Locating Leaks in Water Mains Using Noise Loggers

Abstract

Because of their potential danger to public health, economic loss, environmental damage, and energy waste, underground water pipelines leaks have received more attention globally. Researchers have proposed active leakage control approaches to localize, locate, and pinpoint leaks. Noise loggers have usually been used only for localizing leaks while other tools were used for locating and pinpointing. These approaches have resulted in additional cost and time. Thus, regression and artificial neural network (ANN) models were developed in this study to localize and locate leaks in water pipelines using noise loggers. Several lab experiments have been conducted to simulate actual leaks in a sample ductile iron pipeline distribution network with valves. The noise loggers were used to detect these leaks and record their noise readings. The recorded noise readings were then used as input data for the developed models. The ANN models outperformed regression models during testing. Moreover, ANN models were successfully validated using an actual case study.