Study of the Unsupervised Extraction Method of Transformer Vibration Features

DOI

10.2991/seeie-19.2019.45How to use a DOI?

Keywords

transformer; vibration signal; feature selection; wavelet packet; mutual information

Abstract

A mutual information-based unsupervised feature extraction method of transformer surface vibration is proposed in the paper. Wavelet packet analysis is used to extract surface vibration signal frequency band-energy features of the transform in operation and complete its signal representation. Relevance and redundancy of mutual information measurement are considered to judge importance degrees of features, and feature importance ranking and selection are completed based on unsupervised minimum redundancy and maximum relevance criteria. The analysis of measured signals at different measuring points of the transformer indicates that this method can accurately select important features of transformer vibration signals while effectively reducing data dimensionalities.

Copyright

© 2019, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Wei Xu
AU  - Chi Kang
AU  - Xiangjing Du
AU  - Bin Zhong
AU  - Naihui Wang
AU  - Zhong Jin
AU  - Yang Jing
PY  - 2019/05
DA  - 2019/05
TI  - Study of the Unsupervised Extraction Method of Transformer Vibration Features
BT  - Proceedings of the 2019 2nd International Conference on Sustainable Energy, Environment and Information Engineering (SEEIE 2019)
PB  - Atlantis Press
SP  - 194
EP  - 199
SN  - 2352-5401
UR  - https://doi.org/10.2991/seeie-19.2019.45
DO  - 10.2991/seeie-19.2019.45
ID  - Xu2019/05
ER  -