Research On Transient Leak Detection Algorithm For Water-pipeline Based On The Pipeline Model

With the rapid development of China’s economy,pipeline transportation,as a main mode of transportation,is playing an increasingly important role in the national economy,national defense industrial and people’s production and life.Pipeline transportation is widely used because of its safety,reliability,economy and practicality,which has become the lifeblood of national economic development,but it does not mean that they are risk-free.In recent years,due to various reasons such as pipeline corrosion,aging and third-party interference,pipeline leakage accidents occur frequently,which not only lead to waste of resources and huge economic losses,but even cause serious personal injuries and deaths.Therefore,to ensure the safety and reliability of the pipeline infrastructure,detect leakages and confirm leakage points accurately have become a crucial requirement.In order to ensure the safe operation of the pipeline and minimize the harm caused by leakage accidents,it is very important to deeply study the leakage detection technology to raise the sensitivity of detection and accuracy of localization.The main research content of this paper is the study of water-pipeline leakage detection method.This paper proposes a transient leak detection for water-pipeline based on the pipeline model,which realizes the effective leakage detection.This study takes the reservoir-pipeline-valve(RPV)system as the research object.The water hammer differential equations in fluid hydraulics are used to describe the changes of flow and pressure in the process of pipeline flow.The transfer matrix method of frequency domain analysis method is introduced to solve the water hammer equation,and the field matrix of single pipe is derived.The transient hydraulic model of RPV system is established,and then the pressure variation is obtained.In this paper,we propose a transient-based leakage detection method using multiple signal classification-like(MUSIC-Like),which effectively solves the problem of how to locate two leaks whose distance between leaks is less than half the shortest probing wavelength.Drawing lessons from the classical spatial spectrum estimation method to locate the direction of the signal and combining with the transient hydraulic model of pipeline,the spatial spectrum estimation method is used for pipeline leak detection.According to the relationship from the pipeline model between pressure variation and leakage position and pipeline parameters,the optimization problem of the MUSIC-Like leak location algorithm is given,and the calculation formula of locating leaks is solved.In the experiment,the performance of the MUSIC-Like method in the cases of a single leak and multiple leaks is respectively discussed by comparing with three classical spatial spectrum estimation methods.In the single-leak case,the MUSIC-Like algorithm provides precise localization estimation even for a high level of noise.For the multiple-leak case,the MUSIC-Like method is superior to the other three methods.It is capable of identifying all leaks that the leak-to-leak distance is less than half the shortest probing wavelength,nevertheless,the other three methods determine the situation as that of only one leak.Considering that the result of MUSIC-Like leak location algorithm appears more side lobes when the signal-to-noise ratio is low,the side lobe position may be misidentified as the leak position,especially when the number of leaks is unknown.And in the multiple-leak case,the localization performance of MUSIC-Like algorithm is related to the leak locations.According to Bayesian inference theory,this paper proposed a Bayesian inference leak detection algorithm based on transient hydraulic model.The algorithm obtains different competitive models by changing the number and locations of leaks in the transient hydraulic model.The nested sampling algorithm is used for model selection and firstly select which model that is the most appropriate to fit the given measured data from competitive models.Through the process,the number of leaks is first estimated,and then the leak location information are extracted from the selected model.The experimental results show that this algorithm is an accurate tool for leakage enumeration and location estimation.