For a coupled aero-structural analysis I'm working on, I have to extract the eigenvectors/mode shapes of my structure, so that I can run aerodynamic calculations on them. I have a parameter (in this case span length) that I vary, performing this analysis at each parameter step. A problem I have recently encountered is that the mode shapes are inconsistent throughout the range of the parameter (see figures below for examples). I've checked by manually "flipping" the extracted output in my code, and it has a very significant impact on the results.

I've tried each of the eigensolvers abaqus provides, using SIM, and displacement vs. mass normalizing the eigenvectors, but I can't seem to get any kind of consistency in the mode shapes. Because changing the span sometimes affects the order of the modes (and creates/eliminates certain shapes), it's also very difficult to just programmatically check whether the outputs have a minus sign or not. The main question is this: Has anyone encountered a problem with the Abaqus eigensolver giving them this kind of directionally inconsistent eigenvectors before? And, if so, what did you do to resolve it?

I'm also open to any suggestions for how to programmatically characterize the mode shapes to be able to detect whether or not a particular mode is flipped. Happy to provide more details as needed to diagnose the issue.

enter image description here (Some are the same) enter image description here (Some are flipped) enter image description here



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.