MPC for the comprehensive energy system considering wind speed prediction of intermittent wind energy

The development of new energy sources is crucial for reducing carbon emissions and promoting the sustainable ecological development. However, the inherent stochastic fluctuations in renewable energy integration pose significant challenges to grid frequency stability. Based on this, a comprehensive energy system is constructed, integrating doubly-fed wind generators, a high-order pumped storage unit, hydropower generating units, and coal/nuclear power units. Within the comprehensive system, a detailed model of the high-order pumped storage regulation system is proposed, along with integrated inertia control with adaptive coefficients in the wind generator. To mitigate the wind-induced disturbances, a new microscale wind speed prediction model, PiTransformer, is developed and embedded within the model predictive control (MPC) framework, enabling the formulation of robust control strategies that account for intermittent wind power. Specifically, in PiTransformer, Gaussian noise is introduced at the embedding layer, a gated linear unit (GLU) is integrated, and the residual structure in the feedforward layer is enhanced through an adaptive scaling factor. Finally, the comprehensive energy system is implemented in the IEEE 39-bus case, demonstrating effective and rapid frequency regulation under random load and wind energy. The frequency regulation capabilities of doubly-fed wind generator and pumped storage are also validated.