I have collected multiple accelerometer response channels, while performing a sine sweep on an ED shaker, and I'm attempting to calcuate Transmissibility on some of the data.
For instance: I have two response channels on my UUT, and I'd like to calculate the transmissibility between those two channels. I tried using the Time Signal Calculator, while the data was being collected, but totally forgot about the signal crossing zero, so all I have are a bunch of very low/high numbers as the frequency sweeps from 20Hz to 2kHz.
I did not use SimCenter to control the shaker; I used a shaker controller software we had purchased with the shakers. I simply recorded (8) response accels using SimCenter, and would like to post-process multiple trans channels from that data.
If that is doable, please advise.
Thx in advance.
Solved! Go to Solution.
First, it is recommended to use the Siemens Controller. If the Transmissibility is important and part of your standard practice, this is a function that is collected by default when the option is selected in the test setup using the Siemens controller. Any other channel, including a control channel, can be used as the amplitude and phase reference.
In case the transmissibility is required and not measured, then the time data recording can be used to calculate this. If you have Testlab tokens, or a license for the Testlab “Offline Sine Data Reduction”, then you can easily use this for the calculation to process the data which will duplicate the results as if you measured the function in real time during the control test.
It is easiest to use this method if you also recorded the COLA signal. If not then there is more to do.
i totally agree with Bill, best way is surely to perform both measurement an processing within Simcenter Testlab. However, there migth be a way to get the transmissibility also from time-histories only.
If you have tokens, you could enable the Signature Throughput Processing add-in.
Once the time data are there, you could process them into transmissibilities as you would do for FRFs. WIn the acquisition settings, choose free run and set the number of averages to a very high number (i.e. 1000 or 10000) and the overlap to 90% (the high number of averages is only to ensure you use all time data available and do not stop earlier). You can then select in the Fixed Sampling part the block settings. Once this is done, in the Channel Processing settings, for the type of channel you have, you can select FRF and choose the reference channel you want to use.
Keep however in mind that by doing this process (which is the same averaging you would use for a random signal) you need to define a window (normally Hanning) which migth introduce some bias in your estimates.
Thank you Bill and smanzat. I will attempt that process this morning.
I'm a little concerned with using Siemens to control the shaker. I was initially told that Test.Lab could be used for Open Loop Control, which isn't an option when controlling expensive shakers. Yes, the shakers themselves have multiple fail-safes, but I'd rather stick with closed-loop/feedback control I have with the shaker controllers. This option may have changed with the V18 update, but I haven't yet had enough time to learn the options available. Plus, the sweeps are completed, and I'd rather not repeat those (not fun).
I will attempt to calculate using your suggestions. Thank you for the info. I'm hoping that the V18 Simcenter, if pushed out as advertised, has opened up the Offline Sine Data Reduction application, since we do have plenty of token licenses, but not a specific license for that application. I can easily use Signature Throughput Processing to try calculating using smanzat's suggestion. I use that add-in frequently, just not for this specific purpose.
Thanks to both of you! I will report back when I have some results.
I agree with you that a closed loop controller is required for your situation. We have different software for “controlling” the shaker excitation, and depending on the situation, this can be construed as either “open-loop” control or “closed-loop” control. I was referring to “Closed-loop” control in my suggestion. You are lucky in that with the new licensing scheme in Testlab 18, (for your future tests) with your tokens you can also have access to the “Closed-loop” form of sine control.
For this current situation, with your tokens and Rev18, you can also reproduce the Transmissibility calculation using the “Offline Sine Data Reduction”, using the recorded time history data. There are different ways as Simone suggests too, but since you have tokens, I think OSDR will be the easiest to reproduce the data you need, and the software is set up exactly for this particular scenario. Hopefully the COLA was recorded too from your other controller. If you need help with any of this please contact our support group.
Excellent fellas!!! Thank you so much for your time this week!!
I did the post-processing, as Simone (smanzat) described, using the FRF method and setting various reference channels (since I was collecting data at multiple response locations and required trans between various points), and set everything as you described, and it worked fantastic.
I had two specific sweeps, and the mass of the sample varied between them. I could see, from the time-history data, the lighter sample went into resonance and peaked at 90Hz, and the heavier sample at 68Hz. The transmissibility data peaked at those same, first mode, values.
Again, thank you both for your time this week.[/b] The next time I perform these types of tests on the ED shaker, I will use your solution Bill, and let you know how that works, and if I can save some time overall (as opposed to post-processing each curve).