COMPOUND FAULTS DIAGNOSIS FOR A CENTRIFUGAL COMPRESSOR USING REDUNDANT SECOND GENERATION WAVELET PACKET WITH SLIDING WINDOW DENOISING
Compound faults of rotating machinery always take place synchronously and affect each other. As a result, it is a challenging task to separate and extract compound fault features form vibration signals effectively. In this paper, a new method based on redundant second generation wavelet packet (RSGWP) with sliding window denoising is proposed for compound faults diagnosis. The good time-frequency multiresolution property of RSGWP may preferably decouple compound fault features to different frequency bands and its waveform is beneficial to extract feature components of vibration signals. Otherwise its sliding window denoising piecewisely disposes detail signals based on N kinds of the rotational frequency and local threshold is chosen according to signal features and noise intensity in every sliding window. It overcomes the lack of traditional threshold denoising and better extracts weak fault characteristics in each rotating period. The proposed method is applied to the compound faults diagnosis for a centrifugal compressor. Analysed results show that it is effective to extract and identify a large area friction and crushing of bearings, multi-site tooth breaking and some other compound faults. This method helps to accurately estimate fault positions and shorten the maintenance time.
Redundant Second Generation Wavelet Packet (RSGWP) Sliding Window Denoising Compound Faults A Centrifugal Compressor
Jing Yuan Yanyan Zi Zhengjia He Zhen Li Wei Cheng
School of Mechanical Engineering, State Key Laboratory for Manufacturing System, Xian Jiaotong University, Xian710049
国际会议
北京
英文
1908-1915
2008-10-27（万方平台首次上网日期，不代表论文的发表时间）