Bayesian Network Combined Fuzzy C-means Methodology for Turbine Blades Fatigue Performance Evaluation

Abstract

In this paper, a fatigue performance evaluation model for steam turbine blades based on Bayesian network combined fuzzy c-means algorithm was proposed. Bayesian network was viewed as a classification technique to evaluate fatigue performance. Fuzzy c-means algorithm was applied to perform cluster analysis of fatigue performance values and made them discrete. Low-cycle fatigue tests on certain kind of steam turbine blades were performed. Experiment results well examined the validity of the evaluation model. The proposed methodology significantly provided a possible approach to assist operators and engineers in carrying out online monitoring of blades’ fatigue degradation.