Hybrid Evolutionary Optimization/Heuristic Technique for Water System Expansion and Operation

Abstract

This paper presents a methodological solution to The Battle of Background Leakage Assessment for Water Networks (BBLAWN) competition. The methodology employs two constrained multiple-objective optimization problems and is implemented in the context of a software application for the generic hydraulic optimization and benchmarking of Water Distribution System (WDS) problems. The objectives are the combined infrastructure and operational costs and system-wide leakage, both to be minimized. In order to accelerate the evaluation of potential solutions, a distributed computing approach permits multiple EPANET solutions to be evaluated in parallel. A pressure-driven demand extension to EPANET assists the optimization in accurately ranking near-feasible solutions and to dynamically allocate leakage demand to nodes. Pressure Reducing Valves (PRVs) have been located in two ways: a priori, with respect to the optimization analysis and a posteriori after the infrastructure optimization to reduce excess pressure and pipe leakage. The latter demonstrates better overall fitness, leading to optimal configurations dominating those obtained with the former. Several temporal resolutions for PRV settings have been evaluated to contrast the optimal solutions with the computational effort required.