Parameter identifiability of ductile fracture criterion for DP steels using bi-level reduced surrogate model

Hou, Yuliang [Université de Technologie de Compiègne] Zhao, Qiaoli [Zhengzhou University] Sapanathan, Thaneshan [UCL] Dumon, Alexandre Rachik, Mohamed [Université de Technologie de Compiègne]

Generally, a large number of fracture tests are required to characterize the material parameters of the modified Mohr-Coulomb (MMC) fracture criterion. It takes enormous amount of time and expenses to prepare specimens and analyze experimental measurements. However, manufacturing industries are keen to seek fast and cost effective approaches. Therefore, a bi-level reduced surrogate model has been proposed as an alternative method for the parameter identification. Using this approach, the identification process becomes feasible and it requires only a limited number of experimental tests. The surrogate model is used in the current framework to empirically capture the non-linear evolution of material parameters for the fracture onset under the uniaxial loading condition. It assembles local critical elements associated with the global 3D finite element (FE) models. The surrogate model of fracture strain is constructed using Diffuse Approximation and the local elements to identify the unknown parameters. Then, global fracture simulations which consider the identified parameters, are preformed to update the target fracture strain and to compute the corresponding failure onset displacement. Satisfactory results are obtained by adopting the concept of successive Design of Experiments (DOE). The identification protocol is validated, and it is capable of calibrating ductile fracture parameters for different DP steels. Finally, a parametric study is performed to analyze the sensitivity of each free parameter and the influence of the used polynomial basis within the identification approach.