CIRTEN UNIPI RL 1510-2013
Contents
1. 13 Pee PN 16 Figure 5 Outer Coupling Junction Box Calling Points 17 Figure 6 Inner Coupling Junction Box Calling POINTS 18 Figure 7 CRCoupler main form general information tab 22 Figure 8 CRCoupler main form general information tab File options from the Tool Strip Menu 23 Figure 9 CRCoupler main form general information tab File options from the Tool Strip Menu handling the loading saving Of a CaSe ceseeeseseceeccecccsceeeceeeeueeeeesseeeeecceceeceeseeuuueeeeessedeeeceeesesseeeeeeeaaaages 23 Figure 10 CRCoupler main form general information tab selection of RELAP and CFX solvers 24 Figure 11 CRCoupler main form Coupling interfaces tab before any user iINput 25 Figure 12 CRCoupler Interface creation secondary form a before update b after update 26 Figure 13 CRCoupler Interface creation secondary form combo boxes for selection of sender code a CFX boundary b RELAP boundary c exchanged variable d 26 Figure 14 CRCoupler Interface creation secondary form check to avoid duplicate names 27 Figure 15 CRCoupler main form Coupling interfaces tab check before interface removal 27 Figure 16 CRCoupler main form Coupling interfaces tab example with a list of user defined NEO I E EE E E E E ES 27 Figure 17 CRCoupler mai2. OUTLET1 OUTLET2 re O d 2 a dn d 2 D E Time s Figure 38 Test04 Temperature at Boundaries 49 INLET 1 INLET2 OUTLET1 OUTLET2 3 8 E _ Time s Figure 39 Test04 Velocity at Boundaries MA mi A 18 Figure 40 Test04 Flow Features 1 50 Temperature Time Value 0 50007 s Time Value 0 800066 Contour T 373 39 368 03 367 68 357 33 351 97 346 62 341 26 335 91 330 55 1325 20 319 84 314 49 509 14 303 78 298 43 293 07 Time Value 1 20007 s Temperature Time Value 1 00007 Contour T 373 39 368 04 362 68 357 33 391 97 346 62 341 26 335 91 1 330 55 325 20 318 84 314 49 309 13 Bu 303 78 Bu 298 42 293 07 K t kn a 0 030 Zico m Di 0 040 OI e Figure 42 Test04 Flow Features 3 51 Temperature Time Value 1 40007 s Contour T 373 39 aal a D 368 04 362 58 157 33 351 97 i A 346 62 341 26 335 91 330 55 f 325 20 310 84 314 490 309 13 303 78 298 42 293 07 i i 1 ii ERT 0 130 fm 1023 JD Time Value 1 34107 1 025 Figure 43 Test04 Flow Features 4 lemperature Time Value 1 94118 5 Contour T 373 39 368 04 362 58 357 33 346 62 335 91 330 55 325 20 319 84 314 48 309 13 303 78
3. 150 Coupled Steps Figure 23 Test01 2 CFX RMS Residuals In Figure 24 and Figure 25 the comparison between a RELAP stand alone case and the coupled cases Test01 1 Test01 2 Test01 3 is shown Discrepancies in results are below 3 relative errors and mainly due to some differences in the friction modelling between CFX and RELAP 35 Relative Error lt 3 RELAP stand alone Coupled Test01 1 Coupled Test01 2 Coupled Test01 3 T E 8 gt 5 Time s Figure 24 Test01 Comparison Pipe Velocity at 2 5 meter RELAP stand alone Coupled Test01 1 Coupled Test01 2 Coupled Test01 3 Temperature C 5 Time s Figure 25 Test01 Comparison Pipe Temperature at 2 5 meter 36 3 2 Test02 Pipe Double Coupling Boundary In order to test both the In Flow and Out Flow boundary types together the same problem analysed in Test01 time response of a 10 m pipe is tested with a central part modelled in CFX 5 m long and two side parts in RELAP 3 2 1 Set up of the CFD model The CFX simulation set up is the same of the previous case except for the boundary conditions which are given below BOUNDARY CONDITIONS e Inlet velocity and temperature imposed through a function linked to the cel input subroutine e Outlet outlet pressure controlled condition set by a function linked to the cel input subroutine e Wall Adiabatic no slip condition with auto
4. 298 42 er 293 07 K 2 130 mi 1025 3 075 351 97 E 341 76 0 05 0 100 fre Time Value 2 11246 5 1 025 Figure 44 Test04 Flow Features 5 0050 0 075 0 100 rr 52 o 0 050 0 130 fm a 00850 0 100 re 0024 1 015 0 085 0 075 l Figure 45 Test04 Flow Features 6 53 4 Conclusions and Future Development The activity described in the present report performed by GRNSPG in the framework of PAR 2012 on behalf of CIRTEN had the objective of developing improving and testing a software tool capable of handling coupled CFD TH system code simulations for nuclear reactor analysis applications In particular such work was carried out as a continuation of that previously performed by the same scientists in the frame of PAR 2011 The work had the following particular goals based on the open issues identified from the previous activity 5 To improve the coupling methodology e g robustness stability and efficiency issues 6 Toassess the influence of sensitivity factors 7 To develop a GUI 8 To perform V amp V The above objectives have been achieved as far as possible consistently with the limits of the resources available for this task and the developed Coupling Tool can be effectively used for coupled simulations within the range of tested conditions As already pointed out in Ref 2 the thorough development and qualification of such a tool is a complex
5. e Wall Adiabatic no slip condition with automatic turbulence wall treatment and no wall roughness 3 1 2 Set up of the RELAP model Test01 1 The RELAP model consists in a PIPE connected to two TIME DEPENDENT VOLUMES through two SINGLE JUNCTIONS PIPE 104 is 9 5 m long the remaining 0 5 m being modelled in CFX and with a diameter of 0 1 m It s divided in 19 equal elements Absolute roughness is set to zero and no additional concentrated loss factors or abrupt area change is modelled TIME DEPENDENT VOLUME 100 fixes the pressure and the temperature according to the laws shown in Figure 18 while component 108 is used to realize the coupling Namely the pressure in such component is imposed via a CONTROL VARIABLE updated by CFX within the input of every restart The maximum time step for the transient simulation is set to 0 001 seconds the coupling step being 0 05 s A subroutine of the Master program extracts from the output file the value of the velocity in the SINGLE JUNCTION 106 that models the connection with CFX domain Test01 2 The RELAP model consists in a PIPE connected to two TIME DEPENDENT VOLUMES through two SINGLE JUNCTIONS The only difference in the setting of the Test01 2 case is that PIPE 104 is 5 m long the remaining 5 m being modelled in CFX and with a diameter of 0 1 m It s divided in 10 equal elements Figure 21 shows the nodalization of the Test01 2 RELAP model which is qualitatively equal to the Test01 1 cas
6. for example from here http www perl org get htmliwin32 The one used on the computers on which the CRCoupler and the coupling tools have been developed and tested is ActivePerl 5 by ActiveState However any interpreter should be working fine 2 4 3 Main features CRCoupler is intended to perform the following main tasks 1 Get user information about both the RELAP input file and CFX definition file to be coupled as well as the respective results files necessary for initialization purposes 2 Create the CCL file if necessary Get user selections about code versions 4 Parse RELAP input file to gather information on boundaries eligible for the creation of coupling interfaces on control variables etc Parse CFX definition file to gather information on eligible boundaries 6 Interact with the user to define the coupling interfaces RELAP boundary CFX boundary sender code control variables if needed etc Get user additional input such as coupling scheme selection under relaxation etc Provide the above information to the coupling engine Manage and monitor the coupled calculation as well as standalone test RELAP and CFX calculations Some pictures are provided in the following to show the features available through the GUI The GUI is based on a main form and a secondary form the latter used for coupling interface definition The main form see Figure 7 consists of the following three tabs Gene
7. removal by the discharge line This type of problem is inherently three dimensional and difficult to simulate only using a single code In fact the complex flow patterns within the mixing box are difficult to be modelled with simple components like branches with tuned loss coefficient values and nearly 1D approaches On the other hand the overall system can be quite expensive to be solved completely with CFD tools Moreover complex components as the pump are more easily simulated by a system code 3 4 1 Set up of the CFD model The CFX simulation set up is the same of the first case Test01 except for the following settings TRANSIENT ANALYSIS transient simulation with adaptable time step Minimum and maximum time steps set to 10 sec and 10 sec respectively Minimum and maximum target coefficient loops set to 4 and 6 respectively CONVERGENCE CRITERIA Convergence criterion based on the following two conditions is applied e Maximum in the first part and Root Mean Square after the Temperature step normalized values of the equation residuals must drop by at least 4 orders of magnitude The convergence criteria change was needed since the temperature step brought some convergence issues mainly due to the quickly changing properties of water during the fast 100 C temperature rise e Average pressure temperature and velocity at inlet and outlet boundaries are monitored in order to control convergence and coupling issues BOUNDARY
8. the coupling 19 2 4 The Graphical User Interface 2 4 1 General information A Graphical User Interface GUI has been developed to simplify and automatize the use of the coupling tool and reduce the sources of errors and mistakes in the code coupling process The GUI has been named CRCoupler or more briefly CRC which stands for CFX RELAP Coupler CRCoupler is a windows form based program written in Microsoft Visual Basic according to the object oriented programming approach Freely available Microsoft Visual Studio 2010 Express Ref 5 was adopted as a development environment which in turns is based on the Microsoft NET Framework 4 0 Ref 6 The above choice allowed to take advantage of the availability of many development and debugging tools as well as a huge library of pre defined classes methods functions etc and to obtain a final tool whose quality is comparable to that of common commercial software The current version of CRCoupler is identified as 0 0 that is the very first release and is pretty much open to improvements and upgrades however it is already functional within the limits of its capabilities and pre requisites 2 4 2 Pre requisites The following pre requisites are necessary for the installation and utilization of the current version of CRCoupler 1 Tested operating system Microsoft Windows versions Seven 64 bit and XP 32 bit e The software is expected to be running also on different M
9. CRCoupler finds the name and number of the control variable associated with that boundary and which will be used for variable exchange purposes Those 24 name and number thus appear on the corresponding text boxes which otherwise are kept disabled The type of the selected RELAP boundary is shown either pipe junction time dependent junction or volume If the exchanged variable is Velocity or Temperature and the sender code is RELAP then the selection of a cross sectional profile is enabled Moreover several checks are performed to ensure the consistency and the adequacy of the input information e g see Figure 14 An existing interface can be deleted by the Remove interface button A check is made to avoid inadvertent deletion Figure 15 Figure 16 shows the table of interfaces for one of the sample cases previously studied The initial values indicated are dummy File Edit Show Data Tools Velocity Initial value ie Figure 11 CRCoupler main form Coupling interfaces tab before any user input Control variable name Control variable name Initial value Sl units Initial value SI units Velocity Temp profile Velocity Temp profile 25 a b Figure 12 CRCoupler Interface creation secondary form a before update b after update Sender code Sender code CFX boundary CFX boundary RELAP boundary RELAP boundary RELAP boundary type
10. RELAP boundary type Exchanged variable Jl Exchanged variable Control variable Control variable Control variable name Control variable name Initial value Sl units Wa Initial value 51 units Velocity Temp profile i m Velocity Temp profile Interface name Interface name Sender code Sender code CFX boundary CFA boundary RELAP boundary RELAP boundary RELAP boundary type Sie i RELAP boundary type Exchanged variable ES Exchanged variable Control variable Control variable Fr Temperature 200000000 Control variable name Control variable name 1 Initial value SI units Initial value 51 units 1 ann VelocityTemp profile 9n2 Velocity Temp profile Eng Greck updao Done Zeg Figure 13 CRCoupler Interface creation secondary form combo boxes for selection of sender code a CFX boundary b RELAP boundary c exchanged variable d pei SEZ FAT Ae CRCoupler That name has already been used for another interface please enter a new one 26 Figure 14 CRCoupler Interface creation secondary form check to avoid duplicate names File Edit Show Data Tools Velocity Initial value profe se men e emm e es m m im zm em o mmm een Jess Dm ei o e ow e en em keen ees ee es he zs ms Dm e eo
11. di articoli scientifici facilmente reperibili sui principali motori di ricerca specializzati Lorenzo Mengali Ingegnere Aerospaziale collaboratore dal 2008 presso il Gruppo di Ricerca Nucleare S Piero a Grado Universita di Pisa quale esperto di Fluidodinamica Computazionale Autore di rapporti tecnici interni e di articoli scientifici facilmente reperibili sui principali motori di ricerca specializzati Fabio Moretti Ingegnere Nucleare 2004 Dottore di Ricerca in Sicurezza Nucleare e Industriale 2009 collaboratore dal 2008 presso il Gruppo di Ricerca Nucleare S Piero a Grado Universit di Pisa quale coordinatore dell area Fluidodinamica Computazionale e responsabile di attivit sperimentali e di training Autore di rapporti tecnici interni e di articoli scientifici facilmente reperibili sui principali motori di ricerca specializzati 56
12. for calculation initialization o The system default text editor can optionally be invoked by pressing a button for possible CCL file editing If no CCL file exists then a button allows its creation from the def o Following possible editing of the CCL file the definition file can be updated accordingly by pressing the appropriate button ai CRCoupler v0 0 File Edit Show Data Tools General information Coupling interfaces Calculation management l Working directory Please specify working directory ES Case name case RELAP RELAP solver RELAP53Dv242 e E Restart RELAP input file Please specify RELAP imput file 2 RELAP restart file Please specify RELAP restart file 2 CFA CFX version vi40 CFX definition fle Please specify CFXdefintion fle 2 Edit CCL CFS results file Please specify CFS results file 2 Create CCL from def Update def from CCL Figure 7 CRCoupler main form general information tab Part of the above features is included also in the Tool Strip Menu at the top of the form Figure 8 In addition the Load case and Save case functions are available for proper saving of user input information and re use of pre defined input The information is written onto an ASCII file that is named name of the case crc In order to check for inadvertent or unwanted operations interactive message boxes are called such as those shown in Figure 9 22 User selection among multiple
13. in the SINGLE JUNCTION 106 and the temperature in 104 05 to models the connection with CFX domain The maximum time step for the transient simulation is set to 0 001 seconds the coupling step being 0 05 s 3 3 3 Results Explicit Coupling scheme fail to converge for this case For this reason the semi implicit scheme was tested Under relaxation was used within the inner loops However no distortion in the time response of the coupled results was expected since the coupling code waits for convergence of the inner loops before advancing to the next time step Figure 30 and Figure 31 show the CFX convergence behaviour of Test01 2 MAX Figure 30 and RMS Figure 31 residuals drops by at least 4 order of magnitude in most of the coupled steps A better convergence behaviour can be identified with respect to the previous cases thanks also to the improved coupling scheme adopted 41 mb MAX Rsiduals NR ATE o Onh AS RER Coupled Steps Test03 MAX Residuals Energy d MAX Residuals gt H Energy 150 Coupled Steps Figure 30 Test03 CFX Max Residuals The simulated flow velocity in the closed loop system was compared to results from a RELAP stand alone case Figure 32 shows a quite good comparison with errors below 2 and mostly due to differences in friction models between the two coupled codes No significant distortion from the under relaxation is added to the coupled results as e
14. order to be suited for coupling RELAP components available for couplings are Time Dependent Volume TDV and Time Dependent Junctions TDJ where coupling variables from CFX can be set and Pipes PIPE Branches BR and Single Junctions SNGJ where flow variables are taken and sent to the CFX sub domain In particular pressure and temperatures are exchanged in the volumes of PIPE BR and TDV components while velocities or mass flow rates are left to TDJ or SNGJ COUPLING BOUNDARY TYPES The possible information flow can be summarized in two different patterns patterns names refer to the CFD sub domain chosen to be the Master process In Flow Fluid flow from RELAP to CFX For Hydraulic Simulations energy equation not relevant velocities or mass flow rates are passed to CFX while the pressure is given back to RELAP In Thermo Hydraulic Simulations also the Temperature or Enthalpy is passed to the CFX sub domain Scalar variables transported with the flow are also supplied to CFX Out Flow Fluid flow from CFX to RELAP The opposite of the previous case Velocity or mass flow rate Temperatures and scalar quantities are passed to the RELAP sub domain while the pressure is given back to CFX 12 COUPLING INTERFACE CFX boundary interface 1D to 3D CFX RELAP SYSTEM SYSTEM 1D to 3D Figure 3 Schematic Sketch of Coupling Boundaries and Variables DATA INCONSISTENCIES In the data transfer process between the tw
15. task requiring considerable R amp D effort possibly in the context of a mid or long term and adequately supported program The work performed so far represents a good basis for such future development which should focus on the following aspects 9 Implementation of a new boundary type for thermal coupling between CFX walls and RELAP heat structures 10 Implementation of additional features and numerical options e g for tuning the numerical scheme 11 Further enhancement of numerical issues 12 Improvement of the GUI and of automatic handling of data and files 13 Systematic V amp V with particular emphasis on validation against experimental data and on application to GEN IV specific problems 54 References 1 2 3 4 EI 6 7 8 9 ENEA e Ministero dello Sviluppo Economico Accordo di Programma sulla Ricerca di Sistema Elettrico Piano Annuale di Realizzazione PAR 2012 Marzo 2013 L Mengali M Lanfredini F Moretti F D Auria Stato dell arte sull accoppiamento fra codici di sistema e di fluidodinamica computazionale Applicazione generale su sistemi a metallo liquido pesante CIRTEN Universita di Pisa Gruppo di Ricerca Nucleare di San Piero a Grado GRNSPG Report RdS 2012 1509 31 Luglio 2012 Versione 0 CERSE UNIPI RL 1509 2011 Lavoro svolto in esecuzione dell Attivita LP3 C1 C AdP MSE ENEA sulla Ricerca di Sistema Elettrico PAR 2011 ANSYS CFX 14 0 User Manual 2012 embe
16. CONDITIONS e nlet1 velocity and temperature imposed through a function linked to the cel input subroutine e nlet1 velocity and temperature imposed through a function linked to the cel input subroutine e Qutlet1 outlet pressure controlled condition set by a function linked to the cel input subroutine e Outlet outlet pressure controlled condition set by a function linked to the cel input subroutine e Wall Adiabatic no slip condition with automatic turbulence wall treatment and no wall roughness 44 Figure 33 shows the CFD computational domain for the Test04 case The Cartesian geometry was chosen in order to obtain simpler and more regular grids The overall size of the component is also quite small in order to reduce the required computational costs The main dimensions of the developed model are shown in the Figure along with some grid parameters INLET 2 OUTLET 2 Figure 33 Test04 CFX Computational Domain Mixing Box 3 4 2 Set up of the RELAP model The RELAP model consists in a PIPE connected to a TIME DEPENDENT VOLUME through a TIME DEPENDENT JUNCTION and TO A BRANCH The same BRANCH is connected to two TIME DEPENDENT VOLUME via another BRANCH and a SINGLE JUNCTION The circuit is modelled with other two part consisting in two TIME DEPENDENT VOLUME a TIMEDEPENDENT JUNCTION and a BRANCH PIPE 312 is 2 5 m long with area of 0 0004 m and has a square section It is divided in 5 equal elements Absolute roughn
17. ENER Agenzia nazionale per le nuove tecnologie l energia e lo sviluppo economico sostenibile Ricerca di Sistema elettrico Accoppiamento di codici CFD e codici di sistema L Mengali M Lanfredini F Moretti F D Auria Report RdS 2013 048 ACCOPPIAMENTO DI CODICI CFD E CODICI DI SISTEMA L Mengali M Lanfredini F Moretti F D Auria GRNSPG Settembre 2013 Report Ricerca di Sistema Elettrico Accordo di Programma Ministero dello Sviluppo Economico ENEA Piano Annuale di Realizzazione 2012 Area Produzione di energia elettrica e protezione dell ambiente Progetto Sviluppo competenze scientifiche nel campo della sicurezza nucleare e collaborazione ai programmi internazionali per il nucleare di IV Generazione Obiettivo Sviluppo competenze scientifiche nel campo della sicurezza nucleare Responsabile del Progetto Mariano Tarantino ENEA Il presente documento descrive le attivit di ricerca svolte all interno dell Accordo di collaborazione Sviluppo competenze scientifiche nel campo della sicurezza nucleare e collaborazione ai programmi internazionali per il nucleare di IV generazione Responsabile scientifico ENEA Mariano Tarantino Responsabile scientifico CIRTEN Giuseppe Forasassi Sigla di identificazione Distrib Pag di ENER Ricerca Sistema Elettrico ADPFISS LP2 018 L 58 Titolo Accoppiamento di codici CFD e codici di sistema PAGINA DI GUARDIA Descrittori Tipologia del documen
18. SIS transient simulation with fixed time step of 0 05 sec and total time of 10 Sec 30 SOLVED EQUATIONS Mass balance Continuity Momentum balance Reynolds averaged Energy Balance Total Enthalpy Transport of turbulent kinetic energy BI Transport of turbulent eddy frequency El The turbulence is accounted for with the k w based Shear Stress Transport SST model 7 and 8 WORKING FLUID Weak compressible liquid water from IAPWS tables with a reference pressure of 1 MPa DISCRETIZATION SCHEMES Spatial discretization through the recommended High Resolution advection scheme implemented in ANSYS CFX which locally blends between an upwind and a second order scheme depending on the local gradients of the solution Time discretization through the second order Backward Euler scheme Turbulence terms with the first order upwind scheme CONVERGENCE CRITERIA Based on UNIPI experience and common best practices in CFD analysis 7 and 9 a convergence criterion based on the following two conditions is applied Maximum or if not possible Root Mean Square RMS normalized values of the equation residuals must drop by at least 4 orders of magnitude Average pressure temperature and velocity at inlet and outlet boundaries are monitored in order to control convergence and coupling issues PARALLELIZATION The simulation is run in serial mode single processor Figure 19 and Figure 20 show the computational do
19. a sicurezza nucleare e collaborazione ai programmi internazionali per il nucleare di IV generazione List of Contents List of Contents List of Figures Om 1 Introduction o o 2 The Coupling Tool 2 1 Selected Codes 2 2 Coupling strategy 2 2 1 Main Features 2 2 2 Coupling Boundaries and Variables 2 2 3 Coupled Initialization 2 3 Coupling Routines 2 3 1 Coupling Manager 2 3 2 CFX routines 2 3 3 RELAP modifications 2 4 The Graphical User Interface 2 4 1 General information 2 4 2 Pre requisites 2 4 3 Main features 3 Test Cases 3 1 Test01 Pipe Single Coupling Boundary 3 1 1 Set up of the CFD model 3 1 2 Set up of the RELAP model 3 1 3 Results 3 2 Test02 Pipe Double Coupling Boundary 3 2 1 Set up of the CFD model 3 2 2 Set up of the RELAP model 3 2 3 Results 3 3 Test03 Closed Loop with Pump _ 3 3 1 Set up of the CFD model 3 3 2 Set up of the RELAP model 3 3 3 Results 3 4 Test05 Complex 3D re circulating component _ 3 4 1 Set up of the CFD model 3 4 2 Set up of the RELAP model 15 15 19 20 20 20 21 29 30 30 33 34 37 37 37 38 40 40 40 41 44 44 45 3 4 3 Results 4 Conclusions and Future Development References Curriculum Scientifico del Gruppo di Lavoro 54 35 56 48 List of Figures Fieufe 1 EXDIICIE Coupling Scheme a iaia 10 Figure 2 Semi Implicit Coupling en TEE 11 Figure 3 Schematic Sketch of Coupling Boundaries and Variables
20. choices e g for code solver is allowed by the use of combo boxes as shown in Figure 10 i Ga DE Arce E ffe EE VELLE Edit Show Data Tools Working directory aces Calculation manag Case name Solver RELAP input file CFA definition file as Load case Save case Exit RELAP restart file Please specify RELAP restart file CFA CFA version v14 0 LEX definition file Please specify CFX definition file Figure 8 CRCoupler main form general information tab File options from the Tool Strip Menu O Do you want to save the current case b Figure 9 CRCoupler main form general information tab File options from the Tool Strip Menu handling the loading saving of a case 23 P LAP solver RELAP5 3 3 al RELAPS 3D v2 4 2 D AP inputfile RELAPS 3D v4 D 3e Prova NCC ims P restart file a sion v14 0 k in file Vv 14 5 prova ts file b Figure 10 CRCoupler main form general information tab selection of RELAP and CFX solvers The Coupling interfaces tab is devoted to the definition of the coupling interfaces i e on which information needs to be exchanged between the codes at which boundaries in which direction etc The tab shows like in Figure 11 before any user input First of all a parsing of the RELAP input file and of the CFX CCL file needs to be performed by
21. ct access to shared memory areas is more fast and efficient However it requires major modifications of the codes structure and is less versatile Moreover considering the typical CFD computational times the I O routines contribution to the overall coupled simulation computational time can be considered negligible For these reasons data transfer through I O routines was chosen In addition to the file management a synchronization manager was also developed It controls the synchronization points coupling steps and manages the permissions for the advancement of the two coupled codes preventing errors in I O routines addressing data not yet calculated Currently all these tasks Coupling Interface are carried out by two separate components the Coupling Manager mainly addressing RELAP related functions see Section 2 3 1 and the CFX User Fortran routines see Section 0 2 2 1 Main Features The main features of the developed coupling technique are described here below following the classification defined in Ref 2 Non overlapping domains The two coupled codes will exchange information coupling variables only through specific interfaces CFX boundaries and RELAP dummy components Time Dependent Volumes Time Dependent Junctions and Pipes Branches specifically modified for coupling Each code solves its equation in its domain taking data from the other code only as imposed boundary conditions In line coupling Taking into account t
22. dded in the software package Idaho National Laboratories RELAP5 3D Code Manuals Appendix A RELAP5 3D Input Data Requirements version 4 0 INEEL EXT 98 00834 V2 March 2012 downloadable from http www inl gov relap5 r5manuals htm http www microsoft com visualstudio ita products visual studio 2010 express http www microsoft com it it download details aspx id 17718 ANSYS CFX 14 0 User Manual 2012 embedded in the software package D Wilcox Turbulence Modelling for CFD DCW Industries Inc Griffin printing California 2000 F Menter CFD Best Practice Guidelines for CFD Code Validation for Reactor Safety Applications EU FP5 ECORA Project Evaluation of computational fluid dynamic methods for reactor safety analysis EVOL ECORA D01 Germany February 2002 55 Curriculum Scientifico del Gruppo di Lavoro Francesco D auria Professore Ordinario di Termoidraulica e di Ingegneria del Nocciolo Moduli dell insegnamento Termoidraulica e Ingegneria del Nocciolo Cod 42411 per il Corso di Laurea Magistrale in Ingegneria Nucleare Universit di Pisa Autore di oltre 100 articoli su rivista e numerose altre pubblicazioni arp unipi it listedoc php ide 5808 Marco Lanfredini Laureando in Ingegneria Nucleare collaboratore dal 2011 presso il Gruppo di Ricerca Nucleare S Piero a Grado Universit di Pisa quale utilizzatore di codici termoidraulici di sistema Autore di rapporti tecnici interni e
23. e Figure 16 CRCoupler main form Coupling interfaces tab example with a list of user defined interfaces Finally the Calculation management tab Figure 17 is used to perform the following tasks Getting additional used information such as o Selection of coupling numerical scheme either Explicit or Semi implicit o Under relaxation factor Run and Stop simulations o Coupled calculation o Standalone CFX calculation o Standalone RELAP calculation Showing output and log ASCII files from RELAP CFX and coupled calculations 21 File Edit Show Data Tools 7 Calculation management Coupling scheme Explicit File shown temporary dirfile 001 out Underrelaxation factor 1 i Di il im bi bi D I fra ca DO OCH D Oo n LO Me bi Pl Write coupling input file Update CFA definition file Figure 17 CRCoupler main form Calculation management tab 28 3 Test Cases In this chapter some of the tests simulated with the Coupling Tool are described Simple verification tests are described in sections 3 1 and 3 2 where a circular pipe is tested against a pressure and temperature ramp In section 3 3 a closed loop is studied addressing the increased instability issues of this type of configuration Finally in section 3 4 an actual 3D case with mixing and re circulating loop is analysed 29 3 1 Test01 Pipe Sin
24. e 33 Figure 21 Test01 2 RELAP Model Test01 3 The RELAP model consists in a PIPE connected to two TIME DEPENDENT VOLUMES through one SINGLE JUNCTIONS and one TIME DEPENDENT JUNCTION PIPE 104 is 5 m long and with a diameter of 0 1 m It s divided in 10 equal elements Absolute roughness is set to zero and no additional concentrated loss factors or abrupt area change is modelled TIME DEPENDENT VOLUME 100 and TIME DEPENDENT JUNCTION 102 are used to set the temperature and the velocity received via a CONTROL VARIABLE updated by CFX within the input of every restart A subroutine of the Master program extracts from the output file the value of the pressure in the VOLUME 104 01 that models the connection with CFX domain The maximum time step for the transient simulation is set to 0 001 seconds the coupling step being 0 05 s Figure 22 shows the nodalization of the RELAP model Figure 22 Test01 3 RELAP Model 3 1 3 Results All the analysed cases were coupled with the implemented explicit coupling scheme For the long configurations some under relaxation is also used Figure 23 shows the CFX convergence behaviour of Test01 2 RMS residuals drops by at least 4 order of magnitude in most of the coupled steps The other cases have similar trends 34 Test01 2 RMS Residuals Mass and Momentum le RMS Rsiduals a Coupled Steps Test01 2 RMS Residuals Energy k af RMS Residuals gt
25. e of the art and in a preliminary development work aimed at a coupling interface between the CFD code ANSYS CFX and the system code RELAP5 including a demonstrative application Such preliminary development work brought to the definition of a coupling strategy and of the related needs in terms of software and information technology development to the availability of a first rough but effective coupling tool and to the identification of technical open issues and needs for further development and improvement work such as convergence issues V amp V aspects practical usability of the coupling tool etc Those outcomes constituted the starting point for the present activity which had the objective of improving and further developing and testing the coupling tool Specific objectives of the present work are 1 To improve the coupling methodology with particular reference to robustness stability and efficiency issues 2 Toassess the influence of sensitivity factors and parameters previously identified 3 Todevelop a User Graphical Interface GUI for a more intuitive and efficient use of the coupling tool 4 Toextend the scope of Verification and Validation V amp V Section 2 describes the developed coupling tool from the general strategy through the details of the implementation of the coupling engine and of the GUI Section 3 focuses on the verification and testing of the coupling tool against several test cases of differe
26. ed First Call Read Mesh Timestep Loop the Outer Look in Tinestep Loop Coefficient Loop gt works Explicit Couplin Linear Solution Start of Run User Input master only User Start Start of Timestep transient only User Input Start of Time Step Start and End of Rigid Body Solution in some cases Start of Coefficient Loop User Input Start of Coefficient Loop Start and End of Rigid Body Solution in some cases Start of Linear Solution Y Abort End of Coefficient Loop Error Condition End of Linear Solution End of Timestep transient only Write Solution End of Run User Output master only Figure 5 Outer Coupling Junction Box Calling Points 17 N Ich init DI jcb read First Partitioning Call only if partitioning is performed DI Ich out Start of Partitioning only H partitioning is performed a j cb end End of Partitioning only if partitioning is performed First Call Start of Run User Input master only User Start Start of Timestep transient only User Input Start of Time Step Start and End of Rigid Body Solution in some cases Start of Coefficient Loop User Input Start of Coefficient Loop Start and End of Rigid Body Solution in some cases Start of Linear Solution Linear Solution a 2 oO Timestep Loop Error Condition ni m End of Linear Solution End of Coefficient Loop End of Ti
27. ess is set to zero and no additional concentrated loss factors or abrupt area change is modelled The BRANCH component has an area of 0 0004 m anda length of 0 1 m Also those components have a square section The maximum time step for the transient simulation is set to 0 0005 seconds the coupling step being assigned CFX through the adaptable time stepping approach Figure 34 shows the nodalization of the RELAP model 45 Figure 34 Test04 RELAP Model TIME DEPENDENT VOLUME 200 fixes the temperature according to the law shown in WW hi E O O Temperature vol 200 Temperature K 2 3 Time s Figure 35 and TIME DEPENDENT JUNCTION 202 imposes a velocity according to the Figure 36 46 LA Lu A E O O Temperature vol 200 Z OO ln 3 sl E U E U F 2 3 Time s Figure 35 Test04 Imposed Temperature of the first feeding line co Velocity jun 202 Velocity m s hu T n O 0 5 1 1 5 Time s Figure 36 Test04 Imposed Temperature of the first feeding line Temperature of TIME DEPENDENT VOLUME 308 follows the value in 304 01 and the velocity in 302 is constantly equal to the one in 310 Pressure in TIME DEPENDENT VOLUME 306 is constantly imposed at the value of pressure in 312 01 minus the head of the pump The pressure component 106 and 206 the temperature in 300 and the velocity in 302 are imposed via a CONTROL VARIABLE updated by CFX withi
28. face Section 0 which allow the user to set up the coupled simulation in a more user friendly fashion Finally some improvements were also needed to set parameters and data areas needed by the synchronization manager A description of the different User FORTRAN routines implemented in the Coupling Code is given here below together with a sketch Figure 4 of their role in the coupling strategy only for the main ones jcb read cel input and jcb out Ich init Junction Box Routine that initialise the coupled simulation In particular it creates all the required data areas for variables handling and set their values to the ones assigned by the user Moreover it set initial parameters for synchronization management 15 Ich read junction box routine called at the start of every coupled CFX step either a new time step for explicit coupling scheme or a new inner iteration for the semi implicit scheme It waits for permission from the Synchronization Manager to read RELAP result files from the previous coupled step After the needed conversions due to possible data inconsistencies it writes the read values in dedicated data areas of the MMS previously defined with the jcb init routine cel input user CEL routine Takes values stored in the dedicated data areas by the jcb read subroutine and use them to compute suitable boundary conditions e g fully developed profiles can be constructed from mass flow rates and geometric data Under relaxatio
29. gle Coupling Boundary Several tests were made on a pipe configuration with only one Coupling Boundary These exercises were intended for testing the explicit coupling scheme and the general features of the Coupling Code In order to address the previously identified instabilities linked to the unsteady inertial terms and to the size of the CFX liquid inventory see Ref 2 two CFX configurations were implemented i a short pipe of 0 5 m and ii a long pipe of 5 m Among the various simulations three were chosen as representative of this test case and are described in this section I Test01 1 RELAP CFX coupled pipe testing the n Flow Boundary Type with short CES configuration II Test01 2 RELAP CFX coupled pipe testing the n Flow Boundary Type with long CFX configuration Il Test01 3 CFX RELAP coupled pipe testing the Out Flow Boundary Type with long CFX pipe All these cases study the same phenomena the time response to a 1 sec ramp in pressure 0 25 bar and temperature 25 C in a 10 m long circular pipe of 0 1 m diameter The imposed upstream pressure and temperature conditions are shown in Figure 18 Upstream Conditions Total Pressure Total Pressure bar Time s Upstream Conditions Total Temperature Total Temperature C Time S Figure 18 Upstream Conditions Test01 amp Test02 3 1 1 Setupofthe CFD model All the three cases have in common the following settings TRANSIENT ANALY
30. he CFD code ANSYS CFX and the system code RELAP5 including a demonstrative application Such preliminary development work brought to the definition of a coupling strategy and of the related needs in terms of software and information technology development to the availability of a first rough but effective coupling tool and to the identification of technical open issues and needs for further development and improvement work such as convergence issues V amp V aspects practical usability o the coupling tool etc Those outcomes constituted the starting point for the present activity which had the objective of improving and further developing and testing the coupling tool Note Rapporto emesso da Gruppo Ricerca Nucleare San Piero a Grado GRNSPG Universit di Pisa Autori L Mengali M Lanfredini F Moretti F D Auria GRNSPG Copia n In carico a nome M Targntigo na m Tarantins 25 09 13 A gt ema AVA TT CIRTEN Consorzio Interuniversitario per la Ricerca TEcnologica Nucleare UNIVERSITA DI PISA S PIERO A GRADO NUCLEAR RESEARCH GROUP ACCOPPIAMENTO DI CODICI CFD E CODICI DI SISTEMA Autori L Mengali M Lanfredini F Moretti F D Auria CERSE UNIPI RL 1510 2013 PISA 6 Settembre 2013 Lavoro svolto in esecuzione dell Attivit LP2 C1 d AdP MSE ENEA sulla Ricerca di Sistema Elettrico Piano Annuale di Realizzazione 2012 Progetto B 3 1 Sviluppo competenze scientifiche nel campo dell
31. he strong feedbacks required in coupling 1D TH with 3D CFD codes the codes will run concurrently with a continuous exchange of information in both ways Partitioned solution The use of proprietary codes does not allow access and modification of the source codes for this reason the choice is restricted to a partitioned approach Despite the lower efficiency with respect to a monolithic strategy the partitioned solution offers an improved versatility for the possibility of a modular approach Moreover no further development will be required to use the Coupling Code with future versions of the TH and CFD codes Sub cycling Both codes are allowed to make their own sub cycling between coupling steps Moreover these synchronization points coupling steps can be set a priori fixed approach or calculated during the coupled simulation adaptable coupling step mainly following the CFX adaptable time stepping features This configuration is the most versatile allowing both codes to use the most suitable time step for an efficient solution of the problem Sequential Coupling Considering that CFD computational times are usually substantially larger than TH ones the time efficiency improvement using a parallel approach can be considered negligible For this reason a Sequential Coupling was chosen which is simpler to implement and also more stable Numerical Schemes In order to solve the convergence issues identified in the previous work a se
32. icrosoft Windows systems but this has not been checked yet Anyway it is envisaged that with relatively little extra developing effort the software portability can appropriately be extended to other Microsoft Windows releases e g Vista an Eight 2 Microsoft NET Framework 4 0 or later e The package can freely be downloaded from Ref 6 and installed Alternatively the CRCoupler installing wizard itself detects the missing passage if the case and looks after its installation 3 ANSYS CFX version 14 0 or 14 5 or 15 0 e Different releases can straightforwardly be included as options by minor modifications to the program e Version v15 0 is not available yet and is expected to be released by ANSYS by the end of 2013 e The code must be installed in the default installation folder that is C Program Files ANSYS InclvXXX with XXX 140 145 or150 The infrequent case in which the installation unit is not C has not been handled 4 RELAP5 versions 3 3 gl 3D v2 4 2 3D v4 0 3ie e Different releases can straightforwardly be included as options by minor modifications to the program These ones are those available on the computers on which the CRCoupler and the coupling tools have been developed and tested 20 5 PERL interpreter e The PERL scripting language is used by the engine of the coupling tool which is in turn invoked by CRCoupler e PERL interpreters for Microsoft Windows can freely be downloaded
33. in this case the limitation of variables change is carried out within inner couplings However since convergence of coupled variables has to be achieved in order to complete the time step under relaxation will not distort the time response of the coupled solution The execution sequence for the semi implicit coupling scheme is shown in Figure 2 New Coupled Step YES EXPLICIT COUPLING SCHEME Figure 1 Explicit Coupling Scheme New Coupled Step Inner Iteration YES End Coupled Simulation SEMI IMPLICIT COUPLING SCHEME Figure 2 Semi Implicit Coupling Scheme 11 2 2 2 Coupling Boundaries and Variables A Coupling Boundary is the combination of components boundaries specifically modelled to transfer data between the two coupled sub domains Coupling variables can travel both ways from and to each sub domain depending on the type of the coupling boundary A schematic sketch of the coupling Boundaries and Variables is shown in Figure 3 while a description of the exchange patterns is given here below CFX All fluid boundaries in CFX inlets outlets openings can be coupled with only minor adjustments in their set up All thermodynamic variables e g pressure temperature and velocities can be exchanged through the same boundary in both directions RELAP On the other hand the RELAP sub domain need some geometric modifications adjustments or insertion of dummy components as explained better in Section 2 3 3 in
34. initial setting of the problem in both the sub domains When the coupled steady state solution is reached restart files are written for both the codes and can be used as improved initialization conditions for any transient coupled simulation having the same initial settings 14 2 3 Coupling Routines 2 3 1 Coupling Manager A Perl routine was developed to manage the coupling of the two codes The functions of this program are described hereafter The Coupling Manager read the input file provided by the GUI that contains information about the interfaces of coupling and other setting chosen by the user to perform the calculation The RELAP nodalization is modified to add the minor edit request needed for coupling After these modifications the script performs a RELAP standalone calculation to initialize this domain and reset the time to zero The Coupling Manager organizes the information exchange between the two codes providing in a suitable format the files containing the needed information Namely the output data from CFX are manipulated to obtain a restart input file for RELAP and the output file of RELAP in scanned to find the information to be passed to CFX in the minor edit section In case of under relaxation request the Coupling Manager manipulates the data to apply this technique Log files to monitor the status of the calculation and the exchanged variables are created 2 3 2 CFX routines To add additional features and physical mode
35. ls to CFX it is possible to write subroutines in FORTRAN and have the CFX Solver call them through a source code interface CFX supports user subroutines written in FORTRAN 77 or FORTRAN 90 These allow the user to access the Memory Management System MMS of the code thus giving a very fine control over the simulation access to resolved field variables and allowing the introduction of almost any external user made routine Two kinds of routines are available User CEL CFX Expression Language routines that can be used to introduce user defined functions in addition to the predefined ones available in CFX and Junction Box Routines JCB that can be called at several points during the solution in order to perform tasks defined by the user These features are essential for the creation of a coupling with external codes In order to transfer data between codes access to the internal data structures of the CFX solver is needed User FORTRAN routines allow internal data manipulation through the use of the CFX Memory Management System MMS utilities These utilities are implemented in the code and described in the ANSYS manuals 3 With respect to the previous work PAR2011 some additional CFX routines were developed jcb_init and jcb end More importantly all the routines were improved to allow for different coupling schemes and a more generic and versatile handling of the coupling variables This was essential for the development of the Graphic Inter
36. mains for the short and long pipe configurations respectively In order to reduce the computational times rather coarse grids were developed with number of nodes below 20 000 The overall grid quality is good and the distance of the first node from the wall is 2 mm for the short pipe and 3 mm for the long one 31 N of nodes ICEM quality Wall node distance OUTLET MOTO im Figure 19 Test01 CFX Computational Domain Short Pipe Configuration N of nodes ICEM quality Wall node distance OUTLET ooro mi WS 1 000 m Figure 20 Test01 CFX Computational Domain Long Pipe Configuration BOUNDARY CONDITIONS Different conditions were imposed depending on the coupling boundary type under testing In Flow boundary type Short configuration Test01 1 and Long configuration Test01 2 e Inlet velocity and temperature imposed through a function linked to the cel input subroutine 32 e Outlet outlet pressure controlled condition with zero relative pressure absolute pressure equal to the reference pressure of 10 bar e Wall Adiabatic no slip condition with automatic turbulence wall treatment and no wall roughness Out Flow boundary type Long configuration Test01 3 e Inlet Total relative pressure and Total temperature imposed following the defined transient conditions see Figure 18 above e Outlet outlet pressure controlled condition set by a function linked to the cel input subroutine
37. matic turbulence wall treatment and no wall roughness 3 2 2 Setup of the RELAP model The RELAP model consists in a PIPE connected to two TIME DEPENDENT VOLUMES through two SINGLE JUNCTIONS and a PIPE connected to two TIME DEPENDENT VOLUMES through one SINGLE JUNCTIONS and one TIME DEPENDENT JUNCTION PIPE 104 and 204 are 2 5 m long and with a diameter of 0 1 m They are divided in 5 equal elements Absolute roughness is set to zero and no additional concentrated loss factors or abrupt area change is modelled TIME DEPENDENT VOLUME 100 fixes the pressure and the temperature according to the law shown in Figure 18 while component 108 is used to realize the coupling Namely the pressure in such component is imposed via a CONTROL VARIABLE updated by CFX within the input of every restart A subroutine of the Master program extracts from the output file the value of the velocity in the SINGLE JUNCTION 106 that models the connection with CFX domain The maximum time step for the transient simulation is set to 0 001 seconds the coupling step being 0 05 s Figure 26 shows the nodalization of the RELAP model Figure 26 Test02 RELAP Model St 3 2 3 Results The coupled simulation is again tested with the explicit coupling scheme Convergence of the coupled simulation was somehow worse and a higher under relaxation factor was used With these settings the simulation runs quite smoothly reaching the CFX imposed convergence criteria within few i
38. mestep transient only Write Solution End of Run User Output master only Semi Implicit Coupling works in the Inner Loop Coefficient Loop Figure 6 Inner Coupling Junction Box Calling Points 18 2 3 3 RELAP modifications A special attention has to be posed in developing a RELAP input suitable for coupling The critical points are the ones in which the information have to be exchanged between the two codes The interfaces have to been modelled using TIME DEPENDENT VOLUME and TIME DEPENDENT JUNCTION components where the thermo hydraulic condition are controlled via CONTROL VARIABLE to allow CFX to pass data to RELAP Namely at every coupling time with the restart input the value of the CONTROLVAR constant is changed To exchange data from RELAP to CFX a devoted subroutine of the Master program read at each coupling time the output file of RELAP and extract the value that have to been passed The use of RELAP 3D 4 0 3 required to find the values needed for the coupling in the output because the strip procedure is not properly working after the restart procedure in this version of the code To model a pump the suction line have to be modelled with an interruption using two TIME DEPENDENT VOLUME where the one connected to the pump have the temperature set constantly at the same value of the volume connected at the other added TIME DEPENDENT VOLUME This necessity is due to the presence of the TIME DEPENDENT JUNCTION required by
39. mi implicit coupling was developed To improve convergence both the implemented schemes explicit and semi implicit can use under relaxation Explicit Coupling In the explicit coupling the CFD tool is the master code and the TH code is the slave After completing initialization of the coupled variables the Master Code CFX performs the first time step The calculated solution variables are put in dedicated text files to provide boundary conditions for the Slave process RELAP With these updated coupling variables new time step values the Slave Code solves its domain closing the time step If ending conditions are satisfied the coupled calculation is stopped otherwise a new coupled step is calculated This type of coupling scheme is more prone to instabilities The main reason for this is that the time step size is limited by the Courant Friedrich Levy CFL limit Large time steps in coupled explicit runs can lead to numerical instabilities or inconsistent results and solution divergence Under relaxation can be used to improve the convergence of the solution However since under relaxation limits the change of coupled variables between coupling steps its use in transient problems can distort the time response of the solution For this reason the use of under relaxation with explicit coupling is recommended only for steady state or very slow transient problems Advantage of such scheme is the faster computational times since boundary conditions a
40. n form Calculation management Tab 28 Figure 18 Upstream Conditions Test01 amp Test 30 Figure 19 Test01 CFX Computational Domain Short Pipe Configuration 32 Figure 20 Test01 CFX Computational Domain Long Pipe Configuration 32 Fe 21 Testo RENN das 34 Fielire 22 Tee KC RELAP NO linia 34 FeUe 23 Test01 2 CFX RMS Resina diia 35 Figure 24 Test01 Comparison Pipe Velocity at 2 5 Meter 36 Figure 25 Test01 Comparison Pipe Temperature at 2 5 meter 36 PEN FO PT 37 Figure 27 Test02 Comparison Pipe Velocity at bmeter 38 Figure 28 Test02 Comparison Pipe Velocity at bmeter 39 Fe TED EPP atan outs 40 Figure 30 Test03 e Max NNN PR 42 Figure 31 Test03 CFX RMS Residuals ococonoononconorocnonaciononacionoonoronnonaricnonasroonorornonariononasironconarennonasos 43 Figure 32 Test03 Comparison Loop VEIOCity ii 43 Figure 33 Test04 CFX Computational Domain Mixing Box 45 Fleure 34 Testa RELAP NRL oia 46 Figure 35 Test04 Imposed Temperature of the first feeding line ce eeecccessseccceesececeeeceeseeeceteuneces 47 Figure 36 Test04 Imposed Temperature of the first feeding line cc eecccccesseccceesececeeeceeeaeseceeeeneees 47 PeuUr e3T Test04 CEX Max RESIGUAIS ua 48 Figure 38 Test04 Temperature at Boundaries rernaransnsvrvnnrerensnerensrerennvenansvrvnnnerenervnnnerensverensvevsnsv
41. n of some of the transferred values is possible Ich out junction box routine called at the end of every coupled CFX step It computes suitable averages of the variables to be exchanged from the MMS data areas and writes them along with the time step size and other synchronization parameters to specific CFX result files If needed the routine can also make conversions in order to transfer consistent values of the exchanged data i e conversion from velocity to mass flow rates or conversion from relative to absolute pressure values Under relaxation of some of the transferred values is also possible in order to improve stability Ich end junction box routine When ending conditions either from the internal solution or from the Synchronization Manager are identified this routine stops the coupled simulation ending the CFX simulation and sending a stop message to the Coupling Manager COUPLING INTERFACE Data Areas P from RELAP RELAP results txt file he CFX boundary interface RELAP RESTART Input file _ FORELAP _ Figure 4 CFX Coupling Routines Calling points for the different Junction Box Routines are shown in Figure 5 and Figure 6 for the explicit and semi implicit coupling scheme respectively 16 N Ich init Ich read E jcb out N Ich end First Partitioning Call only if partitioning is performed Start of Partitioning only H partitioning is performed End of Partitioning only if partitioning is perform
42. n the input of every restart A subroutine of the Master program extracts from the output file the value of the velocity in the JUNCTION 103 01 and 204 01 the temperature in 104 01 the pressure in 304 01 and the temperature in 204 01 that models the connection with CFX domain 47 3 4 3 Results The semi implicit coupling scheme was adopted for this case since there is a closed loop in the system together with other possible instabilities issues Under relaxation was used through the inner loops Figure 37 the CFX convergence behaviour of Test04 RMS residuals drops by at least 4 order of magnitude in most of the coupled steps In particular quite low residuals can be achieved in part of the transient not affected by the strong temperature gradients Test03 MAX Mass and Momentum MAX Rsiduals 0 200 400 600 800 1000 1200 1400 1600 1800 2000 Coupled Steps WEE Ee Gg CO n op v 10 0 00000000084 ee nee cece e eessen eene ee e e een ee l keeeeeeeesssseeeeeeee Geld e see ee e H Energy 00000 ree ee ee ee ee 0 keeeessssseeeeeeeesse i e o oon onnnono oo E 3T i 3 rs gt 10 E A ARA CRI IA EEN St 10 0 200 400 600 800 1000 1200 1400 1600 1800 2000 Coupled Steps Figure 37 Test04 CFX Max Residuals Figure 38 and Figure 39 show the time response of temperature and velocity for the fluid boundaries of CFX two inlets and two o
43. nt levels of complexity 2 The Coupling Tool This chapter describes the developed in house Coupling Tool between a system code and a CFD code Section 2 1 presents the employed commercial codes Section 2 2 describes the main features of the coupling strategy while in section 2 3 a description of the developed coupling routines is given Finally the Graphical Interface is described in section 2 4 2 1 Selected Codes For the development of the Coupling Tool commercial codes available and regularly used at GRNSPG are chosen in particular RELAP5 and RELAP5 3D as the system code and ANSYS CFX as the CFD code Description of the main features of each code can be found in respective code manuals e g Refs 3 and 4 or more briefly in the report of the previous activity Ref 2 2 2 Coupling strategy Considering the coupling issues identified in the preliminary work carried out in the previous PAR2011 Ref 2 the Coupling Code was improved and tested against some simple cases see Section 3 for further details In particular more robust numerical schemes were developed and all the coupling routines were modified to allow for a more generic and intuitive set up of the coupled simulation An external Coupling Interface was developed to manage the information exchange between the two codes In general data transfer can be addressed in two different ways through shared memory areas and through I O routines handling text files A dire
44. nternal iterations The simulated flow is tested against the same RELAP stand alone case showing rather good comparison as shown in Figure 27 and Figure 28 below A little higher distortion in the first phase of the coupled simulation is found with respect to the previous cases Test01 probably due to the higher under relaxation factor used together with the explicit coupling scheme E 2 8 3 Time s Figure 27 Test02 Comparison Pipe Velocity at 2 5 meter 38 LA La all Aen E al b d a d ken Ja un d CH oo on Figure 28 Test02 Comparison Pipe Velocity at 2 5 meter 39 3 3 Test03 Closed Loop with Pump Coupled systems are usually more prone to instabilities in closed loops Test03 consists in a pump start up within a closed loop with a pressurizer fixing the pressure in one point of the system to 1 MPa Isothermal conditions 20 C were imposed in order to test the hydraulic strong coupling Velocity Pressure strong feedbacks in closed loops without additional sources of uncertainties 3 3 1 Set up of the CFD model The long pipe geometry used in the previous case was adopted also in this case The CFX simulation set up is the same of the previous case Test02 including the boundary conditions which are repeated here below for simplicity BOUNDARY CONDITIONS e Inlet velocity and temperature imposed through a function linked to the cel input subroutine e Outlet outlet pressure cont
45. o sub domains some inconsistencies can rise i not consistent variables e g Relative Pressure Vs Absolute Pressure Velocity Vs Mass Flow Rates ii not consistent profiles e g from 1D data to 3D profiles The Coupling Code can handle such discrepancies either i making the required transformations or ii with the possibility of impose Fully Developed velocity profiles in the CFD sub domain Coupling Boundaries types components and exchanged variables are summarised in Table 1 below Table 1 Coupling Boundaries Components and Variables Temperature Enthalpy PIPE BR Default Uniform Profile Velocity Mass Flow Rate Inlet Opening Possible 1D gt 3D profiles GET Inlet Opening SEND u a Outlet peng Pressure PIPE BR GET n TDV Outlet Opening Temperature Enthalpy 3 Velocity Mass Flow Rate 5 TDI 13 2 2 3 Coupled Initialization During the testing phase some initialization issues were identified More specifically depending on the assigned starting values for the coupling variables strong oscillations can occur in the initial part of the coupling simulation These can lead to divergence of solution or to incorrect prediction of the initial transient response For this reason a COUPLED INITIALIZATION feature was added to the Coupling Code Starting with the assigned initialization values a steady coupled solution is calculated with strong under relaxation and using the
46. pressing the appropriate buttons By such operation CRCoupler gathers information on the boundaries that can be used as coupling interfaces and on the control variables that may be associated with RELAP time dependent junctions and volumes and which may be necessary for data exchange from CFX to RELAP etc The Add interface button is used for the creation of a new coupling interface It invokes the Interface creation form shown in Figure 12 Such secondary form allows the user to input the following information Name of the interface Sender code i e the code the information is transferred from either CFX or RELAP Figure 13 a CFX boundary among those recognized from CCL parsing Figure 13 b RELAP boundary among those recognized from RELAP input parsing Figure 13 c Exchanged variable either Pressure Velocity or Temperature Figure 13 d Initial value for the exchanged variable in SI units Velocity or temperature profile either uniform or fully developed only when RELAP is the sender code and thus an assumption is needed for the cross sectional profile at the CFX boundary Before the creation of the interface is confirmed by pressing the Done button the user must perform the Check Update operation by the corresponding button which gives the following results see Figure 12 b If CFX is the sender code and the RELAP boundary is a time dependent junction or a time dependent volume then
47. ral information Coupling interfaces Calculation management The General information tab allows the user to input the following information by several controls such as dialogs combo boxes check boxes and text boxes Path of the working directory in which all input and output files are placed before and or after the calculation runs CCL stands for CES Command Language and a CCL file is an ASCII file which contains all information about the setup of a CFX calculation except the spatial discretization 1 e the mesh according to a sort of scripting language which allows easy manipulation for calculation setup modification or update A CCL file is exported from a definition file i e the binary file with def extension that contains mesh and calculation setup upon user request and can be imported onto a def file in order to update the setup Details on these aspects can be found in Ref 3 21 Name of the coupling case case1 is adopted by default RELAP o Selection of the solver mod 3 3 gl 3D v2 4 2 3D v4 0 3ie o Restart mode o Path of the input file either i or inp o The system default text editor can optionally be invoked by pressing a button for possible input file editing o Path of the restart file rst for calculation initialization CEA o Selection of the solver v14 0 v14 5 v15 0 o Path of the definition file def o Path of the results file res
48. re not changed within one time step The execution sequence for the explicit coupling scheme is shown in Figure 1 Semi implicit Coupling Inner iterations steps are added within each time step In particular the Master and Slave codes keep solving the same time step until specified convergence criteria are reached Once these conditions are met the current time step is closed and the execution of the next one is initiated Also in this case the CFX code works as the master code initiating the simulation and passing boundary conditions to the Slave process If the values of the exchanged variables calculated by the Slave process are not consistent with the ones calculated by the Master process a new inner iteration is initiated The main advantage of the semi implicit coupling scheme is that consistent coupling variables are found and mass momentum energy conservation is more easily achieved Moreover any disturbance or change in a coupled variable within one sub domain will be limited by the immediate feedback of the solution in the other sub domain This is the main reason why semi implicit schemes tend to increase the numerical stability of coupled simulations with strong physical feedbacks Such schemes are very efficient in case of steep increase or decrease of thermal hydraulic variables In order to improve further the stability of the solution under relaxation can be used within inner iterations In contrast with the explicit coupling scheme
49. rnnnene 49 Figure 39 Test04 Velocity at Boundaries iii 50 Figure 40 Test0A Flow Features Lua 50 Heure 41 Test04 Flow Features Lapa 51 Fieure 42 Test04 Fo FCatures Ja 51 Fig re 43 FestO4 Flow Features dinar is 52 Fietire 44 Test0d ON Feed se 52 FIgure 45 KR Flow Features oi lla 53 1 Introduction This report describes the work performed by the Gruppo di Ricerca Nucleare di San Piero a Grado GRNSPG of the University of Pisa as member of CIRTEN consortium in the frame of the Accordo di Programma MSE ENEA sulla Ricerca di Sistema Elettrico Piano Annuale di Realizzazione 2012 Ref 1 In particular this report constitutes the deliverable LP2 c 1 d of the corresponding activity scheduled in Linea Progettuale 2 of Project B 3 1 The activity deals with the development and improvement of coupling techniques between Computational Fluid Dynamics CFD codes and thermal hydraulic TH system codes techniques that are expected to enhance the analysts and designers capabilities to simulate and predict the behavior of nuclear reactors including those belonging to Generation IV during normal and abnormal operating conditions The work described in this report is a follow up of that previously performed in the frame of PAR 2011 and documented in Ref 2 That initial work consisted in a review of the existing and accessible literature and technical documentation on the same subject so as to identify the stat
50. rolled condition set by a function linked to the cel input subroutine e Wall Adiabatic no slip condition with automatic turbulence wall treatment and no wall roughness 3 3 2 Setup of the RELAP model The RELAP model consists in a PIPE connected to two TIME DEPENDENT VOLUMES through one SINGLE JUNCTIONS and one TIME DEPENDENT JUNCTION and a PIPE connected to two TIME DEPENDENT VOLUME using a SINGLE JUNCTION and a PUMP component Figure 29 shows the nodalization of the RELAP model Figure 29 Test03 RELAP Model 40 PIPE 104 and 204 are 2 5 m long and with a diameter of 0 1 m They are divided in 5 equal elements Absolute roughness is set to zero and no additional concentrated loss factors or abrupt area change is modelled PUMP component 102 have a length of 0 2 m and a volume of 2 e 3 m The velocity of the pump change from 0 to 5 0 rad s in 5 seconds The temperature of TIME DEPENDENT VOLUME 100 follow the value of volume 204 05 The pressure in the TIME DEPENDENT VOLUME 108 the temperature in TIME DEPENDENT VOLUME 200 and the velocity in TIME DEPENDENT JUNCTION 202 are imposed via CONTROL VARIABLE updated by CFX within the input of every restart Component 208 is required to interrupt the circuit This solve the issue coming from the necessity to impose the velocity in the TIME DEPENDENT JUNCTION 202 in addiction at the one imposed by the pump A subroutine of the Master program extracts from the output file the value of the velocity
51. to Rapporto Tecnico Collocazione contrattuale Accordo di programma ENEA MSE su sicurezza nucleare e reattori di IV generazione Argomenti trattati Analisi di sicurezza Termoidraulica dei reattori nucleari Termoidraulica del nocciolo Generation IV reactors Sommario This report describes the work performed by the Gruppo di Ricerca Nucleare di San Piero a Grado GRNSPG of the University of Pisa as member of CIRTEN consortium in the frame of the Accordo di Programma MSE ENEA sulla Ricerca di Sistema Elettrico Piano Annuale di Realizzazione 2012 In particular this report constitutes the deliverable LP2 c 1_d of the corresponding activity scheduled in Linea Progettuale 2 of Project B 3 1 The activity deals with the development and improvement o coupling techniques between Computational Fluid Dynamics CFD codes and thermal hydraulic TH system codes techniques that are expected to enhance the analysts and designers capabilities to simulate and predict the behavior of nuclear reactors including those belonging to Generation IV during normal and abnormal operating conditions The work described in this report is a follow up o that previously performed in the frame of PAR 2011 That initial work consisted in a review of the existing and accessible literature and technical documentation on the same subject so as to identify the state of the art and in a preliminary development work aimed at a coupling interface between t
52. utlets The mixing effect and re circulating features of the problem under study can be seen in the boundaries temperature evolution OULTET1 and OULTET12 temperatures start to rise after about half a sec to values lower than the INLET1 temperature step At about 1 sec the inlet 1 starts to rise in temperature too following the re circulated flow from outlet 1 In figure 32 the shut down of the first feeding line can be identified with the related decrease in velocities in the other lines with the OUTLET2 line decreasing more than the INLET1 and INLET2 lines The detailed flow field for different values of simulation time is shown in figures from Figure 40 to Figure 45 It is interesting to note the mixing of the thermal jet coming from INLET 2 The jet is Somehow distorted upward by the flow field and eventually reach OUTLET1 near 1 sec After that the Temperature step is finished and the introduced energy continue to be mixed lowering the temperature peaks After nearly 2 sec from the simulation start the higher temperature wake is re introduced by INLET1 at a lower peak value After 1 5 seconds the INLET2 starts to slow down giving the possibility to the mixed temperature zone to partly enter 48 the INLET2 region After 2 5 seconds the domain is quite mixed and temperature at INLET1 begins to decrease again These complicated flow features would have been very difficult to model in a RELAP stand alone simulation INLET1 INLET2
53. xpected 42 MAX Rsiduals ve LIM fett LAD A ri tai 100 150 200 Coupled Steps Test03 RMS Residuals Energy a d MAX Residuals gt H Energy 150 Coupled Steps Figure 31 Test03 CFX RMS Residuals RELAP stand alone coupled Testos l 5 Time s Figure 32 Test03 Comparison Loop Velocity 43 3 4 Test05 Complex 3D re circulating component This last case finally tests the coupling capabilities in an actual 3D problem The CFD sub domain consists in a Mixing Box connecting four flow lines modelled in RELAP through two inlets and two outlets More specifically there are two feeding line a discharge line and a closed loop with a pump reconnecting one of the outlets with one of the second feeding line to the second inlet Moreover a Temperature step 1 sec is imposed to the main feeding line in order to address mixing issues After 1 5 seconds from the start the main feeding line is closed with a descending ramp of 0 5 seconds and the flow continues to circulate driven only by the second feeding line and by the circulating loop with the pump In this configuration the temperature step will be mixed within the domain and will be partially transported to the re circulating loop Eventually the step will come back reduced in magnitude due to the mixing through the connected inlet continuing to circulate through the system until its complete
Download Pdf Manuals
Related Search