Water Distribution Network operation optimization: an industrial perspective

A significant portion of the operating costs associated with drinking water distribution networks is related to energy usage, which is mostly employed to drive pumps. One strategy to improve energy efficiency and to reduce energy cost is to operate water pumps in an optimal manner that allows a reduction in energy consumption. This produces also environmental benefits, since decreasing energy consumption contributes to the reduction of the associated greenhouse gas emissions, helping utility providers to reduce their carbon footprint and to reach sustainability goals. However, numerical optimization of water distribution network operation is a difficult problem to solve, given the combination of non-linear hydraulic dynamics and the presence of discrete decision variables, corresponding to pumps and valves having on/off or open/closed characteristics. In this work, we address such problem from an industrial perspective, reformulating the mathematical program that is at the core of such operation optimization solutions using complementarity constraints to transform the resulting mixed integer nonlinear program into a nonlinear program having only continuous variables. This allow us to obtain a tractable optimization problem, that could be solved in a short amount of time even for large-scale water networks, making it compatible with industrial implementation and real-time optimization.