Abstract:  ABSTRACT：    To smooth minute-level wind power fluctuation and make wind power satisfy the grid standard, this paper proposes a control strategy of hybrid energy storage system （HESS） with the power allotment levels adjusting dynamically based on wavelet analysis. And this paper designs a method of frequency division that the pwer allotment levels adjust according to the result of wind power short-term prediction and the state of charge （SOC） of HESS, which can change the low-frequency target power into power grid, as well as the middle and high frequency power absorbed by battery and super capacitor respectively. Then, this paper establishes a multi-objective optimization function considering the control cost and smoothness of HESS, and applies improved NSGA-Ⅱ algorithm to achieve rolling optimization, calculate the optimal power allotment levels and obtain the optimal control strategy. This paper improves elitism strategy to overcome the shortage of basic NSGA-Ⅱ algorithm in maintaining population diversity, the lack of uniformity and large population size problem. Finally, case studies through the actual observation data of a wind farm in the United States demonstrate that the proposed control strategy takes advantage of different types of energy storage and smooths the fluctuation quickly and economically.

Key words:  hybrid energy storage system（HESS）, smoothing wind power fluctuation, wavelet analysis, rolling optimization, NSGA-Ⅱ