Early Warning of Incipient Faults for Power Transformer Based on DGA Using a Two-Stage Feature Extraction Technique

Early warning for transformer faults is valuable for maintenance decision-making. However, limited work has been done in this area due to the difficulty of the model establishment. This paper proposes a two-stage feature extraction method for early warning of power transformer faults, where a novel...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on power delivery 2022-06, Vol.37 (3), p.2040-2049
Main Authors: Zhang, Yang, Chen, Hong Cai, Du, Yaping, Chen, Min, Liang, Jie, Li, Jianhong, Fan, Xiqing, Sun, Ling, Cheng, Qingsha S., Yao, Xin
Format: Article
Language:English
Subjects:
Decision making
Dissolved gas analysis
early warning
Faults
Feature extraction
Gases
Genetic algorithms
Genetic programming
Maintenance engineering
Oil insulation
Power transformer insulation
transformer diagnosis
Transformers
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Early warning for transformer faults is valuable for maintenance decision-making. However, limited work has been done in this area due to the difficulty of the model establishment. This paper proposes a two-stage feature extraction method for early warning of power transformer faults, where a novel feature extraction process is applied by combining feature ranking and genetic programming (GP). In the first stage, the data is labeled as normal and fault states and the feature extraction is evaluated on the data. Then, extracted key features and their growth rates are relabeled as normal and warning states, after which the feature extraction process is re-evaluated on the relabeled data. Obtained features and logic expression can finally be used for early warning. The proposed framework can implement an early warning with about 100 days in advance for transformer faults and is verified through 8 sequences of data. The comparisons with two recently reported methods show the superiority of the proposed method.
ISSN:0885-8977
1937-4208
DOI:10.1109/TPWRD.2021.3103455