In automobiles, the brake system is an essential part responsible for control of the vehicle. Any failure in the
brake system generates subsequent catastrophic effects on the vehicle cum passenger’s safety. Hence condition
monitoring of the brake system is indispensable. This study focuses on the condition monitoring of a hydraulic brake
system through vibration analysis. A machine learning approach was used for this vibration analysis. A hydraulic brake
system test rig was fabricated. Frequently occurring fault conditions were simulated. Under good and faulty conditions of
a brake system, the vibration signals were acquired using a piezoelectric transducer. From the vibration signal statistical
features were extracted. The best feature set was identified for classification using attribute evaluator. Selected features
were then classified using K Star algorithm. The classification accuracy of such artificial intelligence technique was then
compared with the decision tree (DT) and Locally Weighted Learning (LWL) algorithm. Comparative results for fault
diagnosis of a hydraulic brake system were reported and discussed. For brake fault diagnosis, K Star performs better and it
gives the maximum classification accuracy as 98.55%. The model built can be used for condition monitoring of a
hydraulic brake system.
Keywords: Attribute evaluator, confusion matrix, C4.5 decision tree algorithm, entropic distance measure, K star, locally
weighted learning, statistical features.
Rights & PermissionsPrintExport