The Simulation Research Of Leaking And Locating For Petroleum Pipeline

The detection and location of pipeline leakage is one of the important researching tendency in engineering area. Although across several decades of developments, all methords still have not been improving,at present there all many problems need to be solved, the chief one is to decline the error alarming rate.This paper use the methord of the wavelet transform and neural network detecting the leakage of pipeline,sent some datas nearby minimum in noise reduction signal to network to train,consist leakage recognization base of data in order to judge the leakage ,this methord improves the accuracy of negative pressure wave detecting.The paper introduces some pipeline leak detection technologies in our country and others, studies the meliorated Negative Pressure Wave detection technology and the correlative theory. Analysis is done on the change rules of liquid's temperature along with the pipeline's distance. This paper points out the defects of the const pressure wave velocity method and do some improvements on it rebuilding the expressions of the pressure wave velocity in accordance with the Simpson's rule. Combining with the time difference value of the pressure wave getting to both ends of the pipeline, adopts half divide search method to search the leak site of the pipeline.A means is presented which can eliminate the dynamic response time diversity between pressure sensors in the beginning and end of the pipeline, by adopting Global Position System (GPS) to unify the system time between data collecting systems in two ends.This thesis analyzed and discussed the wavelet transform and leakage detection used in system from principle to centre algorithm realized in computer. MATLAB is chiefly used to process the data and it is a powerful signal processing tool. Analyzing the collecting data contains noise reduction of original data and the location of leakage. At the process of eliminating interrupting pressure wave, I adopted the algorithm of BP artificial neural network, take advantage of ability of automatic organization and learning. Firstly, training the data of leakage opening and closing the valve, transmission in opposite direction in order to remember their character, according to the different character, we can distinguish the interruption of work and pressure wave generated from the leak and judge the leak consequently.