FATIGUE LIFE AND RELIABILITY OF RANDOMLY EXCITED STRUCTURES

An improved stochastic model is used to investigate the fatigue failur e of a structural or machine component under random excitations. The f ailure is treated as a classic first-passage event in which a dominant crack grows to the critical length. Two specific improvements are mad e over previous analyses in that the critical crack length is treated properly as an absorbing boundary for the crack growth process, and th at the correlation among the successive stress cycles which induce the crack growth is taken into account. Analytical and numerical results are given of the probability density and statistical moments of the fa tigue life, and of the reliability function in the case of the Gaussia n stress process.