This content has been marked as final. Show 2 replies
That is interesting, Steve. I don't believe the Simulation module calculates a relative strain energy when extracting mode shapes. I have not seen the technique you described before now. I believe you have to use the latter, i.e., decide on a "normalized" displacement field such as one from Frequency and use that for a static analysis to get a "normalized" strain field. Perhaps the normalization can be done by setting the maximum displacement in a particular mode shape to one, then run a static with non-zero displacements and show the relative strain energy. The issue becomes, I am unsure if other mode shape displacements are scaled relative to each other so you may not know how to scale non-zero displacement values for each mode shape. To be sure it may be best to use the Linear Dynamic module and just use a baseline white noise (constant) amplitude excitation across a bunch of frequencies. Then, for sure, the mode shapes will all be relative to each other, but at that point you could just plot strain energy, anyway.
Thanks Anthony for the reply. One clarification: I don't need to normalize across different mode shapes - I just need to ratio the total strain energy of one component to the total strain energy of the assembly for a particular mode shape. The ratio is a measure of the relative "work" being done by one component vs. another. The problem with using a static or even linear dynamic analysis is the inability to recreate a mode shape using an applied force/displacement or set of forces/displacements.
Is there a way to "copy" a displacement field for a set of nodes from the frequency analysis and apply it to a static analysis? I remember Ansys having that ability as a way to do a substructure analysis. You could define a boundary and extract the node displacements at that boundary for the big assembly solution. Then you could take the node displacements and apply them to a detailed analysis that only considered the components inside the boundary.