Presentation Details
  1. Home
  2. Presentation List
  3. Presentation Details
New Features Derived from Dissolved Gas Analysis for Fault Diagnosis of Power Transformers Based on Membership Function
ID：4 View Protection：ATTENDEE Updated Time：2022-09-12 10:34:05 Hits：343 Oral Presentation

Start Time：Pending（Asia/Shanghai）

Duration：Pending

Session：No Session »

Presentation File

Tips: Only the registered participant can access the file. Please sign in first.

Abstract
Condition-based maintenance has become an important approach to maintain long-term stable and safe operation of power equipment. With the rapid development of machine learning technology, it can effectively improve the accuracy of transformer fault diagnosis. In this paper, based on the state evaluation criterion, the membership function of DGA is established, and the membership degree of each DGA parameter is calculated as a new feature parameter. SVM and random forest algorithm are used to train and diagnose IEC TC 10 fault database respectively. The results show that the membership degree of DGA as a feature parameter is increased by 2.4 % in SVM and 33.03 % in random forest algorithm, which verifies the effectiveness of the algorithm
Keywords
power transformers, fault diagnosis, gas ratios dissolved in oil, membership Function, IEC TC 10 database
Speaker
Chao Gao
Ltd.；Major equipment management office； China Nuclear Power Operations Co.

Zhongqing Yang
Department of equipment management, Suzhou Nuclear Power Research Institute

Submission Author
Chao Gao Ltd.；Major equipment management office； China Nuclear Power Operations Co.
Erya Gao Ltd.；Major equipment management office； China Nuclear Power Operations Co.
Zhongqing Yang Department of equipment management； Suzhou Nuclear Power Research Institute
Yuhui Feng Ltd.；Major equipment management office； China Nuclear Power Operations Co.
Bing Song Ltd.；Major equipment management office； China Nuclear Power Operations Co.
Qian Li Ltd.；Major equipment management office； China Nuclear Power Operations Co.
Submit Comment
Verify Code Change Another
All Comments
Important Date

  • Conference Date

    Sep 25

    2022

    to

    Sep 29

    2022

  • Aug 15 2022

    Early Bird Registration

  • Sep 10 2022

    Contribution Submission Deadline

  • Nov 10 2022

    Registration deadline

  • Nov 30 2022

    Draft paper submission deadline

  • Nov 30 2022

    Final Paper Deadline

Sponsored By
IEEE DEIS
Organized By
Chongqing University
Contact Information
Login to view full contact details
Previous Conferences