The Research Of Fault Identification And Intelligent Diagnosis For Rotating Machinery

Rotating machinery internal structure is complex, involving more parts. The failure in rotary type mechanical parts accounted for 70% of overall equipment failure ratio, and the rotating machinery fault caused by the working condition of gear problem only accounted for 60%. The vibration signal of running rotating machinery are periodic, non-stationary, the fault diagnosis efficiency and fault extraction of the traditional analysis method are lower. How to effectively extract information and diagnosis fault are particularly important. In view of this, this article main research work is as follows:First of all, introduce the reasons of fault identification as the example of the common fault types of rotating machinery and the gears vibration mechanism. Introduce the equipment required to build the fault test bench of the transmission and the reducer, and introduce the common analysis of vibration signal.Secondly, introduce the EMD algorithm and the relationship of EEMD algorithm and EMD algorithm. In the situation of fuzzy entropy, extracte the fault feature vectors, and diagnosis the fault of transmission.Thirdly, introduce the endpoint effect causes in the EMD algorithm, and the IMF components caused by abnormal situation is formed in the endpoint problem. Study the application of GRNN algorithm and that applied in the endpoint prediction, and study the BLCC and optimization method. The two methods integrated in the EMD algorithm. Compared with nomal EMD algorithm, the new method has the advantage in the optimization of gearbox fault data.Then, introduce the wavelet analysis theory and the lift wavelet theory, studies the principles of adjacent coefficients of noise reduction method. Fuse the method in the lift wavelet decomposition. The research method is applied to the experimental data of the reducer fault, and the fault characteristic frequency in the reducer of easy extraction.Finally, introduce the principle of SVM, the linear support vector machine principle and nonlinear support vector machine principle. Introduced the choice of kernel function of support vector machine, and the genetic algorithm was used to obtain the parameters of kernel function advantage. SVM has good effect in small sample classification. Apply the aboved algorithms to data of rotating machinery, and use GA-SVM to intelligent diagnosis.