Identification Algorithm for Simultaneous Leaks in a Pipeline

Abstract

This work addresses the problem of simultaneous leaks’ identification in a main pipeline by assuming a constant friction factor. The solution is based on the invariance of the physical leak parameters for any operation point. By considering a quadratic function of the errors between one and two leaks in a pipeline with a steady-state behavior, a new procedure is proposed for the estimation of four parameters associated with the two leaks. In comparison with the algorithm reported in Peralta and Verde (2023), the new one simultaneously adjusts both unknown positions with their gradients. The advantages of the new algorithm are the reduction of maneuvers for obtaining satisfactory estimated parameters and the straightforward estimation of the parameters associated with the orifices through direct nonlinear functions. The results with synthetic data show a considerable improvement regarding the estimation previously reported.