基于模拟故障数据库的GIS局部放电反演方法  被引量:12

Partial Discharge Inversion Method for GIS Based on Simulated Fault Database

在线阅读下载全文

作  者:王淼 罗林根[1] 钱勇[1] 王辉[1] 宋辉[1] 盛戈皞[1] WANG Miao;LUO Lingen;QIAN Yong;WANG Hui;SONG Hui;SHENG Gehao(Department of Electrical Engineering,Shanghai Jiao Tong University,Shanghai 200240,China)

机构地区:[1]上海交通大学电气工程系,上海200240

出  处:《高电压技术》2022年第5期1663-1672,共10页High Voltage Engineering

基  金:国家重点研发计划(2020YFB1709701)。

摘  要:基于特高频信号的时差定位法是目前针对气体绝缘封闭开关设备(gas insulated switchgear,GIS)局部放电定位的主要技术方案之一。然而,该方法主要应用于带电检测,由于需要高精度的同步采样装置,限制了该方法的在线监测应用。更重要的是放电源的定位结果中不包含特高频幅值或是放电量信息,不利于对具体缺陷的严重程度进行判断。针对以上不足,提出了一种基于模拟故障数据库的GIS局部放电反演方法,通过在与实际GIS设备尺寸参数相同的数字虚拟模型中开展局放模拟实验,获取GIS内部不同位置发生局放时观测点处的特高频信号幅值,从而生成包含局放源坐标与观测点信号参量的模拟故障数据库。进一步利用智能学习算法对坐标信息与信号参量的非线性映射关系进行拟合,形成故障反演定位网络。当局部放电发生时,根据实测特高频信号幅值及反演定位网络实时给出局放源的定位结果,并在数字空间内推演电压信号与放电量的关系,从而给出放电定量估计结果。T型GIS母线仿真结果表明:所提方法的平均定位误差为13.52cm,放电量定量估计的平均相对误差为11.88%。实验中以直筒型GIS母线段为例,得到平均定位误差为11.52 cm,定量的平均相对误差为16.70%。该研究可为基于数字化模型的GIS在线监测内部局部放电源定位及进一步的绝缘缺陷分析提供解决思路。The time difference of arrival method based on ultra high frequency(UHF)signal is one of the main techniques for partial discharge(PD)localization in gas insulated switchgear(GIS).However,it is mainly used in the live detection,and its application in online monitoring is limited due to the need of high precision synchronous sampling device.More importantly,the localization results of the PDs do not contain UHF amplitudes or discharge quantities,which is not conducive to judging the severity of specific defects.In view of the above shortcomings,this paper proposes a partial discharge inversion method based on the simulated fault database for GIS.The PD simulation experiments are carried out in a digital virtual model with the same dimension parameters as the actual GIS equipment,and the UHF signal amplitudes at observation points are obtained when the PD occurs at different positions in GIS.Thus,a simulated fault database is generated containing the coordinates of PD sources and the signal parameters of observation points.Then,an intelligent learning algorithm is used to fit the nonlinear correspondence between coordinate information and signal parameters as fault inversion localization network.In the event of PD,the real-time localization results of PDs are given according to the measured UHF signal amplitudes and the inversion localization network.Moreover,the relationship between voltage signals and the discharge quantity is deduced in the digital model,and the quantitative estimation result is obtained.The simulation results in T-type GIS bus show that the average localization error of the proposed method is 13.52 cm,and the average relative error of the quantitative estimation is 11.88%.In the experiment,a straight tube type bus section of a GIS is taken as an example.The average localization error is 11.52 cm,and the mean relative error of quantification is 16.70%,which further proves the effectiveness of the proposed method.The research of this paper provides a solution for internal PD localization and the

关 键 词:GIS 数字空间模型 故障数据库 故障反演 局放定位与定量 

分 类 号:TM595[电气工程—电器] TM855

 

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