New method to estimate leakage rates in drinking water distribution systems

摘要

Leaking pipes are a serious problem for many water utilities. This paper presents a promising new method to detect the magnitude of leakage in branching pipes of a drinking water distribution system. It is assumed that nearly continuous measurements of pipe flow rates are available for a certain period of time (say at least one hour). The basic procedure involves computing various sample statistics from the set of measured flow rates as these flows are truncated progressively from the bottom. Results yield trajectories of the mean, variance, minimum and maximum of the measured pipe flow rate versus the level of data truncation. In the presence of leaks, the trajectories of the sample statistics diverge from their expected theoretical path when plotted on a standardized graph. The point of departure on the graph indicates where the standardized truncation threshold matches the network leakage rate. Computed trajectories of the sample statistics are compared against theoretical predictions derived from a mixed truncated standard normal distribution. An example demonstrates application of the proposed leak detection method to a real pipe network.