I am hoping someone can please help with a Vibration query.
A client is seeking to monitor a low speed bearing in some mining equipment.
They wish to use their existing wireless sensors (mems) to monitor vibration and acoustic issues. I am unsure yet if the equipment is variable speed or fixed though speed ~ approx. 6 RPM.
Client is wishing to monitor bearing faults, (at least) having some recent failures.
The Client has been advised by Party A) that while it may seem a low frequency, high sensitivity accelerometer is the best sensor for these low speed applications, this is not correct. Party A) has also advised that a stud mounted sensor (when using high frequency feature extraction techniques) with a standard accelerometer (100 mv/G) performs better in the majority of cases.
Currently Client owns triaxial Mems sensors with:
- Accelerometer sensitivity of 0.031mg/LSB.
- Fmax of 8kHz in X+Y, 5.1kHx in Z
- FFT Frequency Resolution of 0.0625 for 500Hz Fmax and 0.625Hz for 5kHz Fmax and
- Spectral data configurable 100 - 24,000 lines of resolution.
The sensor also possesses an acoustic sensor with Fmin to Fmax of 50Hz to 80kHz and a peak sensitivity of 26dB FS +/-1dB.
Firstly, assuming the 100 mv/g comment is correct, unsure how 0.031mg/LSB equates to this? Is 0.031 a good sensor and suitable for this application?
To complicate matters, Party B) the original supplier of the Client's wireless MEMS sensor suggests that:
"the issue with 6 RPM is the need to achieve 0.01Hz frequency resolution, which means even at a sampling rate of 4000 Hz, we would need 400,000 samples, in a single direction. Means, 1.2 million data points. What can be done is we can just take data points in 1 direction to mainly detect bearing and gear faults. Because at such low RPM, nobody in the world will detect shaft-related faults, because the RPM is so low that even with a shaft fault existing such as misalignment or unbalance, the machine can live for years."
Please absolve the long post.
Seeking your learned guidance please as a 'cold-eyes' third party of how best to resolve the Clients use case and ideally use the MEMS wireless sensor they already own if possible.
I am hoping someone within this incredible network of minds can support.