High-precision fast feature extraction method for partial discharge based on adaptive tracking fast S-transform
Ruoyu Wang, Baiqiang Yin, Lifen Yuan, Shudong Wang, Dalong Cheng, Yage Liu, Bing Li, Lei Zuo 等 9 位
Hefei University of Technology State Grid Corporation of China (China)
阅读操作
确认中在文库中上传 PDF 后可生成中文音频讲解。
摘要与影响
Partial discharge (PD) signals are often affected by various interferences in actual acquisition. The existing time-frequency feature extraction methods take too long to calculate, resulting in low practical application value. To solve those problems, this paper proposes a method for extracting partial discharge features based on two-dimensional time-frequency analysis using adaptive tracking fast S-transform(ATFST). The proposed approach adaptively adjusts its parameters according to the time–frequency characteristics of different PD signals, thereby achieving higher time–frequency resolution and enabling accurate feature extraction. Furthermore, it selectively identifies and extracts key feature intervals while effectively filtering out noise and irrelevant segments. This not only enhances feature quality but also reduces computational complexity and accelerates the extraction process. Experimental validation was conducted on five types of PD signals, with the extracted features evaluated using four widely adopted deep learning models. All models achieved recognition accuracies exceeding 98.25%, with a maximum accuracy of 99.6%. Compared with the traditional S-transform, the ATFST method has higher time-frequency resolution and faster feature extraction speed, and demonstrates good recognition performance under current experimental conditions, providing multidimensional features for artificial intelligence algorithms such as deep learning.
逐年被引趋势
暂无年度引用数据
关键指标
同类平均 = 1
同领域 · 同年份 · 同类型
Google Scholar 与 OpenAlex 的被引统计范围不同,数值存在差异属正常。
AI 辅助阅读
依据:摘要
可就本文提问;依据不足时会说明。
学术脉络
学科主题
工程High voltage insulation and dielectric phenomena
Machine Fault Diagnosis Techniques · Electrical Fault Detection and Protection