Operation of Multiple Pumped-Water Sources with No Storage

Abstract

The cost of pumping is one of the most significant operational expenditures in a water distribution network. When a network has multiple water sources, which are associated with pumping stations, it may be possible to optimize those costs. One way to minimize costs is to determine the optimal flow rate for each pumping station and for every point of the temporal demand curve, while keeping the energy cost minimized. This also requires that the required minimum pressure at the critical point in the network be satisfied. This paper introduces a principle known as the setpoint curve as the key component of the network optimization methodology. A direct search algorithm based on the Hooke–Jeeves approach is tested in two case studies. Two key cost factors, the electric tariff and the cost of water production, are considered for each water supply source. The model also considers the pressure-dependent consumption, which directly influences the setpoint curve. To test the methodology, a software application is implemented using the EPANET toolkit. The two case studies demonstrate the benefits of the approach in developing optimal pump operating policies which otherwise would be difficult to infer.