摘要
The number of electric vehicles (EVs) has been growing rapidly in recent years, and if their potential as mobile energy storage is fully utilized, they could provide substantial support to the power grid. However, effective aggregation regulation faces significant challenges due to the heterogeneity of individual EVs. To address this issue, this paper proposes two Vehicle-to-Grid (V2G) power flexibility quantification strategies based on the approximate inner-box approach: conservative and aggressive strategies to assess the flexibility capability of EVs in future time periods. To address the long computation times required for solving the power flexibility of large-scale EVs, this paper develop a deep learning-based surrogate model to replace traditional optimization methods, significantly enhancing computational efficiency. Finally, the proposed method is validated using real data, with experimental results demonstrating its significant effectiveness and superiority.
