在本文中,我们考虑了一个样本的线弹性非均匀复合材料的弹性动力学平衡,并提出普遍定理,提供下限的总弹性应变能加上动能,和总的补充弹性能量加上动能。对于一个一般的非均匀样品,在一个单一的频率进行谐波运动,我们表明,在所有一致的边界数据,产生相同的平均应变,均匀应力边界数据呈现总弹性应变能加上动能的绝对最小值。我们还表明,在所有一致的边界数据,产生相同的平均动量的样品中,匀速边界数据呈现总的弹性余能加上动能的绝对最小值。在我们的处理中,我们不假设统计均匀性或材料各向同性,尽管它们不被排除在外。这些普遍定理是已知的静态情况下的普遍定理(Nemat-Nasser and Hori 1995)的动态等价物。据设想,本文中提出的总能量的界限将被用来制定可计算的界限上的整体动态性能的线性弹性异质复合材料的任意微观结构。
英文摘要:
In this paper we consider a sample of a linearly elastic heterogeneous composite in elastodynamic equilibrium and present universal theorems which provide lower bounds for the total elastic strain energy plus the kinetic energy, and the total complementary elastic energy plus the kinetic energy. For a general heterogeneous sample which undergoes harmonic motion at a single frequency, we show that, among all consistent boundary data which produce the same average strain, the uniform-stress boundary data render the total elastic strain energy plus the kinetic energy an absolute minimum. We also show that, among all consistent boundary data which produce the same average momentum in the sample, the uniform velocity boundary data render the total complementary elastic energy plus the kinetic energy an absolute minimum. We do not assume statistical homogeneity or material isotropy in our treatment, although they are not excluded. These universal theorems are the dynamic equivalent of the universal theorems already known for the static case (Nemat-Nasser and Hori 1995). It is envisaged that the bounds on the total energy presented in this paper will be used to formulate computable bounds on the overall dynamic properties of linearly elastic heterogeneous composites with arbitrary microstructures.
Multiple scattering theory for strong scattering heterogeneous elastic continua with triaxial inhomogeneities: theoretical fundamentals and applications
Physics-Informed Machine Learning and Uncertainty Quantification for Mechanics of Heterogeneous Materials
本文提出了一种基于物理信息神经网络(PINNs)的非均质固体弹性形变求解模型及相应的不确定性量化(UQ)。对于本研究,PINNs框架-模数由Nvidia开发的,其中我们实现了一个模块的力学非均质固体。我们使用PINN近似动量平衡,假设各向同性的线弹性本构行为对损失函数。与控制方程一起,相关的初始/边界条件也柔和地参与损失函数。分析了其中异质性表现为基质中的空隙(低弹性模量区域)和纤维(高弹性模量区域)的固体,并对从商业有限元(FE)分析包获得的解决方案验证了结果。本研究还表明,PINNs可以捕捉到精确的材料界面处的应力跳变。此外,本研究探讨了通过在几何形状和材料属性的变化与替代功能在PINN相关的优势。所提出的UQ研究表明,PINNs解决方案的平均值和标准差与Monte Carlo FE结果吻合良好。PINN预测的单个代表性空隙和单纤维复合材料的有效杨氏模量与FE预测的杨氏模量相比非常好,这建立了PINN制剂作为一种有效的均匀化工具。