A stochastic modelling and analysis methodology for quantification of fatigue damage Original Research Article

Fatigue damage accumulation in engineering materials is usually quantified using a suitable damage parameter that can serve as a reliable descriptor of damage development. Tests are conducted on material samples in order to collect sample data about damage development. A computational methodology is developed in the present paper for modelling and analysis of these data, based on a rigorous stochastic approach. From sample data, the true probability distributions of damage parameter are extracted based on an analytical approach. The correlation properties of damage development process are also determined in terms of autocorrelation functions. A systematic way of characterizing the damage modes of the material, in terms of the correlation characteristics is also developed. The damage development is modelled as an embedded Markov process. A recursive stochastic matrix equation that characterizes the damage development using the true probability distributions and correlation properties of test data, is formulated. Application of the present methodology to a composite laminate is demonstrated.