Pipeline parameter identification and leak localization using experimental data

Pipeline systems are important in many fields of real life to distribute fluids from one location to another. Leakage from such pipeline systems poses serious problems from the technical, environmental and economic points of view. Early leak detection and localization is, therefore, important for real life applications. In this paper, an experimental study is conducted to collect the pressure head measurements at a number of nodes along a pipeline carrying oil for both the healthy and leaky cases. The experimentally measured data are utilized to identify the leak factor and coefficient of friction of the considered pipeline. The identified parameters are utilized by the propose localization scheme to determine the leak location. The identification is implemented by a window marching technique that uses the collected pressure head measurements and seeking the minimum objective function that represents the mismatch between the measured and numerically modeled pipeline variables. Monte Carlo simulation results are reported to demonstrate the effectiveness of the proposed parameter identification and leak localization techniques.