| Rotating machinery account for large proportion in modern machinery equipment. The condition monitoring and fault diagnosis of rotating machinery has become a significant research topic. Mechanical fault feature extraction and fault diagnosis has been the key of the research. Local Mean Decomposition(LMD) with advantages to analyze the mechanical vibration signal is widely used in fault feature extraction. However, there are some deficiencies to be improved. The deficiency and improvement method of LMD is mainly studied, and the pattern recognition methods of fault type and the development and application of fault diagnosis system is researched in this paper.Firstly, the end effect causes of local mean decomposition is analyzed, and an improved method, the maximum similarity coefficient method, is put forward. With comparison analysis of simulation and experiment research, the validity of the method is verified.Secondly, aiming at the difficult extraction of weak high frequency signal, as well as the false frequency of LMD when extract fault feature of rotating machinery, a fault diagnosis method based on differential local mean decomposition(DLMD) is put forward. The simulation verifies the feasibility and effectiveness of the method. With study and analysis of an engineering signal with compound faults, the feasibility of this method in practical application is validated.Then, in view of the pattern recognition of rotating machinery fault types, a rotating machinery fault diagnosis method is proposed, which combine local mean decomposition with sample entropy and fuzzy clustering. The method processes vibration signal of rotating machinery with LMD, and calculates the sample entropy of the product function components, which are used as the feature vector to establish fuzzy matrix for fuzzy clustering analysis and pattern recognition. Then the classification and diagnosis of machinery faults are realized.Finally, combined MATLAB with Lab VIEW, a platform for fault diagnosis of rotating machinery is developed. Using the advantage of graphical programming language of Lab VIEW and the powerful data processing ability of MATLAB, a mechanical fault diagnosis interface is designed and used to process fault data. |
