I have a clamped-clamped beam, with base excitation. Its response is very nonlinear (backbone curves) to base excitation. Is it possible to simulate such nonlinear backbone curves (Y-axis is response amplitude and X axis is frequency sweep) using ANSYS? If yes, please let me know. Thanks in advance.
No, you can not obtain the backbone curve directly from ANSYS ; just get the reduced order model (ROM) of your structure using FE softwares (for your simple case, ROM can be calculated even without using FE package) and then apply continuation techniques for this ROM +nonlinear terms to obtain the backbone curve.
Dear Claudio ,
Actually, we can not get ROM directly from FE packages. We just use the FE package to get full size of K M matrices ,then calculate the ROM using available methods like Craig-Bampton .
This paper might be of interest to you.
http://portal.research.lu.se/portal/files/1668392/4378098.pdf
Kind regards.
Payam
Claudio Pedrazzi
Thanks for the suggestions. I have been able to simulate nonlinear response of a structure using ANSYS WORKBENCH 18.1.Initially I tried using modal + transient modules (connected together) which does not allow large deflections (geometric nonlinearities to be turned ON).
Modal + harmonic also is a linear analysis which does not allow large deflections to be turned ON.
I had my success with just the Transient Analysis module. However if you select Acceleration>Base excitation then also the large deflection option is unavailable. One needs to select acceleration then select the entire body/beam then select direction of acceleration and enter the equation directly [A*sin(2*3.142*f*time)], where A, f are replaced with numerical values. 'time' is automatically accepted by ANSYS as the time for simulation. The 'time' duration and step size needs to be adjusted (keeping in mind the desired range of frequencies and Nyquist criterion) in the 'Analysis Settings'.
In the 3rd file attached you can see that the response plot is nonlinear!
I have attached snippets of the Workbench simulation properties here. What I am looking for/working on now, is to automate the entire process to run through a sweep of different frequencies, f. :)
P.S. A big thanks to Mr. Peters in (ANSYS community moderator, who helped me)
- Badges
- Science topic