When assisting wind farms operation

energy storage systems (ESSs) are constrained by temporal coupling characteristics and capacity limits. Persistent deviations in wind power forecasts may result in insufficient energy reserve capacity

reducing the ESS's ability to mitigate wind power fluctuations. To address this issue

a real-time smoothing method for wind power fluctuations that simultaneously considers power and energy reserves of ESSs is proposed. First

based on the conventional ESS operation model

a power reserve model incorporating six reserve modes is constructed

and an energy reserve model considering temporal constraints is established. Second

the energy reserve constraints are embedded into the real-time optimal operation framework of the ESS

forming a real-time wind power smoothing model that integrates both power and energy reserves

thereby improving the ESS's adaptability to stochastic wind power fluctuations within the prediction horizon. Since the core of model solution lies in determining the capacities of power and energy reserves

a day-ahead optimization model is established and solved via stochastic integer programming to obtain reference reserve values

providing a basis for online smoothing control. Finally

simulations are carried out using actual wind farm operational data and the assessment criteria specified in the two detailed rules of power grids. The results verify the feasibility and effectiveness of the proposed method.