Mid-Term Stochastic Scheduling of a Price-Maker Hydro Producer With Pumped Storage

A yearly stochastic self-scheduling model for a price-maker hydro producer is presented. In the short term, the producer aims at maximizing profits in the day-ahead market. Residual demand curves model the producer's interaction with his competitors and the load demand. A modification of the residual demand curve is proposed to enable the definition of optimal pumped-hydro bids. Multistage stochastic programming provides hedging against the mid-term uncertainty of inflows, load demand, and competitors' offers. The proposed method models the impact of short-term profit maximization decisions on mid-term scheduling within a compact stochastic mixed integer linear programming (MILP) approach. A test case with a three-stage, 90-scenario stochastic tree and data of the Greek Power System is presented. Results provide insight on both mid-term reservoir management and short-term market-based operation and indicate the feasibility of solving a very large-scale scheduling problem as a unique MILP using a commercial solver. The role of forward contracts in mitigating market power is also analyzed.