Optimal Coordinating Polocies of Pumped Hydrostorage Plants in the Presence of Uncertainty

This paper describes a two level stochastic optimization procedure for the computation of optimal coordination strategies of pumped hydrostorage plants with integrated electric power systems. At the first level a quadratic programming procedure optimizes the operation of pumped hydrostorage plants over a period of one day for specified terminal stored energy levels. At the second level a dynamic programming successive approximations procedure determines optimal stored energy levels at the beginning of each day. Uncertainty resulting from generating unit forced outages, electric load and run of the river variations is incorporated at this level. The method has been implemented using sparcity techniques. Results obtained with a realistic 46-generating unit, 11,900 MW peak load system are included. The results indicate that the impact of generating unit forced outages and load uncertainty on storage plant scheduling is substantial. In general, thermal unit outages amplify the economic benefits from storage plants. Also for a given electric power system there exists an optimal total capacity of storage plants. The methodology of this paper is applicable for scheduling other storage plants and can be used for operations planning as well as expansion planning.