Residual diverrgence / mesh quality

Andrea159357 · February 18, 2019, 05:33

Hello, i'm really struggling understanding how fluent works.
I'm running a thermal fluid problem with
different meshes but i cannot reach convergence.

The first mesh is obtain with cutcell and even though max skew is 0.992 and max asp Ratio is 67, i improved quality in fluent and i reach perfect convergence of FLUID equations solution and monotonous behaviour of residuals. Then i turn on ENERGY equation and everything blows up. I got message of " Limited temperature in xxx cell of 1 K " and " Limited temperature in xxx cell of 5000 K ".

I thought was for bad quality mesh so i remade it.
The second one is a normal all TET mesh + inflation layer.
This one has max skew 0.88 and max asp Ratio 123. (COOL i thought).
Though even the fluid solution can't reach convergence.

How is it possible?

sufjanst · February 18, 2019, 05:39

I had a similiar problem. For some reason, fluent sets the fluid temperature to 1K. Of course, the simulation will crash.

Fix: Just activate the energy equation and patch your flowfield to a temperature of 300K (or whatever your conditions are). That should do it.

Andrea159357 · February 18, 2019, 06:08

already done. I patched all domain to 1000 K (solution is around 1050 K ), first 30 iteration residual converge to e-9. Afterwards it starts growing unitil it explode.
I stop the solver to the min residual i got and solution is good. But i was wondering why after 30 iter blows up but the thing is why with a mesh better than the previuos one even the fluid can't reach convergence?
how fluent works?

sufjanst · February 18, 2019, 06:18

That's indeed pretty strange. Can you share some information on your problem and your case setup?

Andrea159357 · February 18, 2019, 06:22

YEAH. What do you need?
It is not a set up probelm since i already run a simulation with a slightly different geom and it worked perfectly.

MKuhn · February 18, 2019, 06:40

It is possible to use the case set up from the previous simulation? (read mesh --> replace mesh --> discard data). If divergence occurs again then it should a mesh problem. Stop the solver after the energy begins to grow up. Locate the cells where the temperature is out of your expected range. An inproper face match between to different cell regions could be the problem as well (where you assume a heat flux, but there is no heatflux possible because of inproper face match).

Andrea159357 · February 18, 2019, 07:03

I did like you said, thats why it is not a set up problem but is a mesh one. I even did what you said, i mark and adapt those cell but nothing ggod happened, so i remash.
The cells with high skew are not in a single region but scattered all around the geom (15/20 cell) and they influence the cells around.

With an improper face match U mean between cells or between domains, e.g. solid and fluid?

MKuhn · February 18, 2019, 07:07

Quote:

Originally Posted by Andrea159357

With an improper face match U mean between cells or between domains, e.g. solid and fluid?

I mean domains, where you have to connect different volumes in your CAD-Model.

Andrea159357 · February 18, 2019, 07:26

that's possible. I used share topology in Spaceclaim but maybe i generate connection using "find contact" in meshing that screw up everything.
Thanks a lot for your help, i check now.

February 18, 2019, 05:33	Residual diverrgence / mesh quality	#1
Andrea159357 Member Andrea Join Date: Mar 2018 Posts: 62 Rep Power: 8	Hello, i'm really struggling understanding how fluent works. I'm running a thermal fluid problem with different meshes but i cannot reach convergence. The first mesh is obtain with cutcell and even though max skew is 0.992 and max asp Ratio is 67, i improved quality in fluent and i reach perfect convergence of FLUID equations solution and monotonous behaviour of residuals. Then i turn on ENERGY equation and everything blows up. I got message of " Limited temperature in xxx cell of 1 K " and " Limited temperature in xxx cell of 5000 K ". I thought was for bad quality mesh so i remade it. The second one is a normal all TET mesh + inflation layer. This one has max skew 0.88 and max asp Ratio 123. (COOL i thought). Though even the fluid solution can't reach convergence. How is it possible?

February 18, 2019, 05:39		#2
sufjanst Member Join Date: Mar 2016 Posts: 73 Rep Power: 10	I had a similiar problem. For some reason, fluent sets the fluid temperature to 1K. Of course, the simulation will crash. Fix: Just activate the energy equation and patch your flowfield to a temperature of 300K (or whatever your conditions are). That should do it. Andrea159357 likes this.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
HeatSource BC to the whole region in chtMultiRegionHeater	xsa	OpenFOAM Running, Solving & CFD	3	November 7, 2016 06:07
Extrusion with OpenFoam problem No. Iterations 0	Lord Kelvin	OpenFOAM Running, Solving & CFD	8	March 28, 2016 12:08
Floating point exception error	lpz_michele	OpenFOAM Running, Solving & CFD	53	October 19, 2015 03:50
Micro Scale Pore, icoFoam	gooya_kabir	OpenFOAM Running, Solving & CFD	2	November 2, 2013 14:58
calculation stops after few time steps	sivakumar	OpenFOAM Running, Solving & CFD	7	March 17, 2013 07:37

February 18, 2019, 06:08		#3
Andrea159357 Member Andrea Join Date: Mar 2018 Posts: 62 Rep Power: 8	already done. I patched all domain to 1000 K (solution is around 1050 K ), first 30 iteration residual converge to e-9. Afterwards it starts growing unitil it explode. I stop the solver to the min residual i got and solution is good. But i was wondering why after 30 iter blows up but the thing is why with a mesh better than the previuos one even the fluid can't reach convergence? how fluent works?

February 18, 2019, 06:18		#4
sufjanst Member Join Date: Mar 2016 Posts: 73 Rep Power: 10	That's indeed pretty strange. Can you share some information on your problem and your case setup?

February 18, 2019, 06:22		#5
Andrea159357 Member Andrea Join Date: Mar 2018 Posts: 62 Rep Power: 8	YEAH. What do you need? It is not a set up probelm since i already run a simulation with a slightly different geom and it worked perfectly.

February 18, 2019, 06:40		#6
MKuhn Senior Member Moritz Kuhn Join Date: Apr 2010 Location: Germany, Dresden Posts: 218 Rep Power: 17	It is possible to use the case set up from the previous simulation? (read mesh --> replace mesh --> discard data). If divergence occurs again then it should a mesh problem. Stop the solver after the energy begins to grow up. Locate the cells where the temperature is out of your expected range. An inproper face match between to different cell regions could be the problem as well (where you assume a heat flux, but there is no heatflux possible because of inproper face match).

February 18, 2019, 07:03		#7
Andrea159357 Member Andrea Join Date: Mar 2018 Posts: 62 Rep Power: 8	I did like you said, thats why it is not a set up problem but is a mesh one. I even did what you said, i mark and adapt those cell but nothing ggod happened, so i remash. The cells with high skew are not in a single region but scattered all around the geom (15/20 cell) and they influence the cells around. With an improper face match U mean between cells or between domains, e.g. solid and fluid?

February 18, 2019, 07:26		#9
Andrea159357 Member Andrea Join Date: Mar 2018 Posts: 62 Rep Power: 8	that's possible. I used share topology in Spaceclaim but maybe i generate connection using "find contact" in meshing that screw up everything. Thanks a lot for your help, i check now.