Stepped sine simulation on nonlinear component

[lawful]

Hello,

I have a structure with a lap joint which I tested at different forces and found being nonlinear (stepped sine excitation revealed softening behaviour in the FRF).

Now I would like to be able to predict the forced response of my FEmodel** by extracting the very same FRF I extracted via stepped sine.

In my understanding, I should do the following things:
a) choose a force level
b) perform a time-domain analysis (NL transient) using sine excitation as input
c) cut out the transient and keep just the steady-state response
d) calculate FFTs and produce a single FRF point
e) steps b) to d) covering all the frequency range
f) steps a) to e) covering all the force levels

That seems a HUGE task to me, and I am wondering if there was a better/quicker way of doing it.

**FEmodel has been linearly validated with prestressed normal modes using contact, friction and bolts preload.


Thanks in advance!

 

[GregLocock]

Why not use a swept sine signal rather than a single frequency?

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[lawful]

with swept sine excitation the response may be corrupted by transients.

The stepped sine excitation is more conservative in this sense - I could stitch together several sinewaves to create a discrete-swept-sine excitation though.

My problem is that either way it would take just too long to retrieve some useful results and I wonder if there is a workaround to extract frequency responses with nonlinear analysis or at least speed up the process.

I can think of a MATLAB script that generates the NASTRAN input files, feeds them to NASTRAN and processes their results, but perhaps you have better ideas.

Thanks again!

 

[GregLocock]

If you mean what I think you mean that is a solvable issue. working in the time domain you filter out all but the excitation frequency at each point in time, and then do the usual FFT.

In the real world that's called a tracking filter.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?