Fault Diagnosis of Ball Bearing Using Walsh–Hadamard Transform and Random Tree Classifier

Bearing failure may result in the breakdown of machinery or possibly damage the human being operating the machinery. It is therefore necessary to diagnose bearing faults at an early stage. This paper presents the application of Walsh–Hadamard transform and tree-based classifier for detecting bearing faults. An experimental was conducted from the Case Western Reserve University bearing data center. At the initial stage, coefficients from inner race fault, outer race fault, ball fault, and healthy bearing were calculated from the measured signal. Twenty-five statistical features are calculated from the acquire signals with different speeds and fault sizes. Feature vector form is used for classification purpose using various classifiers, i.e., random tree and logistic model tree. Training and testing of classifier are done, and the result revealed that 100% fault identification accuracy is obtained with random tree for both training and testing. Similarly, 98.43% training and 100% fault identification accuracy are obtained with logistic model tree. The result shows that the methodology adopted is effective to diagnose various bearing faults.