Regression model based on Markov chain theory for composite fatigue curve approximation

Abstract

Model of fatigue damage accumulation, which is founded on the bases of Markov chain theory, is used as nonlinear regression model for processing of S-N fatigue curve data of laminate. A simple method of approximate estimation of model parameters is suggested. A numerical example is given. Some parameters of this fatigue curve model can be considered as local static strength distribution parameters. This model can be used for prediction of fatigue curve changes as consequence of the static strength distribution function parameter changes and for prediction of the cumulative distribution function of fatigue life for arbitrary stress cycles consequence.