Surface Roughness Characterization for Stress Concentration Factor Predictions: A Bayesian Learning Approach

Abstract

The surface roughness has an important influence on the fatigue life of the structures. The fatigue life reduces due to the stress concentration caused by surface roughness. The stress concentration governs the fatigue crack initiation and propagation. The accurate acquisition of the stress concentration factor of rough surfaces is a key issue in determining fatigue life. Nevertheless, semi-empirical models may be biased for various machining processes. Besides, finite element method simulations cannot give explicit expression of the stress concentration factor. Bayesian learning can construct accurate prediction models which offering a number of additional advantages. In this paper, based on several data pairs constructed by finite element method, the correlation expression between the stress concentration factor and statistical roughness parameters of surfaces is established quickly through Bayesian learning. Compared with some other semi-empirical models, the accuracy and stability of the proposed method are certified. This paper provides a simple and effi-cient approach to determine the stress concentration factor for rough surfaces under different processing conditions.