A Stochastic Dynamic Programming Approach To Balancing Wind Intermittency With Hydropower

Hydropower is a rapid response energy source and thus a perfect complement to the intermittency of wind power. However, the effect wind energy has on conventional hydropower systems can be felt, especially if the system is subject to several other environmental and non-power use constraints. The goal of this paper is to develop a general method for optimizing short-term hydropower operations of a realistic multireservoir hydropower system in a deregulated market setting when there is a stochastic wind input. The approach used is a modification of stochastic dynamic programming (SDP). The methodology is applied to a representation of multiple projects in the Federal Columbia River Power System, which is currently being dispatched by the Bonneville Power Administration. Currently, studies on hydropower operations optimization with wind have involved linear programming orstochastic programming, which are based on linearity of the objective function and constraints. SDP, by contrast, is a stochastic optimization method that does not require assumptions of linearity of the objective function or the constraints. The true adaptive and stochastic nonlinear formulation of the objective function can be applied to multiple timesteps, and is efficient for many timesteps compared to stochastic programming