Showing results for 
Search instead for 
Do you mean 
Reply

How to reproduce a vibration beam test?

Hello everybody,

I performed a experimental modal analysis und a vibration test of a cantilever beam and meassured the eigenfrequencies.
Then I excited the cantilever beam with the resonance frequency of the first mode via a electro magnet and performed a decay test. The wave form of the excitation was a sinus.
The data of motion was recored by a accelerometer.

Now I wanted to reproduce this experiment in NX Nastran 7.5 with a SEMODES 103 Response Simulation.

If I perfomed a semodes 103 simulation the values of simulation are similar to the experiment.
But if I tryed to performe a response simulation with a enforced motion location for my recorded acceleration as excitation, I lost the first eigenfrequency and also some other frequencies do not have the same value as i know from testing an simulation with semodes 103.

I tried to excite my sytem with a nodal force location and everything was fine.

So do have someone an suggestion to solve my problem?
This would be very helpfully!

Best regards,
Martin
4 REPLIES

Re: How to reproduce a vibration beam test?

Hello Martin,

Just some question:

1. when you do response simulation, the enforced motion direction is same as the direction of your accelerometer?
2.Your accelerometer is measuring in 1 direction only?
3. your data of motion is the output, and you are using the output data as an input to your simulation?


Best regards,

Tuw

Re: How to reproduce a vibration beam test?

Hello Tuw,
1. yes, the direction is the same.
2. the accelerometer is measuring in all 3 directions, but the signal I use is just in normal direction to the surface of the beam.
3. The meassured data is my input then.
Best regards,
Martin

Re: How to reproduce a vibration beam test?

Hello Martin,

I assume you are doing a frequency response using Sol103-RS.

Please try to increase the spectral lines, during 'New event dialog box'. Additional lines will give you more resolution, while too few spectral lines may miss a significant response peak.

Also, try to increase the number of modes obtained from the initial normal mode analysis (Sol103). The more modes you have, the more mass participation for RS to compute peak.

Next, you can try to toggle from Data recovery method from mode displacement to mode acceleration. Mode displacement have problem when involve large static component in a dynamic analysis.

Hope this helps.

Best regards,

Tuw

Re: How to reproduce a vibration beam test?

Hello Tuw!
No, i perform a transient analysis.
The GTAC support told me that for Sol103-RS it's not possible to excite a model with an acceleration, velocity or displacement. Only Nodal Force Locations are supportet by NX.
The differences between Sol103 and Sol103-RS is caused by a different constraint from the enforced motion location.
Maybe a different solver will provide a solution for my problem.
If anyone has an suggestion this would be very helpful.
Best Regards,
Martin