Rayleigh Damping Question

[Bruce Maison]

I am a new FEAP user and am having difficulty using Rayleigh damping.  See results in attached files.  Please indicate what I am doing wrong in the analysis command sequence.
I plan to do nonlinear static and dynamic earthquake analysis with more complex models, and the command sequence shown in the attachments is what I plan to use.
Thank-you
Bruce Maison

[Prof. R.L. Taylor]

I do not see any non-zero density.  What does your output file report as the material data?  Are there any differences in the results based on what you specify?  Probably running with Newmark method first will be helpful since it does not have any numerical damping for the default parameters.  THen you can tell what the effect of your different inputs will be.  I do see quite a bit of damping if I add/remove the rayleigh damping line from IT7 (after adding the boundary conditions back in and using Newmark for the transient solution).

[Bruce Maison]

The model has zero material density and the output file prints a warning note.  The model has point masses.
It appears that the Rayleigh damping command does not use the stiffness proportional coefficient (a1).  See the attached file DampNote.doc.
File IT11 uses mass proportional damping and generates the correct results
File IT12 uses stiffness proportional damping and generate incorrect results of zero damping.
Is this a bug in FEAP or am I not using command properly?
Thanks
Bruce Maison     

[Prof. R.L. Taylor]

Can you please send the const.TH file too -- maybe it is not needed but just to be safe!

[Bruce Maison]

Per your request, attached is the file. 
Thanks again for you help with the damping bug.
Bruce Maison

[Prof. R.L. Taylor]

Rayleigh damping is not implemented for all the elements.  Since you asked for mass damping and had only a lumped mass that works fine.  However, the default element type does not have the damping implemented.  If you add to your material record: SHEAR OFF you will get the standard bending element using Hermite cubic interpolation for the displacements -- which has no shear deformation.

If you look at the coding in this element (frams3d.f) you should be able to modify the other elements should you need one with shear deformation.

We suggest also that you use multiple batch-end sets in your data file to help separate the information that follows the 'end'.  Also, it was necessary to remove the 'tol,,1.e-02' record since no iterations were done with it in.  A modified IT12 file is attached.

With the changes the damping is the nearly the same for both mass and stiffness damping.

[Bruce Maison]

Thank-you. The damping is now working properly per your directions.
We have questions about which FINITE FRAME element formulation is best for our analysis problem.  Please see the question titled "Frame Elements with FINITE displacements"
We appreciate your comments
Bruce Maison