Turbine Fault Diagnosis Based on Fuzzy Theory and SVM

  • Authors:

  • Fei Xia;Hao Zhang;Daogang Peng;Hui Li;Yikang Su

  • Affiliations:

  • College of Electric Power and Automation Engineering, Shanghai University of Electric Power, Shanghai, China 200090;College of Electric Power and Automation Engineering, Shanghai University of Electric Power, Shanghai, China 200090;College of Electric Power and Automation Engineering, Shanghai University of Electric Power, Shanghai, China 200090;College of Electric Power and Automation Engineering, Shanghai University of Electric Power, Shanghai, China 200090;Nanchang Power Supply Corporation, Nanchang, China 330006

  • Venue:
  • AICI '09 Proceedings of the International Conference on Artificial Intelligence and Computational Intelligence
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

A method based on fuzzy and support vector machine (SVM) is proposed to focus on the lack of samples in fault diagnosis of turbine. Typical fault symptoms firstly are normalized by the membership functions perceptively. Then some samples are used to train SVM of fault diagnosis. With the trained SVM, the correct fault type can be recognized. In the application of condenser fault diagnosis, the approach enhances successfully the accuracy of fault diagnosis with small samples. Compared with the general method of BP neural network, the method combining advantages of fuzzy theory and SVM makes the diagnosis results have higher credibility.