Search

This paper presents a novel algorithm driven by the minimization of the transport function for the partitioning of water distribution networks (WDNs) into district metered areas (DMAs). The algorithm is based on the linear programming (LP) embedded inside a multi-objective genetic algorithm, which enables engineering criteria, such as the minimization of the boundary pipes and the maximization of the uniformity of DMAs, to be considered in the partitioning. Furthermore, the application of the algorithm on the dual network topology based on segments and valves guarantees that configurations of DMAs that respect the real positions of isolation valves for WDN partitioning are obtained. After being described on a small WDN, it is successfully validated on a large size WDN, proving better performance than other algorithms in the scientific literature for the generation of engineeringly appealing DMA configurations, with almost identical hydraulic performance to the unpartitioned WDN. HIGHLIGHTS A clustering based on transport function minimization is proposed.; The dual topology based on segments and valves enables the real valve positions to be considered in WDN partitioning.; The transport function minimization by LP is embedded in the multi-objective optimization to incorporate engineering judgment criteria.; Engineeringly appealing DMA configurations are obtained in the applications to a real WDN.;