高压电器2026，Vol.62Issue(3)：61-68,8.DOI:10.13296/j.1001-1609.hva.2026.03.008
配电变压器声音检测中基于时频域自相似性去噪方法的可行性分析
Feasibility Analysis of a Time-frequency Domain Slef-silimarity Denoising Method for Sound Detection in Distribution Transformers
摘要
Abstract
The use of sound signals for condition monitoring of power distribution equipment offers the advantages of being low coat and contactless.It,however,also has the drawback of being susceptible to strong ambient noise.Existing study has sucessfully denoised sound signals by leveraging the differences in time-frequency domain self-similarity between ambient noise and distribution transformer operating sounds through the use of a parameter-free clustering algorithm,achieving promising simulation results.The fault noise of transformer under different operating conditions has diefrence.If the denoising method causes the leakage of transformer sound signal samples in differ-ent operating conditions,the sound sample set after screening may not contain the early fault sound samples,result-ing in misjudgment in the subsequent state recognition process.Therefore,in this paper the box transformers in dif-ferent working environments are taken as an example,the sound samples free from the interference of environmental noise are screened out in accordance with the time and frequency domain self-similarity differences between the op-erating sound of distribution equipment and the environmental noise.It is proved through the time distribution char-acteristics of stationary sound segments and distribution transformers in different operating conditionsthat the meth-od can cover all operating conditions of distribution transformers,which provides a strong support for the subse-quent state monitoring of distribution transformers based on sound signal for production practice.关键词
声音信号/状态监测/配电变压器/箱式变压器/消除噪声/样本筛选Key words
sound signal/condition monitoring/distribution transformers/box transformers/noise removal/sample screening引用本文复制引用
龙骧进,刘元,苏盛,陈凤,李彬..配电变压器声音检测中基于时频域自相似性去噪方法的可行性分析[J].高压电器,2026,62(3):61-68,8.基金项目
国家自然科学基金资助项目(51777015) （51777015）
湖南省自然科学基金项目(2022JJ60089). Project Supported by National Natural Science Foundation of China(51777015),Natural Science Foundation of Hunan Province(2022JJ60089). （2022JJ60089）
