Department of Natural Sciences (NAT), Mid Sweden University, Sundsvall, Sweden.
Soft Matter. 2018 Nov 14;14(44):8945-8955. doi: 10.1039/c7sm02561k.
High-porosity, three-dimensional wood fiber networks made by foam forming present experimentally accessible instances of hierarchically structured, athermal fiber networks. We investigate the large deformation compression behavior of these networks using fiber-resolved finite element analyses to elucidate the role of microstructures in the mechanical response to compression. Three-dimensional network structures are acquired using micro-computed tomography and subsequent skeletonization into a Euclidean graph representation. By using a fitting procedure to the geometrical graph data, we are able to identify nine independent statistical parameters needed for the regeneration of artificial networks with the observed statistics. The compression response of these artificially generated networks and the physical network is then investigated using implicit finite element analysis. A direct comparison of the simulation results from the reconstructed and artificial network reveals remarkable differences already in the elastic region. These can neither be fully explained by density scaling, the size effect nor the boundary conditions. The only factor which provides the consistent explanation of the observed difference is the density and fiber orientation nonuniformities; these contribute to strain-localization so that the network becomes more compliant than expected for statistically uniform microstructures. We also demonstrate that the experimentally manifested strain-stiffening of such networks is due to development of new inter-fiber contacts during compression.
由泡沫成型制成的高孔隙率、三维纤维网络提供了实验可及的分层、非热纤维网络实例。我们使用纤维分辨有限元分析研究了这些网络的大变形压缩行为,以阐明微观结构在压缩响应中的作用。使用微计算机断层扫描和随后的体素化为欧几里得图表示来获取三维网络结构。通过对几何图形数据进行拟合程序,我们能够识别出九个独立的统计参数,这些参数对于用观察到的统计数据再生人工网络是必需的。然后使用隐式有限元分析研究这些人工生成的网络和物理网络的压缩响应。从重建网络和人工网络中直接比较模拟结果,在弹性区域就已经显示出显著的差异。这些差异既不能完全用密度缩放、尺寸效应或边界条件来解释。唯一能够对观察到的差异提供一致解释的因素是密度和纤维取向不均匀性;这些因素导致应变局部化,从而使网络比统计均匀微观结构所预期的更具弹性。我们还证明,这种网络的实验表现出的应变硬化是由于在压缩过程中形成新的纤维间接触所致。