Based On Mass Transit Vehicles To Go Troubleshooting Information Fusion Line System

As one of the most important parts of the urban rail vehicles, the traveling system supports the body of vehicle and transmits the braking force and driving force. The traveling system should be in good condition to ensure vehicle running safety. Nowadays the structure of vehicles become more complex than before, which increases the difficulty of extracting fault feature, and the traditional diagnostic methods are not able to meet the requirements. An information fusion method based on the neural networks and the D-S evidence theory is introduced to detect the faults in the traveling system of urban railway vehicles.First, the main components of the traveling system are introduced, and the wheel and rolling bearings are selected as the research objects. The common failure modes and their causes as well as the characteristics of fault signals are analyzed. The vibration signal and the hotbox signal are used as fault information sources. The basic components of traveling system experimental platform and the performance parameters of the sensors are given.Second, the basic principles of the experience model decomposition algorithm and the BP neural network are presented, and the experience model decomposition algorithm is used to process the vibration signal. The intrinsic mode functions are got and then used as the input of the BP neural network, the experiment is conducted with the data from traveling system experimental platform. The experimental results show that the partial diagnosis model has a good performance when detects the single fault, but has a bad performance when comes to composited faults.Then, the principle of infrared hotbox detection method is briefly introduced, and the temperature law of bearing box is analyzed, and the classification method is used to process the hotbox signal. The threshold is got by analyzing the temperature range of bearing in different failure modes. The experiment is conducted, the experimental result shows that the partial diagnosis model has a good performance when detects the single fault, but has a bad performance when comes to composited faults.Finally, the principles and fusion rules of the D-S evidence theory are introduced, the method to correct the deficiencies of DS evidence theory is proposed and an example is given to certify its effectiveness. The fusion diagnosis model is established. The experimental result shows that the accuracy fusion diagnosis method is higher than the partial diagnosis.