Using a constant time-step in MES (Algor V.16)

[jdoggk40]

Hey everyone,

I'm back again with another question regarding a hip impact simulation I am running.

I have what I believe is a pretty accurate simulation running, but the results are not correct. In Algor's results window I am comparing the stresses in the neck of the femur between different simulations (different hip protectors and one without any protection). I want to make sure the analysis runs all the way through so that I can get the max stress at the top of the "bell curve".

I think that the problem is with the time step (which decreases as the results are converging). The simulation really slows down when the time-step changes, so I want to adjust this to a constant time-step. Clicking the checkbox for this option before analyses does not seem to work. The model verifys when run, but will not begin simulation. No error message comes up or anything telling me why it might not have run.

Does anyone know if there are any possible causes for the simulation not wanting to run with a constant time-step?

Thanks,
Jesse

 

[GBor]

Jesse,

Why do you think the results are not correct and how do you think the time-step ties into this error?

One thing you can do is let the time-step drop, but write out the results of all time-steps and see what is happening (where you think it is going wrong).

As for the model not running from the initial analysis start, if you are trying to run a constant time-step, but the simulation can not converge at that first interval, I suspect it will not run. Beyond that, you may have to speak to a developer and they may suggest that you update your software.

Garland

 

[jdoggk40]

Thanks for your response Garland.

When the time step decreases with impact, the simulation gets to a point where it seems to almost stall. At times, if it gets too slow it will give an error for "overflow". I still get results but not all the way through the impact peak (max stress). I need this peak to compare results w/ and w/o the protector, so I thought that if I ran it with a constant time-step, I would be able to run the simulation through impact and through the peak.

I'll try different time steps and hopefully the problem was not being able to converge on the first step.

Thanks,
Jesse