Energy storage systems in wind farms can effectively compensate for the fluctuation of wind energy. How to determine the size of energy storage system in a wind farm is a topic in need of immediate study. A novel method is proposed for designing the optimal size of an energy storage system considering wind power uncertainty. This approach uses the non-parametric estimation method for analysis of the wind power forecast error and the cumulative wind power deviation during the scheduling period. With the knowledge

a cost and benefit analysis model is developed to obtain the optimal size of the energy storage system. A series of wind farm data in California are used as numerical cases

which shows that the algorithm is feasible and successful in designing the optimal size of the energy storage system.