Application study on improved wavelet analysis algorithm for pump rotor fault diagnosis

Wavelet Analysis is a new signal analysis theory, but in the way of the engineering application, the wavelet analysis has the problems such as frequency confusion and amplitude frequency offset distortion. This study makes the use of the frequency compensation, an improved algorithm of wavelet packet, and selects the test data of pump rotor typical faults, i.e., dynamic unbalance, rotor bow and dynamic unbalance. It has been proved that the measure of frequency compensation can recognize the fault of pump rotor effectively. This study shows that the improved algorithm can effectively solve the problems such as amplitude frequency, and confuse and distortion offset. This study established a certain foundation for application of wavelet analysis. (authors)