Bearing fault diagnosis using TQWT based Entropy features

The rolling element bearing plays a significant role in any mechanical manufacturing industry. Bearing fault diagnosis using vibration signal processing has been a vital subject of study over past many years. In present work, an automated bearing fault diagnosis technique using Tunable Q-factor Wavelet Transform based time-frequency decomposition is proposed. Proposed technique is carried out in three methodological steps. Firstly, vibration signals are decomposed into time-frequency coefficients using Tunable Q-factor Wavelet Transform. Further, entropy based features are computed from decomposed time-frequency coefficients as features of vibration signals in second step. In the third step, classification of vibration signals is performed using three soft computing methods viz. Support Vector Machine, Artificial Neural Network and Random Forest Tree classifier. Classification results revealed that presented methodology has potential to design and develop a real time bearing fault diagnosis system. In present work, classification accuracy of 100% is achieved using Support Vector Machine Classifier.