Probabilistic Fatigue Life Estimation of Plate with Multiple Stress Concentration Zones

Fatigue is a complex phenomenon which causes sudden failure of a component subjected to cyclic loading in the elastic region. Fatigue lives of nominally identical specimens subjected to the same nominal cyclic stress display scatter. This phenomenon reflects the stochastic nature of a fatigue damage process. Probabilistic and statistical methods are appropriate for managing the large uncertainties that exist in the fatigue process and provide reliability as a measure of structural performance. In this work, MATLAB was used as a platform to evaluate probabilistic fatigue life of the plate with multiple stress concentration zones using Latin Hypercube sampling. Fatigue reliability was computed by accounting for the scatter present in material properties in terms of random design variables and verified. Among the random design variables, fatigue strength exponent was having higher influence on fatigue life. Fatigue reliability was found to be decreasing with increase in input variation.