Combining model predictive control with constraint-satisfaction formulation for the operative pumping control in water networks

Abstract

This paper proposes a method to combine linear Model Predictive Control (MPC), a Constraint Satisfaction Problem (CSP)formulation and a Network Aggregation Method (NAM) for the predictive operational control of water pumping in DWNs. The proposed method can produce optimal pumping strategies for complex DWNs in short computation times, while avoiding the need for non-linear programming techniques to cater for non-linear flow-head equations. The proposed approach is simulated using Epanet to represent the hydraulic DWNs. The D-Town benchmark water network is used as a case study.