面向阻尼控制的交直流混合系统线性状态空间模型辨识  

Linear State Space Model Identification of AC/DC Hybrid System for Damping Control

作　　者：伍文城[1,2] 王晓茹 宋新立[3] WU Wencheng;WANG Xiaoru;SONG Xinli(School of Electrical Engineering,Southwest Jiaotong University,Chengdu 610021,China;Southwest Electric Power Design Institute Co.,Ltd.of China Power Engineering Consulting Group,Chengdu 610056,China;China Electric Power Research Institute,Beijing 100192,China)

机构地区：[1]西南交通大学电气工程学院,四川省成都市610021 [2]中国电力工程顾问集团西南电力设计院有限公司,四川省成都市610056 [3]中国电力科学研究院有限公司,北京市100192

出　　处：《电力系统自动化》2022年第9期91-98,共8页Automation of Electric Power Systems

基　　金：国家重点研发计划资助项目(2016YFB0901001)。

摘　　要：为将确定性和随机混合系统与实现方法推广应用于包含多回直流的交直流混合系统,并实现自动化的状态空间模型辨识,基于奇异熵和Tanh-Sigmoid饱和函数,提出了一种抗噪能力强、适合密集模态系统的阶次判定方法。在确定阶次的基础上,再根据过去时域、未来时域与阶次关系确定小范围的输入参数测试数据集,并通过比较由各参数获得的辨识模型性能指标,选择出最优的辨识结果。该方法无须人工给定系统输入参数。最后,通过CIGRE Nordic 32母线系统和中国南方电网算例,验证了所提方法的有效性及优越性。To extend the combined deterministic and stochastic systems identification and realization(DSR)method to the AC/DC hybrid systems with multiple DC circuits and realize the automatic state space model identification,a system-order determined method based on singular entropy and Tanh-Sigmoid saturation function is proposed,which has strong anti-noise ability and is more suitable for dense modal systems.On the basis of the determined system-order,a test data set of small-range input-parameter is constructed according to the relationship between the past horizon,future horizon,and the order.By comparing the performance indices of the identification model obtained from each parameter,the optimal identification result is selected.The method does not need to give the system input parameters manually.Finally,the effectiveness and the superiority of the proposed method are verified by the cases of the CIGRE Nordic 32-bus system and China Southern Power Grid.

关 键 词：交直流混合系统 高压直流 系统辨识 状态空间模型 子空间方法 

分 类 号：TM721.3[电气工程—电力系统及自动化]