基于Vinnicombe距离的电力系统低频振荡预警精度研究  

Research on the Accuracy of Low-frequency Oscillation Warning in Power System Based on Vinnicombe Distance

作　　者：于淼 杜蔚杰[1,2] YUMiao;DU Wei-jie(School of Mechanical-electronic and Vehicle Engineering,Beijing University of Civil Engineering and Architecture,Beijing 100044,China;Beijing Engineering Research Center of Monitoring for Construction Safety,Beijing 100044,China;State Key Laboratory of Control and Simulation of Power System and Generation Equipment,Tsinghua University,100084,China)

机构地区：[1]北京建筑大学机电与车辆工程学院,北京100044 [2]北京市建筑安全检测工程技术研究中心,北京100044 [3]清华大学电力系统及发电设备安全控制和仿真国家重点实验室,北京100084

出　　处：《控制工程》2022年第2期339-347,共9页Control Engineering of China

基　　金：清华大学电力系统及大型发电设备安全控制与仿真国家重点实验室基金资助项目(SKLD20M17);北京市高等教育学会项目(YB2021131);北京建筑大学金字塔人才培养工程项目(JDYC20200324);北京建筑大学研究生教育教学质量提升项目(J2021016);北京建筑大学教育科学研究重点项目(Y19-12);北京建筑大学社会实践与创新创业课程项目(SJSC1913);北京建筑大学研究生创新项目(PG2021090);国家级大学生创新创业训练计划项目(202110016052,S202110016122,S202110016123,X202110016177,X202110016178,X202110016179);国家自然科学基金委青年科学基金资助项目(51407201)。

摘　　要：随着我国智能电网技术的发展,低频振荡预警已成为电力系统稳定性研究的重要问题。提出一种基于关键特征广域降维数据Vinnicombe距离的电力系统低频振荡幅值预警指标识别方法,该方法首先对PMU采集的原始大数据进行筛选与降维预处理,生成低频振荡幅值预警指标所需的初始特征量矩阵,然后通过对电力系统监控终端节点和各节点之间的区域进行编号,生成网络关联多特征向量状态检测矩阵,再结合Vinnicombe距离计算传递函数距离,判断是否发生低频振荡,并有效提高低频振荡幅值预警识别精度。最后,通过10机39节点新英格兰系统验证所提方法的正确性与有效性。With the development of smart grid technology in China, low-frequency oscillation warning has become an important issue in the study of power system stability. An identification method of low-frequency oscillation amplitude warning indicator for power system based on Vinnicombe distance of key feature wide-area downscaled data is proposed in this paper. Firstly, the original big data collected by PMU are screened and pre-processed for dimensionality reduction to generate the initial feature quantity matrix required for low-frequency oscillation amplitude warning indicator. Secondly, the node network associated multi-feature vector state detection matrix is generated by numbering the monitoring terminal nodes of power system and the area between the nodes. Then, combined with Vinnicombe distance, the transfer function distance is calculate to determine whether low-frequency oscillation occurs and effectively improve the recognition accuracy of low-frequency oscillation amplitude warning. Finally, the correctness and effectiveness of the method proposed in this paper are verified by the New England system with 10 machines and 39 nodes.

关 键 词：关键特征数据 低频振荡幅值预警 Vinnicombe距离 广域数据降维 预警识别精度 

分 类 号：TM743[电气工程—电力系统及自动化]