A critical challenge to the continued use of engineering structures as they are asked to perform longer than their design life is the prediction of an initiating crack and the prevention of damage, estimation of remaining useful life, schedule maintenance and to reduce costly downtimes and inspections. The research presented in this dissertation explores the cumulative plastic strain energy density and thermodynamic entropy generation up to crack initiation. Plastic strain energy density and thermodynamic entropy generation are evaluated to investigate whether they would be capable of providing a physical basis for fatigue life and structural risk and reliability assessments. Navy aircraft, specifically, the Orion P-3C, which represent an engineered structure currently being asked to perform past is design life, which are difficult and time consuming to inspect from carrier based operations and are currently evaluated using an empirically based damage index the, fatigue life expended, is used as an example in this investigation.

A set of experimental results for aluminum alloy 7075-T651, used in airframe structures, are presented to determine the correlation between plastic strain energy dissipation and the thermodynamic entropy generation versus fatigue crack initiation over a wide range of fatigue loadings. Cumulative plastic strain energy and thermodynamic entropy generation measured from hysteresis energy and temperature rise proved to be valid physical indices for estimation of the probability of crack initiation. Crack initiation is considered as a major evidence of fatigue damage and structural integrity risk. A Bayesian estimation and validation approach is used to determine systematic errors in the developed models as well as other model uncertainties. Comparisons of the energy-based and entropy-based models are presented and benefits of using one over the other are discussed.