Validation of a Crack Initiation Life Analysis Procedure 2000-01-1666

Any analysis tool should be validated to verify computation accuracy. Fatigue analysis processes are inherently complex to validate due to the sensitivity and interdependency of variables. In the current work the authors systematically separate the variables for independent evaluation and verification. The approach is the result of new Neuber notch analysis equations, derived by the authors, and draws upon the experience of the authors as well as several other analysts in developing an approach that reduces interaction between variables. The extension of the Neuber’s Rule into multiaxial format is presented. Each building block of the process is discussed: loads and load histories, stresses, stress concentrations, notch analyses (uniaxial and multiaxial), material properties (stress-strain response as well as life data), and damage accumulation. A method to account for fretting is discussed. Occasionally, during an evaluation, a variable is either overlooked or not quantified well. The deviations in life computations due to the effects of these omissions or inaccuracies are presented. These effects arise from different aspects such as omission of residual stresses, accuracy of stress computation, inadequacy of stress concentration values, and unknown or inaccurate representations of surface finish.