ABSTRACT
A methodology based on solving a large-scale nonlinear programming (NLP) problem is presented for the optimal operation of pumping stations in water distribution systems. The objective function is to minimize pumping cost over a planning horizon, and the constraint set includes system constraints, which account for the hydraulics involved in a water distribution system, bound constraints on decision variables, and other constraints that may reflect operator preferences or system limitations. The optimal control problem for a water distribution system is complicated by the fact that the mathematical problem can be very large in the number of constraints, many of which are nonlinear, and the large number of decision variables that are nonlinear. The state of a water distribution system can be defined by a specification of all pipe flow rates or all nodal heads at any given time of the day.
