Stochastic dynamic programming for unified short- and medium-term planning of hydro power considering market products

In a deregulated market, the operator of a hydro storage power plant has to consider different market products as well as uncertainties in its prices for optimal scheduling. A suitable mathematical framework for such problems is stochastic dynamic programming. The critical aspect there is modeling, since efficient solution techniques are well known. We present a modeling approach, where the actual available electricity market products act as the basis for the modeling. Advantages of this modeling approach are the natural consideration of power Future products and Hourly Price Forward Curves. Further the model is capable of unifying the short and medium-term optimizations. A two-stage mixed-integer stochastic program with variable time periods is obtained, with the first stage deciding on the amount of power Futures to bid and the second stage as hourly recourse action on the day-ahead market. The problem can be formulated dynamically so that it can be efficiently solved in parallel.