To my limited understanding, FEA gives me part deformation under certain boundary conditions. So if I have the CAD model, I will know how much it will bend (for example) if I press my finger on it , or simply due to earth's gravity.
Now if I only know the shape of a part under gravity (and boundary conditions about how it is supported), can I find out the part shape in zero-gravity space? Is there a way to solve this kind of problem?
Hi Xiaobing Feng,
Simply reverse gravity direction and run the problem for deformed part. At least for the linear systems with small strains this will give almost undeformed shape.
Thanks Victor for the answer. Please allow me to clarify my previous question.
I put a part on a 3-point support and measured its shape. I'd like to know how much deformation (or reverse deformation) will occur if it's now floating freely in space.
I'm not sure if simply reversing gravity would work well enough, given there are also three support points. Is there a typical process to solve this kind of problem?
To me it feels like trying to solve the CAD geometry model from FEA results.
Hi Xiaobing Feng,
It is quite reasonable to expect that linear system with small strains returns to almost undeformed shape after applying reverse load to deformed shape without stresses. On the other hand, the problem you are interested in is a kind of inverse (and possibly ill-posed) problems and is connected with minimization of discrepancy between measured and calculated deformations. The most often least mean squares algorithms are used in such problems. Maybe Tikhonov regularization will be needed. Anyway its not a simple problem.
- Badges
- Science topic