Fatigue life prediction by distribution analysis of acoustic emission signals

Abstract An ordered statistics treatment of the fatigue process that predicts a fatigue life distribution of the Weibull type, which is an asymptotic distribution of extreme smallest values, is presented. The distribution of the propagation rates of cracks which lead to catastrophic failure is an asymptotic distribution of extreme largest values. This property is combined with the fact that both the fatigue crack propagation rates and the cumulative number of acoustic emission counts are related to the stress intensity range. The physical meaning of this connection is derived by considering the energy balance during crack propagation and by relating the release of energy to the acoustic emission characteristics. Reference to previously published work lends some support to the theoretical treatment. The theory presented indicates that analysis of the amplitude distribution of the acoustic signals emitted during cyclic stress may provide a non-destructive method of predicting fatigue life.