currently I am looking into the Operational Modal Analysis Add On in LMS Test.Lab. First I tested a small, simple structure which I excited by lots of random impacts with a little stick. The result was quite good and I now am trying a bigger more complex structure. Because of the size and preloads I want to really excite that structure only by operational forces. However in the interesting load case the engine is only running at a specific frequency. Of course there are other orders, but the excitation forces won't be similar to white noise (broad frequency spectrum). Instead I will get distinct peaks at the orders of the engine.
Can I in this case still use the time data to create an operational modal analysis? Or do I need sweeps from the Engine or other random excitation forces? Or what is the common way to determine the modes under operational loads? Because of the preloads and unlinear behaviour I can't really perform a modal analysis. So with an operational modal analysis I can measure with the correct conditions, but sadly my operational excitation forces only have specific orders.
Thank you for any advice on how to tackle such a problem.
Solved! Go to Solution.
thanks for the interesting and relevant question!
In case of engine excitation, we typically suggest two options, depending on the operating conditions.
If your system runs in stationary conditions, obviously you will see all excited orders as dominant and your spectrum will be dominated by harmonics. Consequently, the excitation cannot be anymore be considered as broadband. If the rotational speed of the engine is measured or can be estimated accurately, then you can use the Harmonic Filter (available as add-on) to estimate and remove these harmonics. You can now use Operational Modal Analysis to estimate your structural modes.
Another possibility is to rely on transient operating conditions, normally run ups or coast downs. In these cases, the different orders acts as sine sweeps: once an order crosses a resonance, it will be amplified. If an accurate tacho signal is available, you can track your dominant orders, use the tacho signal as reference and then use Order-based Modal Analysis. This technique is very effective if the run up (or coast down) can be controlled to be slow, so that the system is given time to respond.
In principle, you could also use standard Operational Modal Analysis during run-ups or coast down, but then you would see spurious peaks in the stabilization diagram whenever an order is ending.
If you are interested in having more details on one of the two approaches (harmonic removal + OMA) or Order-based Modal Analysis, i can suggest you the following references:
And if you want more details on how to use them in test.Lab, do not hesitate to contact your local office for more support.
thank you for the fast and very helpful answer. Especially the 2nd article about removing the harmonics of the excitation is helpful as I expect to not get low enough in excitation frequency with the order based OMA. The gasoline engine probably will stall before getting low enough in frequency. So, I'll look into the particular add on.
Further I found literature about transmissibility based OMA. Because of the transfer functions used, OMA gets independent from the type of the excitation mechanism. Is there a way to perform an OMA in LMS with transfer functions?
happy to be of help!
About the transmissibility-based OMA, I know the approach is around since quite some years, and we look at its evolution with interest, but the identification is not compatible with the one we have currently implemented now for half-spectra so it is not possible to use it in Test.Lab rigth now.
Wish you a ssuccessful harmonic removal
I am working on a IFB washing machine. I want to conduct Operational Modal Analysis (OMA) on it. I considered doing an Run-up test,I have a decent tacho available. In which Add-on can i conduct the OMA and i want to obtain a campbell(Waterfall/Color Map). I believe it can be obtained in Signature Acquisition license. But the issue is i am not able to understand what oders are and which orders i have to look into. I am studying mainly the panel vibration. And embossment patterns on washing machine. The machine is assembled back and in working condition. If a video is available for conducting OMA it would be great.
I apologize for ask your help about a little doubt mine respect a test maked here in a restricted metal frame with a common hammer. I show you the pics:
In the 4 channels showed, the first impulse (result the impact with the hammer) have a weird shape and I'd like to know the reason, I hope you can help me with that.
Good Morning friend.
I send you the pictures from the impact test maked to the metalic frame.
In the measurement channels appears a shape weird at the first impact, when We used a common hammer (without transducter) and the second impact the shape is more normal and the third have the same weird shape. I hit the frame very hard to try to excite many mode in the frame, I'm thinbking about only impact not so hard with the hammer the frame. The acelerometers were fixed to the frame with tape not with a magnetic base but in the same channel appears the weird shape and after the normal shape of the impact.
Thanks for advance.
Are you performing impact Testing in Test.Xpress or Test.Lab? Here's an article to explain how to perform impact testing in Test.Lab: https://community.plm.automation.siemens.com/t5/Testing-Knowledge-Base/LMS-Test-Lab-Impact-Testing/t...
Did you enter the correct sensitivty value for hammer and accelerometers in channel setup, define the x, y or z directions? Also, have you try metal tip on the hammer? Usually metal tip goes to higher frequency than plastic or rubber tips. How do you mount the accelerometers on the structure? Did you use super glue or dental cement?
Well I used other software to see the measurements and it seems that the impact was very hard, we didn't use modal hammer only we used a common hammer with metal tip.
We try to make the impact test again friend.
Thanks for the help.