Simulation-based optimization of pump scheduling for drinking water distribution systems

Abstract

Efficient Water Distribution Systems (WDSs) are crucial for modern society. Their operation requires large amounts of energy with significant financial impact for the utility providers. Existing solution methods are often oversimplified and can only solve the problem for very small, schematic networks. This article studies a pumping scheduling problem for WDSs in which, besides scheduling the operation of pumps over a planning horizon, several constraints regarding hydraulic properties are considered. The goal is to provide a pumping plan of minimum cost that satisfies all demand and respects operational and hydraulic constraints. This work proposes a nonlinear and non-convex formulation as well as a high-performance heuristic. The physical hydraulic behaviour is ensured via hydraulic simulation software. The present method significantly improved the best solutions for several benchmark instances by up to 17%. The solutions also reduce the energy consumed during peak periods, when the electrical grid is most strained.