Making peridynamic models fit for real-world multiphysics applications

点击翻译按钮获取中文摘要

Peridynamics (PD) is a recent nonlocal theory receiving increasing attention. It has been widely exploited as a promising tool to model challenging phenomena in fracture analysis such as nucleation, growth and propagation of cracks in solids. More recently, also application of PD to corrosion problems has attracted considerable attention. However, PD modeling is still in its infancy compared to classical local theories and associated computational methods such as the finite element method. As a consequence, while PD has been thriving in the academic community over the last years, its application to complex real-world problems still remains sparse. One of the main reasons for this is that PD models are suffering at the moment from a considerably higher computational cost than established approaches like the finite element method because the tools and efficient implementation schemes that have been developed for the latter over several decades are at the moment still lacking for PD. In this project, we will address this key problem of PD modeling by developing novel efficient and robust implementation schemes as well as supplementary tools making PD models computationally more efficient. Specifically, we aim at the development of (1) substantially improved adaptive discretization and integration schemes and (2) of new nonlocal boundary conditions to handle unbounded domains with PD models. These developments are expected to make PD for the first time fit for tackling a broad range of complex real-world multiphysics problems. This will be demonstrated in the last stage of the project, by the development of a PD multiphysics model of the corrosion degradable biomedical magnesium (Mg) implants.