Research on partition strategy of an urban water supply network based on optimized hierarchical clustering algorithm

Abstract

Abstract The partitioning of the urban water supply network can significantly enhance water supply quality. Nonetheless, the bulk of the recently deployed partition approaches overlooked the question of whether the district's fluctuation regulation of flow data is consistent. When the district is modified, it most likely leads to an increase in pressure at a node. To tackle the problem, the flow data from a city's water supply network was evaluated in this article. The random forest approach was also used to extract time-domain characteristics from flow data, and the water supply network split was optimized using the random forest-hierarchical clustering (RF-HC) strategy. Finally, the results were examined and compared. The results suggest that the RF-HC-based water supply network partition technique can better meet the aim of consistent flow changes in the district, as well as offer a theoretical foundation and technological support for the optimal dispatch of press concerning the water supply network.