基于TLS-ESPRIT算法的变压器油纸绝缘等效电路参数辨识及新特征量提取  被引量:6

Parameter identification of oil paper insulation equivalent circuit of transformer and new feature extraction based on TLS-ESPRIT

在线阅读下载全文

作  者:苏凯强 刘庆珍[1] SU Kaiqiang;LIU Qingzhen(College of Electrical Engineering and Automation,Fuzhou University,Fuzhou 350108)

机构地区:[1]福州大学电气工程与自动化学院,福州350108

出  处:《电气技术》2022年第7期89-96,共8页Electrical Engineering

基  金:国家自然科学基金项目(61174117)。

摘  要:应用介质响应法研究变压器油纸绝缘老化特性过程的关键,是准确地辨识出油纸绝缘等效电路参数和反映绝缘老化程度的特征量。针对现有等效电路参数辨识方法存在的局限性,提出一种相邻奇异值之比定阶的总体最小二乘-旋转矢量不变技术(TLS-ESPRIT)算法,实现对扩展德拜等效电路的参数辨识。首先利用去极化电流构建Hankel矩阵;然后根据收敛前的相邻奇异值之比个数确定弛豫支路数;最后通过TLS-ESPRIT算法辨识出等效电路参数。经仿真和实测验证,该方法可以准确、惟一地实现参数辨识,具有良好的抗噪声能力。进一步地,利用该方法提出2个新特征量:去极化能量初始斜率和去极化能量半衰期,大量实测数据的统计分析表明,新特征量能有效反映油纸绝缘的老化程度。The key to study the aging characteristic of oil-paper insulation of transformer by using dielectric response method is to accurately identify the equivalent circuit parameters and the characteristic quantity.Aiming at the limitation of the existing identification methods,a TLS-ESPRIT algorithm based on the ratio of adjacent singular values is proposed.Firstly,the Hankel matrix is constructed by depolarized current.Then the relaxation path number is determined according to the ratio of adjacent singular values.Finally,the equivalent circuit parameters are identified by total least squares-estimation of signal parameters via rotational invariance technique(TLS-ESPRIT)algorithm.Simulation and actual measurement show that this method can accurately and uniquely identify parameters and has good noise resistance.Then,two new characteristic parameters,the initial slope of depolarization energy and the half-life of depolarization energy,are proposed by the method.The statistical analysis shows that the new characteristic parameters can effectively reflect the aging degree of oil paper insulation.

关 键 词:油纸绝缘 总体最小二乘-旋转矢量不变技术(TLS-ESPRIT)算法 参数辨识 去极化能量 老化诊断 

分 类 号:TM855[电气工程—高电压与绝缘技术]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象