Cancel
Showing results for 
Search instead for 
Did you mean: 

RPM Extraction in LMS Test.Lab

Siemens Experimenter Siemens Experimenter
Siemens Experimenter

(view in My Videos)

 

Doing a test and don’t have a tachometer? Or don’t have time to instrument a tachometer? Have a system in which it is impossible to measure the RPM?

 

No problem!   

 

With the Offline RPM-Extraction add-in (Tools -> Add-ins -> Offline RPM Extraction, 20 tokens), it is possible to extract a RPM trace from the vibration or acoustic throughput data you collected.

 

1. Choose the throughput files of the signal you would like to track and replace in the input basket. Selected data could be a vibration or acoustic signal.

1.pngFigure 1: Replacing throughput data in the Input Basket.

 2. Go to the Time Data Selection workbook. Ensure the “Data Source” is set to “Input Basket” then click “Replace”. The throughput channels that were put in the input basket in Step 1 will appear in the Data Set.

 

2.pngFigure 2: Importing the data into the Time Data Selection workbook.

3. Go to the RPM Extraction workbook. Select “Make trace list” to import the throughput data into this workbook from the Time Data Selection workbook. Select which channel to use to calculate the time-tracked colormap by checking the box in the “Selected” column (a blue box will appear). Then click “Calculate spectral map”. In the resulting colormap, it is desired to see a strong order that can be traced to determine the rpm. If necessary, change the processing setting for your map under “Settings…” until the desired order content appears.

 

3.pngFigure 3: Bring in the data and create a spectral map tracked on time.

 4. Extracting the RPM:

 

4.A: Place the cross-hair cursor on an order.

 

4.B: Click “Accept Point”. This will accept this point for the order calculation.

 

4.pngFigure 4: Selecting the order.4.C: Click “Predict Curve”. This will estimate the order trace and a black line will appear on the colormap where the order trace is estimated.

 

4.D: Enter the order number of the order the cursor is placed on. This is essential to get the correct RPM trace.

 

4.E: Click “Calculate RPM”.

 

4.F: The RPM trace will appear.

 

5.pngFigure 5: Calculating the RPM.

NOTE: Under the “Processing” area, there are three method options: the “One point method”, “Two points method”, and “Multiple points method”.

 

  • One point method: predicts the order after accepting one point on the order line. The order curve is then predicted from the beginning of the time trace to the end. In this example, the order is well defined and the one point method works well.

 

  • Two points method: predicts the order after accepting two points on the order line. The order is only predicted between the two accepted points (it is not predicted from the beginning of the time trace to the end like with the one point method).

 

  • Multiple points method: predicts the order after accepting multiple points on the order line. This is useful when the order line is not very well defined. The order is only predicted between the lowest time value accepted point and the highest time value accepted point.

Try playing around with these different settings to see how they differ.

 

5. Using the “Save” button, a new run will be created in the Navigator. Offline RPM extraction will not store data back into the original throughput run. So, if you want all of your original throughput channels to save into the run created in RPM Extraction, you must add all the throughput channels to the input basket and import all the throughput channels into RPM Extraction (Steps 1 and 2).

 

5.A: Name the run.

 

5.B: Save the data.

 

6.pngFigure 6: Save the extracted RPM data.6. View the data in the Navigator workbook. The data will be stored in a folder with the run name that was typed in the RPM Extraction workbook. The extracted trace will always be named channel “5555: RPM Extr”. Drag the extracted RPM into a plot to see what it looks like.

 

7.pngFigure 7: Extracted RPM is stored with the name “5555: RPM Extr”.

For this example, the actual tachometer data was also recorded. Below is the actual tacho data compared with the extracted RPM. You can see they are nearly identical.

 

8.pngFigure 8: The extracted RPM data (5555:RPM Extr) compared with the RPM calculated from the actual tachometer (1:Tacho1).

 

Questions? Contact us!

 

LMS Test.Lab processing links:

LMS Test.Lab acquisition tips:

General Knowledge articles:

Comments
Creator
Creator

Good afternoon friend!

 

I have a question respect the RPM extraction.

 

I'm reading this tutorial for practice this option in the LMS Test.lab, also I have the signal for the turbodata1 and when analyze the option RPM extraction, I'm a little confused respect to the values of the RPM results.

 

Show a follow figure:

 

Captura.JPG

 

I know that to convert RPM to Hertz the convertion factor is 60, so, for the same time point (7,13) I obtain 86,36 Hz in the spectral map and 1287,86 RPM in the RPM curve, but 86,36 Hz equals to  5181,6 RPM and not 1287,86 RPM, like the RPM curve. Im a little confused about that, please help me.

 

Grettings

Alvaro

Siemens Phenom Siemens Phenom
Siemens Phenom

Hello Alvaro,

 

In order to use Offline RPM extraction you need to be able to identify a dominant order in the waterfall / colormap and know what the order number is.  For example, the black curve you've drawn is it 1st, 2nd, 3rd, or 4th order ?  You've entered 4th order and we use that calculate the RPM.  So the speed is calculated as ~ (86.36 Hz (cycles/sec) * 60 (sec/minute)) / 4 (your order specified).

 

If you specify 1st order and re-calculate the RPM you would get the speed you expect based on 1st order.

 

Kevin

Creator
Creator

Very thanks Kevin.

 

Well, is a nice tool to know the RPM and I understanded now about this option, but is necessary to know the order dominant. what happen if the dominat order isn't RPM of the motor? because I did  an experimental test of a motor and the the owner of the motor said me that the RPM is 1800 RPM, but I don't see anyone peaks at this frecuency (I installed 4 sensores) in my FFT's spectrums, the RPM of the motor always will be the order dominant?    

Siemens Phenom Siemens Phenom
Siemens Phenom

Hello Alvaro,  if the order is very quiet or does not vibrate much it may not be the dominant order.  If you can hear it, can you use a microphone rather than an accelerometer?  I would be very suprised if 1800 RPM is the speed of the motor that 30 Hz or it's harmonics do not show up at all.

 

If you are following the tutorial, you can determine the RPM using 1st, 2nd and 4th order of the engine if I remember correctly and later in the tutorial you can determine the speed of the turbo using 1st or 2nd order of  the turbo signal.

 

Kevin

Creator
Creator

Hello Kevin

 

The motor information indicates revolutions at 1770 RPM, however, analyzing the spectrum of accelerometers installed in the motor housing (horizontal and vertical directions) it is not possible to identify peaks or harmonics at that speed of rotation (ie, equivalent in frequency). However, there are peaks identified in other frequencies that would be related to faults in the bearings, belts, etc. But to be sure, it is necessary to determine the rotation of the motor in the FFT. We have not bought a tachometer, would it be better to buy it in these cases? Is it possible that the motor's axis is well balanced and the corresponding peak can not be identified in the FFT?

 

Grettings

Alvaro

Experimenter
Experimenter

Hello

 

Since i don't have a tacho, but i want to extract the RPM and predict the order. I know that my machine runs at 800+/- 10 rpm. How should i collect this data? Should i run a signature analysis using stationary as tracking methos and then follow the steps you have shown?

Contributors