Research On Fault Self-Diagnosis System Of High-Power Tractor Based On Optimized Neural Network

In recent years,with the acceleration of China’s urbanization and the successive introduction of policies such as land transfer,agricultural production has developed in the direction of large-scale and intensified,which has driven the development of agricultural machinery in the direction of high-power and large-scale.At the same time,with the popularization of digital and intelligent technology in agricultural equipment,agricultural machinery is becoming more and more complex and difficult to operate.If a malfunction occurs,it is also more difficult to repair.In addition,due to the characteristics of crop growth,the seasonality of agricultural production is more prominent,which leads to concentrated production and harvesting time.During this period,agricultural machinery is used for a long time and high load,which undoubtedly increases the failure rate of agricultural machinery equipment.Once agricultural machinery equipment fails,it will cause agricultural production to stagnate and cause greater economic losses.As the main power machinery,High-power tractors are used more and more widely.China is already a big country in the production and use of tractors.Based on this,this paper analyzed the fault self-diagnosis system of High-power tractors.The main research contents are as follows:1.The CAN bus of high power tractor and the principle of SAEJ1939 protocol were analyzed.Firstly,the CAN bus technology in tractor was briefly introduced and its principle was analyzed.Then,the SAEJ1939 communication protocol,which is widely used in construction machinery,was introduced through the high-level protocol of CAN bus,and the message format and communication mode of J1939 protocol were introduced in detail.Finally,the parsing process was analyzed according to SAEJ1939 protocol principle.2.The fault diagnosis algorithm of High-power tractor was analyzed.As the basis of this study,BP neural network has some problems,such as slow convergence speed and easy to fall into local minimum.In this paper,particle swarm optimization algorithm was introduced to optimize BP neural network.Then,in order to further improved the diagnostic performance of the network,a competitive multi-swarm cooperative particle swarm optimization algorithm was used to optimize the network based on the standard particle swarm optimization algorithm.3.The performance of four different fault diagnosis algorithm was analyzed.In order to verify the optimized network performance,this paper took Weichai WP6 diesel engine as the research object,and analyzed the diesel engine under five working conditions: normal,low oil pressure,inlet pipe blockage,high pressure oil pump failure and piston ring fracture.The results showed that the recognition accuracy of BP neural network optimized by the competitive multi-population cooperative particle swarm optimization algorithm was the highest,reaching 94.8%.4.According to the established fault diagnosis model of High-power tractor,the fault self-diagnosis system of High-power tractor was programmed.Firstly,the system architecture was designed.Remote fault diagnosis was adopted to store and call CAN message of High-power tractor using MYSQL cloud database.Then,Raspberry Pi 4B,2-CH CAN HAT module and Shift-EC20 4G module were selected to build the data acquisition module.Finally,the fault self-diagnosis system of High-power tractor was programmed by Lab VIEW,which included user login module,data flow display module and fault diagnosis module.5.The new type of KAT2204 High-power tractor made by Xuzhou Kaier and the hydraulic dynamometer YP660 were used to carry out the loading experiment of the tractor.By carrying out experiments the idle speed,medium load and full load working conditions of the High-power tractor were simulated.The data flow display module and fault diagnosis module of the fault self-diagnosis system were verified by experiments under these three working conditions.The results showed that the fault self-diagnosis system of High-power tractor can display the parameters of the tractor accurately and identify the fault of the tractor effectively.