Probability-Based Methods for Fatigue Analysis

Modern fatigue analysis techniques, that can provide reliable estimates of the service performance of components and structures, are finding increasing use in vehicle development programs. A major objective of such efforts is the prediction of the field performance of a fleet of vehicles as influenced by the host of design, manufacturing, and performance variables. An approach to this complex problem, based on the incorporation of probability theory in established life prediction methods, is presented. In this way, quantitative estimates of the lifetime distribution of a population are obtained based on anticipated, or specified, variations in component geometry, material processing sequences, and service loading. The application of this approach is demonstrated through a case study of an automotive transmission component. Finally, the relevance of such a capability to quality assurance issues, through determination of the sensitivity of service performance to design and manufacturing variables, is discussed.