Development of Efficient Model Predictive Control Strategy for Cost-Optimal Operation of a Water Pumping Station

Considering time-of-use electricity pricing, the optimal scheduling problem of a pumping station is reformulated into a control sequence (CS) optimal scheduling problem, for which a reduced dynamic programming algorithm (RDPA) is proposed to obtain the solution. It is shown that the RDPA allows a reduction of the operational cost by about 60% compared to a basic conventional control strategy, in the example investigated. The fast computation feature of the RDPA facilitates the implementation of a model predictive control (MPC) strategy. In the simulations, RDPA within the MPC structure is found to provide robust control and a marginally increased operational cost, given a +/- 10% inflow rate uncertainty and a modest stochastic rainfall variability (up to 20%).