Wind power stabilization and control based on decomposition QPSO-LSTM prediction model

Wind power generation has randomness and volatility, which may potentially jeopardize the stability of the power grid. Hybrid energy storage system (HESS) can suppress wind power fluctuations, ensure stable output of grid-connected power generation, and meet grid fluctuations effectively. Based on Quantum-behaved Particle Swarm Optimization (QPSO), Long Short-Term Memory Networks (LSTM) prediction model and State of Charge (SOC) of hybrid energy storage a fuzzy control strategy is proposed in this paper. Using the decomposed data, the LSTM network is optimized using QPSO optimization algorithm to improve the wind power prediction performance. On the basis of the prediction results, combined with the hybrid energy storage SOC, the fuzzy control strategy is used to control the output of HESS to ensure the smooth integration of wind power into the grid. The HESS control strategy is optimized to reduce the overcharge and discharge, and improve the ability of the HESS to stabilize wind power fluctuations. Compared with the traditional strategy, this control strategy reduces the fluctuation range of wind power output under the same configuration conditions, and reduces the dead time of the HESS, and the comprehensive performance of the system is significantly improved.