A new methodology for optimizing the daily operations of a pumping plant

Abstract

A new methodology for the optimization of the daily operations of a pumping plant has been developed. Two major issues have been investigated in relation to pump operations; maximizing pumping during off-peak electricity tariff periods and minimizing the head against which water is pumped. The new approach presented in this paper incorporates the combination of both a pump schedule and trigger levels to control the system. Different trigger levels are used during different periods of the day, such that peak pumping and pumping head are minimized. Additionally, scheduling is used to maximize pumping in the off-peak period thereby ensuring that the maximum amount of water is available in elevated storage facilities at the start of peak tariff periods. A genetic algorithm that incorporates realvalue and integer decision variables has been developed to determine the best combinations of trigger levels and scheduling for a predicted demand flow for the next day. The methodology has been applied to the Murray Bridge Water Treatment plant in South Australia. The current operating controls based on fixed trigger levels in the upper tank at this plant have been compared with the new combined scheduling/trigger levels operating controls approach. Very promising results have been obtained, with a 20% saving predicted on energy costs during low and moderate demands.