IcoFsiFoam simulation crashes when using a smaller timestep

mathieu · January 6, 2009, 19:24

Hi,

I tried a FSI case (the one available at http://powerlab.fsb.hr/ped/kturbo/OpenFOAM/run/flappingConsoleSmall_HJ_21Mar2008 .tgz) with icoFsiFoam on OpenFOAM 1.4.1-dev. The case worked fine but if I set the timestep to 0.0001 instead of 0.0003, the simulation crashes. I tried many things but I was not able to solve this problem yet.

I am currently working on an "improved" version of icoFsiFoam that handles strongly coupled cases and that problem is still very annoying (for example, I can't do timestep independance tests since I can't reduce the timestep). Any help would be greatly appreciated.

Mathieu

joris · May 17, 2009, 04:54

Hello,

I have studied the effect of the time step on the stability of partitioned FSI simulations. If the fluid/solid density ratio is high (e.g. if you neglect the mass of the structure), it is normal that the convergence becomes more difficult if the time step is reduced. May I suggest a paper of myself on this subject: http://dx.doi.org/10.1016/j.compstruc.2008.05.005

Best regards,
Joris

January 6, 2009, 19:24	Hi, I tried a FSI case (the	#1
mathieu Member Mathieu Olivier Join Date: Mar 2009 Location: Quebec City, Canada Posts: 77 Rep Power: 17	Hi, I tried a FSI case (the one available at http://powerlab.fsb.hr/ped/kturbo/OpenFOAM/run/flappingConsoleSmall_HJ_21Mar2008 .tgz) with icoFsiFoam on OpenFOAM 1.4.1-dev. The case worked fine but if I set the timestep to 0.0001 instead of 0.0003, the simulation crashes. I tried many things but I was not able to solve this problem yet. I am currently working on an "improved" version of icoFsiFoam that handles strongly coupled cases and that problem is still very annoying (for example, I can't do timestep independance tests since I can't reduce the timestep). Any help would be greatly appreciated. Mathieu

May 17, 2009, 04:54	Effect of the time step on stability	#2
joris New Member Join Date: May 2009 Posts: 3 Rep Power: 17	Hello, I have studied the effect of the time step on the stability of partitioned FSI simulations. If the fluid/solid density ratio is high (e.g. if you neglect the mass of the structure), it is normal that the convergence becomes more difficult if the time step is reduced. May I suggest a paper of myself on this subject: http://dx.doi.org/10.1016/j.compstruc.2008.05.005 Best regards, Joris

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Timestep for transient simulation	houman	CFX	2	April 28, 2005 04:44
TIMESTEP PROBLEM IN TRANSIENT SIMULATION	George	CFX	4	September 9, 2003 11:45
Timestep for Free Surface simulation	jh	CFX	0	February 13, 2003 00:55
timestep for combustion simulation	yangqing	FLUENT	0	August 18, 2002 09:46
Can y+ be much smaller than 1 for...	Rodrigo Madariaga	FLUENT	3	June 8, 2001 00:38