CFD Online Discussion Forums - Status of continuous RANS adjoint

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)

- SU2 (https://www.cfd-online.com/Forums/su2/)

- - Status of continuous RANS adjoint (https://www.cfd-online.com/Forums/su2/115909-status-continuous-rans-adjoint.html)

Status of continuous RANS adjoint

Hello
Does version 2.0 have the continuous RANS adjoint with SA model ?

When I try a problem with FROZEN_VISC= YES, then the adjoint seems to run. But with FROZEN_VISC= NO, I get this error

Code:

------------------------------ Begin solver -----------------------------

Single iteration of the direct solver to store flow data.

[cfdlab:07930] *** Process received signal ***

[cfdlab:07930] Signal: Segmentation fault (11)

[cfdlab:07930] Signal code: Address not mapped (1)

[cfdlab:07930] Failing at address: 0x8

[cfdlab:07930] [ 0] /lib64/libpthread.so.0(+0xfab0) [0x7f76f14d0ab0]

[cfdlab:07930] [ 1] SU2_CFD(_ZN11CNSVariable12SetVorticityEv+0xf) [0x6cee3f]

[cfdlab:07930] [ 2]                                                                                  SU2_CFD(_ZN11CNSSolution13PreprocessingEP9CGeometryPP9CSolutionPP9CNumericsP7CConfigt+0x22a)         [0x61a8aa]

[cfdlab:07930] [ 3]                                                                                  SU2_CFD(_ZN21CMultiGridIntegration13FAS_MultigridEPPP9CGeometryPPPP9CSolutionPPPPP9CNumericsPP7CConfigtttmt+0x20e) [0x49a49e]

[cfdlab:07930] [ 4]                                                                                  SU2_CFD(_ZN21CMultiGridIntegration19SetMultiGrid_SolverEPPP9CGeometryPPPP9CSolutionPPPPP9CNumericsPP7CConfigtmt+0x258) [0x497de8]

[cfdlab:07930] [ 5]                                                                                  SU2_CFD(_Z20AdjMeanFlowIterationP7COutputPPP12CIntegrationPPP9CGeometryPPPP9CSolutionPPPPP9CNumericsPP7CConfigPP16CSurfaceMovementPP19CVolumetricMovementPPP14CFreeFormChunkm+0x551) [0x4abee1]

[cfdlab:07930] [ 6] SU2_CFD(main+0x10d1) [0x6bab41]

[cfdlab:07930] [ 7] /lib64/libc.so.6(__libc_start_main+0xec) [0x7f76f114b518]

[cfdlab:07930] [ 8] SU2_CFD() [0x45d7b9]

[cfdlab:07930] *** End of error message ***

Segmentation fault

Quote:

Originally Posted by praveen (Post 419354)

Hello
Does version 2.0 have the continuous RANS adjoint with SA model ?

When I try a problem with FROZEN_VISC= YES, then the adjoint seems to run. But with FROZEN_VISC= NO, I get this error

Code:

------------------------------ Begin solver -----------------------------

Single iteration of the direct solver to store flow data.

[cfdlab:07930] *** Process received signal ***

[cfdlab:07930] Signal: Segmentation fault (11)

[cfdlab:07930] Signal code: Address not mapped (1)

[cfdlab:07930] Failing at address: 0x8

[cfdlab:07930] [ 0] /lib64/libpthread.so.0(+0xfab0) [0x7f76f14d0ab0]

[cfdlab:07930] [ 1] SU2_CFD(_ZN11CNSVariable12SetVorticityEv+0xf) [0x6cee3f]

[cfdlab:07930] [ 2]                                                                                  SU2_CFD(_ZN11CNSSolution13PreprocessingEP9CGeometryPP9CSolutionPP9CNumericsP7CConfigt+0x22a)         [0x61a8aa]

[cfdlab:07930] [ 3]                                                                                  SU2_CFD(_ZN21CMultiGridIntegration13FAS_MultigridEPPP9CGeometryPPPP9CSolutionPPPPP9CNumericsPP7CConfigtttmt+0x20e) [0x49a49e]

[cfdlab:07930] [ 4]                                                                                  SU2_CFD(_ZN21CMultiGridIntegration19SetMultiGrid_SolverEPPP9CGeometryPPPP9CSolutionPPPPP9CNumericsPP7CConfigtmt+0x258) [0x497de8]

[cfdlab:07930] [ 5]                                                                                  SU2_CFD(_Z20AdjMeanFlowIterationP7COutputPPP12CIntegrationPPP9CGeometryPPPP9CSolutionPPPPP9CNumericsPP7CConfigPP16CSurfaceMovementPP19CVolumetricMovementPPP14CFreeFormChunkm+0x551) [0x4abee1]

[cfdlab:07930] [ 6] SU2_CFD(main+0x10d1) [0x6bab41]

[cfdlab:07930] [ 7] /lib64/libc.so.6(__libc_start_main+0xec) [0x7f76f114b518]

[cfdlab:07930] [ 8] SU2_CFD() [0x45d7b9]

[cfdlab:07930] *** End of error message ***

Segmentation fault

Hi Praveen,

We are currently in a development stage for the unfrozen continuous adjoint, but the frozen one is working properly in the latest version of the code (v2.0.2). We intend to release a version including unfrozen viscosity this summer.

Cheers,
Aniket