Integrated Estimation Method for Leakage Change based on a State-Space Model

With aging pipe infrastructure, leakage management has become a critical issue for the sustainable operation of water supply. Several surrogate indicators and analysis methods are independently used to estimate leakage in a water distribution network. However, their accuracy leaves room for improvement due to their strict assumptions, such as linear leakage growth. This paper proposes a method for estimating leakage change by combining different indicators to achieve higher accuracy and consistency for the combined data. The proposed method is based on a state-space model of leakage change and its observation in Bayesian multilevel settings. A Markov Chain Monte Carlo (MCMC) sampler is used to estimate leakage and other parameters of the proposed model. The proposed method was applied to a case, and it was confirmed that the proposed method reduced the level of error in leakage estimation by half compared with an existing method.