Research On Pipeline Leakage Diagnosis And Localization Based On Modal Acoustic Emission

City gas supply system is one of the lifelines in modern cities. The main transport manner is by pipelines. As the gas pipeline becoming longer and longer, security problems are more and more serious.To improve the detecting efficiency, the real-time ability, the localization accuracy and extend the detecting range, a novel pipeline leakage diagnosis and localization method based on modal acoustic emission is presented. The main contents and conclusions are as follows,1. The techniques of wireless sensor networks and data aggregation are introduced to pipeline diagnosis. The deficiencies of existed detecting methods and the advantages of WSN are pointed out based on the research background and developing actualities of pipeline detecting techniques. According to the characteristic of WSN, acoustic emission is chosen as the detecting method. And a pipeline leakage diagnosis and localization system based on wireless sensor networks is designed and implemented.2. Distributed data aggregation architecture is used in the system. At source nodes the original signals are analyzed in multi-levels by wavelet transformation and the single-mode signals which contain the leakage characters are obtained to eliminate noises and the dispersion nature of the signals. At sink node the signals are divided into several groups according to their average range. The final leakage position is obtained via the principle of cross-correlation TDOA localization and the weighted average of all the results form every group.3. A graphical user interface (GUI) of the Pipeline Leak Detection System based on artificial intelligence methods has also been designed and developed. The acoustic emission signal data aggregation process is realized by using wavelet analysis, neural network and D-S evidence theory in three aggregation levels respectively, including data-level, feature-level and decision-making-level. This system can judge the characteristics of the leakage source effectively.The experiments results show that the multi-nodes detecting manner of WSN can solve the problems of miscarriage of justice and difficult decision-making in the system effectively. At the same time, the localization accuracy is improved obviously by data aggregation in multi-levels.