Kovačič Aljaž, Borovinšek Matej, Vesenjak Matej, Ren Zoran
Faculty of Mechanical Engineering, University of Maribor, 2000 Maribor, Slovenia.
Materials (Basel). 2018 Jan 26;11(2):193. doi: 10.3390/ma11020193.
This paper addresses the problem of reconstructing realistic, irregular pore geometries of lotus-type porous iron for computer models that allow for simple porosity and pore size variation in computational characterization of their mechanical properties. The presented methodology uses image-recognition algorithms for the statistical analysis of pore morphology in real material specimens, from which a unique fingerprint of pore morphology at a certain porosity level is derived. The representative morphology parameter is introduced and used for the indirect reconstruction of realistic and statistically representative pore morphologies, which can be used for the generation of computational models with an arbitrary porosity. Such models were subjected to parametric computer simulations to characterize the dependence of engineering elastic modulus on the porosity of lotus-type porous iron. The computational results are in excellent agreement with experimental observations, which confirms the suitability of the presented methodology of indirect pore geometry reconstruction for computational simulations of similar porous materials.
本文针对计算机模型中重建莲花型多孔铁逼真的不规则孔隙几何形状这一问题展开研究，该模型在其力学性能的计算表征中允许简单的孔隙率和孔径变化。所提出的方法使用图像识别算法对真实材料样本中的孔隙形态进行统计分析，由此得出特定孔隙率水平下孔隙形态的唯一指纹特征。引入了代表性形态参数，并用于间接重建逼真且具有统计代表性的孔隙形态，这些形态可用于生成具有任意孔隙率的计算模型。对这类模型进行了参数化计算机模拟，以表征工程弹性模量对莲花型多孔铁孔隙率的依赖性。计算结果与实验观测结果高度吻合，这证实了所提出的间接孔隙几何形状重建方法适用于类似多孔材料的计算模拟。
