Prognostic Monitoring and Analyzing System for Motors

Abstract

Focusing on the ‘Bottom Line Today,’ the cost of downtime has a big impact on profitability. If equipment starts to wear, it is possible to start producing parts with unacceptable quality, not known for a long time. Eventually, machine wear will seriously affect not only productivity but also product quality. To evade non-productivity and in turn the machine wear, we chose to propose the concept of prognostic monitoring and analyzing system for motors. To be specific, we have designed a system to monitor the health of an induction motor in varying load and faulty conditions. The health is monitored using multiple sensors placed on the various parts of the motor that are critical in diagnosing the condition of the motor. We acquired data from these sensors and analyzed the same on MATLAB using real-time power spectral density. We then developed a machine learning model using principal component analysis on MATLAB to predict the health of the motor based on the aforementioned conditions that the motor is under subjection. Based on the prediction, we have assigned a suitable maintenance program that keeps the motor in optimum working condition. This prediction shows the motor down time under a particular chosen faulty condition.