Pipeline Leakage Detection And Location Based On Extended H_? Filter And Unscented H_? Filter

Pipeline leakage detection and location has been a hot topic in the research of the pipeline.At present,many domestic pipelines have been running for a long time,which lead to leakage accident occurring frequently due to corrosion,aging,poor management,man-made damage and other factors.It brings a great threat to the safety of people's lives,property and the protection of environment.Therefore,finding the accurate location of pipeline leakage timely has a great significance.In general,pipeline leakage detection and location methods are divided into two categories based on hardware and software detection methods.A novel pipeline leakage detection and location method based on extended H_? filter and unscented H_? filter is proposed in this paper.By now,literatures on Kalman filter applied to pipeline leakage detection and location are more,but papers on H_? filter applied to pipeline leakage detection and location are few.The main work of this paper has the following four parts:1.This paper establishes the state-space model of oil pipeline containing leakage,and introduces the principle of the oil pipeline leak detection and location methods.It simulates the pipeline fluid parameters,such as pressure and flow,both at steady state and leakage.After adding noise,this paper does the same simulation.This paper analyzes the influence of noise on the leakage model state simulation results.2.This paper proposes an extended Kalman filter(EKF)based on the standard Kalman and extension principle,simulation is given to verify the EKF algorithm.EKF need calculate the derivatives and Jacobian matrix,which leads to a large amount of computation.The approximating linearization can produce truncation errors and lead to the inaccurate estimation of results which makes the precision be low.Unscented transform is a kind of nonlinear processing methods,an unscented Kalman filter(UKF)based on unscented transform is designed for pipeline leakage detection.Simulation is given to verify the UKF algorithms in the case of Gaussian and Non-Gaussian noises.3.This paper proposes an extended H_? filter(EHF)based on the principle of extension.Kalman filter evaluates the poor effect for the interference of systems with uncertainty and unknown noise,H_? filter has good robustness when the noise energy is bounded.Trying to verify the simulation effect of EHF algorithm used in oil pipeline leakage detection and location.4.This paper proposes an unscented H_? filter(UHF)based on the principle of unscented transform.UT transform can process the nonlinear model,thus UHF can keep the model accuracy and improve the precision of filter effect when applied to pipeline leakage detection and location.This paper simulates the UHF algorithm to verify the detection and location effects in the case of Gaussian noise and non-Gaussian noise.