Combined optimization of pipes and control valves in water distribution networks

Abstract

An optimization model is presented for the combined design of pipes and control valves in water distribution networks. The model uses a genetic algorithm to select pipe diameters and location of control valves, and employs a non-linear programming algorithm to determine the valve settings. The optimal design includes minimization of both pipes and water leakage costs. The optimal valve regulation has the objective of reducing leakage with constraints on minimum pressure heads. The proposed methodology is applied to two networks, with results indicating that the optimal design is dependent on the unit cost of water. An analysis on this parameter shows that if water cost is low, the sizing of the diameters is not influenced by the pressure regulation policy to reduce leakage. As the cost of water increases, the combined optimization increases the pipe diameters, since this magnifies the possibility of pressure regulation by valves.