Simplified Part - Assembly won't merge into one seamless part

[Marcel_]

Hello,

I am experiencing trouble with the simplify function of Inventor 2014. I would like to simplify this assembly into one seamless part for FEM calculation:

The assembly consists out of parts that, in reality, will be welded together. Due to their form, there is no way to place them exactly onto each other, therefore there are slight gaps between the parts:

Now when I try to merge the assembly into one part via "simplify" --> "create simplified part" and choose "single solid body with seams between planar faces merged", Inventor does not fill these gaps (even though it displays only one part in the bottom right corner of my window). Unfortunately, the remaining gaps brings "contact" problems in the FEM simulation.

Do you have advice on how to close the gaps and really merge the assembly into one part?

thanks in advance,

Marcel

[JDMather]

Hard to tell from an image, but it looks like it could be as simple as adding a Thicken feature along that face to merge the disjointed solid bodies.

[nkeeley]

My advice is to redraw as a single solid. It's going to give you the most stable FEA. Simplify is a great tool, but I think what you're trying to do is a little outside it's scope.

[LT.Rusty]

Rather than going through the trouble you went through here, why wouldn't you just leave it as an assembly and add in the welds that you need, and run your analysis on the part as it will actually be built? 

[Marcel_]

Hello,

thank you very much for your advice. I have gone with the option to actually add weldseams. Do the Inventor weldseams give realistic results in the FEM Analysis?

I have not been able to try for myself because I am still experiencing some other problems, which I have to weed out as well.

Marcel

[LT.Rusty]

I do a lot of structural steel stuff which has to go to an independent testing facility for actual load testing before we can advertise the product for a specific ANSI-specified load rating.  On a typical part with a span of ~40 inches or so, and several hundred individual weld beads, I find that the variations between the predictions from the FEA and the real-world testing results are generally within about .01" or so on displacement.  This is pretty good, considering the margin of error of the fabricators, who are doing it all by hand and haven't exactly heard of MIL-TFD-41C. 

[Marcel_]

That is very interesting, thank you.