Brushless DC Motor Fault Classification Using Support Vector Machine Algorithm with Discrete Wavelet Transform Feature Extraction

A motor fault is a common problem in BLDC motors. If the fault is not detected, it can cause malfunctions in the machines. A mechanical motor fault diagnosis using a Support Vector Machine with Discrete Wavelet Transform feature extraction is initiated in this research. It will be compared to the previous study of Borja et al., which utilizes the K-nearest neighbor alongside the Discrete Wavelet Transform in diagnosing motor faults. Paired T-test was used to differentiate the capabilities of SVM-DWT and K-NN-DWT. It shows the accuracy of the two classifiers; the Support Vector Machine gets an accuracy of 98.67%, while K-Nearest Neighbor obtains an accuracy of 97.49%. The results prove that the Support Vector Machine was more capable than K-Nearest Neighbor in diagnosing motor faults.

I. Introduction

The brushless direct current (BLDC) motor technology is high-performance and requires less maintenance. The best motor to use is one of this kind for high-quality and efficient applications. Because this type of motor can withstand the rotational force at its maximum speed and has low noise and heat, even if it is a smaller or lighter motor, it may operate continuously. This is one of the advantages of this motor because the BLDC system controls the motor, uses less energy, generates less heat, operates more silently, and has a longer lifespan [1]. Numerous techniques have been put out in recent years to diagnose electric motor faults, including DFT, STFT, Neuro-fuzzy Transform [2], Fault Tolerant Control [3], and other algorithms that were performed but still need some improvement.