大电网动态安全风险智能评估系统

针对大规模交直流混联系统的快速动态安全风险防控需求,研究了系统动态安全风险智能评估系统的架构与关键技术。结合电力系统安全风险评估的一般流程,提出了基于机器学习的安全风险智能评估系统的总体框架和动态安全风险统一评估模型结构;构建了包含发电负荷运行方式、网络拓扑结构和故障位置特征的训练样本集合,采用样本平衡技术提高评估模型精度;基于深度学习提取动态安全风险高级特征,采用主流机器学习框架构建和更新动态安全风险统一评估模型。以省级电网为例验证了所提动态安全风险智能评估系统的可行性。

Intelligent Assessment System for Dynamic Security Risk of Large-scale Power Grid

The architecture and key technologies of the intelligent assessment system for dynamic security risk of power systems are studied for preventing and controlling fast dynamic security risk of large-scale AC-DC hybrid systems. The overall framework of intelligent assessment based on machine learning and the structure of unified dynamic security risk assessment model are proposed considering the general assessment process. A training sample set containing fault locations, generation and load modes, and network topology structures is constructed. The sample balance technique is used to improve the accuracy of the assessment model. Advanced features for assessing dynamic security risk are extracted based on deep learning, and a unified assessment model of dynamic security risk is constructed and updated with mainstream machine learning framework. A provincial power grid is taken as an example to verify the feasibility of the proposed intelligent assessment system for dynamic security risk. This work is supported by National Key R&D Program of China(No. 2017YFB0902600)and State Grid Corporation of China(No. SGJS0000DKJS1700840).