Models for short term scheduling of hydro-thermal resources in a liberalized electric energy market

In this paper we develop two decision support procedures for the short-term hydro-thermal resource scheduling problem of a power producer operating in a liberalized market. These procedures, based on mixed integer LP models, determine the unit commitment of thermal units and the production levels of committed thermal units and available hydro plants in each hour so as to maximize profits, while satisfying constraints describing the hydro system, the thermal system and the market. The thermal system is modelled in great detail as it allows start-up and shut-down manouvres in every hour of the planning horizon, taking into account minimum up-time and down-time constraints as well as ramp-up and ramp-down constraints. In both models the annual scheduling decisions are given (optimal maintenance plans of hydro and thermal plants, optimal weekly discharge of seasonal basins) as well as forecasts on basin natural inflows. In the first model the power producer is assumed to be a price taker: energy prices do not depend on his own production decisions, i.e. are exogenous to the decision model, and the optimal schedule is determined on the basis of price forecasts. In the second model the power producer is assumed to be a price maker: assuming that competitors’ supply curves can be estimated, the model takes explicitly into account the Market Operator activity of choosing the cheapest bids on the basis of the aggregated supply curve and the hourly energy prices are therefore endogenous variables.