A Multivariate Statistical Model of Water Leakage in Urban Water Supply Networks Based on Random Matrix Theory

In this paper, model construction of random matrix theory is carried out to analyze water leakage urban supply pipeline network, and a multivariate statistical network based on proposed. This paper presents three-level DMA, short-term demand prediction method use patterns, aiming solve the problem inaccurate caused by climate social factors. Correlation analysis factor main factors affecting daily changes. The extracted information influencing historical monitoring data are used establish Elman model. At same time, consumption pattern was correlation analysis, matched with predicted day, combined expected value for momentary allocation, which improved accuracy demand. precise location real-time hydraulic under conditions established, GA NSGA2 designed, nondominated ranking algorithm elite strategy achieve accurate using calibration search. uses estimation A laboratory case pipe built simulate in residential areas, operating confirm that can estimate leakage. applicability wavelet transform calculation compared an engineering example. example confirms amount calculated has less relative error more similar trend than algorithm.