Development of a Robust Framework for Fault Detection and Quantification in Automotive Drum Brakes Using Vibration Measurements and ANN

<div>Passenger safety is of utmost importance in the automotive industry. Hence, the health of the components, especially the brake system, should be effectively monitored. On account of the significance of artificial intelligence in recent times, any brake fault resulting during operation can be accurately detected using a combination of advanced measurement techniques and machine learning algorithms. The current study focuses on developing and evaluating a robust framework to quantify and classify the faults of a general automotive drum brake. For this purpose, a new experiment for a drum brake, which can be operated under a controlled environment with known levels of faults, is developed. The experiment is instrumented to measure the fundamental dynamic signals (such as brake torque, the angular velocity of the brake drum, and brake shoe accelerations) during a braking event. The response signals from several experiments with various faults and operating conditions serve as the input dataset for establishing the fault quantification algorithm. Multiple variants of this algorithm are devised using different subsets of the input dataset. The selection of features in each variant is done through sensitivity-based segregation with the help of artificial neural networks. The performance of all the variants is comparatively evaluated, and the best among them is determined based on the fault quantification error. Finally, fault classification is carried out using the best variant after establishing the classification thresholds based on the confusion matrix. The following are the novel aspects of this work: (i) design and development of a laboratory experiment for drum brakes that can imitate a real-life braking condition; (ii) measurement of the dynamic response of the system during a typical braking event with a controlled type and level of brake fault using appropriate instrumentation; (iii) estimation of the magnitude of multiple brake faults, in addition to their classification; and (iv) identification of the critical vibration measurements necessary for detecting faults in brakes. In addition, the physical insights into the brake system response, selected features, and the fault quantification algorithm are presented. The proposed framework can also be implemented for fault diagnosis in different automotive subsystems by using an equivalent experiment. The goal of the current work is to develop a simple in situ tool for monitoring the health and diagnosing faults in automotive drum brakes. When integrated with other smart diagnostic and prognostic features, this tool can help automotive manufacturers improve passenger safety.</div>