Research On Gear Pump Failure Feature Information Extraction And Diagnosis Based On The LMD And SVM

Gear pump as power hydraulic system. Once the gear pump breaks down, it will lead to the whole hydraulic equipment paralysis, and will bring immeasurable loss. The gear pump is studied as research object in this paper, the content is as follows:(1)The paper analyzes the experiment of noise reduction effect of wavelet transform and dual tree complex wavelet transform, and the results show that the de-noising effect of dual tree complex wavelet transform is better. The decomposition principle of LMD(Local Mean Decomposition) is studied, and the short of this method is analyzed, and the end effecting of LMD and excessive components are resolved by proposing the relevant point correlation-extension method and correlation-selection method. The experiment proves that the improved method is valid. In terms of the advantage of noise reduction effect that based on dual tree complex wavelet transform and the advantage of LMD decomposition that can be obtained more physical meaning of single component signals, the method of DT-LMD(Dual-Tree Local Mean Decomposition) energy feature extraction is put forward. Through the experimental validation of DT-LMD decomposition, the result shows that DT-LMD can more effectively than LMD extract the fault characteristic signal(2)The paper studies the selection of kernel function of SVM(Support Vector Machine). According to the characteristics of the gear pump signal, a simple structure of the kernel function is proposed, and the nuclear parameter is optimized through cross validation. Compared with other kinds of kernel function using the UCI database, the results show that the new kernel function is the highest classification accuracy.(3)This paper adopts the development function of LabVIEW friendly interface and powerful numerical analysis and data processing functions of MATLAB to research and develop the gear pump fault diagnosis system,making full use of LabVIEW’s own MATLAB script node to combine the strengths of the two software together and with the new method realizing the intelligent diagnosis of gear pump.After small sample experiment and testing of the actual data, the fault diagnosis method based on DT-LMD decomposition and the SVM is feasible in theory, can adopt to work in a complex environment for bearing fault diagnosis of the gear pump has important reference significance.