Motor fault detection using vibration patterns

This work presents a novel method for fault detection of electrical motors using vibration signal. Most of the motor faults generate specific patterns in the motor vibration that can be captured and analyzed for diagnosis. Early detection of motor faults can save the motor from subsequent deteriorations into more severe conditions, and thus can save lot of maintenance costs. In our work, an optical mouse was used to capture decently accurate information of the motor-vibration. Features are extracted in time and frequency domain using which an Artificial Neural Network (ANN) called Multi-Layer Perceptron (MLP) was trained to learn different motor conditions such as healthy and faulty. A MATLAB-based user interface was developed to record, monitor, analyze and classify the motor vibration data. This study shows that using simple features and ANN structure can effectively and efficiently classify different types of motor faults. The use of low-cost mouse sensor has made this method very attractive to wide range of applications where a cost-effective solution is desired.