The Design Of Performance Testing And Fault Diagnosis For Metro Braking System Based On Labview

Metro has already been a popular transportation nowadays. Since the first one was born, the manufacture technology keeps improving continuously. With the development of science and technology, the structure complexity of metro trains is increasing, especially the braking system. Although the capabilities of metro train’s braking system are enriched and improved, but failure of one single part may destroy the whole system, even may cause economic losses and casualties. Hence, the development and innovation of fault detection and diagnosis technology for metro train braking system is of considerable importance.In this dissertation, a system is designed for monitoring metro train braking system and diagnosing faults. This braking performance detecting system is able to acquire the pressure of metro train brake cylinder, the voltage and the current of motor, and the temperature of wheel and brake-shoe during braking, based on which such braking performance parameters as initial braking speed, average deceleration, braking distance can be calculated. The fault diagnosis module designed with a combination strategy of expert system and artificial neural network is capable of warning and diagnosing once faults are detected in metro train braking system. The braking performance detecting system is programmed via Labview, supporting by Access database technique and by Matlab which helps to establish and train the radial basis function neural network (RBFNN). The hardware system is set up with appropriate sensors and data acquisition (DAQ) devices in consideration of the metro train type for testing.The fault diagnosis module adopts the knowledge acquiring method of fault tree and production rules-based knowledge representation. A methodology combining RBFNN and production rules for braking system fault diagnosis is carried out. The braking performance detecting system is introduced in details in aspect of working process and module functions. A series of type tests has proved that it is reliable and efficient that the system designed in this dissertation could diagnose the faults in braking system accurately though data acquiring and calculating.