
[Sponsors] 
Low Mach number wing/body junction convergence 

LinkBack  Thread Tools  Search this Thread  Display Modes 
September 15, 2014, 05:56 
Low Mach number wing/body junction convergence

#1 
Member
Zeno
Join Date: Sep 2013
Location: Delft, The Netherlands
Posts: 63
Rep Power: 10 
Hello Everybody,
I am trying to simulate a naca 0015 wing attached perpendicularly to a flat plate. The Mach number is ~0.1 and I am using the compressible solver. My stopping criterion is based on the reduction of the residual of 4 orders of magnitude. I am using a coarse mesh with 7*10^5 cells. However, I have problems with the convergence. Is it because I am using the compressible solver ?( I tried to use the incompressible one and it still doesn't converge). Or is there a problem with the settings of my configuration file? Attached are a plot of the residuals, an image of the mesh and a file containing the info specified in the configuration file and the error I get after ca. 30000 iterations. Any help is appreciated. Thanks, Z Code:
 Physical case definition  Compressible RANS equations. Turbulence model: Spalart Allmaras Mach number: 0.098. Angle of attack (AoA): 6 deg, and angle of sideslip (AoS): 0 deg. Reynolds number: 762000. No restart solution, use the values at infinity (freestream). The reference length/area will be computed using y(2D) or z(3D) projection. The reference length (moment computation) is 0.34. Reference origin (moment computation) is (0.25, 0, 0). Surface(s) where the force coefficients are evaluated: wing. Surface(s) plotted in the output file: wing. Input mesh file name: naca0015_rotated.su2  Space numerical integration  Roe solver for the flow inviscid terms. Second order integration with slope limiter. Venkatakrishnan slopelimiting method, with constant: 10. The reference element size is: 0.34. Scalar upwind solver (first order) for the turbulence model. First order integration. Average of gradients with correction (viscous flow terms). Piecewise constant integration of the flow source terms. Average of gradients with correction (viscous turbulence terms). Piecewise constant integration of the turbulence model source terms. Gradient computation using GreenGauss theorem.  Time numerical integration  Local time stepping (steady state simulation). Euler implicit method for the flow equations. CFL ramp definition. factor: 1.1, every 1000 iterations, with a limit of 10. CourantFriedrichsLewy number: 1 Euler implicit time integration for the turbulence model.  Convergence criteria  Maximum number of iterations: 999999. Reduce the density residual 5 orders of magnitude. The minimum bound for the density residual is 10^(10). Start convergence criteria at iteration 10.  Output information  Writing a flow solution every 250 iterations. Writing the convergence history every 1 iterations. The output file format is Tecplot ASCII (.dat). Convergence history file name: history. Surface flow coefficients file name: surface_flow. Flow variables file name: flow. Restart flow file name: restart_flow.dat.  Config file boundary information  Farfield boundary marker(s): farfield. Symmetry plane boundary marker(s): symmetry. Constant heat flux wall boundary marker(s): wing, flatplate.  Read grid file information  Three dimensional problem. 727478 interior elements including halo cells. [n1204:07091] 47 more processes have sent help message helpmpibtlbase.txt / btl:nonics [n1204:07091] Set MCA parameter "orte_base_help_aggregate" to 0 to see all help / error messages 727478 hexahedra. 677150 points, and 156763 ghost points.  Geometry Preprocessing  Setting point connectivity. Setting element connectivity. Checking the numerical grid orientation. Identifying edges and vertices. Computing centers of gravity. Setting the control volume structure. Volume of the computational grid: 370. Searching for the closest normal neighbors to the surfaces. Compute the surface curvature. Max K: 0. Mean K: 0. Standard deviation K: 0.  Solver Preprocessing  Computing wall distances. Area projection in the zplane = 0.995. Viscous flow: Computing pressure using the ideal gas law based on the freestream temperature and a density computed from the Reynolds number. Note: Negative pressure, temperature or density is not allowed! Force coefficients computed using freestream values.  Input conditions: Specific gas constant: 287.058 N.m/kg.K. Freestream pressure: 99463.3 Pa. Freestream temperature: 288.15 K. Freestream density: 1.20247 kg/m^3. Freestream velocity: (33.1664, 0, 3.48593) m/s. Magnitude: 33.3491 m/s. Freestream total energy per unit mass: 207345 m^2/s^2. Freestream viscosity: 1.7893e05 N.s/m^2. Freestream turb. kinetic energy per unit mass: 4.17061 m^2/s^2. Freestream specific dissipation: 28028 1/s.  Reference values: Reference specific gas constant: 1 N.m/kg.K. Reference pressure: 1 Pa. Reference temperature: 1 K. Reference density: 1 kg/m^3. Reference velocity: 1 m/s. Reference energy per unit mass: 1 m^2/s^2. Reference viscosity: 1 N.s/m^2.  Resulting nondimensional state: Mach number (nondim): 0.098 Reynolds number (nondim): 762000. Re length: 0.34 m. Specific gas constant (nondim): 287.058 Freestream temperature (nondim): 288.15 Freestream pressure (nondim): 99463.3 Freestream density (nondim): 1.20247 Freestream velocity (nondim): (33.1664, 0, 3.48593). Magnitude: 33.3491 Freestream total energy per unit mass (nondim): 207345 Freestream viscosity (nondim): 1.7893e05 Freestream turb. kinetic energy (nondim): 4.17061 Freestream specific dissipation (nondim): 28028 Initialize jacobian structure (NavierStokes). MG level: 0. Initialize jacobian structure (SA model).  Begin Solver  Maximum residual: 1.25842, located at point 41549. Iter Time(s) Res[Rho] Res[nu] CLift(Total) CDrag(Total) 0 1.590721 3.210747 5.593713 1.565292 1.150936 [...] 38618 1.908126 7.419505 6.272192 0.632451 0.035221 38619 1.908131 7.419517 6.272200 0.632451 0.035221 CSysSolve::modGramSchmidt: w[i+1] = NaN CSysSolve::modGramSchmidt: w[i+1] = NaN terminate called after throwing an instance of 'CSysSolve::modGramSchmidt: w[i+1] = NaN terminate called after throwing an instance of 'int' CSysSolve::modGramSchmidt: w[i+1] = NaN CSysSolve::modGramSchmidt: w[i+1] = NaN CSysSolve::modGramSchmidt: w[i+1] = NaN CSysSolve::modGramSchmidt: w[i+1] = NaN int' [n1204:07125] *** Process received signal *** [n1204:07125] Signal: Aborted (6) [n1204:07125] Signal code: (6) CSysSolve::modGramSchmidt: w[i+1] = NaN terminate called after throwing an instance of 'int' [...] [n1204:07137] *** Process received signal *** [n1204:07137] Signal: Aborted (6) [n1204:07137] Signal code: (6) [n1204:07114] *** Process received signal *** [n1204:07114] Signal: Aborted (6) [n1204:07114] Signal code: (6) [n1204:07131] [ 0] /lib64/libpthread.so.0(+0xf710) [0x2b3c9d4af710] [n1204:07131] [ 1] /lib64/libc.so.6(gsignal+0x35) [0x2b3c9d6f2925] [n1204:07131] [ 2] /lib64/libc.so.6(abort+0x175) [0x2b3c9d6f4105] [n1204:07131] [ 3] /usr/lib64/libstdc++.so.6(_ZN9__gnu_cxx27__verbose_terminate_handlerEv+0x12d) [0x2b3c9cdb6a5d] [n1204:07131] [ 4] /usr/lib64/libstdc++.so.6(+0xbcbe6) [0x2b3c9cdb4be6] [n1204:07131] [ 5] /usr/lib64/libstdc++.so.6(+0xbcc13) [0x2b3c9cdb4c13] [n1204:07131] [ 6] /usr/lib64/libstdc++.so.6(+0xbcd0e) [0x2b3c9cdb4d0e] [n1204:07131] [ 7] SU2_CFD(_ZN9CSysSolve14modGramSchmidtEiRSt6vectorIS0_IdSaIdEESaIS2_EERS0_I10CSysVectorSaIS6_EE+0x25e) [0x8b389e] [n1204:07131] [ 8] SU2_CFD(_ZN9CSysSolve6FGMRESERK10CSysVectorRS0_R20CMatrixVectorProductR15CPreconditionerdmb+0x82d) [0x8b536d] [n1204:07131] [ 9] SU2_CFD(_ZN12CEulerSolver23ImplicitEuler_IterationEP9CGeometryPP7CSolverP7CConfig+0x63c) [0x6918dc] [n1204:07131] [10] SU2_CFD(_ZN21CMultiGridIntegration15MultiGrid_CycleEPPP9CGeometryPPPP7CSolverPPPPP9CNumericsPP7CConfigtttmt+0x359) [0x4d7819] [n1204:07131] [11] SU2_CFD(_ZN21CMultiGridIntegration19MultiGrid_IterationEPPP9CGeometryPPPP7CSolverPPPPP9CNumericsPP7CConfigtmt+0x371) [0x4d8801] [n1204:07131] [12] SU2_CFD(_Z17MeanFlowIterationP7COutputPPP12CIntegrationPPP9CGeometryPPPP7CSolverPPPPP9CNumericsPP7CConfigPP16CSurfaceMovementPP19CVolumetricMovementPPP15CFreeFormDefBox+0x67a) [0x4ee4fa] [n1204:07131] [13] SU2_CFD(main+0xb52) [0x735072] [n1204:07131] [14] /lib64/libc.so.6(__libc_start_main+0xfd) [0x2b3c9d6ded1d] [n1204:07131] [15] SU2_CFD() [0x47ce69] [n1204:07131] *** End of error message *** [...]  Physical case definition  Input mesh file name: naca0015_rotated.su2  Output information  The output file format is Tecplot ASCII (.dat). Flow variables file name: flow.  Config file boundary information  Farfield boundary marker(s): farfield. Symmetry plane boundary marker(s): symmetry. Constant heat flux wall boundary marker(s): wing, flatplate.  Read grid file information  Three dimensional problem. 727478 interior elements including halo cells. 727478 hexahedra. 677150 points, and 156763 ghost points. Identify vertices.  Solution Postprocessing  [...]  Exit Success (SU2_SOL)  

September 21, 2014, 21:20 

#2  
Super Moderator
Francisco Palacios
Join Date: Jan 2013
Location: Long Beach, CA
Posts: 404
Rep Power: 13 
Quote:
Visualize the limiter and residuals WRT_RESIDUALS= YES WRT_LIMITERS= YES could be also useful to identify the problem. And do not forget to play with the slopelimiting coefficient. Cheers, Francisco 

September 22, 2014, 05:20 

#3 
Member
Zeno
Join Date: Sep 2013
Location: Delft, The Netherlands
Posts: 63
Rep Power: 10 
Dear Francisco,
Thank you for your reply. It was definitely the value of the slope limiting coefficient. 

October 14, 2014, 12:28 

#4 
Member
Zeno
Join Date: Sep 2013
Location: Delft, The Netherlands
Posts: 63
Rep Power: 10 
For the same test case, I wanted to see if I get better results when RoeTurkel preconditioning is activated.
I therefore set: Code:
%  FLOW NUMERICAL METHOD DEFINITION % % % Convective numerical method (JST, LAXFRIEDRICH, CUSP, ROE, AUSM, HLLC, % TURKEL_PREC, MSW) CONV_NUM_METHOD_FLOW= TURKEL_PREC % % Spatial numerical order integration (1ST_ORDER, 2ND_ORDER, 2ND_ORDER_LIMITER) SPATIAL_ORDER_FLOW= 2ND_ORDER_LIMITER If I change to a first order spatial integration Code:
%  FLOW NUMERICAL METHOD DEFINITION % % % Convective numerical method (JST, LAXFRIEDRICH, CUSP, ROE, AUSM, HLLC, % TURKEL_PREC, MSW) CONV_NUM_METHOD_FLOW= TURKEL_PREC % % Spatial numerical order integration (1ST_ORDER, 2ND_ORDER, 2ND_ORDER_LIMITER) SPATIAL_ORDER_FLOW= 1ST_ORDER Any hint here? Thank you, Zeno 

May 1, 2019, 23:47 

#5 
New Member
Muhammad Hilmi Al Fatih
Join Date: Apr 2019
Posts: 10
Rep Power: 5 
Hey Zeno!
I recently tried to run the low mach number case using a compressible solver too. I am a bit frustrated since it did take very long time to converge. I wonder if you could give any further results on your trial on that problem? Did you finally succeed to get a converged solution using turkelpreconditioning scheme? Also, could you please further elaborate how did you play with the slope limiting coefficient and how was the effect of it? Looking forward to your answer. Cheers, Hilmi 

May 2, 2019, 05:02 

#6 
Member
Zeno
Join Date: Sep 2013
Location: Delft, The Netherlands
Posts: 63
Rep Power: 10 
HI Muhammad,
please note that this is a 5year old thread and SU2 has changed a lot in this time. I haven't been running incompressible simulations in a while so I don't know how SU2 is handling these cases at the moment. By looking at the config files of a simulation similar to the one above, it seemed that I settled for the following options (which may be specified with another syntax in the current version of SU2): Code:
% Reference element length for computing the slope and sharp edges limiters. REF_ELEM_LENGTH= 0.1 % % Coefficient for the limiter LIMITER_COEFF= 0.1 % % Linear solver for implicit formulations (BCGSTAB, FGMRES) LINEAR_SOLVER= FGMRES % % Preconditioner of the Krylov linear solver (JACOBI, LINELET, LU_SGS) LINEAR_SOLVER_PREC= LU_SGS % % Minimum error of the linear solver for implicit formulations LINEAR_SOLVER_ERROR= 1E4 % % Max number of iterations of the linear solver for the implicit formulation LINEAR_SOLVER_ITER= 5 % % Convective numerical method (JST, LAXFRIEDRICH, CUSP, ROE, AUSM, HLLC, % TURKEL_PREC, MSW) CONV_NUM_METHOD_FLOW= ROE % % Spatial numerical order integration (1ST_ORDER, 2ND_ORDER, 2ND_ORDER_LIMITER) % SPATIAL_ORDER_FLOW= 2ND_ORDER_LIMITER % % Slope limiter (VENKATAKRISHNAN, MINMOD) SLOPE_LIMITER_FLOW= VENKATAKRISHNAN Have a look at Venkatakrishnan's paper if you want to understand how to specify limiters. Best, Zeno 

May 3, 2019, 04:51 

#7 
New Member
Muhammad Hilmi Al Fatih
Join Date: Apr 2019
Posts: 10
Rep Power: 5 
Dea Zeno,
Thank you for your prompt reply. It is really helpful. Regards, Hilmi 

Tags 
compressible, convergence criteria, incompressible, junction flows 
Thread Tools  Search this Thread 
Display Modes  


Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
OpenFOAM Low Mach Number Formulation for Combustion  tmu  OpenFOAM  13  December 13, 2017 11:12 
foamextend_3.1 decompose and pyfoam warning  shipman  OpenFOAM  3  July 24, 2014 08:14 
Low Mach Number for Combustion  sahm  OpenFOAM  10  October 11, 2013 20:43 
TVD scheme at low Mach number  Axel Rohde  Main CFD Forum  5  August 6, 1999 02:01 
how calculate the density in low mach number?  Juhee Lee  Main CFD Forum  1  July 31, 1999 15:26 