Reservoir optimization model incorporating inflow forecasts with various lead times as hydrologic state variables

This paper presents two Stochastic Dynamic Programming models (SDP) to investigate the potential value of inflow forecasts with various lead times in hydropower generation. The proposed SDP frameworks generate hydropower operating policies for the Ertan hydropower station, China. The objective function maximizes the total hydropower generation with the firm capacity committed for the system. The two proposed SDP-derived operating policies are simulated using historical inflows, as well as inflow forecasts with various lead times. Four performance indicators are chosen to assist in selecting the best reservoir operating policy: mean annual hydropower production, Nash–Sutcliffe sufficiency score, reliability and vulnerability. Performances of the proposed SDP-derived policies are compared with those of other existing policies. The simulation results demonstrate that including inflow forecasts with various lead times is beneficial to the Ertan hydropower generation, and the chosen operating policy cannot only yield higher hydropower production, but also produces reasonable storage hydrographs effectively.