A Comprehensive Approach for Detecting Brake Pad Defects Using Histogram and Wavelet Features with Nested Dichotomy Family Classifiers

Abstract

The brake system requires careful attention for continuous monitoring as a vital module. This study specifically focuses on monitoring the hydraulic brake system using vibration signals through experimentation. Vibration signals from the brake pad assembly of commercial vehicles were captured under both good and defective conditions. Relevant histograms and wavelet features were extracted from these signals. The selected features were then categorized using Nested dichotomy family classifiers. The accuracy of all the algorithms during categorization was evaluated. Among the algorithms tested, the class-balanced nested dichotomy algorithm with a wavelet filter achieved a maximum accuracy of 99.45%. This indicates a highly effective method for accurately categorizing the brake system based on vibration signals. By implementing such a monitoring system, the reliability of the hydraulic brake system can be ensured, which is crucial for the safe and efficient operation of commercial vehicles in the market.

Signal Processing for the Condition-Based Maintenance of Rotating Machines via Vibration Analysis: A Tutorial

“…Vibration sensors are usually preferred over others because they allow for the early detection of faults [23]. Furthermore, various machine learning techniques, often designated as artificial intelligence methods, have been applied to condition-based maintenance via vibration analysis [24][25][26][27]. For fault classification, these approaches prove highly valuable when sufficient data, including faulty data, are available.…”