Optimising water supply and distribution operations

This paper presents an adaptive optimisation system for dynamic operational control of water supply and distribution networks. Based on the combined use of an artificial neural network for predicting the consequences of different pumps and valve settings and a genetic algorithm for optimisation, the energy cost minimisation system (ENCOMS) is designed to assist water distribution system operators to select optimal operating control settings that will best meet not only the current demands but also the projected ones. ENCOMS identifies maximum cost savings by taking into account the electricity tariff and other system operational constraints that prescribe lower and upper limits on nodal pressures and service reservoirs storages, alternative supply sources and so on. The real-time control process operates continually and is updated at short intervals by data transmitted from the SCADA system and the updated demand forecast. The control system is generic in the sense that it can be applied to any distribution network. It is customised for each application to ensure optimal performance. It is extremely powerful and flexible so that the peculiarities of the specific network and/or tariff structure can be accommodated, as well as the preferences of the operator with regard to the time interval between updates, the operating horizon and so on. Multiple sources of treated water with different production costs can also be incorporated, as can various pressure zones and demand districts.