces in the j direction If the number of computational cells in a given direction is not evenly divisible by the selected decomposition factor the remainder will be evenly distributed among the blocks Setting iaction 1 allows you to control the way that the problem is decomposed for parallel execution which can have significant advantages When iaction 2 you simply specifies the desired number of output blocks using the nbreak flag and allow FCONVERT to determine a parallel decomposition strategy that divides the original file into nbreak output blocks The blocks will be broken such that load balance is maximized This means that FCONVERT will attempt to make all blocks as close as possible to the same size In addition FCONVERT will att
83. R also includes a large selection of generalized realistic surface boundary conditions and hooks to enable efficient loose coupling with external thermal protection system TPS material response and shock layer radiation codes 1 1 DPLR Code Package Using the DPLR Code Package Version 4 01 0 to achieve predictive solutions for hypersonic flow field problems involves completion of five main tasks 1 Creating a structured grid file 2 Converting the grid file to a DPLR readable format with FCONVERT 3 Sending the converted grid file to a number of processors for parallel solution by DPLR2D or DPLR3D 4 Extracting information from the solution needed to create a graphic or data presentation of results with POSTFLOW 5 Creating graphic or data presentations of the predicted flow environment A more detailed description of each of these tasks is presented in Table 1 below Tasks 2 3 and 4 highlighted in Table 1 are completed using applications and software utilities distributed in the DPLR Code Package Version 4 01 0 DPLR Code Version 4 01 0 User Manual 1 2 2 6 09 Overview Table 1 Task Sequence For Predicting Hypersonic Flow Environments with DPLR Code Package 4 01 0 Task Description Tool Input Output 1 Creating a structured grid GridGen or Geometric coordinates for plot3D file that is likely to contain the similar grid object under study serial ASCII shock wave created in the gen
84. STOUT where results for each zone and a sum for all zones are shown not to an output datafile For example if you set ouform 8 and interp 11 in the Neptune Sample Case described in Section 5 3 the on screen output summary might show block 1 nx 32 ny 16 nz 64 gt extracted derivative data from the KMIN surface gt derivative data computed using full viscous fluxes zone t BC19 i 32 j 16 k 1 Fx 9 872234694840E 02 N Fy 3 249055280159E 02 N Fz 3 865734146780E 02 N processing grid variable 1 2 3 processing flow variable 1 2 3 4 5 6 7 8 9 block 2 nx 48 ny 64 nz 64 gt extracted derivative data from the KMIN surface gt derivative data computed using full viscous fluxes zone t BC19 i 48 j 64 k 1 Fx Fy Fz 2 919904481514E 03 N 1 605495325835E 04 N 9 258734449289E 03 N Integrated Surface Quantities DPLR Code Version 4 01 0 User Manual 5 38 2 6 09 5 4 6 Using POSTFLOW Summary Over All Output Surfaces XZ Symmetry Enforced During Final Summation Fx 7 814255901995E 03 N Fy 0 000000000000E 00 N Fz 1 929061572793E 04 N Tech Tips 1 Any ivarp values not included in the list above will be removed from the input deck when ouform 8 2 If aerodynamic forces are selected and iwind is set to either 1 or 2 output forces will be rotated into the wind coordinate system based on either the internal iwind 1 or provided iwind 2 velocity cosines an
85. Timestepping DPLR does offer you the ability to implement local timestepping where the local value of At is used for each computational cell by setting ilt 1 2 but this approach is recommended only for simulations using the full matrix data parallel method FMDP i e iextst 2 CFL Number Ranges When you begin a simulation you should use a very small CFL value to verify that your proposed solution will indeed converge After several hundred iterations if you see that the solution is progressing toward convergence you can increase or ramp the CFL value to specify a larger timestep and speed up the solution process Between each listed CFL number DPLR will perform 20 iterations of the solution If you add an integer to a line with a CFL number DPLR will perform 20 times that integer iterations Note After the first grid adaption DPLR performs only 10 iterations between CFL numbers and 10 times the added integer iteration on the assumption that a post adaption grid is a better starting point and justifies more aggressive CFL ramping For example the following CFL number listing might be appropriate for an initial simulation of a problem 0 01 DPLR performs 20 iterations at a 0 01 timestep 0 05 DPLR performs 20 iterations at a 0 05 timestep 0 10 2 DPLR performs 40 iterations at a 0 10 timestep 0 15 3 DPLR performs 60 iterations at a 0 15 timestep 0 25 3 DPLR performs 60 iterations at a 0 25 timestep 0 50 3 DPLR perf
86. able results for attached flows with a favorable pressure gradient on both blunt and slender bodies Spalart Allmaras Model no compressibility correction Spalart Allmaras Model Catris amp Aupoix comp Spalart Allmaras Model Secundov comp Menter SST Model no compressibility correction Menter SST Model compressibility correction 1 Menter SST Model compressibility correction 2 recommended model for separated flows and flows with adverse pressure gradients Specifies the model to be used for slip wall boundary conditions Allowable values are 0 Disable wall slip recommended setting See Tech Tip below Maxwellian slip model Tech Tip Slip walls are not generally employed for simulations in the hypersonic or supersonic continuum DPLR has currently implemented velocity and temperature slip models but not species density mole fraction slip conditions The slip wall model currently used in DPLR has not been fully validated yet and should therefore be used with caution Specifies the model to be used for blowing wall boundary conditions Allowable values are l or 0 1 2 Disable wall blowing usual setting Specified wall blowing velocity m s Specified unit mass flow rate kg m s 4 12 2 6 09 Using DPLR Tech Tip If blowing wall boundary conditions are taken into account vwal1 gt 0 defines a blowing wall and vwa11 lt 0 defines a sucking wall Pointwise blowing rates can be specified us
87. al vibrational rotational and free electron temperatures respectively Tech Tips 1 The number of unique temperatures depends upon the thermal non equilibrium models chosen with the ivib irot ieex and iel flags 2 At the current time free electron non equilibrium is not supported in DPLR and so will be silently ignored 3 Because only one thermal non equilibrium model may be employed in a given simulation all blocks are assumed to be governed by the same model However each block can have different initial or freestream temperatures by defining multiple freestream specifications and using the initiand ifree flags in each block characterization Specifies a freestream turbulence level for the two equation turbulence models Recommended value 0 0011 4 49 2 6 09 Using DPLR Tech Tip Not used for laminar or Baldwin Lomax algebraic turbulent simulations tkref Initializes the freestream value of turbulent viscosity for the Spalart Allmaras turbulence model itmod 1000 n Recommended value 0 0 freestream turbulent viscosity specified by Ups 2 794 x 107 uja gt 0 freestream turbulent viscosity specified by Ur tkref lt 0 freestream turbulent viscosity specified by UTo tkref Up Tech Tip Not used for laminar or Baldwin Lomax algebraic turbulent simulations subp0O Specifies stagnation pressure in simulations where subsonic boundary conditions are identified subTO Spe
88. arallel Decomposition 8 3 1 Load Imbalance 8 3 2 Decomposition Strategies 8 3 3 Physical Master Blocks vs Virtual Parallel Blocks 8 3 4 Testing for Load Balance 8 3 5 Single Block Input Files 8 3 6 Parallel Recomposition POSTFLOW Output Variables 8 4 1 Grid Related Variables 8 4 2 Mixture Transport Properties 8 4 3 Transport Properties 8 4 4 Mixture Flow Properties 8 4 5 Surface Properties Reference Terms DPLR Code Version 4 01 0 User Manual 8 1 2 6 09 8 1 8 1 1 Appendices Introduction This section of the User Manual contains some reference material and more detailed discussions of content found in previous sections of the publication As the DPLR Code Package is updated additional features and reference information will be added to this section DPLR Code Version 3 06 Utilities The following codes or scripts are provided with the DPLR package in the utilities directory e zbconvert e dpconvert e seqinput Moment e Template This section describes the functions and uses of each of these software tools zbconvert zbconvert is a Perl script that can be used to convert zonal interface files to formats that are readable by e GASP Version 3 0 a commercially available CFD code e SAGe Self Adaptive Gride codE a NASA stand alone grid adaption application that pre dates grid adaption capabilities in DPLR e DPLR The script is run from the command line zbconvert i old inter
89. as necessary at runtime considering their presence during grid generation is usually unnecessary and idummy typically remains 0 idummy should only 1 if you choose to generate your own grid dummy cell coordinates rather than allowing DPLR to do it or you are converting a function file into a restart file Note Input dummy cells will be discarded if mesh sequencing is enabled imseq 1 nborig Specifies the numbers of master blocks in the file Allowable values are e the actual number of blocks in the input file or e the final number of blocks after a grid file recomposition iaction 3 has been performed DPLR Code Version 4 01 0 User Manual 3 7 2 6 09 ouform ouint odummy DPLR Code Version 4 01 0 User Manual Using FCONVERT Specifies the format of the output file See Section 9 2 for more information about supported I O formats Allowable values are 0 Do not generate an output file used for debugging 1 Unformatted parallel file 2 Unformatted plot3d grid or q file 3 Unformatted plot3d grid or function file 11 XDR parallel file preferred for file read into DPLR 21 ASCII parallel file used for debugging 22 ASCII plot3d grid or q file used for debugging 23 ASCII plot3d grid or function file used for debugging Specifies whether an output zonal interface file is to be written and saved for future use Allowable values are 0 Do not write output interface file 1 Write outpu
90. asure of the computational efficiency and thus the operational quality of a grid decomposition strategy Expressed in terms of imbalance this metric is computed as the average amount of wasted CPU time that will result if the proposed grid decomposition strategy is employed to prepare the input grid for parallel processing Ideally as the load imbalance value for a decomposition strategy approaches zero the computational efficiency and thus quality desirability of the strategy increases In practice however things are often more complex and accepting a certain amount of load imbalance can be preferable to a decomposition strategy that introduces an unacceptable number of zonal interfaces or one that requires block breaks in the body normal direction DPLR Code Version 4 01 0 User Manual 3 20 2 6 09 3 4 2 3 5 3 5 1 Using FCONVERT In most cases using the load imbalance metric estimated and reported by FCONVERT whenever a grid file is processed is usually sufficient to provide a first order estimate of the quality of a decomposition strategy Tech Tip To test the load balance for a decomposition strategy before using it in a DPLR run set iaction 0 and nbreak the maximum number of blocks in the strategy FCONVERT will then output the most load balanced way to decompose the input grid into that number of output blocks Parallel Recomposition Although this action is rarely used FCONVERT can be used to recompose
91. ately converge These adjustments can be accomplished in subsequent runs of your simulations using the following technique pslx file revised DPLR inp _ DPLR revised pslx file log file convergence file standard out After DPLR has created a restart file pslx that represents a solution needing some adjustments create a revised DPLR input file by creating and renaming a copy of the dplr inp file you used for your initial run In your new input file enter the following input flag settings iinit 1 igalign 1 nalign 4 ngiter 500 These new settings tell DPLR that the run will be using a restart file that grid alignment is to take place that four such alignments will take place during the simulation and that they will occur every 500 iterations of the run When the revised pslx file is created and data is extracted by POSTFLOW and visualized by a graphics program you can decide if further adjustment to your structured grid is needed to capture the shock wave If so repeat the process above DPLR Code Version 4 01 0 User Manual 7 5 2 6 09 7 2 7 2 1 DPLR Workflow but consider setting a more aggressive CFL ramping schedule i e greater time stepping intervals as you approach a converged solution Workflow Shortcuts Over the years a variety of tools and procedures have been developed to decrease the time spent in creating and running DPLR simulations This section describes several of the more comm
92. atistical mechanics T T 3 Complete thermal equilibrium using NASA Lewis curve fits based on data from NASA s Lewis Research Center s now Glen Research Center computer program CEA NASA Reference Publication 1311 4 18 2 6 09 Using DPLR 4 Two temperature model using LeRC curve fits T Ta T Ty Te not recommended 5 Two temperature model using LeRC curve fits T T T Te Ta NOT WORKING in DPLR 4 01 0 Tech Tip For ieex 1 the contribution of the electronic energy to the total is computed using statistical mechanics based on characteristic temperatures and degeneracies in the chemprops spec file from the cfdinput directory and is assumed to be in equilibrium with the translational mode iel Specifies the model used to compute the free electron energy component of the gas Allowable values are 0 Neglect free electron energy only valid for flows with no ionization 1 Coupled free electron and translational energy Tei T recommended setting assumes that the energy of the free electron gas is governed by the translational temperature 2 Coupled free electron energy and vibrational energy Te Ty NOT WORKING in DPLR 4 01 0 3 Complete thermal equilibrium using NASA LeRC curve fits based on data from NASA s Lewis Research Center s now Glen Research Center computer program CEA NASA Reference Publication 1311 4 Two temperature model using LeRc curve fits T Ta I
93. because ivis 1 defining the problem as a laminar flow simulation and telling DPLR to ignore turbulence and transition related flags prtl 0 72d0 This value is ignored because ivmod 3 not 2 or 12 DPLR Code Version 4 01 0 User Manual 4 59 2 6 09 Using DPLR Global Flags Setting Explanation cont cont cont prtlT 0 90d0 This value is ignored because ivis 1 defining the problem as a laminar flow simulation and telling DPLR to ignore turbulence and transition related flags XXXX 0 0d0 Not used by DPLR at this time XXXX 1 0d0 Not used by DPLR at this time rvr 1 3d0 The viscous overrelaxation parameter for this simulation is 1 3 resmin 1 0d 20 When this simulation converges into a solution the residual will be essentially zero Time Accurate Setting Explanation amp Statistical Options Flags itime 0 Use a 1st order integration of time accuracy imax 5 This value is ignored because itime 1 dttol 1 0d 3 This value is ignored because itime 1 tfinal 9 0d99 Final flow time is essentially infinity tfac 1 0d15 This value is ignored because itime 1 ifstat 0 Do not compute flow statistics iaero 0 Do not compute aerodynamic variables Grid Setting Explanation Adjustment Alignment Morphing Flags igalign 0 Grid alignment will not be performed in this simulation ngiter 500 This value is ignored because igalign 0
94. bout the grid and the flow solution components of the restart file being extracted for processing In this sample case POSTFLOW displays the Keq equilibrium constant limiter that DPLR calculated from ikeq setting in the DPLR input deck restates that the number of species used in the simulation was 5 confirms that the rotational vibrational and translational energies of the flow were ignored for this simulation and verifies that this solution for the Neptune entry probe is based upon a 2 block 3D simulation POSTFLOW then names the output variables that were specified by ivarp in the input deck and confirms that post processing of the restart file is taking place in high memory mode Next POSTFLOW displays a running indicator of block by block progress in processing the restart file while verifying block dimensions Finally the summaries show the name of the output files in this case postpitch dat and postsurf dat while displaying the names of the grid file and the restart file that were used to create the output Extracting Datasets The primary use of POSTFLOW is to extract volume or surface data from the restart file for further post processing or visualization Using the ouform flag in the POSTFLOW input deck data can be saved in two primary output file formats e plot3d ouform 2 3 22 23 32 33 e Tecplot ouform 5 6 25 26 The plot3d format is a standard CFD output format that can be read by most commercial
95. cated by the value in ndr1 DPLR Code Version 4 01 0 User Manual 6 5 2 6 09 DPLR Codes Package Input Output Files ndr2 Specifies the second extent direction of the zonal boundary being described Allowable values are 1 i direction 2 j direction 3 k direction nst2 Specifies the starting point of the interface range in the direction indicated by the value in ndr2 nen2 Specifies the ending point of the interface range in the direction indicated by the value in ndr2 6 2 3 Neptune Zonal Interface File The following zonal interface file was created by hand for the Neptune sample case inint 1 ZONAL BOUNDARY INFORMATION Cell Matching No dummy cells zvers izdum 3 05 0 nblk ninta nintc Boundary 1 nface ndri nstl nenl ndr2 nst2 nen2 1 2 1 16 3 1 64 1 2 1 3 1 64 Boundary nface ndrl nenl ndr2 nst2 nen2 2 2 16 3 1 64 1 2 64 3 1 64 Boundary nface ndrl DPLR Code Version 4 01 0 User Manual 6 6 2 6 09 6 2 4 DPLR Codes Package Input Output Files Input Values in Neptune Zonal Interface File The input plot 3D grid for this problem consists of two master blocks and three interfaces between these two blocks Each interface zonal boundary is described by two lines of data in the interface file The following table explains the meaning of the values entered into the zonal interface file for this sample case Input Value Setting Explanation zvers 3 05 The 3 05 version of
96. ce and moment coefficients Indicates the x y z position in meters as specified in the input plot3d grid file of the moment reference center used for extracting moments and moment coefficients Tech Tip Although extraction of hinge moments for control surfaces is not currently supported in POSTFLOW it can be accomplished using the Moment utility 5 9 2 6 09 Using POSTFLOW imrx imry imrz Specifies planes of symmetry used in the simulation iwind DPLR Code Version 4 01 0 User Manual Allowable values are 0 do not enforce symmetry about this plane 1 enforce symmetry about this plane Tech Tips 1 Possible planes of symmetry are defined as imrx body is symmetric about the yz plane imry body is symmetric about the xz plane imrz body is symmetric about the xy plane 2 POSTFLOW currently supports bilateral symmetry any one of the above flags 1 and quadrilateral symmetry any two of the above flags 1 3 If the symmetry of the vehicle is more complex than a simple bilateral or quadrilateral representation set all of the above flags 0 and compute the symmetry relations off line after post processing is complete 4 If ouform s requesting integrated variable reporting over given surface s and ivarp 600 673 700 773 requesting force or moment coefficients it is important to set aref to specify the full reference area when normalizing these computed forces if the symmetry flags are also used 5 A
97. cessfully installed and run on the following hardware system software configurations Table 2 1 Supported Hardware Software Architectures Architecture Compiler MPI Version Xeon 32 bit Intel Portland MPICH LAM MPI Lahey MPICH Xeon 64 bit Intel Portland MPICH LAM MPI Xeon Dual Core Intel MPICH Opteron 64 bit Intel Portland MPICH LAM MPI Intel Mac 32 bit gfortran MPICH Altix Intel Open MPI Each of these architectures can be specified with the configuration script config during the installation process described in Section 2 3 although on new systems some editing or creation of the makefile comm and include machine h may be necessary DPLR Code Version 4 01 0 User Manual 2 2 2 6 09 2 2 Installation Guide Software Two software packages must be installed before DPLR Code can be installed e Fortran 90 DPLR is written entirely in Fortran 90 running on a UNIX LINUX operating system and thus requires a working f90 compiler on the destination machine e Message Passing Interface MPI DPLR Code uses MPI calls to facilitate inter processor communications so an MPI library must be present in the system Although not strictly required for successfully installing DPLR Code having the following software packages on your system will enhance the utility and or performance of the DPLR Code Package e FXDR fcdr libraries provide a Fortran based interface to th
98. ch then a match at all points of the face pair is likely but the fraction of face cells for which this is true is also calculated and printed in the last column of template con Values less than 1 0 in that last column are a likely sign of gaps or overlaps in the grid 3 Symmetry boundary conditions are considered only after matching block faces are checked for first A somewhat looser tolerance is employed namely min 10 x epsilon 0 001 for measuring the distance of the three pairs of coordinate maxima minima from zero DPLR Code Version 4 01 0 User Manual 8 7 2 6 09 8 2 Appendices These tests can still be fooled by say a flat plate in the z 0 plane or an almost flat surface atx 0 False BC entries of 17 18 or 19 meaning symmetry plane in x y or z respectively should be checked for under such circumstances 4 Template errs in favor of Shuttle Orbiter grids for remaining unassigned block faces These grids are known to contain a single layer of blocks with index k in the radial direction A face not already assigned a flow through BC 20 or symmetry BC 17 19 is marked as BC 26 if it is face 5 k I catalytic radiative equilibrium wall else it is marked as BC 1 if it is face 6 k nk free stream For non Shuttle applications different wall BCs may need to be entered in place of BC 26 Any remaining unassigned face is marked as BC 2 specified inflow or supersonic outflow This choice is appropriate for loca
99. cies 2D simulation the residual in the u momentum component can be tracked with iresv 6 Since DPLR is a 4 26 2 6 09 xscale ils Le Sc DPLR Code Version 4 01 0 User Manual Using DPLR fully coupled code with the exception of some turbulence models convergence of one variable is typically dependent on convergence of the others which limits the utility of single variable residual However this option can be useful for analyzing an unstable simulation because the offending equation will generally blow up before the others do Used to scale the input grid at runtime Allowable values are 1 No scaling recommended f Multiply grid dimensions by this value immediately after read Tech Tip DPLR will print a warning message if xscale is not set to 1 Specifies whether input constants governing laminar Le Sc and turbulent LeT ScT diffusion coefficients are to be interpreted as Lewis or Schmidt numbers Allowable values are 1 Lewis Number 2 Schmidt Number Specifies the value of the laminar Lewis or Schmidt number to be employed in the simulation This parameter is relevant for viscous simulations ivis 0 with constant Lewis Schmidt number diffusion idmod 1 11 Choosing a constant Schmidt number is typically preferred with appropriate values varying with the target destination and entry velocity Recommended values 0 4 0 7 Tech Tip Remember that the preferred approach is to model multi spec
100. cifies stagnation temperature in simulations where subsonic boundary conditions are identified pback Specifies back pressure for subsonic outflow in simulations where subsonic boundary conditions are identified cs An array of input species mass fractions Tech Tips 1 There must be one entry per species in the chosen chemistry model as specified in the input chem file 2 All input mass fractions must sum to 1 0 or DPLR will exit with an error message 3 Input of mole fractions is not supported at this time CFL Number Listing The final entries in the DPLR input deck are a list of Courant Friedrichs Lewy CFL numbers to employ during the simulation DPLR Code Version 4 01 0 User Manual 4 50 2 6 09 Using DPLR CFL numbers are a measure of the explicit inviscid stability limited time step At and are used by convention in CFD codes to enable time advancement to a steady state solution In DPLR the CFL number for a given computational cell is defined as the time it takes the fastest wave in the flow to traverse the thinnest dimension of the cell Global Timestepping For implicit non time accurate simulations this time step is a bit different for every cell in the flow However most DPLR based simulations use the minimum value of At at any cell in the flowfield for all cells in what is called global timestepping a set up approach that has been shown to result in robust solutions and good convergence rates Local
101. creen with place holder default values as shown below To customize the file for your simulation replace flags and values with those you want to change Note that whenever a sign appears in this file DPLR considers whatever follows to be comments and will not parse the information Step 2 Save your control file to your working directory after giving it the same prefix as your solution file and adding the suffix ctr1 DPLR will check for the existence of a control file in your working directory by default every 100 iterations during the solution run If a control file exists and is correctly formatted DPLR will read the new settings for the grid flags and or timestepping flags and continue the simulation using those values DPLR Code Version 4 01 0 User Manual 6 18 2 6 09 DPLR Codes Package Input Output Files Run time control example nplot 200 iplot 1 iteration 500 ngiter 500 nalign 4 gmargin 2 5 Trailing comment iteration 1000 igalign 1 nalign 3 gmargin 3 Another comment end HEHHEAHAEAAHAAEE AEE HAEEEEEEEEAAEEAREEEAEEAE EAE EEEE AEE AAEEEE EEE Iteration number dependent controls HEAHEEHAEEAHAAEE AEE AAEEAEEEEEAEEE AREA EAEEAEEAEEEEE AEE AEE EEE EEE istop nplot iplot nruntime_freq HEAHEAHHHEEAHAEEE AEE AAEEEEEEEEAEEEAREEEHEEAEEAEEEEE AEE AAEERE EEE Grid tailoring controls HEAHEAHAHEEAEAEEE AEE AAEEEEEEEEAAEE PRUHU PUHU UUU EERE AREA AAA EEE igalign ngiter nalign imedge
102. d kbrk for each block in the input file or Set taction 2 enter the number of blocks minimally equal to the number of available processors for the run to decompose the input file into in nbreak then allow FCONVERT to automatically determine the best decomposition strategy for the input grid file Although this method may produce a good result from a CPU load balance perspective it may not produce a good flow solver result Note When parallel decomposition is not being performed setting 1brk 1 on the first line tells FCONVERT not to read additional block decomposition records iseq jseq kseq Specifies sequencing factors in the 7 j and iname DPLR Code Version 4 01 0 User Manual k directions when imseq 1 2 or 2 Tech Tip One set of sequencing factors is required for each block in the input file unless imseq 2 or 2 These settings tell FCONVERT to sequence remove points or upsequence add back points all grid blocks by the same factor so only one set of sequencing factors are required regardless of the number of grid blocks Note When sequcing is not being performed i e imseq 0 setting 1seq 1 on the first line tells FCONVERT not to read additional block sequencing records Specifies the input file name This is the file that will be processed by FCONVERT The filename should be surrounded by single or double quotes and can be specified with either a relative or an absolute path as shown in the exampl
103. d solution Recommended value 1x10 or lower Tech Tip If you prefer to run your simulation to a specified number of iterations irrespective of the residual set istop to the desired number of iterations and resmin to a very small value such as 1x10 Time Accurate and Statistical Options Flags In the past DPLR has been primarily used to solve steady state problems characterized by large but stable time steps With these flags DPLR can be run in a time accurate fashion and thus capable of studying transient phenomena itime lmax dttol tfinal tfac ifstat DPLR Code Version 4 01 0 User Manual Specifies time integration order of accuracy Allowable values are 0 1 order of accuracy l 2 order dual time stepping accuracy Specifies maximum number n of dual time iterations per time step when itime 1 Specifies the residual tolerance f for the dual time iterations when itime 1 Tells DPLR to stop the simulation when flow time reaches this f value Specifies the multiplicative factor on physical time step used to determine dual time step when itime 1 Specifies flow statistics DPLR is asked to compute Allowable values are 0 Do not compute flow statistics Compute mean and RMS disabled in DPLR 4 01 0 1 Reset previously computed flow statistics and start again disabled in DPLR 4 01 0 4 29 2 6 09 Using DPLR 2 Strip previously computed statistics from the restart file di
104. d will be output as lift drag and side forces in addition to the xyz forces otherwise reported Note that this option assumes that the employed grid is in standard aircraft coordinates Freestream Data POSTFLOW can extract freestream data from the restart file You can accomplish this by setting the value of ouform in the POSTFLOW input deck either to 10 to display requested ivarp values with their SI units or to 110 to display a tabular listing of data better suited for direct import to a spreadsheet application In both cases however the results of this operation are only written to the screen in the standard out STOUT not to an output datafile Freestream data are calculated and output for each grid block in the simulation irrespective of any surface extraction or zone specification flags that have been set Separate freestream data are presented for each grid block since DPLR allows multiple freestream specifications to be applied when a simulation is run However in most cases all blocks will have the same freestream information For example if you set ouform 10 and ivarp 110 120 154 58 in the Neptune Sample Case described in Section 5 3 the on screen output summary might show block 1 nx 32 ny 16 nz 64 Freestream Quantities Block 1 p 3 591044259306E 01 Pa T 1 280700000000E 02 K M 3 232180261501E 01 Re L 3 151858720834E 05 1 m DPLR Code Version 4 01 0 User Manual 5 39 2 6 09 5 4 7 U
105. dary nz nface ndrl nstl nenl ndr2 nst2 nen2 DPLR Code Version 4 01 0 User Manual 6 3 2 6 09 DPLR Codes Package Input Output Files Step 3 Enter values that describe the number location extent direction and range of each zonal interface identified in the input plot3D grid file Step 4 Save the file Action At the command line type save filename inter Result The ASCII zonal interface file required by DPLR and identified by the xname flag in the FCONVERT input deck is saved 6 2 2 Input Variables for Zonal Interface Files Input variables required in a zonal interface file are discussed below in the order they appear in the file zvers Specifies the version number of the interface file This is used by FCONVERT to automatically upconvert older interface files when they are read thus assuring full backward compatibility Allowable values are the real numbers of the major and minor releases of the DPLR Code Package from 2 31 through the current version number 4 01 0 izdum Specifies whether dummy cells are accounted for in the interface file Allowable values are 0 Input file does not include dummy cells 1 Input file includes dummy cells Tech Tip This option is meant for developers to use in debugging If izdum 1 FCONVERT will automatically strip the dummy cell information before processing the zonal interface file which could have unwanted results nblk Specifies the number of master b
106. dea is to leave at least the outer two points of each radial grid line beyond the shock following reconvergence of the flow solution The shock location tends to move inward with each tailoring so erring on the low side with ds_mult is normally safe during early adaptions especially if the initial boundary is far away from the shock An optional multiplier of the outermost radial spacing of the grid normally not needed This permits additional control over the outer grid boundary and may be used to increase or decrease the radial adjustment produced by the normal alignment scheme Although values may be positive or negative real numbers the typical value for this input is 0 Tech Tip This control allows the boundary to expand or contract everywhere if there is reason to believe the current outer boundary is too close to the eventual shock or too far from it If extrapolation is implied the extrapolations are linear Beware of possible crossed grid lines or excessive cell skewness if any existing radial lines are convergent Constant value with different meanings for different settings of imradial and igalign When imradial 1 ds1 sets the minimum allowable wall spacing anywhere in the volume 4 34 2 6 09 cellRe dslmx ds2fr DPLR Code Version 4 01 0 User Manual Using DPLR dsi 0 No minimum allowable wall spacing anywhere in the volume ds1 lt 0 Lower limit for cell Reynolds number smoothing min
107. ditions and send the correct data to all processors in the simulation DPLR will then overwrite the stored grid file to include the dummy cell information The preferred setting for this flag is igdum 0 because there is usually no need to recompute dummy cells unless an error is detected and the boundary conditions change during the simulation 1gdum 1 Seldom needed rarely used Allows you to zero out the body normal added dissipation term in the boundary layer If kb1 is a positive integer the body normal eigenvalue will be zeroed out for the kb1 cells nearest to each solid wall in the simulation and smoothly increased to the specified value Tangential epsilon augmented near axis boundary condition for k lt kbl Must have one or more 1011 1019 or 1011 2019 BCs for this to have any effect For developers only Leave set to 0 Specifies equation of state being used Allowable values are 0 1 Perfect gas 2 Real gas excluded volume NOT WORKING in DPLR 4 0 Controls the residual variable s tracked by DPLR Allowable values are 1 Total density sum of L2Norm of all species densities Velocity sum of velocity components 3 Energy sum of energy equations 4 Turbulence variables sum of turbulence variables for Spalart Allmaras or Menter SST model n Conserved variable n See Tech Tip below Tech Tip You can track the residual of a single equation by using a negative integer for iresv For example for a 5 spe
108. e 1 Constant cell Reynolds number wall spacing wall spacing varies over the body surface from the value set in ds1 to the value set in ds1mx 2 Usea constant wall spacing wall spacing at all surface locations will equal the value set in ds1 Tech Tip If ds 1 0 the current wall spacing will be used Specifies the number of geometrically spaced points to place near the body surface during reclustering Recommended value 2 Tech Tip If ngeom lt 2 a pure two sided Vinokur stretching routine will be used 4 32 2 6 09 ismooth Using DPLR Specifies the type of smoothing to employ at the outer boundary following a grid alignment Allowable values are 0 Do not smooth the outer boundary 1 Smooth outer boundary of grid based on changes in arc length at each body point location 2 Smooth outer boundary of grid based on total final arc length at each body point location 3 Smooth using both method 1 and method 2 preferred for initial fs_scale DPLR Code Version 4 01 0 User Manual adaption from a hyperbolic grid Tech Tips 1 Setting ismooth 1 is generally preferred because it tends to give smoother outer boundaries and will eventually asymptote to zero smoothing when the outer boundary stops moving Also it avoids some problems seen with SAGe when smoothing grids with abrupt changes in surface geometry such as a propagation of the surface geometry to the outer boundary as seen in the Shuttle Orb
109. e Because DPLR maintains backward compatibility the current version of POSTFLOW can be used to post process restart files generated with earlier versions of the DPLR Code Package Running POSTFLOW Step 1 Open the text editor program for your system Action At the command line prompt type path to your post directory post_flow3d_mb inp Result An input file appears on screen with placeholder default values To start with a blank deck delete the default values as shown on the following page DPLR Code Version 4 01 0 User Manual 5 2 2 6 09 Using POSTFLOW Input file for postflow imemmode itruev istat inrest ingrid inbcf ouform iwrtd interp nzones isep istyp iunits lref aref imrx imry imrz iwind cxs cys iexbc lt list of BC numbers to extract from dataset ivarp lt list of variable numbers to extract from dataset Tecplot plot3d zone information iwrt ifac imin imax jmin jmax kmin kmax bkmin bkmax zonetitle stag2d body2d flow2d terminator fname pname gname bname Step 2 Enter problem specific values for each of the input variables or flags See Section 5 2 for a description of input flags and a list of allowable values Step 3 Rename and Save your POSTFLOW input file to your working directory Step 4 Run POSTFLOW Action At the command line prompt type postflow lt yourpostflowfilename inp Result An output file that can be processed by a third party graphics
110. e For example I am expanding this grid would not cause DPLR to change the running simulation in any way e For a generic command type the name of the flag you want to change followed by an sign followed by a numeric value More than one generic command can appear on a line but they should be separated by a comma or colon or semicolon For example nplot 300 iplot 2 would result in DPLR changing the value of those flags when it reads the control file after 100 more iterations DPLR Code Version 4 01 0 User Manual 6 20 2 6 09 6 5 DPLR Codes Package Input Output Files e For an iteration specific command type an sign followed by the word iteration followed by an sign followed by a numeric value Then type the name of the flag you want to change at that iteration an then a numeric value An iteration specific command can also have several flags associated with it but they should be separated by a comma colon or semicolon For example iteration 2000 igalign 3 imradial 2 would result in DPLR changing the value of those flags when it reaches iteration 2000 assuming that iteration had not already been passed when the control file was read When changing a cf1 number listing the following syntax should be used iteration 2000 cfl 100 iteration 2100 cfl 250 iteration 2500 cfl 500 This entry in a control file in your working directory will tell DPLR that when it reaches the first it
111. e Thrust Aerodynamic Forces and Moments Force and moment variables ivarp 600 673 700 773 are usually extracted in conjunction with surface integration in order to generate integrated aerodynamic data and or coefficients Because an accurate surface integration cannot be performed if the data are extrapolated to zone edges POSTFLOW will automatically set interp I1 whenever one or more force and moment variables are specified as output This will result in a surface mesh with gaps along all block boundaries if pointwise surface data are requested unless you perform an off line integration using a utility program such as Moment 600 total force on a face in all directions 601 total force on a face in x direction Fx 602 total force on a face in y direction Fy 603 total force on a face in z direction Fz 604 total force on a face in x direction per unit area Fx_a 605 total force on a face in y direction per unit area Fy_a 606 total force on a face in z direction per unit area Fz_a 610 pressure force on a face in all directions 611 pressure force on a face in x direction Fx_P 612 pressure force on a face in y direction Fy_P 613 pressure force on a face in z direction Fz_P 614 pressure force on a face in x direction per unit area Fx_Pa 615 pressure force on a face in y direction per unit area Fy_Pa 616 pressure force on a face in z direction per unit area Fz_ Pa 5 17 2 6 09
112. e cfdinput directory As information becomes available for other species this file can be updated and expanded Note this option assumes that energy converted to radiation is instantly lost from the control volume 3 To include the surface radiative heating effects in the radiative equilibrium surface energy balance set irad 102 104 and read the pointwise surface radiative heating from a radiation file rname in your working directory See Section 6 7 for more information on radiation files 4 To use an external radiation transport code such as RADEQUIL or NEQAIR to loosely couple flowfield radiation information to DPLR solutions iterate between the two codes as follows a Set irad 0 and run an initial DPLR solution b Extract data using POSTFLOW for input to your radiation transport code c Create a radiation file rname from data generated by 4 20 2 6 09 Using DPLR running your radiation transport code See Section 6 7 for more information on radiation files d Perform a second DPLR run with irad 1 e Repeat until you obtain a fully converged solution f Compare your first and final solutions to determine the significance of the shock layer radiation value ipen Specifies the model used for reaction product energy distribution Allowable values are 0 All species are created destroyed at the internal energy of the mixture 1 9 n 10 percent of dissociation energy for all reactions VOT WORKING in DPLR
113. e at http www tecplot com DPLR Code Version 4 01 0 User Manual 6 27 2 6 09 Chapter 7 DPLR Workflow Contents 7 0 Introduction 7 4 DPLR Workflow Chart 7 1 1 Initial Simulation Run 7 1 2 Subsequent Simulations Runs 7 2 Workflow Shortcuts 7 2 1 Sequence the Grid 7 2 2 Use Runtime Control Files To Adjust Grids and CFL Schedules 7 2 3 Use Template To Create DPLR Input and Zonal Interface Files 7 2 4 Understand Your Computing Resources DPLR Code Version 4 01 0 User Manual 7 1 2 6 09 DPLR Workflow 7 Introduction Using the DPLR Code Package to achieve hypersonic flow simulation solutions can be a complex undertaking Although the main tasks i e grid file generation grid file conversion solution processing and data extraction are essentially sequential in nature they are also iterative and often require concurrent execution to make the most productive use of your time your tool and your computing resources This chapter will suggest a set of actions or flow of work that may help you achieve solutions in a more timely manner Once you become familiar with the use and robust capabilities of the DPLR Code Package however it is likely that you will develop your own customized workflow DPLR Code Version 4 01 0 User Manual 7 2 2 6 09 DPLR Workflow 7 1 DPLR Work Flow Chart Geometry of Interest CAD File W _ v zonal interface file fconvert inp file standard out FCONVERT
114. e below 3 11 2 6 09 xname cname oname DPLR Code Version 4 01 0 User Manual Using FCONVERT ASCIIPlot3Dfilename g Tech Tip The suffix used in the file name is optional FCONVERT will assume the default suffix for the specified file type if not manually entered See Appendix A for a list of file types and associated default suffixes Specifies the name of a previously prepared input zonal interface file when inint 1 See Section 3 4 1 for more information on previous preparation of a zonal interface file The filename should be surrounded by single or double quotes and can be specified with either a relative or an absolute path as shown in the example below YourZonaliInterfaceFileName inter Tech Tip The suffix used in the file name is optional FCONVERT will assume the default suffix for this file type is inter if not entered Specifies the CFD input deck file name if any This file is only used to locate solid walls to assist in decomposing the input grid when iaction 2 and FCONVERT is attempting to automatically break the input grid into blocks for the best possible parallel solution The filename should be surrounded by single or double quotes and can be specified with either a relative or an absolute path as shown in the example below CFDfile inp Tech Tip The suffix used in the file name is optional FCONVERT will assume the default suffix for this file type is inp
115. e native XDR external Data Reference calls on all UNIX LINUX machines XDR enables the creation of platform independent binary files greatly enhancing the portability of generated datasets e g restart and grid files The code can be compiled without the fxdr libraries however in that case all restart and grid files must be written in either ASCII or machine specific native binary format e TECIO A These I O libraries are used by POSTFLOW to create Tecplot binary data files for post processing output Currently DPLR can use Tecplot 360 or Tecplot II libraries However ASCII data in Tecplot can always be written by POSTFLOW regardless of TECIO A availability e LIBGOTO Basic Linear Algebra Subroutines BLAS routines DPLR makes use of several BLAS routines for matrix vector and matrix matrix manipulations Having such libraries on the target machine will result in a 20 25 performance improvement in the overall runtime of DPLR Code 1 fxdr libraries are freely available at http meteora ucsd edu pierce fxdr_ home _page html 2 Tecplot I O libraries are included with Amtec s Tecplot visualization software and may be available for free at http www tecplot com 3 BLAS libraries are generally available from compiler makers as a part of their mathematical libraries for a nominal fee In addition several freeware sources exist In particular for Pentium or AMD architectures a freeware distribution ca
116. e of computational efficiency referred to as load balance Once the desired number of processors to use during the DPLR run has been determined the plot 3D input grid file must then be decomposed into a minimum of one block per processor As discussed in Section 3 2 parallel decomposition of an input grid file can be accomplished in two ways e Setting iaction 1 manually determining the best strategy for decomposing the input grid file then entering one set of decomposition factors in the ibrk jbrk and kbrk flags in the FCONVERT input deck for each block in the input file or e Setting iaction 2 entering the number of blocks to decompose the input file into minimally equal to the number of processors that will be used for the DPLR run in the nbreak flag and allowing FCONVERT to automatically determine the best decomposition strategy for the input grid file Although the choice of setting is dependent upon the situation choosing iaction 1 can have significant advantages including e More direct control over the decomposition strategy to ensure minimal generation of additional zonal interfaces and to avoid breaks in the body normal direction two conditions that support the rapid convergence of DPLR Code solutions e Avoiding the need to generate the DPLR input deck prior to running FCONVERT as is the case when iaction 2 See Chapter 6 for more information on this requirement Load Balance Load balance is a me
117. e the dimensions and number of computational cells in each block and output this information to screen ifile Specifies the type of file to be processed Allowable values are 1 grid file 2 restart flow file 3 boundary condition BC file 4 radiation file DPLR Code Version 4 01 0 User Manual 3 4 2 6 09 idim iinfo ivers DPLR Code Version 4 01 0 User Manual Using FCONVERT Tech Tip Because restart boundary condition and radiation files are never decomposed or recomposed the only functional actions for these file types are iaction 10 format conversion and iaction I file size Specifies the dimension of the input file Allowable values are 2 2D Axisymmetric 3 3D Tech Tip Whatever value idim is set to the input grid file must match it e g if idim 2 the input grid file must be two dimensional if idim 3 the input grid file must be three dimensional However some grid generation programs like GridGen do not support 2D grid generation In this case you need FCONVERT to see a 3D grid file as a 2D file To accomplish this make sure that the k dimension 1 in the input grid file FCONVERT will automatically strip the z coordinates from a 3D input grid file if the k dimension 1 regardless of the idim setting For all other cases FCONVERT will execute the operation if given an input grid file of different dimension than specified by the idim flag but the results may be undesirable
118. ed after 500 iterations are already complete and istop 100 DPLR will run 100 additional iterations reaching completion after 600 total iterations 4 22 2 6 09 DPLR Code Version 4 01 0 User Manual nplot iplot iaxi DPLR Code Version 4 01 0 User Manual Using DPLR Tech Tip DPLR can also be instructed to stop a run when a specified L2norm residual level is reached using the resmin flag In this case termination will occur when the first condition either istop or resmin is met Specifies the frequency of restart file writes DPLR will save a restart file periodically every nplot iterations during the solution run as long as the flag iplot gt 0 Tech Tip A good general value to use for nplot is usually 100 large enough so DPLR does not spend a large percentage of the runtime writing restart files but small enough so a lot of work is not lost if the job quits for some reason Controls the redundancy of restart file writes Allowable values are lor 0 Do not write a restart file use only for debugging 1 Write a single restart file n Save n 1 prior restarts 99 Force restart file write use only for debugging Tech Tip Setting iplot to a positive integer larger than I n causes DPLR to save n 1 previous restart files in addition to the current one To distinguish saved files DPLR will append the iteration number of the restart file to the filename specified in fname For example if fname sample ps1x iplot 3
119. ed for that version of DPLR to function properly Thus removing files from any of the directories is not recommended DPLR Code Version 4 01 0 User Manual 2 5 2 6 09 Chapter 3 Using FCONVERT Contents 3 0 3 1 3 2 3 3 3 4 3 5 Introduction Running FCONVERT Input Flags for FCONVERT Neptune Sample Case 3 3 1 Neptune Input Deck 3 3 2 Neptune Input Deck Settings 3 3 3 Neptune Output Summary 3 3 4 Neptune Output Summary Information Parallel Decomposition 3 4 1 Load Balance 3 4 2 Parallel Recomposition Mesh Sequencing 3 5 1 Sequencing an Input Grid 3 5 2 Upsequencing a Restart File DPLR Code Version 4 01 0 User Manual 3 1 2 6 09 Using FCONVERT 3 Introduction The primary function of FCONVERT is to read plot3d grid files generated by third party software applications such as GridGen and convert them into a format that can be used by DPLR to solve hypersonic CFD problems However you can also use FCONVERT to change the format of a restart file convert a plot3d flow file a k a function file into a restart file process input radiation and boundary condition files and change the number of processors ion which a simulation can be run Finally you can use FCONVERT to manipulate a plot3d grid file through scaling it by a multiplicative factor useful when changing grid units e g from feet to meters and to keep the DPLR solution computationally efficient by e sequencing or coarsenin
120. eeded DPLR Code Version 4 01 0 User Manual 6 11 2 6 09 DPLR Codes Package Input Output Files Figure 6 4 Interior to Interior Zonal Interface Detectable Only with inint 4 DPLR Code Version 4 01 0 User Manual Tech Tips 1 All intra zonal interface boundaries singularity or degenerate axes self closing blocks etc will be detected using any of the detection options inint 2 4 2 Some grid generation programs and post processors introduce round off error in the xyz grid coordinates that can result in the points on either side of an interface being slightly different FCONVERT has a built in tolerance factor to determine when two slightly different points are likely the same but this is not foolproof To determine if all interfaces have been accurately detected you can 1 compute them all by hand as a check case or 2 run the resulting case and look for mismatched interfaces in the resulting solution or 3 monitor the tolerance of each interface found in the input grid as reported by FCONVERT or 4 look for a large number of small patchy interfaces between two faces resulting from an FCONVERT run 6 12 2 6 09 6 3 6 3 1 DPLR Codes Package Input Output Files Boundary Condition Files Boundary condition files also called pointwise boundary condition files provide information to DPLR about the chemical radiation and turbulence conditions that exist at each point on a specified face of a master
121. elocity of this area of the freestream is 31 045 m sec CX 0 8090160044 The cosine of the velocity vector in the x direction is 0 8090160044 i e cx 8 cos 34 2 degrees cy 0 5877852523 The cosine of the velocity vector in the y direction is 0 5877852523 CZ 0 The cosine of the velocity vector in the z direction is 0 because the flow vector is defined as 34 2 degrees in the x y plane Tin 140 3 The translational temperature in this area of the freestream is 140 3 degrees Kelvin Trin 140 3 The rotational temperature in this area of the freestream is 140 3 degrees Kelvin Tvin 140 3 The vibrational temperature in this area of the freestream is 140 3 degrees Kelvin Tein 140 3 The free electron temperature in this area of the freestream is 140 3 degrees Kelvin Turbi 0 001d0 This value is ignored because ivis 1 defining the problem as a laminar flow simulation and telling DPLR to ignore turbulence and transition related flags Tkref 0 00d0 This value is ignored because ivis 1 defining the problem as a laminar flow simulation and telling DPLR to ignore turbulence and transition related flags subp0O 2 650d2 This value is ignored because no subsonic boundary conditions are identified for this simulation subTO 2 650d2 This value is ignored because no subsonic boundary conditions are identified for this simulation pback 1 05d5 This value is ignored because no subsonic boundary conditions are identified for this simulation Cs H2
122. empt to make the blocks as close to cubes as possible by breaking first in the direction s with the most points To do this however FCONVERT requires a valid DPLR input deck to exist one that can be used to determine the locations of body surfaces in the grid file It is a runtime error to set iaction 2 unless a valid DPLR input deck has been specified as input For all decomposition strategies it is important to minimize and preferably eliminate breaking the grid in the body normal direction because DPLR is by default a line relaxation code that solves the Navier Stokes equations through a series of block tridiagonal matrix factorizations This method converges most rapidly when the problem has not been decomposed in the body normal direction Example 1 Consider an input grid file that consists of two blocks The plot3d header record for this case is 2 17 33 129 65 65 129 DPLR Code Version 4 01 0 User Manual 8 13 2 6 09 Appendices Block 1 consists of 65 536 grid cells 16 x 32 x 128 while block 2 consists of 524 288 cells 64 x 64 x 128 Assuming that iaction 2 and nbreak 7 and there are no solid walls specified in the DPLR input deck a portion of the descriptive output for this run will be Input Block 1 size il 16 jl 32 kl 128 65536 cells Input Block 2 size il 64 jl 64 kl 128 524288 cells Largest block is nb 2 original block 2 il 64 jl 64 kl 128 Read input interface file ne
123. eration format flow field when the object application enters at a specified Mach number 2 Converting the grid file into FCONVERT plot 3D file XDR parallel a DPLR readable format leach Secon anne file for use by breaking it first if required Glen ifi DPLR by the problem into a SEG Pe RE E aab e Onal information about the grid blocks file being converted 3 Processing converted grid DPLR2D or FCONVERT XDR output restart file file on a specified number DPLR3D file parallel XDR of processors to work the I xe format problem in parallel naput ne oome performing sufficient problemn seca maon e information about the object calculations to reach a unde Tomay ioy A environment and solution data requirements 4 Extracting information POSTFLOW restart file plot3D and or from the restart file needed Te Sie conan data file s r erp be sage specifications of the data numeric representations of x the predicted flow pouen Doia Pe Ar applications for solution presentation s or further post processing 5 Creating visual TecPlot or plot3D file and or Graphic representations or data similar data file representation report s of solution for use graphics of predicted in further problem analyses application flow and outcome presentations environment DPLR Code Version 4 01 0 User Manual 1 3 2 6 09 1 2 Overview How to Use This Manual This manual is intended to be both a user guide and a reference
124. eration in this list that has not yet been passed it will adopt this timestepping schedule in place of the one in the DPLR input deck and not refer back to that original listing unless the run is stopped and restarted Restart Files The restart file is the DPLR solution file It is named via fname in the DPLR Input Deck and typically has the suffix pslx The first time a simulation is run DPLR writes a restart file as frequently as specified in the nplot input flag and saves as many restart files as specified in the iplot input flag Restart files contain all the input deck values and physical modeling parameters that were used in the simulation Once written a restart file is linked within DPLR to the binary machine readable pgrx grid file that was used for the simulation Tech Tip Although restart files can be written in unformatted parallel psin and ASCII parallel psia formats the preferred format in the DPLR working environment is XDR parallel ps1x a binary machine readable file DPLR Code Version 4 01 0 User Manual 6 21 2 6 09 6 5 1 DPLR Codes Package Input Output Files Converting Function Files to Restart Files The only CFD solution file aka function file that can be used as direct input to DPLR is a ps1x restart file Thus if you want to rerun a solution in DPLR that was originally created in another CFD solver application such as radial_interp or S
125. erfaces zonal interface file for the full face interfaces of the computational grid gasp inputs control file in GASP flow solver format template con connectivity file containing most of the interface and BC information for the FLO107MB flow solver Scanning this can help spot possible problems caused by block faces that don t meet the matching tolerance i e look for too many integer Os in the one line per block output Ancillary Input Files Template looks for two control files both of which are optional Jf either is present in the working directory it will be invoked so beware of unintended usage DPLR Code Version 4 01 0 User Manual 8 6 2 6 09 Appendices If sample inputs is present the header and trailing portions of the sample input deck are transcribed to dplr inputs and dplr inputs 2 Otherwise only the middle sections of those input decks will be produced If template inp 2 is present it can serve either of two purposes or both Initially implemented to control the contents of dplr inputs 2 it can also be used to make the automation of boundary conditions complete for specialized grids of the type developed for rapid analysis of Shuttle Orbiter damage and repair configurations The latter use of template inp 2 is typically confined to workgroups within NASA A sample template inp 2 file to be used with a wing leading edge plug or tile gap filler grid is shown below Sequencing cont
126. ergence these files are usually retained for archival purposes and can be used to plot the rate of convergence of a given aerodynamic simulation Log Files DPLR automatically creates a log file or standard out when a simulation is run and places it in your working directory Log files names with the same prefix as the restart file and the suffix 1log are usually retained for archival purposes and can be parsed to automatically fill out quality check forms if those are used as part of your CFD process The log file contains a subset of the same information that is echoed to your screen in a standard out STDOUT DPLR Code Version 4 01 0 User Manual 6 25 2 6 09 DPLR Codes Package Input Output Files An example of a log file is given below Summary of enabled CPP compiler directives gt AMBIPOLAR 1 gt PARKTEXP 0 50 gt NOHTC Air Mechanism 5 species 5 reactions Park 1990 Model gt Species List N2 02 NO N O gt Reaction rates from air5sp park90 chem gt Reaction Status 11111 gt Keq Fit Used 00000 gt NASA Lewis thermo fits used to find Keq gt Assume molecules created destroyed at mixture Tve Catalytic wall BC enabled gt Constant accomadation coeff gamma 1 000 Rotational Equilibrium Fully Excited Vibrational Non Equilibrium SHO Electronic Energy Neglected Laminar Navier Stokes Simulation gt Gupta Style Collision Integrals amp Yos Mixing Rule g
127. ersion 4 01 0 User Manual 8 8 2 6 09 Appendices File Formats Supported in the DPLR Code Package Format Description File Type Suffix 1 Unformatted Parallel grid pgrd restart psin BC pbcf radiation prdf 11 XDR Parallel grid pgrx restart pslx BC pbcx radiation prdx 21 ASCII Parallel grid pgra restart psla BC pbca radiation prda 2 Unformatted Plot3D grid gu flow qu Format Description File Type Suffix 12 XDR Plot3D grid gx flow qx 22 ASCII Plot3D grid g flow q 32 Gzipped ASCII Plot3D grid gz flow qz 3 Unformatted Plot3D grid gu flow fu 13 XDR Plot3D grid gx flow fx 23 ASCII Plot3D grid g flow f 33 Gzipped ASCII Plot3D grid gz flow fz 5 Binary Tecplot Block plt 25 ASCII Tecplot Block dat DPLR Code Version 4 01 0 User Manual 8 9 2 6 09 Appendices 6 Binary Tecplot Point plt 26 ASCII Tecplot Point dat 8 2 1 Format Numbers The first digit if any of the file format number specifies the data storage type as follows 0 Written as machine specific unformatted files This type of file should be avoided if portability is desired because an unformatted file created by one machine type usually cannot be read by another l Written in XDR format XDR files are binary written to be read on any machine and the recommended storage type for large files including grid and restart files See Tech Tip 1 2 Written as an ASCII file ASCII files are much larger than binary files and sh
128. etitively stop and restart simulation runs See Section 6 4 for more information on creating and managing Runtime Control Files Use Template To Create DPLR Input and Zonal Interface Files Manually creating DPLR input and zonal interface files can be a time consuming task However by using the Template utility created by Scott Thomas and David Saunders and distributed with the DPLR Code Version 3 06 Package these two tasks can be automated plot 3d grid file sample inputs file TEMPLATE dplr interfaces dplr inputs gasp inp template con To use Template to automatically create zonal interface files and block specific areas of the DPLR input file perform the following steps DPLR Code Version 4 01 0 User Manual 7 7 2 6 09 7 2 4 DPLR Workflow Step 1 Rename the generic inp file in the cfdinput directory as sample inputs and save it to your working directory Step 2 Place the structured plot3d grid file of your object of interest in your working directory Step 3 Run TEMPLATE Your working directory now contains four new files dplr inputs dplr interfaces gasp inputs template con When you open the dplr inputs file you will see that Template has created content for the block specific areas of your DPLR input file You may use this content as a guide to enter the values manually or simply copy and paste it into the DPLR input file you are creating for your simulation run When you open the dp
129. example ivarp 1000 tells POSTFLOW to output species densities for all species in the simulation relieving you of the need to identify each species by its order number in the chem file Whenever macro values are used only those variable relevant to the simulation will be extracted so ivarp 0 will automatically extract x y and z coordinates for a 3D flow but only x and y for a 2D or axisymmetric flow 3 Variables prefaced with an asterisk are defined as surface specific quantities and are extracted with respect to a given surface direction as defined either with the i fac flag in the zone specifications see below or automatically determined when extracting surfaces with the iexbc flag 4 All extracted variables are output in SI units Units for dimensional output variables are echoed to the screen when one of the standard output formats are specified Tecplot Plot3D Zone Specification Flags The flags in this section of the POSTFLOW input deck define the extent of data extraction required for specified flow volumes or zones In general one row of data defines each desired extraction The last line in this group is the terminator line in which iwtr 1 instructs the code to stop reading zone specification information A terminator line must be present or arun time error will occur iwrt DPLR Code Version 4 01 0 User Manual Specifies whether or not POSTFLOW will perform data extractions for that line of the
130. file generated with the coarsened grid Step 1 Open and name a new FCONVERT input file Step 2 Set ifile 2 inform 11 imseq 2 iseq jseq kseq to values used during sequencing iname coarsened restart file name oname new uncoarsened restart file name pslx Step 3 Save file to your working directory Step 4 Run FCONVERT lt new FCONVERT input file Step 5 Open and name another new FCONVERT input file Step 6 Set ifile 1 inform 2 imseq 0 iname original plot3d grid filename oname new unsequenced XDR parallel grid file name pgrx DPLR Code Version 4 01 0 User Manual 3 22 2 6 09 Using FCONVERT Step 7 Save file to your working directory Step 8 Run FCONVERT lt second new FCONVERT input file Result Your working directory now contains an upsequenced restart file that can be used to start a new solution run with the DPLR readable grid file containing the original number of data points Tech Tips 1 Starting a new solution run with a restart file is always more time efficient than starting an initial run Thus this quick method of obtaining a valid restart file can significantly shorten the time you will need to obtain a solution for the first run of a CFD simulation 2 If you use different levels of sequencing to obtain restart files be sure to create and save a new pgrx grid file from the original plot3d grid file to match the number of points in the restart file used for each DPLR
131. for one block in the second zonal interface is 64 and for the abutting block it is also 64 The ending point in the second extent direction for one block in the third zonal interface is 32 and for the abutting block it is 48 6 2 5 Creating Zonal Interface Files Automatically Although it is valuable to fully understand the meaning and origin of data in the zonal interface files it is likely that day to day use of the DPLR Code Package will more often involve automatic generation of these files Currently there are three tools available to automatically compute and generate zonal interface files that are readable by DPLR 1 GASP zbconvert 2 Template 3 FCONVERT GASP zbconvert Some commercial grid generation tools are capable of automatically generating interface information in a format that is readable by the commercial CFD code GASP Version 3 For this reason a utility Zbconvert is included with the DPLR Code DPLR Code Version 4 01 0 User Manual 6 8 2 6 09 DPLR Codes Package Input Output Files Package that can convert zonal interface information from GASP Version 3 to DPLR readable zonal interface files See Section 9 1 1 for more information about the utility zbconvert Template Zonal interface files can also be created by the software tool Template developed by Scott Thomas and David Saunders which automatically generates a DPLR input deck and interface file from a multi block grid Template
132. g 0 1011 180 degree singular axis u u 3D 1012 180 degree singular axis v v 3D 1013 180 degree singular axis w w 3D 1014 Singular x axis v v axi 1015 Singular y axis u u axi 1016 360 degree singular axis 3D 1017 Plane of symmetry u u 1018 Plane of symmetry v v 1019 Plane of symmetry w w 1021 90 degree singular axis v v w w 3D 1022 90 degree singular axis u u w w 3D 1023 90 degree singular axis u u v v 3D Mathematically Adjusted Boundaries 2000 2099 2011 2019 amp 2021 2023 currently support a maximum CFL in the vicinity of standard singular axes axes or symmetry planes as per cflmin each block when kdg 0 4 47 2 6 09 Using DPLR 2011 180 degree singular axis u u 3D 2012 180 degree singular axis v v 3D 2013 180 degree singular axis w w 3D 2014 Singular x axis v v axi 2015 Singular y axis u u axi 2016 360 degree singular axis 3D 2017 Plane of symmetry u u 2018 Plane of symmetry v v 2019 Plane of symmetry w w 2021 90 degree singular axis v v w w 3D 2022 90 degree singular axis u u w w 3D 2023 90 degree singular axis u u v v 3D Freestream Specification Flags The flags in this portion of the DPLR input deck define a set of conditions for an area of the flow from which all relevant fluid dynamic quantities can be computed You can define any number of freestream
133. g directory creates a more productive computational environment for DPLR solutions Depending on the format of the file a standard suffix will be assumed See Section 9 2 for a list of file types and associated default suffixes gname fname bname DPLR Code Version 4 01 0 User Manual Specifies the name of the input XDR parallel grid file and will typically have the suffix pgrx If the file was prepared using FCONVERT the name is specified in the oname flag of the FCONVERT input deck This file contains not only the xyz coordinates of all the grid blocks in the simulation but also information about block connectivity and the desired decomposition for processing This file is required and must already exist when the simulation run begins Specifies the name of the input restart file and will typically have the suffix ps1x This is a required file name If this is a new simulation DPLR will create the file to go with the name specified here If the simulation is a rerun the file specified here should already exist See Section 6 6 for more information on restart files Specifies the name of the input boundary condition file and will typically have the suffix pbca This file is not required to run a simulation but having one gives you increased flexibility in specifying point by point 4 6 2 6 09 Using DPLR parameters as opposed to the standard block face parameters See Section 6 4 for more i
134. g in DPLR 4 01 0 Note Dirtlib and Overset functionality must be available at compile time for this functionality to be enabled iover Indicates enabling of Overset Logic if compiled in DPLR Allowable values are 0 Overset logic is disabled 1 Overset logic is enabled ioint Indicates format of domain connectivity file cname Allowable values are 0 ASCII Suggar type dci output xxxxx Not used in DPLR 4 01 0 Space Marching 1D Implementation Each of the flags in this portion of the DPLR input deck allow for shock capture in 1D only ispace Used for 1D space marching simulations i e shock tube flows Allowed values are 0 Disable space marching 1 Enable space marching Tech Tip Seldom used in practice this option gives you a fast and efficient tool to perform 1D flow simulations with complex models If this option is used many of the other flags in the code are ignored dxmin Sets the minimum x spacing for the space marching routine Only used when ispace 1 DPLR Code Version 4 01 0 User Manual 4 36 2 6 09 slength nxtot Using DPLR Sets the total marching distance for the space marching routine Only used when ispace 1 Sets the total number of cells for the space marching routine Only used when ispace 1 Block Specific Flags The flags in this portion of the DPLR input deck can be set differently for each computational block in the simulation ntx nty ntz iconr isim
135. g the grid in one or more dimensions e breaking or decomposing the grid into multiple pieces to run on a parallel machine 3 1 Running FCONVERT Step 1 Open the text editor program for your system Action At the command line prompt type path to your fconvert directory file_ convert inp Result An input file or deck appears on screen with place holder default values To start with a blank deck delete the default values as shown on the following page 1 Unlike DPLR2D and DPLR3D FCONVERT is a serial code so all pre processing must be done on a single processor DPLR Code Version 4 01 0 User Manual 3 2 2 6 09 Using FCONVERT Input file for fconvert iaction ifile idim iinfo ivers nvers inform inint idummy nborig ouform ouint odummy ncedge imseq iscale sfact imir nbreak Decomposition information for each master block ibrk jbrk kbrk Sequencing information for each master block iseq jseq kseq iname xname cname nsin nerin nevin necin ntbin imirx imiry imirz Tech Tip Although you can add as many sections as you need to specify the decomposition and sequencing instructions for each master block in your input grid take special care to preserve the line spacing within each block specific section and throughout the global areas of the input deck as you enter new values and or replace default values with problem specific ones If lines are added to or subtracted within these areas DPLR wi
136. grid block Pointwise boundary condition files are optional If you prepare one for your simulation you must enter the filename in the bname flag in the DPLR input deck Boundary condition files typically have the suffix pbca If a boundary condition file is not prepared for your simulation you should set bname none A generic boundary condition file named pointwise pbca is distributed with the DPLR Code Package Version 4 01 0 can be found in the cfdinput directory Creating a Pointwise Boundary Condition File Step 1 Open the text editor program for your system Action At the command line prompt type path to your cfdinput directory pointwise pbca Result A generic input file appears on screen with place holder default values as shown below To customize the file for your simulation remove the default values but take special care to preserve the line spacing Specifically there must be three lines shown with signs between lines with value entries Step 2 Enter appropriate problem specific values for the input variables as described in Section 6 3 2 Step 3 Save your boundary condition file to your working directory Tech Tip Although you can add as many lines as you need to specify the sizes and ibc numbers for each master block in your input grid preserve the line spacing within each section and throughout the global areas of the input deck If lines are added to or subtracted inappropriately within t
137. han DPLR s pslx format so that FCONVERT can correctly determine the location of the velocity components in the file Specifies the number of unique vibrational temperatures energy conservation equations to be considered in the CFD solution This is only read if you are trying to create a restart file from an input file other than DPLR s pslx format so that FCONVERT can correctly determine the location of the velocity components in the file Specifies the number of unique electronic temperatures energy conservation equations to be considered in the CFD solution This is only read if you are trying to create a restart file from an input file other than DPLR s pslx format so that FCONVERT can correctly determine the location of the velocity components in the file 3 13 2 6 09 Using FCONVERT ntbin Specifies the number of turbulence variables to be considered in the CFD solution This is only read if you are trying to create a restart file from an input file other than DPLR s pslx format so that FCONVERT can correctly determine the location of the velocity components in the file imirx imiry imirz Specifies mirroring factors acrosss the yz xz and xy axes to be used when imir 1 Allowable values are lor0 No mirroring will take place along this axis 1 Mirroring will take place along this axis Tech Tip It is an error to set all three of these flags to 1 3 3 Neptune Sample Case The sample case used thro
138. he volume line iwrt 1 ignore surface data ifac 0 extract all points in the imin 1 imax 1 j jmin 1 jmax 1 and k kmin 1 kmax 1 directions from all master blocks bkmin 1 bkmax 1 Then with iwrt 1 the terminator line tells POSTFLOW to stop reading zone specification information POSTFLOW will now generate five output zones one for each block which contain the entire volume Each zone will be called volume if a Tecplot output file format is selected by setting ouform 5 6 25 26 in the POSTFLOW input deck Surface Data Surface data can be extracted from a restart file in two ways e Using zone specification lines e Using the iexbc flag DPLR Code Version 4 01 0 User Manual 5 34 2 6 09 Using POSTFLOW Zone Specification Lines Continuing with the 5 block 3D example in Section 5 4 1 assume that all blocks have a body surface at j 1 and that these five surfaces completely define the body The following zone specification lines could then be used to extract data from the entire body surface iwrt ifac imin imax jmin jmax kmin kmax bkmin bkmax zonetitle 1 2 1 l 1 1 1 l 1 1 body 1 0 EE Lj Al Ty el 1 1 terminator Using the shorthand code of 1 to mean everything or all these lines tell POSTFLOW to read the body line iwrt 1 extract the j face i fac 2 and extract the j surface jmin 1 jmax 1 from all blocks bkmin 1 bkmax 1 Then with iwrt 1 the
139. he Moment script is presented below running Moment version 3 05 0 Moment Center Xm 0 000000E 00 m Ym 0 000000E 00 m Zm 0 000000E 00 m Reference Values lref 3 650000E 00 m aref 4 500000E 00 m 2 qdyn 2 784862E 03 Pa Vehicle Symmetries xy plane Wetted Area Area 0 000000E 00 m 2 Force components Fx 1 777037E 07 N Cx 1 418013E 03 Fy 1 165808E 04 N Cy 9 302740E 01 Fz 0 000000E 00 N Cz 0 000000E 00 Moment components Mx 0 000000E 00 N m A Cmx 0 000000E 00 My 0 000000E 00 N m R Cmy 0 000000E 00 Mz 6 500303E 04 N m Cmz 1 421099E 00 At this time there is no error checking in place to ensure that this output format is used correctly So although it is not an error to select other variables as output the DPLR Code Version 4 01 0 User Manual 8 4 2 6 09 8 1 5 Appendices results generated by the Moment utility will be incorrect unless forces per unit area are selected At the current time Moment is only needed for the extraction of hinge moments because all other features of the utility are built directly into POSTFLOW Tech Tips 1 You will need to compile Moment as the installation script that comes with the DPLR Code Package will not automatically install the program on your system Ask your System Administrator for information on how to compile and install this tool in your utilities directory 2 Because Moment was original
140. he grid file can simply be decomposed again FCONVERT will strip out the header information decompose the master blocks as desired and write the new header information into the file No other input file type is altered in any way by changing the number of processors in the simulation Testing for Load Balance Although many processors may be available for a run you should try to choose a number that maximizes load balance in order to maximize the computation efficiency of the simulation You can test the load balance for a series of possible decompositions with FCONVERT Set iaction 0 and nbreak to the maximum number of blocks desired FCONVERT will then loop over all possible output block numbers from the number of input blocks to the value of nbreak and output the most load balanced way to decompose into that number of output blocks Using the same example if iact ion 0 and nbreak 10 FCONVERT will generate the following output Decomposing block 1 into 1 ibrk 1 jbrk 1 kbrk 1 Decomposing block 2 into 1 ibrk 1 jbrk 1 kbrk 1 2 Blocks Total load imbalance 43 75 Decomposing block 1 into 1 ibrk 1 jbrk 1 kbrk 1 Decomposing block 2 into 2 ibrk 1 jbrk 1 kbrk 2 3 Blocks Total load imbalance 25 00 DPLR Code Version 4 01 0 User Manual 8 19 2 6 09 Appendices Decomposing block 1 into 1 ibrk 1 jbrk 1 kbrk 1 Decomposing block 2 into 3 ibrk 1 jbrk 1 kbrk 3 4 Blocks Total load imbalance 16 28 Decomposing bl
141. he input deck settings in this sample Case Input Flag Setting Explanation imemmode 2 Run POSTFLOW in high memory mode so that all the features of the program are available itruev 1 Use accurate 2 order expressions to compute derivative values istat 0 Process flow variables instantaneously inrest 11 The restart file to process is an XDR parallel archival file ingrid 0 Use grid information found in the restart file inbcf 0 Use boundary condition information found in the restart file ouform 6 Output post processed data into a Tecplot point binary file iwrtd 0 Do not reconstruct DPLR input decks from the restart file for storage in a subdirectory of your working directory interp 1 Interpolate grid points to cell centers nzones 10 The maximum number of output data zones to be generated is 10 isep 0 All active output datasets are written to a single file istyp 1 This flag is ignored in DPLR 4 01 0 iunits 1 Include SI units in the output file lref 1 0 Use 1 meter as the reference length to normalize moment coefficients aref 1 0 Use 1 square meter as the reference area to normalize force and moment coefficients xmc 0 0 The moment reference center is located at 0 0 on the x axis ymc 0 0 The moment reference center is located at 0 0 on the y axis zmc 0 0 The moment reference center is located at 0 0 on the z axis imrx 0 Do not enforce symmetry about the yz
142. hese areas DPLR will not be able to read the file accurately DPLR Code Version 4 01 0 User Manual 6 13 2 6 09 DPLR Codes Package Input Output Files 2 3 05 0 Pointwise PBCA File template Version 3 05 neq ns ner nev nee net 13 8 0 1 0 0 nmc nme nmt nmv f2 f3 f4 block sizes 16 12 78 40 40 78 ibc numbers for each block 20 20 18 20 26 60 20 3 18 18 26 60 Profile Data for Block 1 Face 6 1 632708000000000E 01 1 632708000000000E 01 1 632708000000000E 01 and so on Profile Data for Block 2 Face 6 1 632708000000000E 01 1 632708000000000E 01 1 632708000000000E 01 and so on End PBCA Data DPLR Code Version 4 01 0 User Manual 6 14 2 6 09 DPLR Codes Package Input Output Files 6 3 2 Input Flags for Pointwise Boundary Condition Files Input flags for a pointwise boundary condition file are discussed below in the order they appear in the file Note that the first three flags that appear in the file are not labeled Flag 1 ibtyp Flag 2 bvers Flag 3 ibdum nblk idim neq net DPLR Code Version 4 01 0 User Manual It is always 2 and should not be changed Specifies the version number of the DPLR Code Package that contained the file template Specifies if the file contains values for dummy cells Allowable values are 0 The file does not contain values for dummy cells 1 The file does contain values for dummy cells Specifies the number of maste
143. his it is always possible to determine the settings and physical constants used to generate the simulation even if the original DPLR input deck has been altered or misplaced Note Although POSTFLOW can process restart files generated by previous versions of DPLR DPLR input decks reconstructed and saved in the subdirectory created by POSTFLOW will always be generated in the format of the current version of the DPLR Code Package Specifies how cell centered finite volume flow data are represented on a node centered grid Allowable values are 0 move flow data to the lower left cell least accurate 1 interpolate grid points to cell centers See Tech Tip 1 2 interpolate flow data to grid points See Tech Tip 2 11 interpolate grid points to cell centers no boundary points See Tech Tip 3 21 interpolate grid points to cell centers even at boundaries See Tech Tip 4 Tech Tips 1 Generates cell centered grids adding additional face centered 5 7 2 6 09 nzones isep DPLR Code Version 4 01 0 User Manual Using POSTFLOW points to the boundaries Flow quantities are not interpolated to the interior of the grid thereby holding distortion of output data to a minimum Because output grid points lie at the cell centers of the original CFD grid the output grid resulting from interp 1 cannot be used to run further CFD simulations 2 Preserves the location of the CFD grid points and interpolates finite volume data on
144. ically reverse the order of the grid in the i direction in each block if necessary to ensure that the output grids and solutions files remain right handed and determine the new zonal interface definitions that result from this mirroring However you must manually correct the boundary conditions in the DPLR input deck by reversing the BC numbers of the iminand imax faces in each block And when using FCONVERT to create or mirror a restart file from an input file other than DPLR s psix format be sure to set correct values for nsin nerin nevin necin and ntbinso that FCONVERT can correctly determine the location of the velocity components in the file Specifies the number of blocks to decompose the input file into when iaction 2 ibrk jbrk kbrk Specifies grid decomposition factors in the i j and DPLR Code Version 4 01 0 User Manual k directions for each master block when iaction 1 Tech Tip The most common use of FCONVERT is to decompose an input grid file into blocks for simultaneous parallel execution on a number of processors When the number of processors to be used for the solution run has been determined the input grid file must be decomposed into at least one block per processor This can be accomplished in two ways 3 10 2 6 09 Using FCONVERT Set taction I manually determine the best strategy for decomposing the input grid file into master blocks then enter one set of decomposition factors ibrk jbrk an
145. id and can be used with or instead of zone specification extraction as defined below Specifies the flow variables to be extracted from the restart file Entries must be an array of comma or space separated integers Allowable values are Grid Coordinates 0 all grid coordinates l x coordinate x 2 y coordinate y 3 z coordinate z Grid Related Variables 10 all path lengths 11 path length along grid lines in i direction si 12 path length along grid lines in j direction sj 13 path length along grid lines in k direction sk 14 unit outward normal x direction cosine sx 15 unit outward normal y direction cosine sy 16 unit outward normal z direction cosine sz 21 body normal distance dn 22 deviation from orthogonality deg dev 23 face area Area 25 maximum cell aspect ratio CAR Mixture Transport Properties 50 total viscosity mu 51 total kinematic viscosity nu 52 total translational thermal conductivity kap 53 total rotational thermal conductivity kapr 5 12 2 6 09 DPLR Code Version 4 01 0 User Manual 54 55 56 57 58 59 Using POSTFLOW total vibrational thermal conductivity kapv free electron thermal conductivity kape total binary diffusion coefficient D mixture mean free path mfp unit Reynolds number Re L cell Reynolds number Re_c Thermodynamic Properties 60 61 62 63 64 65 66 68 69 ratio of frozen specific heats cp cv G frozen
146. ies diffusion coefficients idmod 3 in which case this flag is not used during the simulation 4 27 2 6 09 LeT ScT prtl number or temperature prtlT XXXX XXXX rvr DPLR Code Version 4 01 0 User Manual Using DPLR Specifies the value of the turbulent Lewis or Schmidt number to be employed in the simulation Relevant for turbulent viscous simulations ivis 2 12 regardless of the model used to compute species diffusion coefficients set by the idmod flag Recommended values 0 5 1 0 Tech Tip A value of 0 7 has been baselined for the Mars Science Laboratory Specifies the value of the Prandtl number to be employed in the simulation Relevant for viscous simulations ivis 0 with constant Prandtl thermal conductivity model ivmod 2 12 ikt 2 Recommended value for low air flows 0 72 Tech Tip These models should only be selected for perfect gas non reacting low temperature flows As such the value of prt1 is not usually relevant Specifies the value of the turbulent Prandtl number to be employed in the simulation Relevant for all turbulent viscous simulations ivis 2 12 irrespective of the turbulence or laminar conductivity model Recommended value 0 9 Not used in DPLR 4 0 Not used in DPLR 4 01 0 Viscous overrelaxation parameter Recommended value 1 3 default 4 28 2 6 09 resmin Using DPLR Specifies the minimum L2Norm residual for DPLR to reach to achieve a converge
147. igenvalue limiter is used in the flux extrapolation epsj 0 3 The magnitude of the eigenvalue limiter is 0 3 in the j direction kflx 4 The Euler flux extrapolation method to use in the k direction is MUSCL Steger Warming with Ap kord 3 The Euler flux extrapolation order of accuracy is third order upwind biased DPLR Code Version 4 01 0 User Manual 4 62 2 6 09 Using DPLR Block 1 Flags Setting Explanation cont cont cont omgk 2 0d0 The value of w to employ in the MUSCL scheme is 2 klim 1 The Minmod flux limiter is used in the Euler flux extrapolation kdiss 0 No eigenvalue limiter is used in the flux extrapolation in the k direction epsk 0 03 This value is ignored because kdiss 0 iextst 1 The time advancement method used when simulating this master block will be implicit data parallel line relaxation nrix 4 Four implicit data parallel line relaxation steps will be used in simulating this master block ildir 0 The lines will be formed automatically in an appropriate direction when simulating this master block ibcu 1 Implicit boundary conditions will be updated during each line relaxation step iblag 1 Implicit boundary conditions will not be lagged when simulating this master block ilt 1 Global timestepping will be employed when simulating this master block ibdir 1 This value is ignored because nb1k 2 cflm 1 0d20 This value is
148. imum Values p max 5 0043E 04 min 3 5910E 01 T max 1 5345E 04 min 1 2807E 02 M max 3 2322E 01 min 0 0000E 00 processing grid variable 1 2 3 processing flow variable 1 2 3 4 5 6 7 8 9 block 2 nx 48 ny 64 nz 64 zone t BC19 i 50 j 1 k 66 Zone Maximum and Minimum Values p max 4 4431E 04 min 3 5910E 01 T max 1 4203E 04 min 1 2807E 02 M max 3 2322E 01 min 0 0000E 00 If you set ouform 17 POSTFLOW displays a longer listing to this onscreen summary which includes the ijk locations of these maximum and minimum values in the zone DPLR Code Version 4 01 0 User Manual 5 37 2 6 09 Using POSTFLOW Tech Tip Note that the ijk location is computed relative to the output zone If ijk values for all blocks are required the entire volume should be selected as output 5 4 5 Integrated Surface Data POSTFLOW can integrate data for the following surface variables e face area ivarp 23 e total heating ivarp 531 e mass flow rate ivarp 594 e thrust ivarp 596 e aerodynamic forces ivarp 600 673 e aerodynamic moments ivarp 700 773 e species mass flow rate ivarp 5000 7 You can accomplish this by setting out form 8 and interp 11 and making sure that all output datasets define surfaces either with the iexbc or the ifac flag As with the computation of minimum and maximum values the results of this operation are only written to the screen in the standard out
149. imum spacing is limited by ds 1 of the current adapted arc length When imradial 2 ds1 sets wall spacing everywhere in the volume ds1 0 Maintains wall spacing in the current grid ds1 lt 0 Wall spacing is ds1 of the current adapted arc length When igalign 20 ds1 is the constant value by which the surface grid should be morphed in the body normal direction ds1 gt 0 Used for a recession in the surface ds1 lt 0 Used for a growth in the surface Specifies the value of the cell Reynolds number when imradial 1 For all other values of imradial celle is ignored Specifies the maximum wall spacing allowed when cell Reynolds number spacing is employed imradial 1 For all other values of imradial ds 1mx is ignored dsimx lt 0 Wall spacing is ds1mx of the current adapted arc length Tech Tip Using arc length based spacing is not generally recommended but can be helpful for certain situations such as the tail root area of the Shuttle orbiter grid Specifies the spacing for the outer boundary of the grid Recommended value 0 35 4 35 2 6 09 Using DPLR Tech Tip ds2fr is expressed as a fraction of the spacing that would be used if an unconstrained one sided Vinokur stretching algorithm were employed XXXXX Not used in DPLR 4 01 0 Overset Grid Implementation The flags in this portion of the DPLR input deck control Overset Grid capabilities of DPLR i e enable Chimera topologies Not workin
150. ing the pointwise boundary condition file pbca See Section 6 3 for more information on boundary condition files Complex blowing models can be characterized using the material response boundary conditions icatmd iregmd twal1l epsr and gamcat discussed below icatmd Specifies the model to be used for wall catalysis Allowable values are 0 3 98 99 100 101 198 200 201 DPLR Code Version 4 01 0 User Manual Disable wall catalysis wall is assumed to be non catalytic and the gradient of all species mole fractions is assumed to be zero Constant y homogeneous model value for y must be with the gamcat flag Constant y fully catalytic to ions but supports only homogeneous surface reactions such as N N NM 0 0 0 Material specific surface kinetics reaction rates for different materials are experimentally obtained and given in the catalysis surf file in the cfdinput directory Input material map allows DPLR to access a surface catalytic map specifying pointwise material properties in a boundary condition file Supercatalytic wall assumes chemical composition at the wall is identical to the freestream resulting in conservative enthalpy estimates appropriate for design studies Supcercatalytic with specified freestream most appropriate for high enthalpy simulations in ground test facilities Mitcheltree CO2 model developed for the Mars like C0 atmosphere models diffusion limited
151. ion 4 01 0 grid files zonal interface files boundary condition files runtime control files restart files chemistry files radiation files convergence files aerodynamic files log files and Tecplot files Although each file type can be written in two or more different formats not all formats are compatible with all parts of the DPLR Code For example FCONVERT can read plot3d formatted grid and function files as input but DPLR2D DPLR3D and POSTFLOW cannot See Section 9 2 for more information on file formats Grid Files Grid files define the discretized computational geometry of the CFD problem Grid files can exist in the following formats Description Suffix unformatted parallel pgrd XDR parallel pgrx ASCII parallel pgra unformatted plot3d gu XDR plot3d gx ASCII plot3d g gzipped ASCII plot3d gz The plot3d files created by third party grid generation software packages such as GridGen or GridPro can be read as input by FCONVERT which typically converts them to the XDR parallel grid file format pgrx to be used in a DPLR simulation run Tech Tip Although FCONVERT can write grid files in all of the formats listed above the preferred format for use in the DPLR working environment is XDR parallel pgrx a binary machine readable file DPLR Code Version 4 01 0 User Manual 6 2 2 6 09 6 2 6 2 1 DPLR Codes Package Input Output Files Zonal Interface Files A zonal interface or zonal boundary
152. is supplied as a utility with the DPLR 4 01 0 Code Package For more information about Template see Section 8 1 5 FCONVERT As previously discussed in Section 3 2 setting inint 2 4 in the FCONVERT input deck tells the program to automatically generate the type of zonal interface data required by DPLR for grid processing However each inint setting option offers different levels of computational speed and accuracy Setting inint 2 results in a rapid detection of full face zonal interfaces as shown in Figure 6 1 by comparing the centroid of each master block face where the centroid is computed by averaging all cells in that face Index directions of the two faces can be arbitrary as long as the centroids of a face pair are within a tolerance determined internally based on grid dimensions and clustering Because this method detects full face interfaces only this option should only be used if it is known that the input grid does not contain sub face interfaces i e areas where one block face abuts only a portion of another block face as shown in Figure 6 2 It is useful to note that the computational accuracy of this setting is comparable to that achieved using the Template software utility DPLR Code Version 4 01 0 User Manual 6 9 2 6 09 DPLR Codes Package Input Output Files Block 1 7X5 Block 2 5X7 1 2 3 4 5 6 7 7 6 5 4 3 2 1 Figure 6 1 Full Face Zonal Interface Block 1 7X5 Block 2 5X7 Figure 6 2 Sub
153. iter tail region On the downside this option will tend to preserve oscillations in the outer boundary if they develop during the solution procedure 2 Setting Ismooth 2 is the only approach that can be used for smoothing a grid without adaption igalign 3 because all of the arc length changes are zero Also this option is preferable for the first adaption when there is an extremely large ratio between the shortest and longest arc lengths in the grid particularly if short arc lengths occur near a region with high curvature such as a wing leading edge 3 Setting ismooth 3 causes the outer boundary of the grid to be smoothed using both algorithms described above producing superior quality outer boundaries for the initial adaption of hyperbolic grids Further adaptions should then be peformed with ismooth 1 Specifies the fraction of the freestream Mach number to pick as the adaption contour Recommended values 0 9 lt fs_scale lt 0 95 4 33 2 6 09 ds_mult gmargin dsl DPLR Code Version 4 01 0 User Manual Using DPLR Tech Tip Values smaller than 0 90 lead to smoother grids but increase the chance that the final outer boundary will not contain the entire shock Specifies as a multiple of the local radial grid spacing at the estimated shock location where to place the realigned outer boundary beyond that location Must be gt 0 Recommended values range 1 0 3 0 with 2 5 being typical Tech Tip The i
154. ith a larger limiter as the default value to avoid computed heat transfer at the vehicle surface being too small By setting 1ikeq 31 39 DPLR scales down the 1 9 model selected by 75 Although extreme situations such as these are unlikely to occur during routine use of DPLR the capability of dealing with them does exist for the advanced user who needs to do so Specifies the model used to compute the vibrational energy component of the gas Allowable values are 0 Neglect vibrational energy 1 Vibrational nonequilibrium single 7 recommended for all planetary and high velocity Earth entry flows 2 Vibrational equilibrium using statistical mechanics 3 Complete thermal equilibrium using NASA LeRC curve fits based on data from NASA s Lewis Research Center s now Glen Research Center computer program CEA NASA Reference Publication 1311 Can be employed for low altitude hypersonic flight or some Shuttle type reentry trajectories where vibrational nonequilibrium is not very important 4 Two temperature model using LeRC curve fits T Tea T not recommended 5 Two temperature model using LeRC curve fits T T T Te Ta NOT WORKING in DPLR 4 01 0 11 Rotational nonequilibrium multiple T VOT WORKING in DPLR 4 01 0 Tech Tip If ivib 3 or 4 the values of irot ieex and iel are ignored because these settings uniquely determine the apportionment of internal energies among the various modes 4 17 2 6
155. l in that direction to start with i j kmax cell in that direction to end with 2 interp 2 interpolate flow data to grid points i j kmin grid point in that direction to start with i j kmax grid points that direction to end with shorthand values independent of interp setting i j kmin cell or grid point to start with i j kmax l extract all values in this direction 2 extract all values in this direction less 1 3 extract all values up to the midpoint in this direction DPLR Code Version 4 01 0 User Manual 5 24 2 6 09 bkmin bkmax zonetitle Using POSTFLOW Tech Tip To extract data from a plane set the min and max values in that direction to be the same Specifies the range of master block numbers from which to extract data Entering the shorthand value of 1 in the bkmax flag tells POSTFLOW that the value of bkmax is the number of the last block in the simulation For example in a simulation composed of four master grid blocks bkmin 2 bkmax 1 tells POSTFLOW to extract data from master blocks 2 3 amp 4 An ASCII string surrounded by single or double quotes that will be used to name the zone if Tecplot output is specified If a zone name is not desired this flag should contain an empty string as shown below Tecplot output name of the zone 6699 Non Tecplot output I O Filenames All filenames must be enclosed with single or double quotes fname pname
156. l damage repair grids that are outside any sonic bubble BC 2 should also be adequate for the supersonic outflow faces of ordinary grids although occasional anomalies have been observed in baseline Shuttle solutions so substituting BC 3 for BC 2 is recommended for such known outflow faces Supported Input Output File Formats The DPLR Code Package Version 3 06 reads and or creates the following six file types e Grid files defining the discretized computational geometry of the problem e Zonal Interface files describing how the blocks in multi block grids abut each other in computational space e Restart or cfd function files saved periodically by the CFD code to be used to restart a problem and or post process the solution e Radiation files enabling loose coupling between DPLR and flowfield radiation analysis tools such as RADEQUIL NEQAIR and HARA e Boundary Condition files specifying various types of pointwise boundary conditions and or TPS material maps e Data files generated by POSTFLOW for use in post processing and data analysis of the solution For a more detailed discussion of each of the file types see Section 6 in this User Manual The above listed file types can exist in different formats File formats supported by the DPLR Code Package are listed in the table below Note that each supported format is assigned a unique number and a suffix which is common across the entire code package DPLR Code V
157. ll not be able to read the file accurately DPLR Code Version 4 01 0 User Manual 3 3 2 6 09 Using FCONVERT Step 2 Enter values for each of the input variables or flags See Section 3 2 for a description of input flags and a list of allowable values Action For each flag type allowable problem specific value Result Input deck contains sufficient information for FCONVERT to process the input grid file and convert it into a DPLR readable file Step 3 Save the file with your problem specific name to your working directory Step 4 Run FCONVERT Action At the command line prompt type fconvert lt yourinputdeckfilename inp Result An output grid file in the format you specified through the ouform flag usually XDR parallel is created along with on screen summary of actions performed by FCONVERT See Section 3 3 for an example of a problem specific FCONVERT input deck and the output summary generated after running the program 3 2 Input Flags for FCONVERT Input variables for FCONVERT are discussed below in the order they appear in the deck iaction Specifies the action to perform Allowable values are 0 test decompose over a range of blocks 1 decompose file according to ijk brk 2 decompose file according to nbreak 3 recompose file into original blocks 10 format conversion only no parallel decomposition or recomposition scaling or sequencing still allowed 11 stop after printing file size determin
158. ll three symmetry flags are valid for 3D flows and none are valid for a 2D or axisymmetric flow Specifies the velocity vector orientation axis aka global wind for calculation of certain output variables i e ivarp values Allowable values are 0 do not alter the raw output data 1 determine sign by a dot product with freestream vector recommended for simulations with only one freestream specification 2 determine sign by a dot product with supplied wind vector recommended for simulations with more than one freestream specification 5 10 2 6 09 cxs CyS CZs iexbc DPLR Code Version 4 01 0 User Manual Using POSTFLOW Tech Tip Because the global wind axis is used either to determine the sign of the output skin friction shear stress or to convert output forces into a wind oriented lift and drag coordinate system iwind will be ignored unless i varp 600 673 700 773 Specifies the direction cosines of the global wind axis in the xyz directions when iwind 2 Tech Tips 1 These are defined as unit metrics such that 2 2 Dies exs cys czs 1 and the components u v and w of the freestream velocity vector V are given by u Vecxs v Vecxy w Vecxz 2 User input values are always normalized by POSTFLOW to ensure that these expressions are valid Boundary condition number s for the wall or surface from which data will be extracted Allowable values listed in Section 4 2 for the DPLR input
159. lled ibgoto is available from http www tacc utexas edu resources software DPLR Code Version 4 01 0 User Manual 2 3 2 6 09 2 3 Installation Guide Installing the DPLR Code Package The 4 01 0 version release of the DPLR Code Package consists of two gzipped tar files designated as follows Step 1 Step 2 Step 3 Step 4 dpcodeV4 01 0 tar gz containing four separate executables samples V4 01 0 tar gz containing a set of sample problems complete with grids input decks and running instructions Unzip the DPLR Code file Action At the command line prompt type gunzip dpcodev4 01 0 tar gz Result The archived file is renamed dpcodev4 01 0 tar Untar the DPLR Code file Action At the command line prompt type tar xvfz dpcodev4 01 0 tar Result A directory structure is created See Section 2 4 for more information on directories and files Run the config script Action At the command line prompt type config Result If you are attempting to compile on a system the config script can recognize amakefile comm file is generated containing machine specific information for your system Otherwise you will need to modify such a file yourself Samples based on known MPI and FORTRAN builds can be found in the defs directory After extracting the archive a blank makefile comm is created that includes descriptions of each of the necessary compiler options and paths Create executables files
160. locks in the input plot3D grid ninta Specifies the number of zonal interfaces in the input plot3D input grid DPLR Code Version 4 01 0 User Manual 6 4 2 6 09 DPLR Codes Package Input Output Files ninte Specifies the number of corner edge zonal interfaces in the input plot3D grid Allowable values are 0 Input grid contains no corner edge zonal interfaces Non zero value meant for debugging Tech Tip This option is meant for developers to use in debugging If nintc gt 0 FCONVERT will automatically strip the corner edge zonal interface information before processing the file nz Specifies the grid blocks that define the common face of the zonal boundary being described nface Specifies the block faces that abut thereby indicating the plane in which the zonal boundary lies Allowable values are 1 imin face imax face jmin face 2 3 4 jmax face 5 kmin face 6 kmax face Tech Tip If the grid is for a 2D or axisymmetric problem nface must 1 4 since such problems are assumed to lie in the ij plane ndr1 Specifies the first extent direction of the zonal boundary being described Allowable values are 1 i direction 2 j direction 3 k direction nst1 Specifies the starting point of the interface range from cell center to cell center in the direction indicated by the value in ndr1 nenl Specifies the ending point of the interface range from cell center to cell center in the direction indi
161. lr interfaces file you will see that Template has created a zonal interface file for use in your simulation This method detects full face interfaces only unlike FCONVERT which has the option of detecting subfacing through different settings of inint Thus the zonal interface file generated by Template should only be used when no sub face interfaces exist in the computational grid See Section 9 1 5 for a more complete discussion of the Template utility Understand Your Computing Resources The efficiency of the DPLR Code Package as a CFD solver depends in part on the number of processors available for parallel solution of your flow problem Thus the more processors you can allocate to your simulation run the less time it will take to achieve a solution In addition to raw computing power however knowing the exact number of processors that can be dedicated to your solution will allow you decompose the plot3d input grid into computational blocks that can be most efficiently handled by your computing resources This is accomplished in the FCONVERT input file by setting iaction 2 nbreak n where n is the number of available processors DPLR Code Version 4 01 0 User Manual 7 8 2 6 09 Chapter 8 Appendices Contents 8 0 8 1 8 2 8 3 8 4 8 5 Introduction DPLR Code Version 4 01 0 Utilities 8 1 1 zbconvert 8 1 2 dpconvert 8 1 3 seginput 8 1 4 Moment 8 1 5 Template Supported I O File Formats 8 2 1 Format Numbers P
162. lues or set kb1 to a positive integer to turn off the eigenvalue limiter within the boundary layer 4 39 2 6 09 eps ijk iextst nrlx ildir DPLR Code Version 4 01 0 User Manual Using DPLR Specifies the magnitude of the eigenvalue limiter to use in the Euler flux extrapolation Recommended values for hypersonic blunt body flow simulations are 0 No added dissipation Recommended for body normal direction 0 3 Recommended for radial and circumferential directions Tech Tip Much lower values of eps ijk on the order of 0 01 can be used in separated flows which have no strong shocks and are much more sensitive to the effects of added artificial dissipation DPLR will print a run time warning if it detects a non zero value of eps 1jk in the body normal direction in any block Specifies the time advancement method to use in the simulation Allowable values are 1 Explicit first order Euler 2 Explicit second order Runge Kutta 1 Implicit data parallel line relation DPLR recommended value for steady state problems 2 Implicit data parallel full matrix FMDP Tech Tip For time accurate calculations only the relatively inefficient second order Runge Kutta Midpoint method is offered at this time Specifies the number of implicit relaxation steps to use when using the DPLR or FMDP time advancement methods iextst 1 or 2 Recommended value 4 Specifies the direction in which the lines are to be
163. lues used during sequencing iname coarsened restart file name oname new less sequenced restart file name pslx Step 3 Save file to your working directory Step 4 Run FCONVERT lt new FCONVERT input file DPLR Code Version 4 01 0 User Manual 7 6 2 6 09 7 2 2 7 2 3 DPLR Workflow Step 5 Open and name another new FCONVERT input file Step 6 Set ifile 1 inform 2 imseq 0 iname original plot3d grid filename oname new less sequenced XDR parallel grid file name pgrx Step 7 Save file to your working directory Step 8 Run FCONVERT lt second new FCONVERT input file Your working directory now contains an upsequenced restart file that can be used to start a new solution run along with the DPLR readable grid file containing the same number of data points as the upsequenced solution file See Section 3 5 2 for more information on Mesh Sequencing Use Runtime Control Files to Adjust Grids and CFL Schedules With DPLR Code Version 3 06 you no longer need to wait until your initial solution run is complete to adjust your first guess grid or change your CFL timestepping schedule and then re run the simulation Using a runtime control file you can dynamically interact with a simulation mid run while monitoring the progress of convergence with concurrently running graphic visualizations of restart files as they are being written during your DPLR run By using this option you may avoid the need to rep
164. ly written as a stand alone tool it has functionality that is not being used in this mode Template Template is a Fortran utility that can be used to automatically generate e zonal interface files from PLOT3D grid files e block specific portions of the DPLR input deck containing boundary condition information Manually creating DPLR input and zonal interface files can be a time consuming task However the Template utility created by Scott Thomas and David Saunders and distributed with the DPLR Code Version 4 01 0 Package can automate some or all of these two tasks depending on the grid complexity Note that FCONVERT can also generate the interface file see inint but not the boundary condition portion of the input deck Overview The name Template derives from its original intent namely generation of most of the connectivity file for the multiblock flow solver FLO107MB Block faces not adjacent to other block faces were left for their boundary conditions e g subsonic outflow to be edited into the one line per block template manually The grid blocks were and still are expected to be point to point matched Grids with subfacing can still be processed but some of the interfaces will not be identified See use of FCONVERT for subface cases The grid may contain more than one layer of blocks but following adaptation for DPLR users Template outputs are most complete for the common case of a single DPLR Code Version
165. mal conductivity Schmidt Numbers 87 laminar Schmidt number Sc 97 turbulent Schmidt number Sc_t The Schmidt number Sc is defined as Soe pD DPLR Code Version 4 01 0 User Manual 8 23 2 6 09 8 4 4 Appendices where u is the mixture viscosity p is the mixture density and D is the binary diffusion coefficient Prandtl Numbers 88 laminar Prandtl number Pr 98 turbulent Prandtl number Pr_t The Prandtl number Pr is defined as Pr uC K where u is the mixture viscosity Cp is the total specific heat at constant pressure and K is the thermal conductivity Mixture Flow Properties Stagnation Quantities 102 stagnation mixture density r_o 112 stagnation pressure p_o 122 stagnation temperature T_o Stagnation quantities density pressure and temperature are computed assuming isentropic relations and thus are not valid for a flowfield with varying isentropic exponent y The stagnation quantities are defined as DPLR Code Version 4 01 0 User Manual 8 24 2 6 09 Appendices ave p ps p pS T T o where S is the entropy defined below Pressure 111 dynamic pressure Q The dynamic pressure Q is simply Q pV 2 114 pressure coefficient C_p The pressure coefficient is defined as p p Q where Q is the freestream dynamic pressure Temperature 121 bulk temperature T_b The bulk temperature is defined as in AIAA Paper No 2001 2886 DPLR Code Ve
166. maras Turbulence Model Automatic Grid Adaption Template Utility SST Turbulence Model Baldwin Lomax Turbulence Model Chapman Viscosity Model User Training User Manual 2 6 09 Contents 1 0 Overview 1 0 1 1 1 2 Introduction DPLR Code Package How to Use This Manual 2 Installation Guide 2 0 2 1 2 2 2 3 2 4 Introduction System Requirements Software Installing the DPLR Code Package Directory File Contents 3 Using FCONVERT 3 0 Introduction 3 1 Running FCONVERT 3 2 Input Flags for FCONVERT 3 3 Neptune Sample Case 3 4 Parallel Decomposition 3 5 Mesh Sequencing 4 Using DPLR 4 0 Introduction 4 1 Running DPLR 4 2 Input Flags for DPLR 4 3 Neptune Sample Case 4 4 Monitoring the DPLR Run 5 Using POSTFLOW 5 0 5 1 5 2 5 3 5 4 Introduction Running POSTFLOW Input Flags for POSTFLOW Neptune Sample Case Extracting Datasets 6 DPLR Input Output Files 6 0 6 1 6 2 Introduction Grid Files Zonal Interface Files DPLR Code Version 4 01 0 User Manual ii 2 6 09 6 3 6 4 6 5 6 6 6 7 6 8 6 9 6 10 6 11 Contents Boundary Condition Files Runtime Control Files Restart Files Chemistry Files Radiation Files Convergence Files Aerodynamic Files Log Files Tecplot Files 7 DPLR Workflow 7 0 7 1 7 2 Introduction DPLR Workflow Chart Workflow Shortcuts 8 Appendices 8 0 8 1 8 2 8 3 8 4 8 5 Introduction DPLR Code Version 4 0 Utilities Supported I O File For
167. mats Parallel Decomposition POSTFLOW Output Variables Reference Terms DPLR Code Version 4 01 0 User Manual ili 2 6 09 Chapter 1 Overview Contents 1 0 Introduction 1 1 DPLR Code Package 1 2 How to Use This Manual DPLR Code Version 4 01 0 User Manual 1 1 2 6 09 Overview 1 Introduction Accurate predictions of the environment that a spacecraft will encounter when entering and passing through a flow field such as the Earth or Mars atmosphere at hypersonic speeds can be of enormous value to aeronautical designers Such predictions become even more critical when the shape of the craft changes during a mission due to extreme surface ablation or unexpected damage The ability to anticipate flow environment changes relative to new craft geometries adds vital data to the situation analyses that inform mission command decisions Data Parallel Line Relaxation DPLR code is a computational fluid dynamic CFD solver that was developed at NASA Ames Research Center to help mission support teams generate high value predictive solutions for hypersonic flow field problems in a minimum amount of time using readily available computational resources The DPLR Code Package is an MPI based parallel full three dimensional Navier Stokes CFD solver with generalized models for finite rate reaction kinetics thermal and chemical nonequilibrium accurate high temperature transport coefficients and ionized flow physics incorporated into the code DPL
168. modeling data restart file format NASA Ames Version 4 01 0 solution run at Thurs Feb 5 08 18 10 2009 run in 500 iterations in 4 23E 03 seconds CPP macro settings enabled during run AMBIPOLAR 1 PARKTEXP 0 50 NOHTC Keq limiter set at 100 00 input ns 5 ner 0 nev 0 net 0 number of blocks 2 file dimension 3 extracting the following BCs 25 26 note that extraction of pointwise BCs not supported yet output variables x y z p T tau qw running in high memory mode processing grid variable 1 2 3 interpolating grid to cell centers processing flow variable 1 2 3 4 5 6 7 8 9 10 block 1 nx 32 ny 16 nz 64 gt extracted derivative data from the KMIN surface gt derivative data computed using full viscous fluxes zone t BC26 i 34 j 18 k 1 processing grid variable 1 2 3 processing flow variable 1 2 3 4 5 6 7 8 9 10 block 2 nx 48 ny 64 nz 64 gt extracted derivative data from the KMIN surface gt derivative data computed using full viscous fluxes zone t BC26 i 50 j 66 k 1 writing tecplot file postsurf dat using grid file neptune 8PE pgrx using flow file neptune pslx Figure 5 POSTFLOW Input Deck for Surface Analysis of Neptune Probe DPLR Code Version 4 01 0 User Manual 5 32 2 6 09 5 3 4 5 4 Using POSTFLOW Neptune Output Information In addition to verifying the values entered into the POSTFLOW input deck the POSTFLOW output summary displays information a
169. n and in the log file log these are both subsets of what is echoed to the screen during a DPLR run To capture all the information created ec during a DPLR run you can set up your own user log file to run in the background by typing the following at the command prompt mpirun np X machinefile machine inp Spath dplr2d or dplr3d lt yourdplrinputfilename gt userlogfile amp This will result in your system returning you to the command prompt and capturing the screen data in the background for your future reference See Section 6 8 6 9 and 6 10 for more information on convergence aerodynamic and log files Controls the computation of grid dummy cell coordinates Allowable values are 0 Only compute if necessary preferred setting l Always recompute use if boundary conditions are changed during simulation e g if a setup error is detected 99 Only compute if necessary output debugging files for use by code developers 99 Recompute and output debugging files for use by code developers Tech Tip When a grid file is created by FCONVERT dummy cell values are not computed because FCONVERT does not have information about the correct boundary conditions to enforce at each grid face DPLR will automatically generate the correct 4 25 2 6 09 kbl kdg istate iresv DPLR Code Version 4 01 0 User Manual Using DPLR dummy cell coordinates for each block based on the supplied boundary con
170. n 4 01 0 User Manual 5 35 2 6 09 5 4 3 Using POSTFLOW iexbc 17 18 19 3 This setting tells POSTFLOW to extract the ivarp specified variables for the x y z planes of symmetry and the first order extrapolation of the supersonic exit surface Extracting data via the iexbc flag is a powerful tool within POSTFLOW and should be used whenever possible to simplify extraction of complex surface datasets Tech Tip Note that the iexbc flag can be used together with the zone specification lines in a single POSTFLOW run to extract BOTH surface and volume datasets By using a combination of these methods it should be possible to extract almost any desired subset of flowfield data Line Data at the Intersection of Two Boundaries The iexbc flag can also be used to extract data at the intersection of two surfaces such as along the vehicle centerline To extract surface intersections specify your two desired boundary conditions and separate them with a forward slash For example if you want to extract quantities on a radiative equilibrium catalytic surface ibc 26 along the xz symmetry plane ibc 18 you would enter iexbe 26 18 The first number is always the reference boundary telling POSTFLOW how to extract desired derivative quantities such as heat flux and shear stress The second number is the boundary condition that you want to intersect with the reference boundary You can request multiple intersections in a single POSTF
171. n and load imbalance Tech Tip Although FCONVERT generated solutions for block decomposition are computationally accurate they may not be the most practical way to handle grids for complex object geometries Therefore most DPLR users choose to keep iaction 1 and determine from their own experience the best way to break the master blocks in the input grid recognizing that there are algorithmic limits on how small parallel blocks should be and thus determining how many processors the problem will require Parallel Decomposition DPLR is a distributed memory parallel code so solutions for each grid block are computed simultaneously rather than sequentially Multi block information transfer is handled through MPI data constructs so it is necessary to run on at east as many processors as blocks in the original computational grid Running on more processors than master grid blocks is often advantageous since the largest blocks can then be split decomposed into smaller pieces increasing computational efficiency and decreasing turnaround time This decomposition if required is the most common reason for running FCONVERT Although the ideal number of processors to use for a given job is a matter of personal preference it is generally a function of the total number of processors that DPLR Code Version 4 01 0 User Manual 3 19 2 6 09 3 4 1 Using FCONVERT are available and the number that are necessary to achieve a reasonable measur
172. n on any number of processors without further processing by FCONVERT Parallel Recomposition FCONVERT can also be used to recompose a grid file that was previously decomposed by setting iaction 3 in the FCONVERT input deck This option can only be used with grid files because restart boundary condition and radiation files are never decomposed in the first place In practice this setting is rarely used because it is unnecessary to recompose parallel archival files As previously discussed when the FCONVERT output file is written in one of the parallel archival formats ouform 1 11 or 21 any decomposition is virtual This means that the file merely contains header information instructing DPLR2D or DPLR3D how to properly decompose the file at runtime eliminating any need to actively recompose the file If iact ion 3 is specified with an parallel archival file as input FCONVERT will only strip the virtual decomposition information from the file header If you do set iaction 3 you will need to specify the number of blocks in the recomposed file with the nborig flag If the input file is in plot3d format you must also provide the input interface file inint 1 containing information about how the original grid or restart file was decomposed Although FCONVERT will recompose an input grid file it does not recreate the zonal interface file for the recomposed problem Therefore be sure to save the original zonal interface file to
173. nalign 1 This value is ignored because igalign 0 imedge 1 This value is ignored because igalign 0 DPLR Code Version 4 01 0 User Manual 4 60 2 6 09 Using DPLR Grid Setting Explanation cont Adjustment cont Alignment Morphing Flags cont imradial 2 This value is ignored because igalign 0 ngeom 2 This value is ignored because igalign 0 ismooth 3 This value is ignored because igalign 0 fs_scale 0 95 This value is ignored because igalign 0 ds_mult 2 5 This value is ignored because igalign 0 gmargin 0 0 The outermost radial spacing of the grid will remain as specified in the grid file dsl 0 0 This value is ignored because igalign 0 cellRe 1 0 This value is ignored because igalign 0 dslmx 1 0d 4 This value is ignored because igalign 0 ds2fr 0 3 This value is ignored because igalign 0 Overset Grid Setting Explanation Implementation Flags iover 0 Overset logic is disabled for this simulation ioint 1 This value is ignored becaue iover 0 XXXX 1 This value is not used in DPLR 4 01 0 Space Setting Explanation Marching 1D Impementation Flags ispace 0 Space marching is disabled in this simulation dxmin 1 0d 5 This value is ignored because ispace 0 slength 1 0d0 This value is ignored because ispace 0 nxtot 1000 This value is ignored because ispace 0 DPLR Code Version 4 01 0 User Manual 4 61 2 6 09
174. nformation on boundary condition files rname Specifies the name of the input surface radiation file and will typically have the suffix prdx This file is optional and read only if volumetric radiation data are input and the irad flag is set to 1 If the file is not required for the simulation use none as the filename cname Specifies the name of the input overset connectivity file and will typically have the suffix dci if ioint 1 This file is only required in iover 1 If overset logic is not enabled use none as the filename dname Specifies the name of the input chemistry file and will typically have the suffix chem This file is required and must exist in the cfdinput directory that is created when you install the DPLR Code Package See Section 2 4 for information on the directory and file structure of the DPLR Code Package and Section 6 7 for more information on chemistry files Tech Tip Unlike other input files the absolute pathname to this file must be specified in the input deck Global Modeling Flags These flags are for values that remain constant for all blocks of the simulation nblk Specifies the number of master grid blocks in the simulation This is the same value as nborig in the FCONVERT input deck and will be less than or equal to the number of processors on which the job is run igrid Specifies the format of the input grid file gname Allowable values are
175. ng initial design analysis Maximum temperature specified with twa11 flag DPLR will automatically switch between an isothermal and radiative equilibrium wall on a pointwise basis if this option is used 201 209 Material specific with a view factor correction allows you to modify the surface emissivity in a pointwise manner to enable a simple view factor correction to the hemispherical emissivities for internal and or cavity flows Tech Tip A radiative equilibrium wall is a common design model that assumes all energy incident to the surface of a vehicle is reradiated to space at a rate consistent with the emissivity of 4 14 2 6 09 twall epsr gamcat XXXX vwall ichem DPLR Code Version 4 01 0 User Manual Using DPLR the wall material As the materials knowledge base increases DPLR can be used to simulate how the overall flow environment contributes to and in turn is affected by radiative emissivity of vehicle surfaces during hypersonic flight Specifies the wall temperature to be used for isothermal temperature capped radiative equilibrium wall simulations Also specifies wall temperature if ireqmd gt 100 Specifies the constant value of emissivity to be used for radiative equilibrium wall simulations ireqmd 1 Specifies the constant value of catalytic effcienty y to be used for catalytic wall simulations icatmd 1 2 Not used in DPLR 4 01 0 Specifies the wall velocity or blowing rate
176. nless used for software development Specifies whether mesh sequencing or coarsening is to be performed and in which computational direction s See Section 3 6 for more information on mesh sequencing Allowable values are 0 Do not sequence the file 1 Sequence according to the values of ijk seq 2 Sequence all blocks using ijk seq values 2 Upsequence a restart file Instructs FCONVERT to scale an input grid file if ile 1 by a constant multiplicative factor sfact before creating a DPLR readable output grid file Allowable values are 0 Do not scale input grid file 1 Scale input grid file by sfact Tech Tips 1 This option is typically used to convert grids to SI units 2 If iscaleis set to l for any file type other than a grid file FCONVERT will silently reset it to zero 3 9 2 6 09 sfact imir nbreak Using FCONVERT Specifies the multiplicative scale factor to use when iscale l Specifies whether to mirror the input grid or restart file across one or more axes Allowable values are 0 Do not mirror input file 1 Mirror input file Tech Tip imir is used primarily to generate a reflected grid or restart file in preparation for starting a full body simulation If imir is set to 1 valid entries must be specified for imirx imiry and imirz mirroring factors across the yz xz and xy axes When mirroring is turned on FCONVERT will mirror the appropriate xyz or uvw variable on output automat
177. nplot 100 then after 1000 iterations the following files will exist sample pslx sample ps x 900 sample pslx 800 Note that older restart files created every 100 iterations as specified by the value in nplot are not saved Enables axisymmetry in a DPLR2D simulation run Allowable values are lor 0 Non axisymmetric 2D 1 Axismmetric about x axis 2 Axisymmetric about the y axis 4 23 2 6 09 ires DPLR Code Version 4 01 0 User Manual Using DPLR Tech Tip DPLR2D simulates axisymmetric flows by solving the Navier Stokes equations in cylindrical rather than Cartesian coordinates This allows for an axisymmetric simulation in about the same total solution time as a 2D result The rotation axis of the problem is always assumed to be either the x or y axis Note that DPLR2D simulations are always in the xy plane so rotation about the z axis is not permitted Specifies the type of residual and convergence data that are tracked and output to the screen and to the convergence file Allowable values are 0 Do not output a convergence file Output nit global residual and At iteration number summed residual over all computational blocks timestep for the iteration number Output nit global residual and CFL number Output nit global residual and CPU time Output nit global residual and flow time nA BW N Output nit global residual CFL number and aero data 11 Output nit block residual and At 12 Output nit block
178. nstant volume 7200 n specific heat at constant volume of species n cv_n 8000 all species gas constants 8000 n gas constant of species n R_n 8200 all species equivalent degrees of freedom nkT 8200 n equivalent degrees of freedom of species n dof n 8400 all species partial pressures 8400 n partial pressure of species n p_n 8600 all species mean thermal speeds 8600 n mean thermal speed of species n cbar_n 8800 all species chemical formation energies per unit mass 8800 n formation energy per unit mass of species n eh_n 10000 all species diffusion coefficients 10000 n diffusion coefficient of species n D_n 10200 all species ambipolar diffusion effectiveness 10200 n ambipolar diffusion effectiveness of species n DaC_n 10400 all species effective Schmidt numbers 10400 n effective Schmidt number of species n Sc_n 10800 all species unit diffusion mass fluxes 10800 n unit diffusion mass flux of species n MD_n Tech Tips 1 A single set of output variables may be specified for a given run of POSTFLOW If a variable that is not permitted by the simulation specifications is selected for extraction such as the coefficient of viscosity from an Euler simulation POSTFLOW will remove it from the ivarp array and echo a message to the screen 5 22 2 6 09 Using POSTFLOW 2 The list of species specific variables includes some italicized macro selections that allow extraction of several related items For
179. nterface files There are three interfaces between these two blocks The following table explains the meaning of the input deck settings in this sample case Input Flag Setting Explanation iaction 1 Break each master block the input grid along the i j and k axes as specified in ibrk jbrk kbrk for each block ifile 1 The input file is a grid file idim 3 The input file is a 3D file iinfo 0 Do not output debugging information ivers 1 Do not attempt to change file version nvers 4 01 0 Release version of the DPLR Code package being used Value ignored when ivers 1 inform 22 Input file is an ASCII plot3D grid file inint 1 Read input interface file idummy 0 Input file does not contain dummy cells nborig 2 There are 2 master blocks in the input grid file ouform 11 The output file will be an XDR parallel grid file ouint 0 Do not write an output interface file one already exists odummy 0 Output file does not contain dummy cells ncedge 1 Compute all edge and corner interfaces imseq 0 Do not sequence the input grid file iscale 0 Do not scale the input grid file sfact 1 0 Ignored value because iscale 0 imir 0 Do not mirror input grid file nbreak 1 Ignored value because iaction 1 ibrk jork kork 1 1 1 Do not break the first master block in any direction Tal Break the second master block 7 times in the direction only DPLR Code Version 4 01 0 User Manual
180. ock 1 into 1 ibrk 1 jbrk 1 kbrk 1 Decomposing block 2 into 4 ibrk 2 jbrk 1 kbrk 2 5 Blocks Total load imbalance 10 00 Decomposing block 1 into 1 ibrk 1 jbrk 1 kbrk 1 Decomposing block 2 into 5 ibrk 1 jbrk 1 kbrk 5 6 Blocks Total load imbalance 7 69 Decomposing block 1 into 1 ibrk 1 jbrk 1 kbrk 1 Decomposing block 2 into 6 ibrk 2 jbrk 1 kbrk 3 7 Blocks Total load imbalance 4 32 Decomposing block 1 into 1 ibrk 1 jbrk 1 kbrk 1 Decomposing block 2 into 7 ibrk 1 jbrk 1 kbrk 7 8 Blocks Total load imbalance 5 26 Decomposing block 1 into 1 ibrk 1 jbrk 1 kbrk 1 Decomposing block 2 into 8 ibrk 2 jbrk 2 kbrk 2 9 Blocks Total load imbalance 0 00 Decomposing block 1 into 1 ibrk 1 jbrk 1 kbrk 1 Decomposing block 2 into 9 ibrk 3 jbrk 1 kbrk 3 10 Blocks Total load imbalance 9 99 Finished with Load Balance Check From this output summary you can see that a perfectly load balanced solution is possible if the problem is decomposed to run on nine processors DPLR Code Version 4 01 0 User Manual 2 6 09 8 3 5 8 3 6 Appendices Single Block Input Files In general parallel decomposition must be performed by FCONVERT However in the special case of a single block grid with no zonal interfaces DPLR2D and DPLR3D can perform parallel decomposition at runtime In this case the input grid file can simply be converted to parallel archival format iaction 10 The resulting file can be ru
181. olation supersonic exit 4 Second order extrapolation supersonic exit 5 RESERVED 6 Subsonic reservoir inlet constant mass flow 7 Periodic 8 Inviscid wall flow tangency 9 Viscous adiabatic wall 10 Viscous isothermal wall 11 180 degree singular axis u u 3D 12 180 degree singular axis v v 3D 13 180 degree singular axis w w 3D 14 Singular x axis v v axi 15 Singular y axis u u axi 16 360 degree singular axis 3D 17 Plane of symmetry u u 4 43 2 6 09 DPLR Code Version 4 01 0 User Manual 18 19 20 21 22 23 24 25 26 27 30 35 36 37 38 39 47 49 Using DPLR Plane of symmetry v v Plane of symmetry w w Zone boundary 90 degree singular axis v v w w 3D 90 degree singular axis u u w w 3D 90 degree singular axis u u v v 3D RESERVED Catalytic isothermal wall Catalytic radiative equilibrium wall Non catalytic radiative equilibrium wall Viscous isothermal wall with blowing Catalytic isothermal wall with blowing Not Working in DPLR 4 01 0 Catalytic radiative equilibrium wall with blowing Not Working in DPLR 4 01 0 Non catalytic radiative equilibrium wall with blowing Not Working in DPLR 4 01 0 inviscid wall with blowing Not Working in DPLR 4 01 0 viscous adiabatic wall with blowing Not Working in DPLR 4 01 0 Viscous isothermal wall with slip Catalytic isothermal wall with slip Not Working in DPLR 4 01 0
182. old inp and convert it to the current version However you can also specify a desired output version number other than the current version using the V option seqinput seqinput is a Perl script that can be used to easily sequence coarsen a DPLR input deck It works by dividing the grid sizes of each block in the input deck by a specified sequencing factor The script is run from the command line seqinter i old inp o new inp s I J K where old inp infile the original file you are converting new inp outfile the sequenced file s slist acolon separated list of sequencing factors in the i j and k directions it is assumed that all blocks are sequenced by the same factors DPLR Code Version 4 01 0 User Manual 8 3 2 6 09 8 1 4 Appendices At runtime the script will generate a new DPLR input deck and rename the input grid and restart files with the suffix sIJK a designation you can change to whatever naming convention your are using Moment Moment is a Fortran code that generates integrated force and moment data from an input set of pointwise surface forces When POSTFLOW is run using ouform 11 POSTFLOW will automatically generate a plot3d grid file a cfd function file and a Moment inp file which is the input deck for the Moment utility Once these files have been generated Moment is run from the command line by typing Moment lt Moment inp A sample of the output from t
183. only used of these workflow shortcuts Sequence the Grid As discussed in Section 3 5 computational grids composed of a large number of data points typically take longer to solve than grids with fewer points As a result grids used for initial solutions of CFD problems are sometimes coarsened or sequenced to reduce the number of points while maintaining the topology of the mesh After an acceptable first guess is acquired the grid is restored in a step wise fashion to its original number of points for final solution and post solution data reporting To sequence or coarsen an input grid open the fconvert inp file and enter the following settings imseq 1 iseq n jseq n kseq n where n is the number of times the grid for that block should be coarsened in the i j k directions See Section 3 5 1 for more information on this technique To restore grid points and refine your solution use FCONVERT to upsequence your restart file and create a new pgrx file that matches the refined level of your upsequenced restart file coarsened restart file 1 revised feonvert inp gt FCONVERT gt upsequenced restart file plot3d grid file 2 revised fconvert inp gt FCONVERT gt refined perx grid file that matches points in the upsequenced restart file This can be accomplished by following the steps below Step 1 Open and name a new fconvert inp file Step 2 Set ifile 2 inform 11 imseq 2 iseq jseq kseq to va
184. ormalized by V u Vin velocity in the y direction normalized by V v Vin velocity in the z direction normalized by V w Vin limited rotational temperature Tr_1 limited vibrational temperature Tv_1 limited electronic temperature Te_1 limited free electron temperature Tel_1 Viscous Derivative Based Quantities 501 502 503 504 507 511 512 517 518 520 521 522 523 524 525 526 skin friction coefficient Cf unit viscous force on a face in x direction tau_x unit viscous force on a face in y direction tau_y unit viscous force on a face in z direction tau_z total wall shear stress tau Stanton number based on wall enthalpy Ch Heat transfer coefficient in mass flux units Chm Stanton number based on freestream conditions St Convective heating coefficient Ct radiative equilibrium heat transfer Qeq total wall heat transfer qw translational wall heat transfer qT rotational wall heat transfer qR vibrational wall heat transfer qV free electron wall heat transfer qEl catalytic wall heat transfer qD 5 16 2 6 09 DPLR Code Version 4 01 0 User Manual Using POSTFLOW 527 velocity wall heat transfer qU 531 total wall heating qwi 581 spacing in wall units y yp 584 inner velocity u up 591 blowing velocity through face vb 594 mass flow rate through face mdot 595 unit mass flow rate through face mdotU 596 thrust through fac
185. orms 60 iterations at a 0 50 timestep 1 0 5 DPLR performs 100 iterations at a 1 0 timestep 2 0 5 DPLR performs 100 iterations at a 2 0 timestep 5 0 5 DPLR performs 100 iterations at a 5 0 timestep 10 0 5 DPLR performs 100 iterations at a 10 0 timestep 20 0 DPLR performs 20 iterations at a 20 0 timestep 50 0 DPLR performs 20 iterations at a 50 0 timestep 100 0 DPLR performs 20 iterations at a 100 0 timestep 250 0 DPLR performs 20 iterations at a 250 0 timestep DPLR Code Version 4 01 0 User Manual 4 51 2 6 09 Using DPLR 500 0 DPLR performs 20 iterations at a 500 0 timestep 1 000 0 DPLR performs 20 iterations at a 1 000 0 timestep 1 No more CFL values are available Although larger CFL numbers imply larger timesteps and faster convergence rates there is usually a maximum CFL number that represents a stability limit for a given problem Using CFL values above this number can result in solution divergence For optimum performance therefore you should run at CFL numbers that approach but do not exceed this limit Over time and with experience you will notice that certain classes of problems are associated with an approximate range of stable CFL numbers like those listed above that were used for a capsule shaped problem By starting your DPLR run with one of these stable CFL ranges you should be able to get close enough to a solution to create a restart file that can then be further customized by editing the CFL number range in the input deck
186. ould be decomposed for parallel processing When both files are available you will run the DPLR file conversion executable FCONVERT to create a structured grid file that can be read by DPLR In most cases the recommended form of the structured grid file you create through this file conversion process will be pgrx In addition to the new grid file FCONVERT will create a screen report called a standard out of the actions taken to create the pgrx file This report file can be saved for archival purposes pgrx grid file dplr inp gt DPLR pslx file log file convergence file standard out When the pgrx file for your problem is prepared you will then create an input file for DPLR that specifies a variety of information about the flow environment your object will encounter conditions of flow entry conditions at the surface of different portions of your object and the timestepping regimen you want DPLR to employ during its solution calculations When both files are available you will run the appropriate DPLR executable DPLR2D or DPLR3D to create a solution or restart file In most cases the recommended form of the solution file will be ps1x In addition to the restart file DPLR will also create a log file a convergence file and another standard out file to capture various aspects of the progress of the solution run All three of these report files can be saved for archival purposes DPLR Code Version 4 01 0 User Manual 7 4 2
187. ould be avoided when possible However ASCH plot3d files are frequently used for grid input because they are portable and can be written by most commercial grid generation packages 3 Written as a gzipped ASCII file This format is currently used only for output of plot3d data from POSTFLOW The second digit if any of the file format number indicates the type of file as follows l Parallel archival I O file for use with DPLR This is the preferred file type for grid restart radiation and boundary condition files that are to be read by DPLR 2 Plot3d grid or q file Plot3d grid or function file See Tech Tip 2 4 Parallel multi file grid or restart file Note This file type is no longer supported by DPLR 5 TECPLOT block file 6 TECPLOT point file See Tech Tip 3 Tech Tips 1 To read or write XDR files the fxdr libraries must be installed on your computer and linked to DPLR during compilation See Section 2 2 for more information 2 Plot3d files cannot be read or written by DPLR2D or DPLR3D but are frequently used DPLR Code Version 4 01 0 User Manual 8 10 2 6 09 Appendices to import data from or export data to other programs 3 TECPLOT data files are output by POSTFLOW for post processing purposes but cannot be read as input by any of the codes in this package In order to create binary TECPLOT files the TECPLOT I O library must be properly installed and linked to DPLR See Section 2 2 for more information
188. post processing tools while the Tecplot format is specific for use with Amtec s Tecplot data visualization software POSTFLOW can write Tecplot ASCII dat files as well as binary p1t files although Tecplot binary output requires linking to the Amtec provided tecio a or tecio64 a runtime library If this library is not available on your machine Tecplot binary files cannot be generated DPLR Code Version 4 01 0 User Manual 5 33 2 6 09 5 4 1 5 4 2 Using POSTFLOW Gzipped plot3d output ouform 32 33 is generated via a system call to the gzip utility provided with UNIX and LINUX systems This option may not be available on Windows systems Volume Data Volume data can be extracted from a restart file using the zone specification lines in the POSTFLOW input deck For example assume that a simulation was performed on a five block 3D volume grid and the desired output variables are pressure ivarp 110 temperature ivarp 120 Mach number ivarp 154 and pointwise residual ivarp 999 The ivarp array would be ivarp 110 120 154 999 The following zone specification lines could then be used to extract data from the entire volume iwrt ifac imin imax jmin jmax kmin kmax bkmin bkmax zonetitle 1 0 1 l 1 l 1 l 1 1 volume 1 0 Ly sl Ty i 1 1 1 1 terminator Using the shorthand code of 1 to mean maximum or all these lines tell POSTFLOW to read t
189. program such as Tecplot is created according to your specifications along with an on screen summary of actions performed by POSTFLOW See Section 5 3 for an example of a problem specific POSTFLOW input deck and the output summary DPLR Code Version 4 01 0 User Manual 5 3 2 6 09 Using POSTFLOW 5 2 Input Flags for POSTFLOW Input variables for POSTFLOW are discussed below in the order they appear in the deck imemmode Specifies the memory mode selected for running POSTFLOW itruev istat DPLR Code Version 4 01 0 User Manual Allowable values are 1 low memory mode 2 high memory mode Recommended Tech Tip Using high memory mode makes all the features of POSTFLOW available However if your computing resources are insufficient for running in this mode i e capable of holding all flow variables for the largest physical block in the simulation at any one point in time then choosing the low memory mode will enable you to use the program but will require significantly longer processing times and make certain program features unavailable Specifies the method to use to compute derivative values such as skin friction or heat transfer Allowable values are 0 evaluate derivatives using a 1 order approximation 1 evaluate derivatives using accurate 2 order expressions Recommended Tech Tip When immode 1 low memory extraction of true derivatives is not possible In this case POSTFLOW automatically sets
190. ptune inter Found 3 valid zonal interface blocks in 2 block grid file Decomposing block 1 into 1 ibrk 1 jbrk 1 kbrk 1 Decomposing block 2 into 6 ibrk 2 jbrk 1 kbrk 3 creating 7 total blocks 7 Blocks Total load imbalance 4 32 Output Blockl size il 16 jl 32 kl 128 65536 cells Output Block 2 size il 32 jl 64 kl 43 88064 cells Output Block 3 size il 32 jl 64 kl 43 88064 cells Output Block 4 size il 32 jl 64 kl 43 88064 cells Output Block 5 size il 32 jl 64 kl 43 88064 cells Output Block 6 size il 32 jl 64 kl 42 86016 cells Output Block 7 size il 32 jl 64 kl 42 86016 cells In this example FCONVERT decomposed master block 2 into six nearly equal pieces while leaving block 1 unaltered The resulting load imbalance was 4 32 DPLR Code Version 4 01 0 User Manual 8 14 2 6 09 Appendices This is the most load balanced solution for nbreak 7 but it may not be the most desirable way to split the problem For example if the k direction is body normal for this problem it would be preferable to select a decomposition that does not break the problem in the k direction This can be accomplished by setting iaction 2 and specifying the correct boundary conditions in the DPLR input deck A portion of the FCONVERT output for this run will be Input Block 1 size il 16 jl 32 kl 128 65536 cells 128 524288 cells Input Block 2 size il 64
191. r master blocks which are a fundamental property of the input grid Keep in mind however that the user interfaces to DPLR2D DPLR3D and POSTFLOW deal only with master blocks and that virtual blocks are automatically converted to and from physical blocks as required during program execution Therefore when setting up a problem to run in DPLR only the master block structure of the problem is important If a two block grid is decomposed into nb1k virtual blocks to run in parallel the problem is set up for DPLR as a two block problem regardless of the actual value of nb1k This means that boundary conditions numerical models etc are only specified for the two master blocks DPLR will automatically convert this information to the virtual values at runtime Similarly when the solution is post processed by POSTFLOW it is treated as a two block problem regardless of the actual number of processors that were used This strategy greatly simplifies the preparation execution and post processing overhead required for parallel jobs The output of FCONVERT provides information on the physical block structure and includes physical block sizes which are required for setting up the DPLR input deck A portion of the output from FCONVERT for the sample problem of the previous section is shown below Summary grid dimensions for CFD input deck Hardwired to run on 7 processors Block 1 nx 128 128 16 ny 32 nz Block 2 nx
192. r an example of a problem specific DPLR input deck and run summary Tech Tips 1 The run command above works for MPICH with a single type of processor Different commands may be required for different MPI implementations or execution on heterogeneous clusters Consult your system administrator for details on MPI program execution on your particular machine 2 Ifa machinefile is required by your computer architecture it consists simply of an ASCII listing of available machine node names followed by the number of processes to start on each node Example node001 2 or node001 slots 2 node002 2 node002 slots 2 node003 2 node003 slots 2 Both machine files show that the system can accept a job requiring up to 6 DPLR Code Version 4 01 0 User Manual 4 5 2 6 09 processors Using DPLR 3 To avoid slow performance hangs or crashes be sure that x Nodes listed in the machine file are available for use and free of other jobs Your job does not require more processors or memory than the machine file says the system can accept 4 2 Input Flags for DPLR Input flags for DPLR are discussed below in the order they appear in the deck Input Filenames These are external input files used by DPLR at runtime Although these files can be specified using relative or absolute pathnames with the exception of the chemistry input file which requires an absolute pathname you may find that placing them in your problem specific workin
193. r blocks in the input grid Specifies the dimensions of the simulation Allowable values are 2 Two dimensional simulation 3 Three dimensional simulation Specifies the total number of coupled equations included in the matrix expression of the boundary conditions The value is calculated as follows neq ns ner nev neetidimt1 where ns number of chemical species in the flow ner number of rotational energy equations nev number of vibrational energy equations nee number of electron electronic energy equations Specifies the number of uncoupled turbulence equations 6 15 2 6 09 nmc nme nmt nmv f2 f3 f4 block sizes DPLR Code Version 4 01 0 User Manual DPLR Codes Package Input Output Files Specifies whether there is a catalytic material map Allowable values are 0 No a catalytic material map does not exist 1 Yes a catalytic material map does exist Specifies whether a surface radiation map exists Allowable values are 0 No a surface radiation map does not exist 1 Yes a surface radiation map does exist Specifies whether a transition map exists Allowable values are 0 No a transition map does not exist 1 Yes a transition map does exist Specifies whether a view factor map exists Allowable values are 0 No a view factor map does not exist 1 Yes a view factor map does exist Tech Tip A view factor map is used to account for the ability of concave
194. rational energy per unit mass of species n ev_n 5 20 2 6 09 DPLR Code Version 4 01 0 User Manual 4800 4800 n 5000 5000 n 5200 5200 n 6000 6000 n 6200 6200 n 6400 6400 n 6600 6600 n 6800 6800 n 7000 7000 n Using POSTFLOW all species electronic energies per unit mass electronic internal energy per unit mass of species n ee_n all species mass flow rates through surface mass flow rate through surface of species n mdot_n all species mass flow rates through surface per unit area mass flow rate through surface of species n mdotU_n all species total specific heats at constant volume total specific heat at constant volume of species n cvx_n all species translational specific heats at constant volume translational specific heat at const vol of species n cvt_n all species rotational specific heats at constant volume rotational specific heat at const vol of species n cvr_n all species vibrational specific heats at constant volume vibrational specific heat at const vol of species n cvv_n all species electronic specific heats at constant volume electronic specific heat at const vol of species n cve_n all species frozen specific heats at constant pressure specific heat at constant pressure of species n cp_n 5 21 2 6 09 DPLR Code Version 4 01 0 User Manual Using POSTFLOW 7200 all species frozen specific heats at co
195. re moment on a face in all directions pressure moment on a face in x direction Mx_P pressure moment on a face in y direction My_P pressure moment on a face in z direction Mz_P 5 18 2 6 09 DPLR Code Version 4 01 0 User Manual Using POSTFLOW 720 viscous moment on a face in all directions 721 viscous moment on a face in x direction Mx_V 722 vyiscous moment on a face in y direction My_V 723 viscous moment on a face in z direction Mz_V 750 total moment coefficient on a face in all directions 751 total moment coefficient on a face in x direction Cmx 752 _ total moment coefficient on a face in y direction Cmy 753 total moment coefficient on a face in z direction Cmz 760 pressure moment coefficient on a face in all directions 761 pressure moment coefficient on a face in x direction Cmx_P 762 pressure moment coefficient on a face in y direction Cmy_P 763 pressure moment coefficient on a face in z direction Cmz P 770 viscous moment coefficient on a face in all directions 771 viscous moment coefficient on a face in x direction Cmx V 7712 viscous moment coefficient on a face in y direction Cmy_V 713 viscous moment coefficient on a face in z direction Cmz_V Debugging Status Information 990 pointwise BC numbers along block edges ibcp 991 net charge should always be zero Qnet 992 sum of mass fractions should always be one Csum 998 zero zero
196. res 3 Yos approximate mixing rules preferred model for all reacting gas simulations 4 Full first order Chapman Enskog multicomponent VOT WORKING in DPLR 4 01 0 11 Blottner Armaly Sutton with an Eucken relation requires composition dependent tailoring of the free parameters for maximum accuracy 12 Keyes Equation and constant Prandtl number should be used only where the freestream temperature is very low lt 100K Tech Tip ivmod 3 has been shown to be a reasonable and general approximation to the true Chapman Enskog fluxes DPLR Code Version 4 01 0 User Manual 4 10 2 6 09 idmod DPLR Code Version 4 01 0 User Manual Using DPLR Specifies the baseline model used to compute species diffusion coefficients An appropriate setting for idmod is required for all multi species viscous simulations Allowable values are 1 Constant Lewis Schmidt number assumes all species have the same diffusion coefficient 2 Bifurcation model developed to model boundary layer diffusion of carbon based ablators requires as input least squares fit coefficients for each species 3 Self Consistent Effective Binary Diffusion preferred model for all multi species calculations but somewhat unstable for separated flows See Tech Tip below 5 Iterative multicomponent VOT WORKING in DPLR 4 01 0 11 Constant Lewis Schmidt number ignore ambipolar diffusion widely used but often inaccurate provided for comparison to heritage code
197. riable listing appears in the corresponding ibc flag entry 2 DPLR reads data in this file in the following order 1 Any pointwise boundary condition numbers first and input profiles second 2 Any surface material maps 3 Any surface radiation maps 4 Any surface transition maps 5 Any surface view factor maps 3 DPLR then loops in the following order Inner loop over the face number 1 6 followed by a loop over the block number 1 nblk then repeat the outer loop over read order listed above 4 Data should be written in standard plot3d like format in the order shown in the file above 6 17 2 6 09 6 4 6 4 1 DPLR Codes Package Input Output Files Runtime Control Files As you are monitoring a DPLR run you may notice that your solution is not working the way you anticipated As of version 4 01 0 of DPLR Code Package you can dynamically interact with a simulation mid run to change timestep settings and DPLR s grid adaption values to correct problems without having to stop the run and start over through the use of a runtime control file A generic runtime control file named generic ctr1l is distributed with the DPLR Code Package Version 4 01 0 and can be found in the cfdinput directory Creating a Runtime Control File Step 1 Open the text editor program for your system Action At the command line prompt type path to your cfdinput directory generic ctrl Result A generic input file appears on s
198. rols 442 Plug blocks BC 2 at jmin will be changed to BC 26 wall 10 13 These controls are entered as line pairs a text line followed by an integer list Each line pair is optional case insensitive and the order does not matter meaning either of the pairs may be entered first or omitted The default grid sequencing is 2 2 1 meaning the grid block cell counts in output file dplr inputs 2 are halved in the i and j directions only whereas the 4 4 2 shown would enable solution with the grid coarsened twice as much Keywords implemented for the second type of input are Cavity and Plug with Plug also being appropriate for protruding tile gap filler cases These controls allow the appropriate faces of the indicated blocks to be marked as walls specifically BC 26 meaning catalytic radiative equilibrium Any reasonable format for the list of block numbers is acceptable as long as they are all on one line Tech Tips 1 You will need to compile Template manually as the installation script will not automatically install it on your system Check with your System Administrator for the system specific steps needed to compile and install this tool into your utilities directory 2 Template detects matching faces by comparing the maxima and minima in x y and z If two faces are found to satisfy the six possible comparisons to within the tolerance provided at run time default epsilon 0 0001 distance units and the face dimensions mat
199. rsion 4 01 0 User Manual 8 25 2 6 09 8 4 5 Appendices lonization 180 degree of ionization zeta C n n Surface Properties Heat Transfer 512 heat transfer coefficient in mass flux units Chm This is the heat transfer coefficient expressed in kg m s for use with FIAT 520 radiative equilibrium heat transfer Qeq 4 eq EOT This is the surface heat transfer as computed using the radiative equilibrium wall formation In this expression is the surface emissivity ois the Stefan Boltzmann constant and 7 is the surface temperature This variable is provided mainly as a sanity check to ensure that the computed heat transfer agrees with the radiative equilibrium value when a radiative equilibrium wall is specified DPLR Code Version 4 01 0 User Manual 8 26 2 6 09 Appendices 8 5 Reference Terms Lewis Number Le Le pDC k yat Schmidt Number Sc pD Turbulent Lewis Number LeT Turbulent Schmidt Number ScT Scr Prandtl Number Pr wuC k Turbulent Prandtl Number Pr pu k C cell Reynolds Number DPLR Code Version 4 01 0 User Manual 8 27 2 6 09
200. s 12 Bifurcation model fits obtained with assumption of ambipolar diffusion developed to model boundary layer diffusion of carbon based ablators appropriate when using coefficients provided by Olynick 13 Self Consistent Effective Binary Diffusion ignore ambipolar diffusion provided for comparison to heritage codes Tech Tip The preferred model for all multi species calculations is the Self Consistent Effective Binary Diffusion SCEBD model of Ramshaw and Chang idmod 3 13 which has been shown to give results in good agreement with exact solutions of the Stefan Maxwell equations This model requires as input collision integral data for each binary interaction in the mixture These data are imported to DPLR via the gupta tran physical model file in the cfdinput directory However it does tend to be unstable for separated flows particularly while the recirculation region is being formed Therefore for separated flows start the solution with idmod 1 and an appropriate Schmidt number then switch to idmod 3 once the flow structures have stabilized 4 11 2 6 09 itmod islip iblow DPLR Code Version 4 01 0 User Manual Using DPLR Specifies the turbulence model to be employed An appropriate setting for itmod is required for all turbulent simulations indicated by ivis 2 12 Allowable values are 0 1000 1001 1002 2001 2002 2003 Laminar Flow non turbulent Baldwin Lomax Model reason
201. s taken as a scaled version of the previous distribution at every body location 4 Two options that were allowed in v3 05 igalign 4 and igalign 14 are deprecated in this release As of v4 01 0 the input flag ismooth is provided in order to give the user more flexibility in controlling the type of smoothing that occurs The functionality of igalign 4 and igalign 14 can be replicated in the current release by setting igalign 1 and igalign 11 respectively along with ismoot h 3 ngiter Sets the frequency at which a grid alignment is performed Recommended values 500 1500 DPLR Code Version 4 01 0 User Manual 4 31 2 6 09 ilstadpt imedge imradial ngeom DPLR Code Version 4 01 0 User Manual Using DPLR Tech Tip The first alignment always occurs on a restart prior to running the first iteration and subsequent alignments with the total number set by nalign are performed every ngiter iterations Specifies the number of iterations before DPLR performs the first grid adaption Specifies the method used to locate the bow shock in the simulation Allowable values are 1 Align to a constant Mach number contour Tech Tip Although other programs such as SAGe or Outbound have other options only Mach number based adaption is supported in DPLR at this time Specifies the type of wall spacing to employ during the reclustering of interior points that takes place during a grid adaption Allowable values ar
202. sabled in DPLR 4 01 0 iaero Tells DPLR to track integrated aerodynamic variables at each iteration Allowable values are Do not track variables 1 Write integrated body forces and moments to file at each iteration disabled in DPLR 4 01 0 Grid Adjustment Alignment Morphing Flags Each of the flags in this portion of the DPLR input deck gives you a range of options to use if you decide to realign your input grid to better capture the shock wave inside DPLR You can accomplish this in two ways e Adjusting the values for the grid adaption flags in the DPLR input deck to be used with the restart file after running an initial simulation e Creating a runtime control file to adjust the values for the grid adaption flags while the simulation in running See Section 6 4 for more information on Runtime Control Files To better capture the shock wave using this grid adaption option you need to 1 Move the outer boundary of the input grid to just beyond the shock location as determined by the initial converged solution 2 Smooth the outer boundary surface controlled by the ismooth flag 3 Redistribute the interior grid points controlled by the imradial flag igalign Enables and sets the type of grid adaption you want DPLR to use Allowable values are 0 Do not perform grid alignment 1 Perform basic grid alignment should only be used with a restart file 2 Recluster grid only no alignment not dependen
203. ser Manual 5 30 2 6 09 Using POSTFLOW postflow NASA Ames Version 4 01 0 Maintained by Mike Wright last modified 02 05 09 kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk Parsing the restart file to get physical modeling data restart file format NASA Ames Version 4 01 0 solution run at Thurs Feb 5 08 18 10 2009 run in 500 iterations in 4 23E 03 seconds CPP macro settings enabled during run AMBIPOLAR 1 PARKTEXP 0 50 NOHTC Keq limiter set at 100 00 input ns 5 ner 0 number of blocks 2 file dimension 3 extracting the following BCs 17 18 19 note that extraction of pointwise BCs not supported yet output variables x y z p T M u v w running in high memory mode interpolating grid to cell centers processing grid variable 1 processing flow variable 1 block 1 nx 32 ny zone t BC19 i processing grid variable 1 processing flow variable 1 block 2 nx 48 ny 64 nz zone t BC19 i 50 j 1 k zone t BC19 i 50 j 1 k writing tecplot file postpitch dat using grid file neptune 8PE pgrx using flow file neptune pslx Figure 4 POSTFLOW Input Deck for Pitchplane Analysis of Neptune Probe DPLR Code Version 4 01 0 User Manual 5 31 2 6 09 Using POSTFLOW kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk postflow NASA Ames Version 4 01 0 Maintained by Mike Wright last modified 02 05 09 kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk Parsing the restart file to get physical
204. sing POSTFLOW block 2 nx 48 ny 64 nz 64 Block 2 p 3 591044259306E 01 Pa T 1 280700000000E 02 K M 3 232180261501E 01 Re L 3 151858720834E 05 1 m Extracting Data for External Codes Several of the output format options ouform in POSTFLOW create files intended for use with third party codes or provided post processing utilities In these cases options are hardwired to the values required by the particular third party software Extraction for Moment As of release version 3 05 POSTFLOW can directly compute moments or moment coefficients However the Moment utility provided as part of the DPLR Code Package can also do this computation Moment requires plot3d grid and function files along with an input moment inp file to be created by POSTFLOW See Section 9 1 6 To tell POSTFLOW to create these files e set ouform 1 1 e set interp 11 e set ivarp to either total forces 604 606 pressure forces 614 616 or viscous forces 624 626 e set output datasets to define surfaces either with iexbe or ifac At the current time the only function of Moment that is not built into POSTFLOW is for the extraction of hinge moments Extraction for RADEQUIL To tell POSTFLOW to output a line of sight file for further processing with the shock layer radiation code RADEQUIL set ouform 28 POSTFLOW will then automatically assume the following hardwired settings interp 1 ivarp 11 12 13 110 120 125 1600
205. t Taking gradients of ev gt Multi Species Binary Diffusion Mole Fraction Gradients gt Binary diffusion coefficients from Gupta Collision Integrals gt SCEBD model used to compute diffusive fluxes Ideal Gas Equation of State Axisymmetric Flow rotate about x axis Implicit Data Parallel Line Relaxation kmax 4 gt Using Global Timestepping RR HR HR ARRERA A HK HK KH BRKKRKKHKHK RRB INFORM saving 2 previous restarts Freestream Reynolds Number 8 260E 06 1 m Freestream Frozen Mach Number 1 087E 00 Freestream Equil Mach Number 1 133E 00 Jk H Tech Tip To avoid having DPLR overwrite archival output files with files generated by re runs of simulation rename your restart file before beginning each run so that all the automatically created output files will indicate which simulation run created them 6 11 Tecplot Files Amtec s Tecplot visualization software is a tool often used to process results of DPLR simulations For this reason POSTFLOW has the capability of writing dataset files in two Tecplot specific formats DPLR Code Version 4 01 0 User Manual 6 26 2 6 09 DPLR Codes Package Input Output Files e Tecplot binary p1t e Tecplot ASCII dat As noted in Section 2 2 5 2 and 5 4 however the Amtec provided tecio a or tecio64 a runtime library must be installed on your system to generate binary plt files and may be available from the Tecplot websit
206. t interface file 2 Write output interface file including dummy cells used for debugging 11 Write output interface file including edges used for debugging 12 Write output interface file including dummy cells and edges used for debugging Tech Tip When inint 2 4 FCONVERT automatically creates zonal input information for the plot3D grid file being processed and embeds the information into the file being produced By setting ouint gt 0 you tell FCONVERT to also write the zonal interface information it creates to a separate file thus eliminating the need to regenerate it a potentially time and resource consuming task if the problem is re run in the future Note however that if you change the grid topology in any future run of the problem the zonal interface files will need to be recalculated Specifies when an output file contains dummy cells Allowable values are 0 Output file does not contain dummy cells 3 8 2 6 09 ncedge imseq iscale DPLR Code Version 4 01 0 User Manual Using FCONVERT 1 Output file contains dummy cells use for debugging Tech Tip The appropriate setting of odummy is nearly always 0 Specifies which edge and corner interfaces should be generated Allowable values are 0 Do not compute and edge and corner interfaces 1 Compute all edge and corner interfaces 2 Compute only edge corner interfaces created by decomposition Tech Tip This flag should always be set to 1 u
207. t order approximations of derivatives as well as that used to define y ivarp 581 and the cell Reynolds number ivarp 59 Deviation from Orthogonality 22 deviation from orthogonality deg dev This is defined as the number of degrees the surface normal grid lines deviate from perfect orthogonality For interp 1 this value represents a local average interpolated to the face center The primary use of this output variable is as a measure of overall grid quality Note Orthogonality is desired at all body surfaces but is generally unimportant at flow through boundaries DPLR Code Version 4 01 0 User Manual 8 22 2 6 09 8 4 2 8 4 3 Appendices Mixture Transport Properties Cell Reynolds Number 59 cell Reynolds number Re _c The cell Reynolds number is defined as _ at V An v Re where a is the sound speed V is the local velocity magnitude An is the body normal distance ivarp 21 and v is the kinematic viscosity The cell Reynolds number is typically used as a way to judge the adequacy of the near wall spacing in a boundary layer Re lt 5 is generally sufficient to ensure accurate heat transfer and skin friction Transport Properties Lewis Numbers 86 laminar Lewis number Le 96 turbulent Lewis number Le_t The Lewis number Le is defined as Le pDC kK where p is the mixture density D is the binary diffusion coefficient Cp is the total specific heat at constant pressure and x is the ther
208. t the block should be coarsened n times in that direction For example a sequencing record of DPLR Code Version 4 01 0 User Manual 3 21 2 6 09 3 5 2 Using FCONVERT iseq jseq kseq 3 2 1 tells FCONVERT to retain one out of every 3 points in the 7 direction one out of every 2 points in the j direction and every point in the k direction of the block being described To sequence every block in the grid by the same set of factors set imseq 2 then enter only one set of sequencing factors in the iseq jseqg and kseq input flags If the input grid has zonal interface information associated with it these data will be automatically sequenced along with the grid file Once you have appropriately sequenced the grid file boundary condition file if any and radiation file if any you can set up and run the problem independently from the fine grid solution Tech Tip Be sure that the sequencing strategy you choose for multi block problems results in a coarsened grid that remains point matched While failure to produce a point matched grid across zonal interfaces will not result in a runtime error in FCONVERT it will cause problems in the DPLR run Upsequencing Restart Files After using a sequenced grid file to achieve a good initial solution in a relatively short period of computing time you can proceed to restoring grid points and refining your solution by using FCONVERT to upsequence the restart file To upsequence the restart
209. t upon a shock wave location can be used in initial simulation but do not use with imradia1 1 setting unless the problem has been previously converged 3 Smooth outer boundary only no alignment not dependent upon a shock wave location can be used in initial simulation DPLR Code Version 4 01 0 User Manual 4 30 2 6 09 Using DPLR 5 Perform basic grid alignment but hold first 40 of grid points in the body normal direction fixed should only be used with a restart file permits rapid alignment of grid to shock wave and more aggressive CFL ramp between alignments but should only be used after coarse alignment has been achieved still experimental Not Tested in DPLR 4 01 0 11 Automatic grid alignment basic Allows DPLR to determine when to perform grid adaptions not extensively tested so use with caution 20 Surface morphing constant amount Allows you to morph the surface of the body by an amount specified in the ds1 flag used primarily for surface recession calculations not extensively tested at this time so use with caution 21 Surface morphing variable delta specified in a pointwise boundary condition file Not Working in DPLR 4 01 0 Tech Tips 1 If you set igalign 1 before a restart file from an existing solution exists a runtime error will occur 2 When igalign 5 ngeomand imradial are ignored 3 When igalign 20 the remainder of the flags in this block are ignored and the body normal distribution i
210. ted gt Assuming free electrons are coupled with T Laminar Navier Stokes Simulation gt Gupta Style Collision Integrals amp Yos Mixing Rule gt Fickian Diffusion Mass Fraction Gradients Schmidt Number 0 50 gt SCEBD model used to compute diffusive fluxes Ideal Gas Equation of State 3 Dimensional Flow Implicit Data Parallel Line Relaxation nrlx 4 gt Using Global Timestepping Estimate 187MB stack memory required per PE FF RH HF HR ARRA KK KB KR Akk KKK ARAARA Reading grid file neptune 8PE pgrx DPLR Code Version 4 01 0 User Manual 4 69 2 6 09 Using DPLR gt Reading block 1 grid cell size 32X 16X 64 gt Reading block 2 grid cell size 48X 64X 64 gt Total number of grid cells 229376 gt Computing grid dummy cells Freestream Reynolds Number 8 024E 04 1 m Freestream Frozen Mach Number 3 715E 01 Freestream Equil Mach Number 3 715E 01 nit 1 rmsres 1 0000000000000E 00 cfl 1 0E 05 nit 2 rmsres 9 9999996847695E 01 cfl 1 0E 05 nit 3 rmsres 9 9999993691276E 01 cfl 1 0E 05 nit 98 rmsres 3 8595952079619E 01 cfl 1 0E 01 nit 99 rmsres 3 8034985848052E 01 cfl 1 0E 01 nit 100 rmsres 3 7508579577703E 01 cfl 1 0E 01 writing restart file neptune pslx solution written at Thurs Feb 5 07 21 13 2009 nit 101 rmsres 3 7017658887905E 01 cfl 1 0E 00 nit 102 rmsres 3 2830292631577E 01 cfl 1 0E 00 nit 103 rmsres 3
211. the DPLR Code package is being used izdum 0 There are no dummy cells accounted for in this interface file nblk There are 2 master blocks in the input grid file ninta There are 3 zonal interfaces in the input grid file nintc 2 3 0 Input grid contains no corner edge zonal interfaces nz 1 2 1 2 1 2 Master blocks 1 and 2 participate in the three zonal boundaries being described nface 1 1 2 1 3 1 The first zonal interface is located at the imin of both blocks placing it in the kj plane The second zonal interface is located at the imax of one block and the imin of the other placing it in the jk plane The third zonal interface is located at the jmin of one block and the imin of the other placing it in the ik plane ndr1 2 2 2 2 3 3 The first extent direction for the first interface zone is j The first extent direction for the second interface zone is also j The first extent direction for the third interface zone is k nst1 1 16 1 49 1 1 The starting point in the first extent direction for one block in the first zonal interface is 1 and for the abutting block it is 16 The starting point in the first extent direction for one block in the second zonal interface is 1 and for the abutting block it is 49 The starting point in the first extent direction for one block in the third zonal interface is 1 and for the abutting block it is also 1 nen1
212. tic energy per unit mass eU velocity in the x direction u velocity in the y direction v velocity in the z direction w velocity magnitude Vel frozen Mach number M frozen speed of sound a mean thermal speed cbar normalized velocity in the x direction u Vel normalized velocity in the y direction v Vel normalized velocity in the z direction w Vel momentum per unit volume in the x direction rhou momentum per unit volume in the y direction rhov momentum per unit volume in the z direction rhow total energy per unit volume re total rotational energy per unit volume rer total vibrational energy per unit volume rev total electronic energy per unit volume ree total free electron energy per unit volume rel total chemical formation energy per unit volume reh total kinetic energy per unit volume reU entropy S pointwise unit radiative emission Erad degree of ionization zeta 5 15 2 6 09 DPLR Code Version 4 01 0 User Manual 181 182 183 194 195 196 197 202 204 250 251 252 324 325 326 327 Using POSTFLOW debye length lam_D Tstar Tstar electron charge ec total energy per unit mass in rotational Eqn er_B total energy per unit mass in vibrational Eqn ev_B total energy per unit mass in electronic Eqn ee_B total energy per unit mass in free electron Eqn el_B delta velocity at wall Del_V delta temperature at wall Del_T velocity in the x direction n
213. tics Developmental in DPLR 4 01 0 88 Subsonic exit specify static pressure pback extrapolate others Uses method of characteristics disallows backflow Developmental in DPLR 4 01 0 Pointwise Twall Boundaries 100 199 100 corresponding isothermal boundary condition number 110 No slip viscous isothermal wall 125 Catalytic isothermal wall 130 No slip viscous isothermal wall with blowing 135 Catalytic isothermal wall with blowing Not Working in DPLR 4 01 0 140 Isothermal wall with slip 145 Catalytic isothermal wall with slip Not Working in DPLR 4 01 0 Pointwise Twall and Blowing Boundaries 200 299 200 corresponding non blowing boundary condition number 230 No slip viscous isothermal wall 235 Catalytic isothermal wall Not Working in DPLR 4 01 0 Pointwise Blowing Boundaries 300 399 300 corresponding non blowing boundary condition number 330 No slip viscous isothermal wall with blowing 335 Catalytic isothermal wall wall Not Working in DPLR 4 01 0 4 46 2 6 09 DPLR Code Version 4 01 0 User Manual Using DPLR Tech Tip To specify a pointwise boundary for any cell face you must first set up and include a pbca with the appropriate data in your current working directory Mathematically Adjusted Boundaries 1000 1099 1011 1019 amp 1021 1023 currently support augmented eigenvalue limiters in the vicinity of standard singular axes or symmetry planes when kd
214. to these mesh points Best to use when output data is to be processed using SAGe or a utility such as Outbound to move the outer boundary of the grid or adapt the grid to the computed flowfield 3 Identical to interp 1 except that additional points are not added at the block boundaries so the output grid will have holes in it along those boundaries Best to use for computing integrated forces and moments or for outputting pointwise forces for later offline integration using the Moment utility program See Section 9 1 5 for more information on Moment 4 Identical to interp 1 except that even the points at the boundaries are located using cell centered interpolation Maximum output dimensions using this option is the number of cells in each computational direction plus two points in each direction representing the points added within the boundaries Primarily used for debugging by code developers to gain access to the cell centered values of quantities in the grid dummy cells rather than the face centered values available using interp 1 Specifies the maximum number of output data zones to be generated Recommended value 20 Tech Tip Used by POSTFLOW to size certain output arrays a moderate value such as the recommended 20 should be sufficient However if this value is too small for the output arrays you ask POSTFLOW to generate the program will abort and generate an error message prompting you to increase the nzones value
215. tting of 200 500 Perform 20 iterations of the simulation at a CFL setting of 500 750 Perform 20 iterations of the simulation at a CFL setting of 750 1000 Perform 20 iterations of the simulation at a CFL setting of 1000 2000 Perform 20 iterations of the simulation at a CFL setting of 2000 DPLR Code Version 4 01 0 User Manual 4 67 2 6 09 4 3 3 Using DPLR CFL numbers Setting Explanation cont or timesteps cont for ramping cont 5000 Perform 20 iterations of the simulation at a CFL setting of 5000 1 Stop reading CFL numbers Neptune Output Summary To run this problem in DPLR type a command at the prompt that is similar to mpirun np 42 DPLRBINDIR dpIr3d lt neptune inp Upon execution DPLR will create an on screen summary also known as a standard out of the problem as shown below kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk dplr3d NASA Ames Version 4 01 0 Maintained by Mike Wright last modified 02 05 09 kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk Running on 8 processors gt Allocating 1 nodes to block 1 gt Allocating 7 nodes to block 2 gt Total load imbalance 3 92 gt Input grid file hardwired for 8 processors Fk Sk k SRO Executable Information gt built by twhite on Thurs Feb 5 17 23 43 PST 2009 gt at host m100 gt running Linux 2 6 9 42 0 2 ELsmp x86_64 Makefile Settings gt LD LIBRARY PATH
216. ue of w to employ in the MUSCL scheme is 2 ilim 1 The MinMod flux limiter is used in the Euler flux extrapolation idiss 1 A standard eigenvalue limiter is used in the flux extrapolation epsi 0 3 The magnitude of the eigenvalue limiter is 0 3 in the flow direction jflx 4 The Euler flux extrapolation method to use in the j direction is MUSCL Steger Warming with Ap jord 3 The Euler flux extrapolation order of accuracy is third order upwind biased omgj 2 0d0 The value of w to employ in the MUSCL scheme is 2 jlim 1 The MinMod flux limiter is used in the Euler flux extrapolation jdiss 1 A standard eigenvalue limiter is used in the flux extrapolation epsj 0 3 The magnitude of the eigenvalue limiter is 0 3 in the j direction kflx 4 The Euler flux extrapolation method to use in the k direction is MUSCL Steger Warming with Ap kord 3 The Euler flux extrapolation order of accuracy is third order upwind biased omgk 2 0d0 The value of w to employ in the MUSCL scheme is 2 DPLR Code Version 4 01 0 User Manual 4 64 2 6 09 Using DPLR Block 2 Flags Setting Explanation cont cont cont klim 1 The MinMod flux limiter is used in the Euler flux extrapolation kdiss 0 Flux extrapolation will not be performed in the k direction epsk 0 03 This value is ignored because kdiss 0 iextst 1 The time advancement method used when simulating this master block will be implicit data
217. ughout the DPLR Code User Manual to illustrate how the Code Package works describes a Neptune probe with an ellipsoidal body as shown in Figure 1 This case is an example of aerocapture where drag from the atmosphere is used to decelerate the vehicle and bring it into orbit Figure 1 0 Neptune Probe DPLR Code Version 4 01 0 User Manual 3 14 2 6 09 3 3 1 Using FCONVERT Neptune Input Deck The input deck below shows the problem specific entries to make for FCONVERT to process the serial plot3D grid file of this probe shown in Figure 1 0 into a DPLR readable XDR parallel grid file Input file for fconvert iaction ifile idim iinfo ivers nvers 1 1 3 0 1 4 01 0 inform inint idummy nborig 22 1 0 2 ouform ouint odummy ncedge 11 0 0 1 imseq iscale sfact imir 0 0 1 0 0 nbreak 1 Decomposition information for each master block ibrk jbrk kbrk 1 1 1 7 1 1 Sequencing information for each master block iseq jseq kseq 1 1 1 1 1 1 iname xname cname neptune neptune none oname neptune 8PE nsin nerin nevin necin ntbin 5 0 1 0 0 DPLR Code Version 4 01 0 User Manual 3 15 2 6 09 3 3 2 Neptune Input Deck Settings Using FCONVERT This is a three dimensional problem The input grid is ASCII plot 3D and the output grid file is parallel XDR The original grid consists of two master blocks and must therefore include an interface file See Section 3 4 for more information on zonal i
218. ur cfdinput directory generic inp Result A generic input file or deck appears on screen with place holder default values To start with a blank deck remove the values as shown on the following page Enter appropriate problem specific values for each of the input variables or flags See Section 4 3 for a description of DPLR input flags and a list of allowable values Action For each flag type allowable problem specific value Result Input deck contains sufficient information for DPLR to process an input grid file and develop a solution to the problem Tech Tip Take special care to preserve the line spacing in the file as you enter new values and or replace default values with problem specific ones If lines are added to or subtracted from the input deck file DPLR will not be able to read it accurately DPLR Code Version 4 01 0 User Manual 4 2 2 6 09 INPUT DECK FOR DPLR2D DPLR3D CODE gname fname bname rname cname dname mygridname myrestartname mybcname myradname myconnectname PATH cfdinput air5sp5 chem nblk igrid ivis ikt icatmd ireqmd ichem ikeq itrmod itrans istop nplot igdum kb1 xscale ifstat iaero igalign ngiter imedge imradial fs_scale ds_mult dsl cellRe DPLR Code Version 4 01 0 User Manual irest ikv twall ivib iaxi istate LeT ScT resmin nalign ilstadpt ngeom ismooth gmargin ds 1mx 4 3 Using DPLR
219. want to run a single block 3D problem on eight processors the simplest decomposition would be to break the problem into eight blocks in a single coordinate direction which would generate 7 face interfaces and zero edge or corner interfaces An alternate strategy would be to break into 4 x 2 x 1 blocks which would generate 10 face interfaces and 6 edge interfaces for a total of 16 The most complex decomposition would be 2 x 2 x 2 blocks This strategy would generate 12 face interfaces 12 edge interfaces and 4 corner interfaces for a total of 28 Although each of these strategies are allowed the first would generate the least message passing traffic during run time and would likely result in the most time efficient solution DPLR Code Version 4 01 0 User Manual 8 17 2 6 09 8 3 3 Appendices Physical Master Blocks vs Virtual Parllel Blocks The action taken by FCONVERT during a grid file decomposition depends on the output file format you specify If you select a plot3d output format the input file will be physically split into multiple blocks and written as a multi block file If you select a parallel output file format the input file will be virtually split into a number of blocks for parallel processing but resultant file will retain information about the original physical block structure FCONVERT distinguishes between virtual blocks which are generated purely to facilitate parallel execution and physical o
220. with a maximum CFL limit Tech Tips 1 When the DPLR method of time stepping is chosen iextst 1 only global time stepping ilt 1 or 2 should be used 2 When one or more blocks of a complex simulation are unstable you can specify a maximum CFL number to use only for the problem blocks by setting il 2 and entering a maximum CFL number in the cf1m flag Specifies the grid direction in which to break single block problems for parallel execution Allowable values are 1 i direction 2 j direction 3 k direction Tech Tip When the simulations has only one master block with no zonal interfaces DPLR can perform the necessary 4 42 2 6 09 cflm ibc DPLR Code Version 4 01 0 User Manual Using DPLR decomposition at runtime by breaking the problem into planes in the direction chosen with the ibdir flag Note that DPLR will print a warning message if 1bdir is set such that the grid is broken in the body normal direction Specifies the maximum CFL number to use in the current master block Only used when ilt 2 Specifies the type of boundary condition to use at each of the six faces of each master block i e imin imax jmin jmax kmin kmax Allowable values for each face are 0 Pointwise boundary condition read from input pbca file 1 Fixed at freestream conditions specified by ifree 2 Fixed at freestream if inflow extrapolate if outflow used in rapid analysis process 3 First order extrap
221. y 64 nz 64 gt Finished writing output file neptune 8PE pgrx Done DPLR Code Version 4 01 0 User Manual 3 18 2 6 09 3 3 4 3 4 Using FCONVERT Neptune Output Summary Information In addition to verifying the steps undertaken by FCONVERT the output summary also provides specific information about how the master blocks from the input grid file were decomposed into a set of output blocks suitable for parallel processing by DPLR In this sample case FCONVERT was told how to break the master blocks when iaction was set to 1 andthe ibrk jbrk and kbrk values for each master block were entered into the input deck See Section 3 3 1 1 These settings told FCONVERT to leave the first master block alone to make 1 block and break the second master block 7 times in the 7 direction to make 7 blocks This created a total of 8 blocks in the output XDR grid file and thus required or hardwired the file to be run on a minimum of 8 processors If iaction had been set to 2 the user would have had to know in advance how many processors their system could dedicate to running the problem and enter that value into nbreak FCONVERT would then have calculated the best numerical solution for breaking the input grid into at least the number of blocks equal to the value in nbreak and displayed the resulting block dimensions and load imbalance information in the output summary See Section 3 5 for more information on parallel decompositio
222. yer in place Block by block initialization using iconr flag Restart from saved file reset nit and etime Tech Tip When you are running a simulation for the very first time set iinit 0 Thereafter set iinit 1 unless problem specifics require use of one of the other available options 4 8 2 6 09 Using DPLR ivis Specifies the equation set to solve Allowable values are 0 Euler simulation neglect Navier Stokes terms 1 Laminar full Navier Stokes simulation 2 Turbulent full Navier Stokes simulation 11 Laminar Navier Stokes simulation thin layer 12 Turbulent Navier Stokes simulation thin layer Tech Tips 1 DPLR is a full Navier Stokes solver and recommended to be used as such By setting ivis 0 you can run the problem in Euler mode but the run time per iteration will increase significantly and viscous boundary conditions should not be specified 2 Running DPLR in thin layer mode is also not recommended because there are no time or memory savings in doing so 3 The turbulence model to be employed when ivis 2 is determined by the itmod flag ikt Specifies the model used to compute translational thermal conductivity An appropriate setting is required for all viscous simulations Allowable values are 1 Use the model that is consistent with ivmod 2 Use constant Prandtl number expression Tech Tip ikt 1 is the preferred setting for all practical applications ikv Specifies the model used to compute vibrational

(DPLR) Code Version 4.01.0 User Manual

Contents

Download Pdf Manuals

Related Search

Related Contents