PIMPLE not converging after simulation restart

pawlo · December 4, 2017, 07:45

Hello!

I am doing a simulation with rhoReactingFoam in OpenFOAM v2.3.0. I am also using swak4Foam to define some of my boundary conditions. The case is described in another thread i posted a few days ago:

Is it possible to define a reference pressure cell somewhere in the internalField?

My problem is that the simulation was working very good and i decided to stop it for some intermediate evaluation. Now i tried to restart it from the latestTime (and also a few checkpoints before) but suddenly PIMPLE is not converging anymore. Before i stopped the simulation, PIMPLE needed approx 6 outer loops to converge. Now it reaches 50 (this is the maximum number i defined) iterations without convergence. I did not change anything in my controlDict or other dict files.

Here is an example:

I have the checkpoints at times 10.226 s and 10.227 s with a time step of 0.0005 s. Before i stopped the simulation it took about 30 real time seconds to reach the state at 10.227 from 10.226 within 2*6 outer PIMPLE loops. Now, when i restart from 10.226 it does not converge, PIMPLE reachtes 50 iterations and steps forward to the next point in time...

Does someone know why this is happening?

alexeym · December 4, 2017, 17:08

Hi,

From both descriptions it seems your "swak4Foam routine" is not quite happy about restart. Since there is not much details in both of your messages, it is impossible to find the reason for divergence.

In fact, not many people use 2.3.0, even less use swak4Foam, and even less would like to guess your boundary conditions and swak4Foam magic used. So finally, number of people, able to answer you question in its current form, is around 0.

December 4, 2017, 07:45	PIMPLE not converging after simulation restart	#1
pawlo New Member Join Date: Oct 2017 Posts: 8 Rep Power: 9	Hello! I am doing a simulation with rhoReactingFoam in OpenFOAM v2.3.0. I am also using swak4Foam to define some of my boundary conditions. The case is described in another thread i posted a few days ago: Is it possible to define a reference pressure cell somewhere in the internalField? My problem is that the simulation was working very good and i decided to stop it for some intermediate evaluation. Now i tried to restart it from the latestTime (and also a few checkpoints before) but suddenly PIMPLE is not converging anymore. Before i stopped the simulation, PIMPLE needed approx 6 outer loops to converge. Now it reaches 50 (this is the maximum number i defined) iterations without convergence. I did not change anything in my controlDict or other dict files. Here is an example: I have the checkpoints at times 10.226 s and 10.227 s with a time step of 0.0005 s. Before i stopped the simulation it took about 30 real time seconds to reach the state at 10.227 from 10.226 within 2*6 outer PIMPLE loops. Now, when i restart from 10.226 it does not converge, PIMPLE reachtes 50 iterations and steps forward to the next point in time... Does someone know why this is happening?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Simulation "Blows-Up" After Restart	fatirishman53	OpenFOAM Running, Solving & CFD	4	September 4, 2015 01:16
pimpleFoam Simulation aroud a cylinder does not converges	ASo	OpenFOAM Running, Solving & CFD	12	May 7, 2015 10:33
Restart simulation	Zerit	FLOW-3D	1	February 5, 2015 06:19
restart simulation	darookie	CFX	8	January 14, 2013 03:18
Error when restart simulation	zebu83	OpenFOAM	0	October 20, 2009 05:30

December 4, 2017, 17:08		#2
alexeym Senior Member Alexey Matveichev Join Date: Aug 2011 Location: Nancy, France Posts: 1,938 Rep Power: 39	Hi, From both descriptions it seems your "swak4Foam routine" is not quite happy about restart. Since there is not much details in both of your messages, it is impossible to find the reason for divergence. In fact, not many people use 2.3.0, even less use swak4Foam, and even less would like to guess your boundary conditions and swak4Foam magic used. So finally, number of people, able to answer you question in its current form, is around 0.