WATER LEAKAGE DISCRIMINATION IN WATER PIPELINES FOCUSING ON DETERMINISTIC PROPERTIES

Abstract

 Leakage of water pipelines lead to regional water reduction and outages, secondary damage such as water supply shortage and road collapse. Monitoring using sensors and networks is expected for on-line leak detection, because it enables automatic measurement and transmission to the server. In this study, we focus on the difference of deterministic properties between the background noise and the water leak sound. Furthermore, we used chaos analysis to determine background noise and water leak sound by quantifying the deterministic property, and demonstrate about the difference depending on the pipe type also. The data used in the experiment are the actual background noise and water leak sound measured on an outdoor test course. In order to visualize the characteristics of the water leak sound, it is necessary to convert time series data into geometrical data. After checking whether it is a stochastic data or not, the sound is determined based on the strength of the deterministic property. From the result, the deterministic property of the leak sound is stronger than the background noise, it is clarified that the deterministic property is a useful index for leak detection.