Use of Supervisory Control and Data Acquisition for Damage Location of Water Delivery Systems

Abstract

Urban water delivery systems can be damaged by earthquakes or severely cold weather. In either case, the damage cannot easily be detected and located, especially immediately after the event. In recent years, real-time damage estimation and diagnosis of buried pipelines attracted much attention of researchers focusing on establishing the relationship between damage ratio (breaks per unit length of pipe) and ground motion, taking the soil condition into consideration. Due to the uncertainty and complexity of the parameters that affect the pipe damage mechanism, it is not easy to estimate the degree of physical damage only with a few numbers of parameters. As an alternative, this paper develops a methodology to detect and locate the damage in a water delivery system by monitoring water pressure on-line at some selected positions in the water delivery systems. For the purpose of on-line monitoring, emerging supervisory control and data acquisition technology can be well used. A neural network-based inverse analysis method is constructed for detecting the extent and location of damage based on the variation of water pressure. The neural network is trained by using analytically simulated data from the water delivery system with one location of damage, and validated by using a set of data that have never been used in the training. It is found that the method provides a quick, effective, and practical way in which the damage sustained by a water delivery system can be detected and located.