Multi-Objective Optimization of Water Distribution Network Accounting for Pipe Sizing and Isolation Valves Placement

Water distribution networks (WDNs) are critical infrastructures playing crucial role for the development of cities and population health by delivering water to end-users. Ensuring an efficient water supply system represents an important task for the water utilities as compromise solution between system reliability and cost (investment/maintenance) for appropriate pipe diameter sizing and isolation valve placement. Generally, these two tasks are addressed separately, and valve positioning is defined on an already designed WDN. In this work, a novel coupled multi-objective optimization approach is proposed for simultaneously define optimal pipe diameter and optimal valve placement. To solve this problem, the evolutionary genetic algorithm was combined with the hydraulic simulation software EPANET in the Matlab environment. Two objective functions were adopted, the average demand shortfall related to segmental isolation (as a surrogate for WDN reliability) and the total investment cost (pipe and valve costs). The proposed approach is applied to the WDN of the city of Goro (Italy). Design solutions are compared with those obtained by applying traditional approach for the valve placement. Results show the effectiveness of the proposed method in defining more beneficial solutions in terms of total cost, reliability and hydraulic performance.