convergence problem in simulating self-heating ignition (Version 0.8186)

[Han Yuan]

Dear all,

I am doing a self-heating simulation using Gpyro V0.8186. The model is 1D for classic hot plate experiment configuration (fixed temperature and impermeable boundary condition at the bottom and free surface at the top). It is found that the case can run to simulate the onset of self-heating ignition and initial spread of smoldering. But it gets stuck and doesn't converge once the sample self-ignites and the temperature starts to increase quickly. When it gets stuck, the information provided by Gpyro is "Gas Yj not converged" (see the attached Figure).

I have tried to change the convergence tolerance and relaxation parameter for YJG, but this doesn't help.
I was wondering what is the underlying cause of the problem. It would be great if any of you can have a solution to this problem. 

The input file is attached.

Many Thanks!

Best regards,

Han

[Han Yuan]

Dear all,

I have tried to simulate the same problem (with the same tolerance of convergence) using version 0.700 of Gpyro. The case can run to the completion in Version 0.700.  It seems like for 1D simulation, Version 0.700 seems to be more robust than the Version 0.818x. I have talked to Xuanze and he had similar experiences before that Version 0.818x gets stuck for some cases (1D) that can run smoothly in Version 0.700.

Best regards,

Han  

在 2019年9月19日星期四 UTC+1上午11:19:50，Han Yuan写道：

[Rein, Guillermo]

Dear Gpyro users,

 

We think Han’s issue is related to the BC (when the smouldering front reaches the free surface). He is exploring this.

 

Any help would be greatly appreciated. We have been suffering with these instability for several years and my group does not seems how to remove it.

 

Cheers,

G.

 

Prof. Guillermo Rein

Imperial College London, SW7 2AZ
http://www.imperial.ac.uk/hazelab

--
You received this message because you are subscribed to the Google Groups "Imperial Gpyro Working Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to gpyro-working-g...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/gpyro-working-group/c00c2a1e-49fe-41e1-80cf-11aa127fd34c%40googlegroups.com.

[Han Yuan]

Dear all,

After investigation, I have found where the issue is. The information provided by Gpyro when it gets diverged is "Gas Yj max residual: 10". This is a debug flag imbedded in Gpyro, indicating the term-YJSUM is not converged ( not equal to 1). 

That is because in the version after V0.8172, Gpyro has a tighter YJSUM error checking tolerance(0.9999 <YJSUM < 1.0001) than previous versions  (0.9995 <YJSUM < 1.0005). If we still use previous YJSUM checking tolerance(0.9995 <YJSUM < 1.0005), we can get the case run to completion with no problem. 

Best regards,

Han

在 2019年9月27日星期五 UTC+1下午3:13:53，Han Yuan写道：

[Han Yuan]

To show the successful simulation, the animation of temperature change is attached here.

在 2019年9月19日星期四 UTC+1上午11:19:50，Han Yuan写道：

Dear all,

[Franklin Cândido Costa]

Dear All,

I am facing the same convergence problem using issue Gpyro V0.8186. 

I have changed the YJSUM parameter in gpyro_pyro.f90 to the recommended range 0.9995 <YJSUM < 1.0005 and recompiled the code again, but even though it did not converge. Also, I have tried to change the relaxation factor without success. 

So, Do I have to change this range in another file of Gpyro code? If Yes, which file do I have to modify?

Best Regards!

Thanks!