Damping, which stiffness matrix?

[mattfox]

Dear Seismosoft,

When the damping matrix is computed with a stiffness proportional component and geometric nonlinearities are included, which stiffness matrix is used? Is it the stiffness matrix including or excluding the geometric nonlinearities? My concern is that the latter may lead to negative damping forces.

My apologies if I have missed this somewhere in the SeismoStruct help.

Kind regards,

Matt Fox.

[seismosoft]

Matt Fox, hi.

You are correct, the complete stiffness matrix is used, which may indeed lead to negative damping forces.

This is an issue that we are currently investigating further, and we do hope to be able to propose some valid alternatives in the future.

Seismosoft Support

[jpacheco]

Hi Matt Fox,

Your concern is correct. Also (I wonder if that can help you in any way...), recall the occurrence of softening - i.e., in the components of the stiffness matrix related to the material behaviour - can lead to unrealistic negative damping forces (unrealistic energy input in the structure).

Regards,

João

[Stelios_Antoniou]

Negative damping forces may appear only when tangent, rather than initial, stiffness is employed and in the case of serious softening of the structure.
If you are concern about such behaviour simply revert to initial-stiffness-proportional damping.