No abstract

Leakage detection is an important task to ensure the operational safety of water distribution networks. Leakage characteristic extraction based on high-frequency data has been widely used for leakage detection in experimental networks. However, the accuracy of single-feature-based methods is limited by the interference of background pressure fluctuations in networks. In addition, the setting of leakage diagnosis thresholds has been insufficiently studied, but influences leakage detection performance greatly. Hence, a new method of leakage detection is proposed based on multi-feature extraction. The multi-features of leakage are composed of instantaneous characteristics (ICs) and trend characteristics (TCs), which constitute comprehensive leakage information. The levels of the instantaneous and trend pressure drops in background pressure fluctuations in network environments are quantified for the setting of leakage diagnosis thresholds. In addition, ICs and TCs are used for leakage degree prediction. The proposed method was applied to an experimental network. Compared with the single-feature-based method and the cumulative sum (CUSUM) method, the proposed method achieved increases of 6.01% and 13.66% in F-Scores, respectively, and showed better adaptability to background pressure fluctuations in complex network environments.

Pressure sensor placement is critical to system safety and operation optimization of water supply networks (WSNs). The majority of existing studies focuses on sensitivity or burst identification ability of monitoring systems based on certain specific operating conditions of WSNs, while nodal connectivity or long-term hydraulic fluctuation is not fully considered and analyzed. A new method of pressure sensor placement is proposed in this paper based on Graph Neural Networks. The method mainly consists of two steps: monitoring partition establishment and sensor placement. (1) Structural Deep Clustering Network algorithm is used for clustering analysis with the integration of complicated topological and hydraulic characteristics, and a WSN is divided into several monitoring partitions. (2) Then, sensor placement is carried out based on burst identification analysis, a quantitative metric named “indicator tensor” is developed to calculate hydraulic characteristics in time series, and the node with the maximum average partition perception rate is selected as the sensor in each monitoring partition. The results showed that the proposed method achieved a better monitoring scheme with more balanced distribution of sensors and higher coverage rate for pipe burst detection. This paper offers a new robust framework, which can be easily applied in the decision-making process of monitoring system establishment.