We are evaluating Inventor for possible use as we look to expand into new markets. As we are totally new to the product, we have been going through the tutorials on-line, which were apparently created for use with Inventor 2010 (?). We are using the demo of the 2013 product for evaluation.
Interestingly enough, some of the tutorials in the Stress Analysis section of the tutorials produce wildly different results when compared to the step-by-step instructions. Case in point is the "Caulk gun" tutorial, titled "Stress analysis contacts." This is a good example because the entire model is set up to run without any changes by the user. (Trying to eliminate the "idiot user" variable from the equation!)
When we run the simulation on this model without any change whatsoever, the software generates two errors: 1. It indicates that there are multiple independent objects in the model, and 2. it introduces a "Soft spring," ostensibly to allow the analysis to continue. The results, as might be expected, are wildly off - more than an order of magnitude. The correct answer is 291 MPa max stress in the tutorial script, vs. our 3158 MPa in the real simulation.
We have encountered other strange behavior, somewhat more subtle (and therefore more concerning). For instance, the "Robot Base" model ("FEA Assembly Optimization") does not seem to properly distribute load into the subframe beneath the upper plate. You can see this clearly by comparing the stresses, which seem hardly at all to manifest in the framework. Also, displacements are mostly in the plate itself but not in the framework.
Does anyone know what is causing this behavior? I might be able to discover a new modeling or contact requirement if I hunt for long enough, but I figure that others might want to explore these tutorials and may have the same issues. Hence I bring this all up here and would be grateful for your feedback.