Force Spectral Density from Random Vibration

[Sanjeev.mehta]

We are trying to get Force Spectral Density (in N or lbf) against frequency results from Random Vibration analysis. Can this be achieved using APDL? Would you be able to guide us with a script or example to get these results?

[dlooman]

Would this be the same as a force Response PSD?  The units of that are N^2/hz.  A characteristic of a force response psd is that if you integrate it and take the square root of the integral you get the 1 sigma value of the force associated with the input psd.   In the APDL postprocessor POST26 any quantity can be used to create a response psd.  So for example, the commands below perform a simple psd analysis and then produce a response psd for the FY reaction force.  If you are creating a commands object in Workbench Mechanical you would only need to use the commands starting with /post26.

/prep7

et,1,188

sectype,1,beam,rect

secdata,1,1

mp,ex,1,3e7

mp,nuxy,1,0.3

mp,dens,1,0.000754

n,1

n,2,40

e,1,2

d,1,all

fini

 

/solu

antype,modal

modop,lanb,2

mxpand,2,,,yes

solve

fini

 

/solu

antype,spectrum

spopt,psd,,yes

psdunit,1,acel

psdfrq,1,1,10,100

psdval,1,1000,1000  

psdres,disp,rel

d,1,uy,1.0

dmprat,.05

pfact,1

psdcom

solve

fini

 

/post26

store,psd,10

rfor,2,1,F,Y

rpsd,3,2 

/show,png

plvar,3

 

[Sanjeev.mehta]

Thanks for the response. We are trying to get the response in N/Hz units, which might be reaction force from nodes as function of frequency response.

[dlooman]

Can you provide a reference that explains the meaning of a quantity like N/Hz?   Is it like an FFT result where you get the force content in bins of a certain frequency width?

[Sanjeev.mehta]

Apologies, I take that back. We are looking for N^2/Hz. Thanks.

[dlooman]

OK, good!