Diagnostics of Combined Mechanical and Electrical Faults of an Electromechanical System for Steady and Ramp-Up Speeds

Purpose This study presents the diagnosis of combined or multiple mechanical and electrical faults in an electromechanical system based on support vector machine (SVM). Methods Time and continuous wavelet transform (CWT)-based features have been extracted from vibration and current signals acquired from ten combined fault conditions from an experimental test rig. Then, SVM algorithm based methodology has been developed for diagnosis the fault conditions. Results Results show that the combination of vibration and current signal improves the performance of the diagnosis especially at high-load condition in comparison to vibration or current signal alone. The diagnosis is found to be effective with both time domain and CWT features; however, it is slightly better with the CWT features in comparison to time domain-based features. Conclusion The performance of the diagnosis is found to be better at higher load with both time domain and CWT features. Moreover, the present methodology does not perform well with ramp-up speed conditions.