Optimal pump scheduling by linear programming

Energy is one of the largest water utility expenses. This cost is greatly associated with the pump operation. An optimum pump schedule can reduce the cost significantly while maintaining hydraulics in an acceptable range. In this paper, the pump station relationships are linearized using relationships among energy required, pumping flow, demand factors and tank storage or tank water levels. A linear programming (LP) optimization problem is formulated and solved for a single tank system for the optimal pump schedule to minimize energy cost. This LP model can be solved very quickly for a "near optimal" solution. Available online data can be then incorporated to update the schedule each hour in real-time. This paper describes preliminary methodologies that can be used to optimize the pump schedule and reservoir control. Future work will extend the approach to multiple tanks and pump station systems, better pump curve representations, and spatially variable demands.