Optimal district metered areas design of water distribution systems: A comparative analysis among hybrid algorithms

By subdividing the Water Distribution Networks (WDNs) into District Metered Areas (DMAs), it is possible to regulate pressures, reduce the time it takes to supply water to consumers, mitigate leaks and achieve a uniform distribution of demand. This process of DMAs creation is characterized by two stages: the partitioning of the system and the allocation of control devices. However, improving the management of supply networks is often only possible through the optimal allocation and operation of control devices in the pipelines. In this sense, this study proposes the implementation of different DMAs configurations with objective of costs, pressures, and water age minimization. For this work, hybrid algorithms are developed, consisting of the partitioning algorithms already established in the literature, Fast-Greedy and Louvain, and of the optimization algorithm Particle Swarm Optimization (PSO). The hybrid algorithms were implemented in the Python and Matlab programming languages. In this study, the Modena benchmark network has been used as a case study and the establishing DMAs have been made for single-objective optimizations. The results obtained are compared by analyzing hydraulic criteria, cost, modularity (connectivity) and frequency of valves allocated to the pipes. In general, the hybrid algorithms formed by Louvain algorithm obtained the best results.