Séminaire de Sylvia Feld-Payet

Predicting structural failure is essential to ensuring proper structural design, particularly in the aerospace industry. The literature is rich in models for predicting crack propagation in Linear Elastic Fracture Mechanics, but as soon as plasticity is no longer locally negligible, there is still a need for models supported by experimental observations.
To model damage and fracture in structures undergoing plastic deformation, it is first necessary to ensure the uniqueness and accuracy of the computation. To do this, several aspects must be addressed, such as the type of element and the adaptation of the mesh size. The goal is to predict the evolution of damage, its localization, and its impact on material behavior leading to crack initiation using a continuous model. At this stage, it becomes useful to introduce a true discontinuity into the model to better represent crack propagation. This presentation addresses regularization element formulations and mesh optimization based on an error indicator for modeling coupled damage in a structure. The elements necessary for a continuous-to-discontinuous transition strategy will also be presented. 
Validation of damage models leading to crack initiation and propagation—or even identifying the parameters involved—requires relying on tests to estimate quantities that are difficult to measure because their impact remains localized. Under these conditions, image analysis proves to be a particularly well-suited tool. However, the presence of significant nonlinear phenomena renders methods associated with Linear Elastic Fracture Mechanics unsuitable. This presentation discusses the use of image analysis to identify the parameters of a damage model and demonstrates the value of new methodologies—which are independent of both material and loading conditions—for detecting crack initiation and tracking crack propagation.