Problem diverging when using non-uniform grid
 

Hi,

I am running a hypersonic external flow ramp problem.

It works fine with uniform grid.

To resolve the wall boundary layer better, I need to use non-uniform grid. However, my solution kept diverging, at around 8.e-6, shortly after it starts running

Comparing between the uniform and non-uniform output, it has problems solving for "e":

[5] #0 Foam::error::printStack(Foam::Ostream&) at ??:?
[5] #1 Foam::sigFpe::sigHandler(int) at ??:?
[5] #2 ? in "/lib64/libc.so.6"
[5] #3 Foam::hePsiThermo<Foam::psiThermo, Foam::pureMixture<Foam::sutherlandTransport<Foam:: species::thermo<Foam::hConstThermo<Foam::perfectGa s<Foam::specie> >, Foam::sensibleInternalEnergy> > > >::calculate() at ??:?
[5] #4 Foam::hePsiThermo<Foam::psiThermo, Foam::pureMixture<Foam::sutherlandTransport<Foam:: species::thermo<Foam::hConstThermo<Foam::perfectGa s<Foam::specie> >, Foam::sensibleInternalEnergy> > > >::correct() at ??:?
[5] #5 ? at ??:?
[5] #6 __libc_start_main in "/lib64/libc.so.6"
[5] #7 ? at ??:?
[vis01:21813:0] Caught signal 8 (Floating point exception)
==== backtrace ====
2 0x000000000005a42c mxm_handle_error() /var/tmp/OFED_topdir/BUILD/mxm-3.5.3092/src/mxm/util/debug/debug.c:641
3 0x000000000005a59c mxm_error_signal_handler() /var/tmp/OFED_topdir/BUILD/mxm-3.5.3092/src/mxm/util/debug/debug.c:616
4 0x0000000000032510 killpg() ??:0
5 0x0000000000032495 raise() ??:0
6 0x0000000000032510 killpg() ??:0
7 0x000000000018c7da _ZN4Foam11hePsiThermoINS_9psiThermoENS_11pureMixtu reINS_19sutherlandTransportINS_7species6thermoINS_ 12hConstThermoINS_10perfectGasINS_6specieEEEEENS_2 2sensibleInternalEnergyEEEEEEEE9calculateEv() ??:0
8 0x00000000002277c5 _ZN4Foam11hePsiThermoINS_9psiThermoENS_11pureMixtu reINS_19sutherlandTransportINS_7species6thermoINS_ 12hConstThermoINS_10perfectGasINS_6specieEEEEENS_2 2sensibleInternalEnergyEEEEEEEE7correctEv() ??:0
9 0x0000000000427fe7 main() ??:0
10 0x000000000001ed1d __libc_start_main() ??:0
11 0x0000000000424059 _start() ??:0
===================

Why is this so? I checked both solution just before the divergence but they look very similar.

I have changed the fvSchemes and fvSolution but it doesn't help. I don't think it's the boundary condition. Interestingly, changing the initial conditions makes little difference at all.

Any recommendation? I am thinking of using mapField to copy at output at maybe t = 0.0001 (10% of targeted endTime) from the uniform to non-uniform grid, and continue from then on. Wonder if it's a good way...