Near Infrared Sensor-Based Condition Monitoring and Fault Diagnosis of Induction Motors using machine learning algorithms

Abstract

Induction motors are extensively used in industrial applications because of their rugged construction, reliability, and low maintenance requirements. However, stator winding faults, rotor defects, insulation degradation, and bearing failures can significantly affect motor performance and operational reliability. Early fault detection is essential to reduce downtime and maintenance costs. Conventional monitoring methods utilize current, vibration, acoustic, and thermal measurements for fault diagnosis. Recent advances in optical sensing technologies have introduced Near Infrared (NIR) sensors as an effective non-contact solution for induction motor condition monitoring. This review presents a comprehensive study of induction motor fault diagnosis techniques with particular emphasis on NIR sensor-based monitoring systems. Signal processing methods including Discrete Wavelet Transform (DWT), Dyadic Wavelet Transform (DyWT), Rational Dilation Wavelet Transform (RADWT), and Tunable Q-Factor Wavelet Transform (TQWT) are reviewed and compared. The advantages, limitations, and future research opportunities of NIR-based monitoring systems are discussed. The study concludes that NIR sensor-based monitoring provides a promising alternative for online diagnosis of winding and insulation faults in induction motors.