Results not matching in structural analysis - For Full model analysis and a portion of the model.

[Uday029]

[/indent]I have a complicated FE Model of a box type condenser with large no of DOF for static structural analysis. It has been solved in NX Nastran with all the loads to get the displacements and rotations through out the model.Then, the required portion of the model(design area) has been cut to fecilitate more iterations on small portion of the model and the displacements and rotations at the respective node numbers near the cut have been applied. But with this approach,the stresses are 10 times more than the previous case(where the entire structure has been modeled). However, the displacements along with the behaviour is matching in both cases. The modulus of elasticity also seems to be OK. Where I could have went wrong?

 

[mrfemap]

Applying displacements to a cut mesh can often lead to stress peaks. The enforced displacements, if not truly smooth and continuous will introduce some significant stress risers. From the FEMAP side, we've recently added a Global/Local modeling tool that helps you bring forces from the larger model via a Free Body Diagram, and then RBE3 connecting them into the local model. The RBE3 approach takes care of the "smoothness" issue since it distributes the forces better.

 

[Uday029]

Thanks for your reply mrfemap. However, as per my observation, there are no peak stresses observed due to the enforced displacements applied on the cut edges. But as a whole, the stresses on the entire cut FE model are exactly 10 times the stresses reported at the same region during the analysis of full assembly. Also, the displacements are exactly matching for both the cut model and the full assembly.
Also,Can I have the Global/Local kind of tool in NX CAE?

 

[mrfemap]

I'm the development manager for FEMAP and can only control and comment on what goes into FEMAP. If you send me your two NASTRAN input files, myself, or one of my NASTRAN guys might be able to figure out what's causing the differences.

sherman dot mark at siemens dot com