IoT Vibration monitor
Vibration monitoring has a wide range of applications from analysing machine health to mitigating theft.
This sensor head is designed for monitoring gas and oil pipeline integrity.
Jamtech were approached by one of their clients to devise a remote vibration monitor to operate in a harsh environment. We were able to produce the electronics and high integrity enclosure for housing the sensor. The vibration monitor needed to be able to distinguish abnormal events from background noise. This is achieved by analysing vibrations in the frequency domain and comparing them to the characterised ‘normal signature’.
How does the vibration monitor work?
Vibrations in this case are converted to an electrical analogue signal by a sensor, referred to as a geophone. These sensors are very sensitive to vibration and normally operate in one axis. The analogue signal from the geophone is initially recorded in the time domain and then converted to the frequency domain using a Fast Fourier Transform (FFT). The magnitude plot of the FFT serves as the signature of the vibration being measured and can then be compared to predefine normal vibration signatures.
To help illustrate this, consider the simplistic example of background noise that has a single frequency. Shown below is the time domain plot a) having 32 sample. When parsed by an FFT algorithm the frequency magnitude plot is produced in b). Let’s call this background noise.
Now we’ll add abnormal noise within the background noise into the time domain a). It’s clear the time domain has changed, but exactly how? This becomes clear when converted to the frequency domain b).
It’s now a relatively simple matter to compare the frequency domain plots, above, to determine the presence of an abnormality. Predefined alarm levels can be set to allow the system to automatically send a message from a remote location that an abnormality has taken place.
The above is a very simplistic illustration of vibration monitoring. In a real world application the noise is spread over a much wider spectrum and detection of abnormal signatures may also involve other analytical methods such as moving averages.
This system can be used for a whole range of applications from machine monitoring to theft mitigation.
- National Instruments, “FFT Fundamentals” http://zone.ni.com/reference/en-XX/help/372416B-01/svtconcepts/fft_funda/.
- Lyons, “Understanding Digital Signal Processing”, 2004