Algorithm Research Of Traction Fault Detection And Property Identification

With the increasing development of electrified railway, high-speed railway will quickly become a major construction project of the railway, the scope of its operations will be more widespread. As an important part of the traction power supply system of the catenary. Because it is usually erected along, and no spare,traction power supply system is relatively weak link, in case of failure, the railway transportation will cause great losses. It is necessary to have more stringent requirements to traction power supply system.In this paper, failure mutation detection of high-speed railway traction network and fault type judgment are the object of study, drawing on previous studies based on the experience, according to some of the problems that exist now, with today’s technology and numerical analysis, design new algorithm. Hoping to improve traction fault detection levels. Based on the above considerations, Tianjin South Wharf substation actual test as the background, combining the characteristics of electrified railway catenary fault traveling wave, this paper carried out the following tasks:(1) Fault traveling wave signal is analyzed by using frequency domain analysis and wavelet packet analysis. Master the wave component and cover the useful signal wave band of high pressure line failure when towing nets.(2) On the basis of spectral analysis and wavelet packet analysis, design three fault mutations detection algorithm, namely: based on wavelet modulus maxima fault mutation detection, based on maxima of wavelet packet fault mutation detection algorithm and derivation techniques fault mutation detection algorithm,and the algorithm is applied to electrified railway traction network fault singularpoint detection research work.(3) Established a simulation system in the experimental system. Analog systems build a simulation model by a transient electromagnetic simulation software ATPDraw, simulated electrified railway catenary on the ground fault circuit and open-circuit fault conditions. Due to interference signal in the high-speed railway traction network fault signal is very much, so use ATP-EMTP software for simulation analysis, and then combined with the three proposed algorithms for fault waveform singular point detection. In actual fault detection, use the actual failure data which is collected from warehouse 217 # feeder Tianjin South substation, use three fault detection algorithm proposed to have a comparative analysis and research.By processing the data can be drawn that the proposed algorithm has a certain academic depth and practical value.(4) Detailed analysis of the transmission line ground fault line wave equation,line fault transient wave propagation process, combined with on this basis wavelet packet energy spectrum analysis techniques to determine the use of the initial traveling wave and the reflected traveling wave polar head and wavelet packet energy spectrum in the energy distribution of the different characteristics of each band in a different fault, design the fault character recognition algorithm,experimental data and actual line data from Tianjin male warehouse substation were analyzed to verify the identification algorithm.