Why won't my shear-loaded orthotropic material buckle?


Setting BUCOPT,LANB,5,100,10000 may correct the problem. It is documented that if you have trouble in a buckling analysis, you should try inputting reasonable values for the SHIFT/LDMULTE value on the BUCOPT command

There has been an improvement in the solver for this case at 12.0. The Block Lanczos routine needs to find a starting point (starting shift value), and when the user does not input anything, then the solver uses a theoretical lower-bound estimate based on the diagonals of K and M (Ks). In other words, it tries to use a first shift that is just below the first eigenvalue. However, this estimate assumes that the matrices are diagonally dominant. Normally, in FEA this is the case, but we have learned in the last few years that certain beam/shell elements with no axial stress in the prestress analysis may NOT be diagonally dominant. There may be other cases of this as well. When this is the case, Lanczos gets fooled into choosing a first shift that is very, very high. We have fixed the logic in 12.0 to guard against this case and pick a better initial shift.

So in 12.0, the solver is able to find the first five eigenvalues without inputting a SHIFT/LDMULTE value on the BUCOPT command. At 11.0, you need to simply input a SHIFT value near (within a few orders of magnitude) the lowest expected eigenvalue.





Show Form
No comments yet. Be the first to add a comment!