Aluminum 7075-T6 Fatigue Data Generation and Probabilistic Life Prediction Formulation.

The life extension of aging fleet aircraft requires an assessment of the safe-life remaining after refurbishment. Risk can be estimated by conventional deterministic fatigue analysis coupled with a subjective factor of safety. Alternatively, risk can be quantitatively and objectively predicted by probabilistic analysis. In this investigation, a general probabilistic life formulation is specialized for constant amplitude, fully reversed fatigue loading utilizing conventional breakdown laws applied to the general probability damage function. Experimental data was collected both as a bench mark data base, as well as an example of the implementation of probabilistic fatigue life prediction. Fully reversed, sinusoidal fatigue testing under load control was carried Out at load levels giving high cycle fatigue lives from 1 x 104 to 5 x 106 cycles. The number of replications at each stress level is greater than currently available in the literature, thereby increasing the confidence of predictions in the long-life domain, as well as extending the statistical basis for probabilistic inference. The load level data sets are interpreted by the probabilistic damage function for life location as well as life shape parameters using maximum likelihood analysis. Homologous life ranking and the minimum entropy hypothesis are investigated as well.