Novel leakage detection and water loss management of urban water supply network using multiscale neural networks

Abstract Due to the acceleration of urbanization, water supply pipe networks often lack the planning. Using the series number of the pipes as category label will result in too many classification categories, and requires many training data to achieve detected accuracy. Therefore, this paper proposes novel leakage detection model based on density based spatial clustering of applications with noise (DBSCAN) and multiscale fully convolutional networks (MFCN) (DBSCAN-MFCN) to manage the water loss. In order to reduce the number of categories, a large water network is divided into a number of zones by the DBSCAN. The zones are used as the learning labels, which reduces the size of the output matrix of the MFCN. Then the leakage detection model is built based on the proposed method to detect the leakage area. Compared with support vector machine (SVM), Naive Bayes Classifier (NBC) and k-Nearest Neighbor (KNN), the accuracy of the proposed method is improved by 78%, 72% and 28%, respectively. Meanwhile, the proposed method can solve the problem of leakage area detection, improve leakage detection efficiency and reduce water loss.