A Digital Partial Discharge Defect Model Based on K-Means Clustering
A digital defect model for partial discharge detection training of substation GIS equipment is proposed in this paper, The fixed-duration discharge data of five typical partial discharges were recorded and six eigenvalues of all discharge points were extracted, followed by clustering the initial samples with the K-means clustering algorithm in use of the Euclidean distance, then the overall sample feature clustering algorithm is adopted to optimally select the typical sample points labeled with special discharge type, and the partial discharge samples are expanded with variable white noise to achieve the reproduction of typical partial discharge feature map.
Authors
- Lina Chen (ORCID: https://orcid.org/0000-0002-4576-6814)
- Guoliang Huang (ORCID: https://orcid.org/0000-0002-6238-6127)
- Aijing Li (ORCID: https://orcid.org/0000-0002-1515-0330)
- Yanjie Zhang (ORCID: https://orcid.org/0000-0002-8229-799X)
- Yanyan Xu (ORCID: https://orcid.org/0000-0002-5019-0848)
- Man Cheng (ORCID: https://orcid.org/0009-0002-2322-0551)
- S. L. Hu (ORCID: https://orcid.org/0009-0005-9422-7445)
- Xi Wang (ORCID: https://orcid.org/0009-0004-1136-2215)
Institutions
- Central South University of Forestry and Technology (CN)
- Central South University (CN)
- State Grid Corporation of China (China) (CN)
- Jiaxing University (CN)
- Department of Science and Technology of Shandong Province (CN)
- First Hospital of Jiaxing (CN)
Publication Details
- Published
- 2025-04-28
- DOI
- https://doi.org/10.1109/icaidt66272.2025.00059
- Primary Topic
- Water Quality Monitoring and Analysis
- Type
- article
- Field-Weighted Citation Impact
- 0.00