Electric Locomotive Fault Analysis System Based On Sensor Network And Distributed Diagnosis

SS4B electric locomotive,as a main type of heavy-haul freight,has harsh operating environment and heavy load,and the current SS4 B locomotive has a long service life,so it is very prone to locomotive failure.Electric locomotive is a complex electromechanical composite structure.When a fault occurs,it is difficult to correctly diagnose and find the cause of the fault only by relying on manual experience.Therefore,this paper analyzes the common failure modes of SS4 B heavy-duty freight locomotives,and designs a locomotive failure analysis system.Aiming at the common faults of the locomotive electrical system,a fault diagnosis method is proposed.At the same time,the problems of data acquisition and data compression in the diagnosis system are studied.Firstly,the fault mechanism and diagnostic requirements of grid-side harmonic current,traction rectifier diagnosis and speed sensor are analyzed.Combined with the operating characteristics of SS4 B heavy-haul freight locomotives,the design of the fault analysis system for electric locomotives is completed,and the data of the diagnosis system are presented.Data compression scheme during acquisition and transmission.Secondly,according to the characteristics of large data collection and fast data transmission in the diagnosis system,the method of data compression is used to realize the data processing.The effects of three lossless compression algorithms including Huffman algorithm,LZ77 algorithm and LZW algorithm are analyzed and compared.The locomotive analog signal is collected by the method of compressed sensing,and its compression and noise reduction functions are studied.The analysis selects the appropriate sparse basis,measurement matrix and reconstruction algorithm,and completes the acquisition and compression of the locomotive signal.In addition,the failure mechanism of grid-side harmonic current,traction rectifier and speed sensor is mainly analyzed.The grid-side current harmonics during the operation of the locomotive are detected and analyzed,and the grid-side current harmonic characteristics under normal conditions and fault conditions are compared,and the gridside current harmonic analysis results are used to assist the fault diagnosis of the traction rectifier.The fault mode of the traction rectifier rectifier device when the open circuit occurs is analyzed,the fault features are extracted by the output voltage,and the fault diagnosis of the traction rectifier is realized by the method based on FPA-SVM.The fault types of the DF16 photoelectric speed sensor are studied,and an on-line diagnosis method based on RBF neural network is proposed to diagnose the speed sensor’s faults such as missing pulses and exceeding the duty cycle.Finally,the software and hardware design scheme of the on-board diagnostic system is proposed,and the hardware circuit of the on-board acquisition node is designed;the software of the on-board intelligent diagnostic terminal is optimized.A simulated fault diagnosis system was built,and the data acquisition and transmission functions,as well as the software functions of the vehicle-mounted intelligent diagnosis terminal,were verified.The diagnosis of the speed sensor and the effect of compressed sensing are verified by simulation.There are 132 figures,15 tables and 70 references in this thesis.