Phenomenological probabilistic models for fatigue and fracture prediction


Seminar Fraunhofer IWM, 18. July 2018, Seminarraum des Fraunhofer-Instituts für Werkstoffmechanik IWM in Freiburg, Dr. Alfonso Fernández Canteli, University of Oviedo



Was:                        Seminar Fraunhofer IWM: Phenomenological probabilistic models for fatigue and fracture prediction

Wann                      18. Juli 2018, 10:15 Uhr

Wo:                         Seminarraum W9a des Fraunhofer Instituts für Werkstoffmechanik IWM in Freiburg

Vortragender:        Dr. Alfonso Fernández Canteli, University of Oviedo

A unified probabilistic approach for fracture and fatigue failure analysis is, or should be, an objective of major interest for material scientists and practicing engineers concerning structural integrity. Phenomenological models, in particular those based on extreme value statistics and compatibility, are possible and suitable candidates for probabilistic prediction of fracture and fatigue failures of materials in the design of components. In this work, the derivation of the basic Weibull regression fatigue model and that of crack growth models are presented followed by some extensions and enhancements of them. Such models facilitate planning of the experimental program by providing the adequate test strategy that proves their adequacy for assessment, interpretation and further application of experimental results. As a complement of the foregoing models, the generalized local model (GLM) provides the primary failure cumulative distribution function (PFCDF), as derived from the experimental failure cumulative distribution function (EFCDF) of the test data, which ensures transferability of the lab test results to the component design. The adequate choice of the generalized parameter and applicability of the weakest link principle are the only requirements to achieve the objective characterization of the material. Some examples are presented aiming at illustrating the usefulness of this methodology in practical design of components.