How to calculate gE Gravitational Enveloping value?

[Navalandrevs]

Hello!

I am trying to make a VI where the input be acceleration signal from an industrial accelerometer, and the output be the gE value, that one showed in SKF analyzers, that is for bearings diagnostics.

These value, Could I get from the OAT Envelope Detection VI? It gives me an enveloped curve, and I guess that the gE value is the RMS of this signal? or Do I still have to low pass filter the signal?

Also, as the envelope waveform is entirely above 0 g, if I perform the FFT, it will have a start value (in 0Hz) too high. Did someone try to do the same thing and got that questions?

Thank you

[Bob_Schor]

I'd never heard of gE, or acceleration enveloping, but Google explained it is a technique used in vibration analysis to detect faulty bearings.  The basic technique seems fairly simple, but (as several articles note) the interpretation of the results is best left to experts.

Here is a description of the method:

"How acceleration enveloping works

Acceleration enveloping is a two stage process. The first step is to apply a band pass filter to the mix of low and high frequencies of a defective  bearing’s unfiltered waveform.  This isolates only the frequencies in which the signal of interest is hiding. The filtered output will identify repeating,
high frequency signals. On paper, this process would be represented as a series of spiking energy bursts energy;  these are the impacts from the
rolling elements hitting the defect of the rotating bearing.

The second step in the process is to pass the filtered output through an enveloper, which rectifies (or demodulates) the waveform, by inverting the negative part to positive, and extracts the repetition rate of the energy bursts. This ‘envelope’ is now used as a true vibration signal – helping it to stand out from the noise.

The envelope helps to contain regularly spaced signals, such as a single defect on a raceway, but other causes of noise, such as shaft rub, are random – so will not produce evenly spaced peaks.  It is important to point out that some experience is required to ensure the steps in this process is completed correctly, especially when selecting the correct high- and low-pass frequencies in the first stage of the process."

From the above description, it is relatively straight-forward to write LabVIEW code to carry out this process, but choosing the parameters to capture the raw accelerometer data, filter it properly to identify the "bearing noise" signals of interest, and then passing it through an "envelope" transform (basically subtract the mean and full-wave rectify) and interpreting the results may require ME experience that I do not possess.

Bob Schor

[santo_13]

Looks like Sound and Vibration toolkit has some VIs with the term OAT in that, it is worth trying it out

https://www.ni.com/docs/en-US/bundle/labview-2018-sound-and-vibration-toolkit/page/sndvibtk/oat_extr...

[Navalandrevs]

Thank you Bob for your explanation.

I do have some experience, but I only need to know if the envelope curve I have to pass by a high pass filter again cause of this signal is positive above 0 g an It will give me high level around 0 Hz.

[Navalandrevs]

Dear Mikky

Here 17 years after, I have the same issue than you had.

I am using the OAT Envelope Detection VI from Sound and Vibration Toolkit.

My question is, if I got the envelope curve, How do I calculate the gE value (envelope acceleration measured in SKF). Because the envelope curve is an all positive curve, so the gE peak to peak value is eeh diference between highest and lowest values of the envelope curve? or it is the highest value? or it is the RMS value of that curve? Or do I have to take the DC value out of the envelope curve first to perform the p-p value?

I hope you can see it,

Thank you

[wiebe@CARYA]

---

@Navalandrevs wrote:

Dear Mikky

Here 17 years after, I have the same issue than you had.

---

Mikky? 17 years after what? Re: bearing?

I'm confused...