Fault detection for Gearboxes and Bearings

Description:

This software system detects faults at a much earlier stage than current systems, allowing the operator to better optimize system maintenance. Early detection of faults in gearboxes and bearings is crucial in order to minimize expensive repairs with significant down time. Our first application for this technology is in wind turbines.

Dr. Ming Liang has developed software that detects faults at their earliest possible stages. The algorithms analyze the vibration signals obtained from off-the-shelf sensors (accelerometers) placed on the machine and data collected by many of the existing systems. The software performs multiple functions in a very efficient manner by deploying multiple techniques including the joint time-frequency analysis, multi-modulation extraction and adaptive spectral kurtosis analysis. The proprietary algorithms enable greater efficiency and sensitivity to obtain an assessment of machine health compared to competitive methods which do not effectively synthesize these elements.

Benefits:

Minimal hardware costs to install as the system uses off the shelf sensors, often already installed.
Early detection to minimise down time and repair costs through a more effective analysis.
Reliable and easy to implement as it is less susceptible to changes in operating conditions. It simplifies the estimation of bandpass filter parameters required by nearly all high frequency resonance methods.