01-25-2017 07:31 AM
Hello everyone,
I can calculate SRS (shock response spectrum) using "SVT Shock Response Spectrum VI" of acceleration-time data,
But our customer want to convert shock response spectrum to acceleration-time data. How can I convert ?
Regards...
01-25-2017 09:31 AM
There is (to my knowledge) no generic method to achieve what you are requesting. The output of the Shock Response Spectrum VI is an array of amplitudes for logarithmically spaced frequencies and the result may fit many different input pulse shapes. You are for example missing phase information that could help estimating relative delay between the different analyzed frequencies.
Also the result depends on several input parameters like input waveform duration and critical damping factor, would you know what values were used for the response calculation?
Can you tell us about your use case? Do you roughly know the shape and polarity of your original acceleration pulse? Any other information you can share?
01-25-2017 12:14 PM
Thanks for detailed answer. We are developing earthquake simulator and control software. The earthquake simulator can simulate any acceleration, displacement waveform. Our customer want to simulate acceleration waveform generated from any SRS(Frequency-Acceleration pairs). User will enter to multicolumn listbox frequency and acceleration pairs . Acceleration data will be generated using these pairs.
Regards...
01-25-2017 11:57 PM
01-26-2017 08:19 AM
You'll have to set some restrictions to your pulse shape to reduce the number of degrees of freedom.
For example if you can accept that your pulse is phase-linear (symmetrical) you could define the time domain waveform as the sum of N weighted cosine bursts (referenced to the middle position of your pulse), one for each of your N specified natural frequencies. Then you could try to adjust the amplitudes of your N sine bursts so the SRS output matches your requirements.
I would suggest to try something like the Nonlinear Curve Fit.vi (with or without contraints) with an appropriate start guess (too avoid getting stucked near local extrema). Check eventually the example Ellipse fit.vi to see how to implement the feature.
Note that the SRS VI expects a logarithmic distributed frequency list but you do have access to the corresponding frequency list on the diagram so it should be relatively easy to modify the VI to use your own list.
01-26-2017 08:52 AM
01-26-2017 09:42 AM
The task is more demanding than just creating a small sample VI, and it is normally not the purpose of this forum to create more or less complete solutions but rather to assist with specific questions and issues.
May I suggest you to experiment with the different ideas we've discussed and eventually post your code and results with specific questions if you get stucked. Alternatively you can request assistance with a partner or consultant for more in-dept development work.
There may also be somebody else on the line that can suggest a different approach to your task (??)
Best Regards,