Study On Hydraulic System Fault Diagosis Of Hydro-mechanical Continuously Variable Transmission

Hydro-mechanical Continuously Variable Transmission(HMCVT)combining the advantages of hydraulic transmission and mechanical transmission has efficiency and powerful transmission and wide speed range.The design of more gearshifts can realize high speed ratio range.Compared with the traditional mechanical transmission,the structure of HMCVT is complex and its variable transmission is adjusted by the hydraulic system.Therefore,the hydraulic system directly affects the shifting quality of the transmission.Once the hydraulic system failed,the normal operation of the tractor will be greatly affected.It is extremely necessary for the hydraulic system of HMCVT to design a fault diagnosis system,because the hydraulic system works in a closed environment,so that its fault is not easy to intuitive and measure.Through the analysis of the hydraulic system fault,the paper tries to establish a diagnosis model for the hydraulic system fault being based on the fractal theory,the single BP neural network and the genetic BP neural network,respectively.The main researches of the paper are as follows:First of all,the purpose and the significance of the research in the paper were put forward through the review of fault diagnosis technology developments and hydraulic system fault diagnosis.The paper introduced the transmission scheme briefly and highlight introduced the volume speed control hydraulic systerm of variable pump and fixed displacement motor and wet clutch in the hydraulic system,which provide theoretical support for fault diagnosis.The effect of speed control valve flow and the main oil supply pressure on the section quality is introduced to prepare for fault diagnosis feature extraction.Secondly,according to the operation condition of transmission test bench,wet clutch failures which often occur,such as the spring failure,the seal failure and the oil blockage fault are analysed and simulated,and the experimental data is achieved by data acquisition system which is based on Labview.Then,researching the fractal theory,the fractal dimension is applied to hydraulic system fault diagnosis.The research shows that the box dimension has predicton to the fault pattern recognition,but it has no direct effect on the conclusion of identification.Studying on artificial neural network and genetic algorithm,the genetic algorithm is used to reduce the dimension of BP neural network and optimize its weights and threshold.Compared the diagnosis effect of single BP neural network with genetic BP neural network,we come to conclusion that the fault diagnosis recognition rate of single BP neural network is high and the fault diagnosis efficiency of genetic BP neural network is high.Finally,the box dimension as a feature is used in the genetic algorithm BP neural network model to diagnose fault.Found that both the recognition rate and efficiency of fault diagnosis are improved.