normal stress issue of random vibration

[zengmin]

Hello guys. I built a space frame structure model with beam 188 element and run random vibration analysis on it. But the normal stress distribution seems to be wrong. Can u help me find the possible reason ? 

I choose 9 bottom point as fixed support and apply displacement psd on Zdirection. 

[dlooman]

Beam postprocessing, especially after a psd analysis, can be difficult.  As a sanity check, could you try displaying the axial stress with the user defined result below?: 

[zengmin]

Thanks for your reply. The result is below:

[zengmin]

I also added distributed mass and made prestress for modal analysis.But it doesnt effect the normal stress distribution. Here is all my keyopt. They are all default.

[zengmin]

My beam sections are several kinds of I section with different parameters and tubular section with different parameters.

[zengmin]

I form a part with all line body in designmodeler and chose edge joints as shared topology method

[dlooman]

Prestress doesn't always change beam results.  There is a nicer capability for displaying beam results on the cross-section in APDL:

/post1 

set,last

/graph,power

/esha,1 

/show,png 

plnsol,s,x        ! beam axial stress will be displayed over a 3D image of the beam