The Research And Design Of Heavy Haul Trains Pipeline Leak Diagnosis System

The heavy-haul combined train is an important vehicle of the freight transportation in our country. The air pipe leakage of braking system seriously affects the safety operation of trains. The diagnosis and localization of air pipe leakage is the key problem to be solved. In order to develop a diagnosis system for pipe leakage of the heavy-haul combined train, this thesis captures the real-time pipe status signal and uses the multi-sensor co-location technique.Taking Datong-Qinhuangdao transport line 20,000 tons heavy-haul combined train as the research object, the structure, characteristics and working principle of the braking system are analyzed. By discussing the causes and diagnostic methods of the train pipe leakage, the overall design scheme for the leakage diagnosis system is put forward.Due to the interference in the monitoring environment and the difficulty in diagnosing small leakage position of the train pipe, this thesis comes up with a signal-correlation-enhanced algorithm of capturing signal singularity. First, the pre-filtering and adaptive filtering techniques can be used to filter the disturbance and noise existing in status signals of the pipe. Then the signal can be divided into the high frequency detail and low frequency profile by using the modulus 2 extracted wavelet transform. The signal's self-correlation can be enhanced by the product operation of the adjacent high-frequency signals. Capturing signal singularity is the basis for determining the time of the occurrence of small leakage.Because of the complex structure, lots of controlled source, long control domain of train pipe, we propose a multi-sensor cooperative iterative algorithm, which will accurately locate the concurrent multi-spot leakage. In order to diagnose the leakage of train air pipe, multiple sensors are used in the combined detection of pressure and flow. According to the captured singularity of negative pressure wave signal of long pipe, the relative occurrence time of leakage can be determined. The sampling time of multi-sensor can be synchronized by introducing the mechanism of time calibration. On this basis, the location of concurrent multi-spot leakage can be accurately got by the iterative computation of sensors monitoring information in the overlap region.Finally, this thesis uses the modular design concept. The implementation of software and hardware of the diagnosis system is achieved. The effectiveness of the scheme and diagnostic method is verified.