Nonlinear Random Vibration Analysis 1

[hly11111]

Hi,

I am performing a random vibration analysis in LS Dyna. My model has two parts connected by elastic beams. The overall model is linear, but with the introduction of elastic beam connection, this implies nonlinearity.
As the result, I obtained nonlinear solutions. I'm studying the connection/joints effects on random vibration analysis to understand its limitations in LS Dyna.

My question is that how is it possible for a linear dynamic analysis such as random vibration analysis to have nonlinear solutions? it seems not very intuitive for me that you can start out with a linear analysis and end up with nonlinear solutions. Any comments or elaborations on this would be much appreciated. Any tips from simulation experts (Especially LS Dyna experts) on performing random vibration analysis would be very helpful for me.

Best Regards,

 

[GregLocock]

I found it difficult to understand your post. It is trivially obvious that in the real world nonlinear systems excited randomly give results. So you need to define randomly and results and nonlinear in your case, properly.

Cheers

Greg Locock


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

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

 

[hly11111]

Hi Greg,

Thank you for your prompt response. I am fully aware that in the real world situation, most of the structures are non linear, and when excited randomly, it would give results. However, I'm not asking from the real life testing point of view.
I will try my best to explain my question better.

Before performing random vibration analysis, we need to perform the modal analysis to obtain the natural frequencies and mode shapes of the system. After that, we use modal superposition method as the basis to perform random vibration analysis. Since random vibration is based on modal superposition method, which requires linear system, random vibration is a dynamic linear analysis.

When nonlinearity is introduced in the system by defining connections, contacts or joints, the random vibration is no longer linear and will give nonlinear results. So then my question is that doesn't the nonlinear random vibration violate the superposition method that it uses as the basis?

I hope I explain myself better. Once again, Thanks for your answer.

 

[GregLocock]

Yes that's a decent question. Again in the real world when we do a modal analysis we first have to linearise the system. In practice this means wedging the doors shut, stopping rattles, ans so on. If we still can't get clean measurements then we linearise by testing with a swept sine signal, at a constant force level, and ignoring the off-frequency response.

So you can't really do a modal on a non linear system, you are really doing an operating deflection shape, and the forcing is part of the analysis set, whereas in a linear modal the forcing is theoretically irrelevant.

Cheers

Greg Locock


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

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