Multi-time Scale Optimal Dispatch for the Wind Power Integrated System with Demand Response of Data Centers Based on Neural Network-based Model Predictive Control

Data centers (DCs) are energy consumers with high electricity demand. Due to their Spatio-temporal demand response (DR) capabilities, DCs are crucial DR participants. In view of the difference in real-time requirements and continuity requirements of computing jobs, this work builds a detailed DR model of DCs. To make full use of the latest wind power predictive information, a multi-time scale optimal dispatch based on model predictive control (MPC) is proposed for wind power accommodation improvement and system operating security enhancement. As the real-time optimal dispatch is a quadratic programming problem and the total number of dispatching periods is large, the BP neural network is applied in this work to improve the computation speed. Finally, the proposed model is tested on an IEEE 30-bus power system with wind farms and DCs. Simulation results verify that DCs' DR participation plays an important role in promoting wind power accommodation and system load adjustment. Besides, it is proven that our proposed BP neural network-based MPC method can obtain optimal dispatching results with low computation costs.