期刊
COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING
卷 278, 期 -, 页码 423-444出版社
ELSEVIER SCIENCE SA
DOI: 10.1016/j.cma.2014.05.020
关键词
Impact loading; Computational homogenization; Multi-scale model; Quasi-brittle materials; Representative volume element; Dispersion effect
资金
- Netherlands Technology Foundation (STW) [10615]
A dispersive multi-scale model is presented to model failure in heterogeneous quasi-brittle materials under high frequency loading conditions. In the dispersive multi-scale model, the heterogeneous model undergoing localized failure is replaced by a homogeneous macro-scale model with a cohesive crack and a meso-scale model with diffuse damage. Each material point of the macro-scale model is linked to a heterogeneous meso-scale model. The macro-crack is modeled as a strong discontinuity and the gradient-enhanced damage model is used to model diffuse damage in the meso-scale model. The constitutive law for the bulk material is obtained from the meso-scale model analysis using a standard computational homogenization scheme. The cohesive law for the macro-crack is obtained using a continuous discontinuous homogenization scheme which is based on a failure zone averaging technique. In the dispersive multi-scale model, at the macro-scale, a dynamic analysis is performed and the meso-scale model is solved as a quasi-static problem. The meso-scale inertia forces are taken into account via a dispersion tensor which only depends on the meso-scale model material properties and the heterogeneity of the material. The meso-scale inertia effects appear as additional body forces in the macro-scale model and cause dispersion of the propagating wave. The effect of dispersion on the cohesive cracking is captured via a rate dependent cohesive law. The dispersive multi-scale model is verified against a direct numerical simulation and the objectivity of the scheme with respect to the representative volume element size is shown. (C) 2014 Elsevier B.V. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据