I thought I'd pick up on some of the discussion happening last week around FEA Al Dean and Deelip Meneze had a couple posts about the barriers to entry for this type of analysis. Kenneth Wong and Derrek Cooper did a couple post that talked about Solid Edge Simulation. With Inventor 2011 Simulation launching recently, we will likely see more on this topic. (BTW, the links above lead to all the articles).
The picture above is of Pascal Devantine's automobile window as we drove around France visiting Simulation customers (yes, I'm still here - can you say "Eyjafjallajökull"). Because the crack creates a stress concentration, it continues to grow across his window every time there is a bump or some vibration. Most of us have seen this before in windows or sidewalks.
One of the points being debated is exactly where FEA fits in for the designer. I think this is far from a black and white issue so where you stand often depends on the type of problem and the level of the designer. I’d visit one of the post above if you want to join into the dabate. For me, I thought I’d throw out a more concrete example.
Consider the simple bar below which is constrained on the left side and has a force of 1000 lbfs on the right side. PMI is in inches. You might think the maximum stress is at the thinnest part and would be 1 ksi (1000lbf/1in2).
However, because of the sharp corner, the maximum stress is more than twice this. If you are a mechanical engineer, you know there are tables you can use to look up the stress concentration around this corner.
While most of us are aware of the problems with stress concentrations (often from personal experience when cracks in your sidewalk or window seem to keep growing), the the tables or FEA results allow us to understand how much bigger this stress will be and how we can reduce it. Clearly, from the table, a larger radius fillet is a good idea. Less intuitive is that reducing the height of the thicker part (H) will also reduce the stress.
The advantage of tools like Solid Edge Simulation is that, unlike the table I used above, FEA can work with all types of geometry. We could add additional forces or other changes in the geometry to determine how this affects the maximum stress of the part. These types of problems are also simple enough that non-FEA experts can do them with little training.
Why not “no training at all”? I’ll talk about that in my next post.