Energy Storage Operation with Wind Uncertainty

Thesis (Ph.D.)--University of Washington, 2017-06With environmentally oriented energy policies, wind power penetration sharply increases in the recent years. Although operating cost is low, wind power's intrinsic uncertainty and variability have brought another layer of operating challenges for the power system operators and system participants. Since traditional operation is not specially designed for these type of characteristics, it motivates us to revise the current operating frameworks to better embrace a more sustainable generation mix. Other than wind energy, energy storage systems (ESS) are another key players in this smart grid era. They provide additional operational flexibility to enhance the whole grid operation and planning. Recently, decreasing installation and operation costs for ESS furthers drives the interests to develop advanced operation decision making processes for grid and market applications. Therefore, in this dissertation, our work focuses to understand how to better integrate renewable generation and energy storage with respect to wind uncertainty. We first improve wind modeling with ensemble approach. The proposed data mining concept provides not only better point forecast but also scenario set and uncertainty region for stochastic optimization. To reduce the computation burden, we compare the scenario reduction techniques and propose a submodular scenario reduction algorithm to further improve the computational efficiency. By optimally selecting representative scenarios, we reduce the stochastic programming solution time without sacrificing solution quality. In terms of ESS, we first stand from the system operator's perspective and analyze the benefits of ESS operation in the centralized scheduling with uncertain wind. Stochastic programming based operation strategies well balance the trade-off between the system economics and reliability with the assistance from energy storage. Moreover, we also stand from the merchant storage owner's perspective to derive bidding strategies to increase profits through participating the energy market. A look-ahead strategy fully explores the arbitrage opportunities across the time and space