I was just coming to post an identical question, (in fact I inadvertently posted it under the solidworks forum not realizing there was a separate one for COSMOS, mea culpa.)
Here is the text of my post:
I've been wrestling with a model that gets about 90% finished solving when I get an "internal error #10" and COSMOS fails.
Running SW2005/COSMOSExpress 2005 with all the latest service packs on a Dell Precision 650 with 4 Gb of memory (with 3GB switch enabled) under XP Pro SP2.
Can anyone tell me what that error means? My VAR is either stumped or won't tell me.
"The FFE solver issues this message when it is not converging after 5000 iterations. This means that the problem is hard to converge. Many causes can lead to this situation, including but not limited to: Large difference between Young's modulus of the materials, portions of the model with very low stiffness compared to other portions, etc...
You may want to use a different solver like the FFEPlus solver, or the Sparse solver. The sparse solver being a direct solver, convergence problems are irrelevant. But the required memory and time are much higher."