Research On Diagnosis Method Of Pipeline Leakage Based On Fuzzy Model

With the development of national economy, pipeline is becoming more and more important in the field of oil transportation because of the special advantage pipeline is playing in the long distance transportation of crude oil or refined oil. But with the continuous development of pipeline transportation, aging, corrosion, man-made damage, theft and other reasons lead to the frequent occurrence of pipeline leakage, which result in significant economic loss, environmental pollution and security problems. To detect the leakage of pipeline timely and accurately has great social and economic significance for the protection of state property and the natural environment.Based on the theory of fault diagnosis in the pipeline transportation, methods are proposed to detect the oil pipeline leakage by the theory of T-S fuzzy model and generalized fuzzy hyperbolic model (GFHM). The similar negative pressure wave may be generated by the adjustment of the pipeline working condition and pipeline leakage. Therefore, it is prone to emerging error alarms when the leakage detection method based on negative pressure wave is used in pipeline. The source of negative pressure is distinguished by T-S fuzzy model or GFHM in order to prevent error alarms resulting from the adjustment of working conditions. Neural network technology is be used to identify the T-S fuzzy model and GFHM. The main research work is described as follow:Firstly, in the view of the complex conditions and error alarms based on negative pressure wave method, this paper puts forward the T-S fuzzy model method to approximate the pipeline system and eliminate error alarms. The inputs of the T-S fuzzy model are the variable quantities of parameters. According to the output of the T-S fuzzy model, the negative pressure wave is classified to judge that the changes of parameters are generated by the leakage or not. The error back-propagation neural network is set up to identify the parameters of the T-S fuzzy model.Secondly, due to the complexity of parameters identification of T-S fuzzy model, the GFHM is proposed on the pipeline. The neural network of GFHM is established and BP algorithm is used to identify the parameters of GFHM. According to the output of the model the operational status of the pipeline is judged.On account of the transportation of mixing fluid in the pipeline, the leak location algorithm based on negative pressure wave is improved. The traditional method of leak detection and location based on negative pressure wave only considers one kind of situation that the pipeline contains only one medium. However, there is mixed transportation of different fluid, in this case the paper gives the improved leak location algorithm based on negative pressure wave.Finally, the fuzzy model combined with negative pressure wave is used for pipeline leak detection and location. Error alarms that are caused by the adjustment of transfer pump and valve can be excluded. A large number of simulation experiments are carried out with the software of MATLAB and data obtained from real pipeline.