Fault Detection In Locomotive Traction Motors By Stator Current Dual-Tree Complex Wavelet Analysis

The diagnosis of faults in railway traction systems has a significant importance on both safety and reliability, which can avoid train crashes. In order to extract the fault features effectively, the dual tree complex wavelet is used in stator current signal analysis of traction motors. The contributions and conclusions are made as follows:Compared to conventional monitoring techniques such as vibration monitoring or temperature monitoring, stator current-based monitoring offers significant economic savings and implementation advantages. An overview of induction motors signature analysis as a medium for fault detection is presented. The state-of-the-art and development tendency of this field are also shown. The frequency components of the stator current are theoretical and experimentally investigated. The frequency components induced by abnormal electrical conditions, such as breakage of rotor bars and end rings, short circuit of stator windings are analyzed. The stator current high-resolution spectral analysis is proposed for traction motor early defect detection.Rolling element bearing failure account for a large majority of the mechanical faults in a locomotive traction system. Motor bearing failure induces vibration, resulting in the modulation of the stator current. According to the effect the fault has on the measurable machine parameters, traction motor bearing failures are categorized as either single-point defects or generalized roughness in this research. Both types of faults directly affect the machine vibration, and the effects reflected into the stator current, although there they are typically very subtle. The relationship between the vibrational and current frequencies is described. The analytical model for the influence of rolling-element bearing faults on induction motor stator current is illustrated. Consequently, a framework for bearing fault detection by means of motor current signature analysis is presented.Wavelet transform offers an efficient decomposition for signals containing both transient and non-stationary components. The dual-tree complex wavelet transform (DTCWT) is known to exhibit better shift-invariance than the conventional discrete wavelet transform. The characteristics of DTCWT are illustrated using computer-generated signal. As a result, fine frequency resolution may be achieved and fault feature can be extracted successfully via DTCWT analysis of motor current.Based on the investigations mentioned above, a novel fault diagnostic system for locomotive traction motors was developed. The proposed system was effectively applied to HXN3 type locomotive on-line running tests, and the faults of traction motors were diagnosed successfully.