摘要
大规模新能源接入使电力系统的频率动态特性更加复杂,数据驱动的暂态频率安全评估广受关注,但仍存在特征挖掘不足、先验知识未能充分利用等局限性。为此,文中提出了基于独立时空特征提取与分组增强的暂态频率安全评估方法。首先,引入自适应时空特征提取模块,通过独立的全连接邻接矩阵,表征发电机间的关联关系,提升特征提取能力;其次,设计了基于物理先验知识的分组特征增强模块,依据系统惯性中心频率指导特征融合,促进初始输入特征的交互、增强表征能力;最后,应用多尺度卷积模块融合所有特征获取最终的系统表征,通过模型训练和参数优化实现频率安全评估。标准算例仿真结果表明,文中模型的精度和鲁棒性均优于现有数据驱动模型。
The large-scale integration of renewable energy has made the frequency dynamic characteristics of power systems more complex.The data-driven transient frequency safety assessment has attracted widespread attention,but there are still limitations such as insufficient feature mining and underutilization of prior knowledge.To this end,a method for transient frequency safety assessment based on independent spatial-temporal feature extraction and grouped enhancement is proposed.First,an adaptive spatial-temporal feature extraction module is introduced to characterize the correlations among generators through independent fullyconnected adjacency matrices,thereby enhancing the capability of feature extraction.Then,a grouped feature enhancement module based on physical prior knowledge is designed,which guides the feature fusion by the frequency of system inertia center,promotes interactions of initial input features,and enhances representation capabilities.Finally,the multi-scale convolutional module is used to fuse all features to obtain the final system representation,and the frequency safety assessment is achieved through model training and parameter optimization.Simulation results of the standard case demonstrate that the accuracy and robustness of the proposed model in this paper are superior to those of existing data-driven models.
作者
杨文娴
石访
宋雪萌
田硕硕
聂礼强
程定一
YANG Wenxian;SHI Fang;SONG Xuemeng;TIAN Shuoshuo;NIE Liqiang;CHENG Dingyi(National Joint Engineering Laboratory of Power Grid with Electric Vehicles(Shandong University),Jinan 250061,China;School of Computer Science and Technology,Shandong University,Qingdao 266000,China;School of Electrical Engineering,Shandong University,Jinan 250061,China;State Grid Shandong Electric Power Research Institute,Jinan 250001,China)
出处
《电力系统自动化》
2025年第18期170-181,共12页
Automation of Electric Power Systems
基金
国家重点研发计划资助项目(2021YFB2400800)。
关键词
新型电力系统
暂态频率
安全评估
特征增强
注意力机制
时空特征
new power system
transient frequency
safety assessment
feature enhancement
attention mechanism
spatialtemporal feature
作者简介
杨文娴(2001-),女,硕士研究生,主要研究方向:电力系统频率安全。E-mail:2761100147@qq.com;通信作者:石访(1982-),男,博士,副教授,主要研究方向:电力系统稳定分析与控制、电力系统同步相量测量技术与应用。E-mail:shifang@sdu.edu.cn;宋雪萌(1990-),女,博士,副教授,博士生导师,主要研究方向:信息检索和社交媒体分析。E-mail:songxuemeng@sdu.edu.cn。
