I am trying to reproduce permeation data from an experiment (reference: https://doi.org/10.1016/j.fusengdes.2008.05.016) for validation purposes. The problem can be reconducted to a two-layers composite solid structure, where on one side a tritium partial pressure is imposed and on the other side the concentration is null. The two layers one models a nickel plate, where Sievert's law applied, and the other a layer of stationary molten FLiBe, where Henry's law applies.
My issue is with the interface BC between the two regions. I want to use the recently updated InterfaceSoprtion BC but I get a segmentation fault on the first timestep of the simulation. It probably is a machinie-specific bug since the same input runs fine on another pc. It shouldn't be an environment problem since the problem persists even after conda and TMAP( have been reinstalled.
This is the trace I get on MPI_Abort, running in debug mode:
Thread 1 "tmap8-dbg" hit Breakpoint 1, 0x00007fffdf53e4f0 in PMPI_Abort () from /home/federico/miniforge/envs/moose/lib/libmpi.so.12
(gdb) bt
#0 0x00007fffdf53e4f0 in PMPI_Abort () from /home/federico/miniforge/envs/moose/lib/libmpi.so.12
#1 0x00007ffff5fec3bc in moose::internal::mooseErrorRaw (msg=..., prefix=...)
at /home/federico/projects/TMAP8/moose/framework/src/base/MooseError.C:88
#2 0x00007ffff5fead92 in MooseBase::callMooseError (this=0x555555f79e90, msg=..., with_prefix=true)
at /home/federico/projects/TMAP8/moose/framework/src/base/MooseBase.C:39
#3 0x00007ffff4a3f5a7 in MooseBaseErrorInterface::mooseError<char const (&) [61]> (this=0x555555f79f10)
at /home/federico/projects/TMAP8/moose/framework/build/header_symlinks/MooseBaseErrorInterface.h:33
#4 0x00007ffff5f3edb0 in IterationAdaptiveDT::computeFailedDT (this=0x555555f79e90)
at /home/federico/projects/TMAP8/moose/framework/src/timesteppers/IterationAdaptiveDT.C:328
#5 0x00007ffff5f45680 in TimeStepper::computeStep (this=0x555555f79e90)
at /home/federico/projects/TMAP8/moose/framework/src/timesteppers/TimeStepper.C:87
#6 0x00007ffff5be9040 in Transient::computeDT (this=0x555555ed27a0)
at /home/federico/projects/TMAP8/moose/framework/src/executioners/Transient.C:337
#7 0x00007ffff5be81dd in Transient::execute (this=0x555555ed27a0)
at /home/federico/projects/TMAP8/moose/framework/src/executioners/Transient.C:290
#8 0x00007ffff5fd712e in MooseApp::executeExecutioner (this=0x555555907590)
at /home/federico/projects/TMAP8/moose/framework/src/base/MooseApp.C:1172
#9 0x00007ffff5fde4f7 in MooseApp::run (this=0x555555907590)
at /home/federico/projects/TMAP8/moose/framework/src/base/MooseApp.C:1554
#10 0x0000555555558e7e in Moose::main<TMAP8TestApp> (argc=3, argv=0x7fffffffcc88)
at /home/federico/projects/TMAP8/moose/framework/build/header_symlinks/MooseMain.h:47
#11 0x00005555555586e6 in main (argc=3, argv=0x7fffffffcc88) at /home/federico/projects/TMAP8/src/main.C:15
(gdb)
I tested the same input swapping the BC with the ADPenaltyInterfaceDiffusion BC and the case runs.
There might be a better way to model the BC, any idea is appreciated
Thanks!
Steps to Reproduce
What I did was simply a fresh TMAP8 installation and I run the case n°1 below. It doesn't run on my machine. Then I swapped the BC in the case n°2 and it runs.
Impact
This issue is preventing me from proposing two validation cases for TMAP8 on the V&V list.
Bug Description
Hello!
I am trying to reproduce permeation data from an experiment (reference: https://doi.org/10.1016/j.fusengdes.2008.05.016) for validation purposes. The problem can be reconducted to a two-layers composite solid structure, where on one side a tritium partial pressure is imposed and on the other side the concentration is null. The two layers one models a nickel plate, where Sievert's law applied, and the other a layer of stationary molten FLiBe, where Henry's law applies.
My issue is with the interface BC between the two regions. I want to use the recently updated
InterfaceSoprtion
BC but I get a segmentation fault on the first timestep of the simulation. It probably is a machinie-specific bug since the same input runs fine on another pc. It shouldn't be an environment problem since the problem persists even after conda and TMAP( have been reinstalled.This is the trace I get on MPI_Abort, running in debug mode:
I tested the same input swapping the BC with the
ADPenaltyInterfaceDiffusion
BC and the case runs. There might be a better way to model the BC, any idea is appreciatedThanks!
Steps to Reproduce
What I did was simply a fresh TMAP8 installation and I run the case n°1 below. It doesn't run on my machine. Then I swapped the BC in the case n°2 and it runs.
Impact
This issue is preventing me from proposing two validation cases for TMAP8 on the V&V list.
Case n°1 - InterfaceSorption BC
Case n°2 - ADPenaltyInterfaceDiffusion BC