Positive real parts found in eigenvalues of structural state-space FEM model

[Lysandros]

Hello together,

I have modelled the stiffness K, damping D and mass M matrices of a 500-D.O.F. structure in Matlab and created a state-space dynamical system (x'=Ax+Bu) from these in order to carry out simulations in the time domain. However, I notice that the highest eigenvalues of the system have a very small (1e-12), but positive real part. Only the 8 highest eigenvalues have this issue, all the other 492 have a negative real part. This leads me to believe that the model might lack stability, which cannot be because it the elastic structure is fixated.
The stiffness and damping matrices are positive semi-definite and possess only positive eigenvalues.

Has anyone experienced similar issues?

Thanks




Kind regards
Lysandros

 

[karachun]

Look like model is poorly constrained and have coincident nodes.