From my experience, I know that amplitude vs frequency curves are generated as output from measurement points in a shaker test. (Amplitude can be displacement, velocity or acceleration. Acceleration is obtained by default from the accelerometer, which can be internally converted into displacement or velocity by the data acquisition system).
I have seen certain test results where curves are generated for different engine orders and these have been plotted on the Campbell diagram against the engine rpm. I have also observed that for the same frequency, at the same measurement point, the amplitude of acceleration/displacement is different for different engine orders. How is this data extracted from the default amplitude vs frequency data ?
Hi Arun
In the past, you used sweeping filters. Nowadays, you use FFT analysis. https://en.wikipedia.org/wiki/Fast_Fourier_transform
There are a few ways to do this.
Perhaps you want to take a look here?
http://qringtech.com/2010/09/15/rotating-machinery-angle-order-analysis/
If you want true orders, you should use revolution rather than time as the basis. However, in doing so, to have 1/N order resolution, you must average over N revolutions which suppress intermittent things. If you are more interested in intermittent matters, you use time as basis.
From time, you can use Kalman-Vold filters to cut out orders but it will not be as exact as when using revolutions as basis as you then use a different number of samples per revolution for each order.
http://www.iiav.org/archives_icsv_last/2016_icsv23/content/papers/papers/full_paper_514_20160408233251984.pdf
HP has some, I believe still patented, very clever order filters that are hardware implemnted can do the job in real time.
http://www.keysight.com/upload/cmc_upload/All/6C06DATAACQ_ORDER.pdf?&cc=SE&lc=eng
and look here wehere you can see the difference between RPM and True Order Maps
http://www.dynamic-measurements.com/datasheet/data_sheet1.html
If you look here
http://qringtech.com/learnmore/recommended-reading/
HP application note 243-1 is all about machine signature analysis. Some of the other material covers Frequency Response Measurement etc.
Have fun
Claes
Hello Claes, thank you for your answer.
When we get the shaker output in the form of amplitude vs frequency, why is an FFT needed? (correct me if I am wrong)
Since the amplitude values are already in the frequency domain, what would we get as output if we further perform a FFT? (Sorry if my testing basics are a little scratchy)
And like I have mentioned in the question, how can amplitude value at the same measurement point, for the same frequency, be different for different engine orders?
Hi Arun
You might get away without FFT if you use a sine sweep with your shaker but not for other signal types.
I encourage you to read the material listed here (start from the top)
http://qringtech.com/learnmore/recommended-reading/
This material is designed for self study without prior knowledge on the topic and will answer most of your questions and provide an overview of various sensors and measurement techniques.
The material is old but it is still the best available, so give it a chance. I promise that it is worth your effort.
SIncerely
Claes
I have been wanting to add this for a long time, but kept procrastinating......In the question that I asked above, the test results which I found were not from shaker test. The Campbell diagrams were plotted from an engine test where the input was in the form of engine revolutions per minute (from a roller dynamometer and car engine). Now I could understand how the amplitudes (for same frequency) were different for different engine orders.
- Badges
- Science topic
- Similar topics
- Force