SimSolid Users

shawn_crocker

Anyone out there using simSolid?  We are currently evaluating it.  After a few days of use, this seems like an incredibly powerful tool.  The power I think it stemming directly from how fast it solves a study.  I am solving things in 10-20 seconds that would be 30mins-1 hour in other systems I have tried.  What I am curious of hearing about is if simSolid's whole "meshless" solver is truly accurate?  Of course, anything I have seen is either claiming perfection or claiming absolute fail but I would love to hear if people are actually using it and finding the possible trade off in accuracy for speed noticeable.

S1mon

How does it compare with Onshape's Simulation? I know it does more things, but for basic linear FEA, how is it? We did benchmark Onshape against work done by an external consultant using Ansys and were able to get very similar results (once all the loading was set up correctly).

With any simulation, you'll ultimately need to validate it with real world testing. Do you have other test results you can try to duplicate?

shawn_crocker

ATM We have some real world items with certain fatigue issues.  We are able to set up the study according to the real world and expose the real world issue with what seems like pretty close accuracy.  Comparing with onshape sim on the static level, I sort of feel like I'm seeing a bit better accuracy in onshape.  There are some areas that intuitively seem like very small amounts of stress should be showing up but in simSolid its not quite as clearly there.  In onshape the same area is clearly showing the stress.  In simsolid however, the sim takes about 15 seconds(on an ancient i5 processor), in onshape, the sim cannot fully converge in the one hour cutoff provided by the beta.  It gets about 1/2 way there, so without full convergence, its also sort of hard to compare side by side.

sebastian_glanzner

I have used SimSolid in the past, and it is great if you have a lot of parts in a big assembly!
The last time I used SimSolid was a few years ago, so some settings might have changed since then.

I made it a habit to import the geometry with a "fine" resolution. This helped to get the correct curvature into SimSolid and improved the stress accuracy.

I also activated "adapt to thin features" and "adapt to features". You can also refine the stresses display in the results view. 

There is also the possibility to apply local solution settings for groups of parts. 

shawn_crocker

Thanks @sebastian_glanzner
I have been getting results very similar to other mesh based systems by tweaking those settings as well.