An Intelligent System to Classify Leaks in Water Distribution Pipes

in water distribution network are generally detected and located using conventional methods that is not able to give previously any idea about leak size. Current paper draw proposal plan to discriminate leaks according to its size, using a modular system in four blocs: 1) Signal preprocessing, 2) Representation using modified Mexican hat wavelet 3) Dimensionality reduction using random projection combined to principal component analysis and 4) Classification with artificial neural networks.