Approximate Modeling Approaches for Estimating the Parameters of Fatigue

A material undergoes progressive damage that creates micro-cracks and induces its breaking. To determine its life time, very costly experimental trials are achieved. To overcome this problem, different strategies have been developed using experimental results. These strategies have given different models that predict the mechanical behavior of materials with few and low costs tests. Most of the existing methods for modelling the fatigue of a material have focused on the estimation of the coefficient of fatigue strength ?’f and on the coefficient of ductility in fatigue ?’f . In this work, we study the performance of a dozen of methods proposed in the literature to estimate these coefficients and to predict the life time of the material. We have determined the errors committed in the theoretical models in relation to experimental data. To this end, we have compared the results obtained by these methods with the experimental results on a basis of 82 steels. We have established a protocol to compare the results, which consists in computing the error committed by each model on the estimation of the coefficients of fatigue. The analyses of the results have shown that none of the models can estimate correctly both of the coefficients. We then identified two models that estimate accurately both coefficients ?’f and ?’f separately