A stochastic transmission network planning method considering the virtual power grid modeling of spatial load-shifting capability of data center loads

With the increasing penetration of renewable energy and the growing electricity demand of data centers, traditional transmission network planning faces new challenges in mitigating network congestion and facilitating renewable energy accommodation. To address these challenges, this paper proposes a novel concept of the Virtual Power Grid (VPG), which leverages communication networks to enable workload migration among geographically distributed data centers. By using computing task flows to drive the spatial redistribution of electricity loads, the VPG provides a virtual complement to the physical transmission network. Unlike conventional transmission planning approaches that primarily rely on physical network expansion, the proposed VPG introduces communication-enabled spatial load flexibility as an additional planning resource. Building on this concept, a stochastic transmission network planning model is developed that accounts for multiple uncertainties, including conventional loads, data-center workloads, wind power output, and equipment availability. A Benders decomposition-based method is further employed to enable privacy-preserving coordinated optimization between the power grid and data centers. Case studies on the IEEE 24-bus system demonstrate that the proposed VPG can reduce transmission expansion requirements, alleviate network congestion, decrease wind curtailment and total system costs, providing a new technical pathway for transmission network planning under high renewable energy penetration and large-scale computing loads.