Adaptive Neural Network Control Strategy for Energy Storage-based Wind Power Fluctuation Smoothing

Considering the impact of increasing grid-connected wind power scale on the demand for short-term flexibility regulation capability of power grids, a control strategy using dual battery energy storage systems (BESS) with alternating charge-discharge operation to smooth power fluctuations was designed for wind-storage integrated systems, aiming to extend the cycle life of battery energy storage used for wind power fluctuation smoothing. On this basis, an adaptive neural network controller was proposed for energy storage to smooth wind farm power fluctuations. The controller leveraged the advantage of neural networks in constructing non-mechanistic input-output mapping models for nonlinear systems, addressed the strong randomness of short-term wind power fluctuations, and employed bias correction lock for adaptive adjustment of neuron weights to improve the fast response dynamics of energy storage in smoothing random wind power fluctuations. Simulation results show that the dual BESS alternating operation strategy effectively extends the cycle life of battery energy storage and improves wind power fluctuation smoothing. The designed energy storage controller not only achieves smoothing of short-term wind power fluctuations but also demonstrates good robust adaptability to random wind power fluctuations.