OS

[lbbmhh]

Hi,
I have got a new workstation and I tried to continue my routine simulations but surprisingly all my models crash!
I used the same version of the FEBio solver on my old workstation and the same simulation fully converges there, but on the new machine I get an error that says it finds a zero on a diagonal on a row that doesn't exist in my .feb file at all!!
they only thing which is different is that I use SUSE 11.3 on my old workstation and SUSE 12.1 on the new one coz with the old version my network adapters were not recognized.
could the OS actually effect the solver results?
I'm totally confused! please help
best
Reza

[dsrawlins]

Reza,

I also recently switched from OpenSuse 11.3 to 12.1, but haven't seen any problems. Could you attach your .feb file?

Thanks,

Dave

[lbbmhh]

Hi Dave,
I have tried a live KDE of SUSE11.3 on my new machine and my simulation converges. but when I boot from the hard disck i.e. SUSE 12.1 it crashes! please find the .feb file attached.
thanks for your help in advanced.
best
Reza

[ateshian]

Hi Reza,

I have tried running your model on a Mac and I don't get convergence either. I think the problem arises because of the way the problem is set up, not because of an error in the code or the version of the operating system.

You are performing a contact analysis under load control, even though the two contacting bodies are not initially overlapping. This is a common oversight with contact problem: First you need to push the contacting bodies together under displacement control (only a small amount is enough), so that the contacting surfaces detect each other. Then you can switch to load control. You can use a multi-step analysis to do that.

With only load control, at the first iteration there is no contact force opposing the applied load, since the contact surfaces are not overlapping initially (the contact force is evaluated from the amount of overlap multiplied by the penalty factor). Therefore, with only one force acting on it, the body on which the load is applied cannot satisfy static equilibrium and the analysis fails to converge.

Try it out and let us know if that works.

Best,

Gerard