Study Of The Gearbox Fault Diagnotic Method Based On MATLAB And LabWindows/CVI

With the equipment becoming more and more complicated and automatic, the losses caused by the fault of gearbox to the production and society are increasing, and more likely to endanger to life safety of operation staff, therefore gearbox fault conditon monitoring and fault diagnosis is of great practical significance.Based on Matlab and LabWindows/CVI software platform, a set of fault diagnosis system based on vibration signals of gearbox is developed. Followings are the main research content:Firstly, the typical fault types of gearbox have been introduced, a variety of signal analysis methods are given (digital filter, time domain analysis, frequency spectrum analysis, ZOOM FFT, wavelet analysis, wavelet packet analysis, etc.), and the feature datas were extracted.Secondly, neural network has been used to diagnose the fault type of gearbox, fault diagnosis ability of BP neural network and RBF neural network were mainly studied by using the simulation signal.Thirdly, support vector machine has been used to diagnose the fault type of gearbox. three parameter optimization methods were proposed:CV (cross-validation method), GA (genetic algorithm) and PSO (particle swarm optimization algorithm), and its superiority in dealing with small samples has been proved.Fourthly, the functions of diagnostic test platform was verified by using the gearbox fault database, and the reliability and practicability of the test platform was evaluated.