正交降噪算法在井下变压器局放信号处理中的应用
肖宇
（挖金湾煤业公司，山西 大同 037004）
摘要：为了保障井下供电用干式变压器的稳定安全运行，正确识别其局部放电类型及绝缘状况极其关键。提出了一
种基于正交匹配降噪算法的局放信号处理方法，将现场采集到的局放信号通过正交匹配追踪算法进行初步降噪，再将降
噪后的信号经自回归模型获取其自回归特征值，最后将信号自回归特征值配合随机森林集成分类模型进行放电类型识
别。试验结果表明，该识别模型对变压器局放特征量识别度较高，达到了98%，满足现场工业要求。
关键词：井下变压器；局部放电；正交匹配；自回归量；分类模型
中图分类号：TN911.4     文献标志码：A     文章编号：2095-0802-(2021)01-0207-03
Application of Orthogonal Noise Reduction Algorithm in PD Signal Processing of Underground
Transformer
XIAO Yu
(Wajinwan Coal Industry Co., Ltd., Datong 037004, Shanxi, China)
Abstract: In order to ensure the stable and safe operation of dry-type transformers for underground power supply, it is critical to
correctly identify the type of partial discharge and insulation status. This paper proposed a method of partial discharge signal
processing based on orthogonal matching noise reduction algorithm. The partial discharge signal collected on site was initially denoised
through the orthogonal matching pursuit algorithm, and then the autoregressive feature value was obtained by passing the
de -noised signal though the autoregressive model, and finally the signal autoregressive feature value was combined with the
random forest ensemble classification model to identify the discharge type. The test results show that the recognition model has a
high degree of recognition of the partial discharge characteristics of the transformer, reaching 98%, which meets the requirements
of the field industry.
Key words: underground transformer; partial discharge; orthogonal matching; autoregressive quantity; classification model