Reactive power-voltage hybrid model predictive control method for wind-storage power stations

This paper proposes a hybrid model predictive control strategy for wind farms with energy storage. Building on traditional predictive control, it integrates a neural network-based wind power forecasting model and an energy storage system. By predicting active power, the dynamic reactive power capability of the wind farm collection system is assessed. An analytical method using voltage sensitivity coefficients enables collaborative optimization of reactive power regulation, improving voltage control. Coordination with the energy storage system further enhances reactive power control. In different scenarios, the integration of storage increases flexibility, better meet multi-objective management demands and improve voltage stability. Finally, simulations on the IEEE 14-bus system with 20 wind turbines validate the feasibility of the proposed method. Simulation results demonstrate that the proposed method significantly improves system performance, with the maximum reduction in POC voltage deviation reaching 65.5% and the maximum increase in ESS reactive power capacity achieving 58.3%.