eigenvalue analysis

[Pa.si]

I was doing simulations, between SAP2000 and SeismoStruct. I noticed that the 2 softwares present a minimum error in the resolution, of the eigenvalues ​​in the case of solid sections, while in the case of tubular or double-T sections the evaluation error between the two models is about 5%. Can you tell me why?

[huffte]

The number of possible reasons for discrepancies is too large to detail here. I would suggest that you double-check to ensure that both programs are indeed modeling the exact same structure in terms of:

(1) mass source
(2) element densities
(3) inclusion or exclusion of shear deformation in stiffness formulation
(4) element properties
(5) solution method and tolerances
(6) rigid end joint specification
(7) lumped versus consistent mass formulation of the mass matrix

There are others. but this would be a good start.