基于声纹特征识别的电力变压器运维检测技术研究及性能评估
Research and Performance Evaluation of Power Transformer Operation and Maintenance Detection Technology Based on Voice Print Feature Recognition
申国标 1陈浩 1李德成 1刘寿光 1李章勇1
作者信息
- 1. 云南电网有限责任公司文山供电局,云南文山 663000
- 折叠
摘要
文章开发了一种基于声纹特征识别的电力变压器运维检测技术,其利用梅尔频率倒谱系数结合深度置信网络和支持向量数据描述算法,提升了声纹信号分析的准确度和效率.试验表明,DBN-SVDD算法在变压器缺陷识别中准确率达97.94%,为智能电网的可靠运行提供了技术支持.
Abstract
The article develops a power transformer operation and maintenance detection technology based on voiceprint feature recognition,which combines Mel frequency cepstral coefficients with deep confidence networks and support vector data description algorithms to improve the accuracy and efficiency of voiceprint signal analysis.The experiment shows that the DBN-SVDD algorithm has an accuracy of 97.94%in transformer defect recognition,providing technical support for the reliable operation of smart grids.
关键词
电力变压器检测/声纹特征识别/MFCC/DBN/SVDDKey words
power transformer detection/voiceprint feature recognition/MFCC/DBN/SVDD引用本文复制引用
出版年
2024
NETL
NSTL
万方数据