Leak isolation based on Extended Kalman Filter in a plastic pipeline under temperature variations with real-data validation

Abstract

The present work is motivated by the purpose of considering a more realistic scenario than in former studies on the problem of leak isolation within a plastic pipeline, when the water can be affected by temperature changes. In order to address this situation, a state observer approach based on a model including temperature effect and an Extended Kalman Filter is proposed. Noting indeed that temperature affects some equivalent straight length of the pipe, which is used in the model, the observer estimates it together with the leak coefficients. This approach only considers head pressure and flow rate measurements coming from pipeline ends, likewise, water temperature measurement at upstream tank. Results with real data obtained from a pipeline prototype are shown in two different ways in order to illustrate the performance of proposed leak isolation system, as compared to the traditional approaches found at literature.

Online leak diagnosis in pipelines using an EKF-based and steady-state mixed approach

“…the estimation of̂, the accuracy was improved by 0.69% with respect to the use of the EKF alone, which represents an important improvement with respect to the reported method in Delgado-Aguiñaga et al (2015). Table 2 summarizes the results on the accuracy of the leak location; errors are presented as percentages of pipeline length.…”

Distributed state estimation for leak detection in water supply networks

“…To estimate the leak parameters z l and λ, a discrete time extended Kalman filter algorithm is applied to nonlinear model (5). To do that, this model is discretized by using the Heun s method [14], [17]. In this method, the solution for the initial value problem:…”