Fault diagnosis based on characteristics of the vibration energy on the hydraulic pump

In order to solve the problem that the fault characteristics of hydraulic pump have ambiguity and signal-to-noise of inspected signals is low, a new method based on characteristics of the energy spectrum of vibration signals combing rough set (RS) and principal component analysis (PCA) algorithm is proposed in this paper. This algorithm processes noise abatement for collected vibration signals by wavelet analysis, and decompose-reconstructs characteristic signals at different spectrum; afterwards, eigenvector is constituted by normalized treatment. In addition, it utilizes PCA method to decouple relativity of characteristics among faults in order to reduce dimension of fault characteristics. At last, it establishes rules knowledge base of fault diagnosis through the method of distinguished matrix of the rough set theory for characteristics vector. Moreover, the paper indicates the validity of the method based on experiment of axial piston pump.