Sound-Based Health Monitoring of Induction Motor Considering Load and Measuring Distance Variations Using Frequency Calculation and Statistical Analysis

Abstract

Bearing is an important part of the induction motor, whose function is to help the rotor spin. It contributes the highest percentage of damage compared to other parts. When operated in this condition, it causes overheating, imbalance in the rotation of the rotor shaft, sparks, and noise pollution to the environment. A bearing monitoring system must be implemented and developed to avoid further damage. Furthermore, a non-invasive technique through sound signals was developed in this study. A sound signal is easy to overlap with the noise from other sources. Environmental noise is unavoidable during data collection, affecting health monitoring accuracy (HM). Therefore, this study aims to develop an HM method for sound-based induction motors based on measurement differences, load variation, frequency calculations, and statistics. The distance measured was used as an independent variable of the non-machine noise. The load variations were also applied as required, and the operation of the motor varies according to users’ needs. In an effort to prevent negative environmental impacts, noise monitoring was carried out from the motor operation, and the results showed an HM of accuracy of 83.09%. The best distance for performing HM conditions is 100 cm and 83.59 dB(A). The noise value does not exceed the industrial worker threshold. Therefore, close surveillance of the motor’s condition tends to be conducted with or without a load. It is because the load variation does not affect the accuracy of health monitoring.