Extreme Operating Scenarios Generation Based on Least Squares Loss Function Conditional Generative Adversarial Network

作者：Dan Zhang, Xite Liu, Yan Liu, Xiangyu Sai, Changguo Yao, Yuyouqiang Fu · 年份：2025 · DOI：10.1109/icpst65050.2025.11088979 · 被引用次数：1 · 研究领域：Advanced Algorithms and Applications

Under the "Dual Carbon" strategy, predicting and mitigating the potential security risks brought by the high proportion of new energy access has become a significant issue in constructing a new power system safety defense system. Addressing the lack of research on low-probability, high-risk operation scenarios in power grid operation modes, a method for generating extreme operation scenarios based on the Least Squares Loss Function Conditional Generative Adversarial Network (LS-CGAN) is proposed. First, the extreme operation scenarios and their risk level standards are defined. Second, building upon the original GAN model, the LS-CGAN model is designed by integrating a Least Squares loss function that provides non-saturated gradients and risk labels that guide the generation of scenarios. Specific implementation steps are provided. Taking the IEEE 39-node system with a high penetration rate of new energy as an example for simulation, the results show that the proposed method can efficiently generate extreme operation scenarios based on specified risk levels, providing solid and effective scenario data support for the formulation of more comprehensive control strategies.