LOCAL WAVE METHOD & INDEPENDENT COMPONENT ANALYSIS AND ITS APPLICATION FOR ROTOR SYSTEM IN FAULT DIAGNOSIS
MA Xiao-jiang · Proceedings of the CSEE · 2005
A mechanical fault diagnosis method is proposed combining local wave method and independent component analysis. For the plant vibration signal, the local wave time-frequency image can reflect the energy change in time and frequency axis. For different fault signals, the time-frequency images are distinctly different. So it can be used for fault detection. To realize automatic classification for different fault modes, the independent component analysis is applied to extract the independent source images,then the projection coefficients of source images can be obtained. The Probabilistic Neural Network is used to classify for the projection coefficients. The rotor system fault signals (early rub, loose, misalignment) are investigated using the proposed methods. The result shows that this method is, to a certain effect, effective and practical.