Monday, 24 May 2010

10 Reasons Finite Element Analysis Goes Wrong

Have you ever paid for someone to do some FEA for you and then your component still broke? Here are ten things that frequently go wrong with finite element analysis.

1. Incorrectly applied boundary conditions. Probably the most common reason for inaccurate FEA results.

2. Mesh not fine enough. A course mesh is quick and cheap but gives poor stress results.

3. Not knowing your material well enough. The behaviour of some materials require a different approach to post processing FEA results than others.

4. Applying linear analysis to non-linear FEA problems. Sometimes linear analysis just won't do.

5. Over complicating things. Nonlinear analysis of large models is time consuming and expensive, can the model be simplified?
KISS (keep it simple stupid).

6. Not knowing the difference between real and spurious FEA results.

7. Not getting any backing from management. Use hand calculations to strengthen your argument.

8. Your boss finds a cheaper supplier. Will they deliver the goods at that price or is it too good to be true? Apples and pears?

9. You screw up the units.

10. You get the operating conditions wrong.