Terahertz Imaging Method of Insulation Material Defects in Power Grid Equipment Based on Support Vector Machine Algorithm

WANG Lei; MEI Hongwei; LIU Jianjun; CHEN Dabing; JIANG Huaiyuan; WANG Liming

doi:10.13296/j.1001-1609.hva.2025.03.019

您当前的位置：

首页 >

文章列表页 >

Terahertz Imaging Method of Insulation Material Defects in Power Grid Equipment Based on Support Vector Machine Algorithm

Research & Analysis | 更新时间：2025-05-25

- Terahertz Imaging Method of Insulation Material Defects in Power Grid Equipment Based on Support Vector Machine Algorithm
- High Voltage Apparatus Vol. 61, Issue 3, Pages: 160-167(2025)
- 作者机构：
  
  1.清华大学深圳国际研究生院，广东　深圳　 518055
  2.国网江苏省电力有限公司电力科学研究院，南京　 211103
- 作者简介：
- 基金信息：
  
  Science and Technology Project of State Grid Corporation(5500-202118134A-0-0-00)
- DOI：10.13296/j.1001-1609.hva.2025.03.019
  CLC：
- Received：25 August 2024，
  
  Revised：29 October 2024，
  
  Published：16 March 2025
- 稿件说明：
移动端阅览
WANG Lei, MEI Hongwei, LIU Jianjun, et al. Terahertz Imaging Method of Insulation Material Defects in Power Grid Equipment Based on Support Vector Machine Algorithm[J]. High voltage apparatus, 2025, 61(3): 160-167.
DOI：

WANG Lei, MEI Hongwei, LIU Jianjun, et al. Terahertz Imaging Method of Insulation Material Defects in Power Grid Equipment Based on Support Vector Machine Algorithm[J]. High voltage apparatus, 2025, 61(3): 160-167. DOI： 10.13296/j.1001-1609.hva.2025.03.019.

摘要

电网设备绝缘材料在生产、运输、施工、现场运行等过程中容易产生缺陷，这严重影响了电网的安全运行。为实现对绝缘材料缺陷的无损检测，文中提出了基于支持向量机（SVM）的太赫兹二维成像方法。制备了聚乙烯（PE）、聚氯乙烯（PVC）和硅橡胶3种带缺陷的典型电力绝缘材料平板样品进行测试，得到缺陷识别准确率的不低于92.3%。在成像方面，与传统的单一特征参数成像方法相比，SVM成像方法对缺陷区域与非缺陷区域具有更好的区分度，边缘过渡区域明显变窄，成像质量较高。

Abstract

Insulation materials of power grid equipment are prone to defects in the process of production

transportation

construction and operation

which seriously affects safe operation of power grid. In order to realize the nondestructive detection of insulation material defects

a terahertz 2D imaging method based on support vector machine (SVM) is proposed in this paper. Such Three kinds of insulation materials samples as PE

PVC and silicone rubber are prepared for testing

and the accuracy of defect identification is not less than 92.3%. In terms of imaging

the SVM imaging method

compared with the traditional feature parameter imaging method

has a better distinction between defect and non-defect region

the edge transition region is obviously narrower and the imaging quality is higher.

关键词

Keywords

references

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

No data

Related Author

No data

Related Institution

No data

⁰