Abstract

In order to optimize the energy consumption in a process of pumping water to a tank, a model predictive control (MPC) strategy for the centrifugal pump speed is proposed, simulated and discussed. The hydraulic system studied consists of two equal diameter pipes interconnecting two tanks by means of a variable speed centrifugal pump. Based on a dynamic model of the hydraulic system, the evolution to N steps ahead of the pump operating point is predicted. Considering an estimation of the water consumption and the constraints on system variables (limits on the storage volume, flow and frequency of the centrifugal pump), by using an online optimization algorithm the necessary speed signal is determined, such that the pump delivers the desired flow rate and hydraulic head, while minimizing the energy consumption of the pump system. Mean values of 13.37% in energy savings of the MPC with respect to conventional proportional–integral–derivative controller (PID) are obtained for tank volume set points. To highlight the result, a system with flow control by means of parallel pumps (PP) was added for comparison. The result can be extended for the case of transporting slurry instead water, updating loss coefficients and assuming that the fluid flow is laminar.