Numerical Simulation Study On Leakage Location Of Long Water Pipeline With Bifurcation Structure

With the substantial increase in the demand for water resources in production and life,the contradiction between supply and demand caused by water resources is also increasing,and pipeline leakage accidents due to aging,corrosion and human-induced long-distance water transfer projects often occur.Improving the efficiency and management level of water supply projects.Proposing an accurate and efficient pipeline leak detection and location method has become an urgent problem to be solved.This article takes the long water pipeline with bifurcated structure as the research object,the main conclusions are as follows:(1)Using the transient flow model of the bifurcated pipeline considering the non-constant friction,the pressure data of the pipeline at various moments and along the nodes under different leakage amounts,different leakage positions,and different valve closing laws are calculated.The pressure curve at the end of the fork-line valve is more complicated than the simple straight pipe model.By slowly closing the valve,the peak value of the water hammer pressure introduced by the leak detection can be greatly reduced.The downstream valve valve is controlled by the form of cosine.Although the frequency phase spectrum of the end pressure signal is not obvious Linear relationship,but the difference in different leakage positions can also be used as the basis for judging the leakage position.(2)A discrete gas model was used to study the effect of a small amount of gas on the hydraulic characteristics of the leaking pipeline.The results show that the leakage volume and leakage position in the case of two-phase flow can be calculated by analyzing the pressure signal at the valve end.Under different gas content,the wave velocity of low gas content is much larger than that of high gas content.When the gas content is high,no obvious liquid column separation occurs,and when the gas content is low,the first few phases Significant liquid column separation occurred.(3)Use Flowmaster software to establish the corresponding bifurcated structure model and compare it with the programming results.It was found that both methods can perform leak detection by analyzing the change law of the pressure signal at the end of the downstream branch valve.(4)Using the node parameters at each position and time period simulated by the Flowmaster software as samples,the radial basis neural network was used to train and test the samples,and the new data was used to predict the leak location,and the true value and the predicted value were found to be very close.The feasibility of using radial basis neural network to locate leaks in long water pipelines with bifurcations is proved.