Research On The System Of Pipeline Leak Location Based On BP Network

With the rapid development of pipeline transportation, pollution of the environment and economic loss caused by leaking accident of transportation pipeline are more serious. According to the equipment aging, corrosion and oil stolen, leakage accidents in transportation pipeline occurred frequently in recent years. For the inaccurate position precision and low efficiency of traditional detection technology, an automatic system of pipeline leak location is necessary to be designed. Pipeline leakage detection is a multi-field and cross-subject thesis which involved many subjects, such as sensor technology, signal detection and signal processing. In this paper, the leakage detection technology is studied based on Neural Networks.This paper designed a system of pipeline leak location based on BP Neural Networks about water supply pipeline. Real-time data collection, condition monitoring, and leak point location could be realized by this system.Both home and board pipeline leakage detection technology was summarized in this paper, meanwhile, the principles, difficulties and solutions of pipeline leakage detection technology were described in detail. Neural Network Technology and Wavelet Packet Algorithm were also introduced. Noise elimination and feature vector extraction of pressure singal were based on application of wavelet packet. Using BP network to recognize pipline work station, system could locate leak point as feature vector was regard as input. An acquisition card was designed based on single chip computer; data analysis program was complied by MATLAB, user-interface was designed by Visual Basic. At last, real-time monitoring experiment was complished in the lab; afterward the experimental results were well analyzed. The result shows that leak point can be located accurately, by the system.