Model predictive control applied to the long-term hydrothermal scheduling of the Brazilian power system

This paper presents a case study concerning the application of model predictive control (MPC) to the long term hydrothermal scheduling of the Brazilian power system. According to MPC, the hydro and thermal generation decisions at each stage are provided by a deterministic nonlinear optimization model considering predicted inflows. The model, which is solved by interior point method, also takes into account tie line constraints between interconnected areas. In order to evaluate the performance of the approach several simulations over historical inflow scenarios were performed, and statistics about operation costs, hydro and thermal generation, power flow interchange, reservoir storage, load shortage, among others, are obtained. General results are compared to those from the stochastic model in use in Brazil and the results have shown substantial decrease in expected operation costs and load shortages, as well as an increase on water storage, both cause and effect of the better management of water resources.