高斯牛顿技术求解偶应力反问题

建立了便于敏度分析的偶应力反问题数值求解模型,给出了直接法和伴随法两种敏度计算格式.在反演计算中采用了高斯牛顿技术对未知本构参数进行识别,探讨了测点数目、初值选取和数据噪音对反演结果的影响,数值算例给出了令人满意的结果.     

变几何域传热的表面热流反演方法

变几何域的表面热流反演是一类特殊的热传导逆问题,在再入飞行器烧蚀型防热材料的表面热流反演中具有工程实用价值.本文首先对变几何域传热的正问题计算方法进行了校核验证,然后建立了求解变几何域表面热流反演问题的顺序函数法和共轭梯度法;给出了这两种反演方法的基本思想和算法推导,并针对典型算例进行了仿真.结果表明:两种反演方法都能计算出较好的反演结果,并且算法受测量噪声的影响较小,具有较好的鲁棒性;反演算法能适应不同的几何域变化函数,但几何域变化量的测量误差在表面热流的反演结果中会有较为直接的反映.     

聚合物熔体三维挤出胀大的数值模拟

采用有限元方法分析K-BKZ本构方程描述的聚合物熔体的三维挤出胀大.对于本构方程中偏应力张量的计算,首先给出质点的运动轨迹,分段求出局部的变形梯度张量,再求出整体的变形梯度、Cauchy-Green应变张量和 Finger应变张量,沿轨迹采用分段高斯积分计算应力.把应力作为方程的右端项,给出迭代方法,求解非线性方程组.并根据自由面处的边界条件,迭代得出出口处自由面的最终位置.对轴对称流道和矩形流道进行分析计算,并把结果与二维分析和实验结果进行了比较,显示方法是可行的.     

宏微观耦合本构模型参数识别取值范围研究

宏微观耦合本构模型的参数识别往往通过反分析方法进行,为了使参数识别结果具有高的置信度,需要确定合适的参数取值范围. 基于动态再结晶过程的微观机理以及相应本构方程的数学特征,提出一个确定参数取值范围的方法. 首先详细给出考虑动态再结晶的黏塑性本构模型,并根据模型构造物理机理,提出通过6 步确定该模型参数取值范围的方法;其次,对300M 低合金钢进行不同温度、应变速率下的热变形试验,测试宏观的流动应力-应变数据及微观的组织数据;然后应用提出的方法,依据试验数据,确定参数取值范围;最后,基于确定参数取值范围中获得的知识,对模型进行局部修改,使模型模拟结果更接近实验结果.      

基于机器学习软骨细胞的时间依赖性力学行为及本构参数反演

探究软骨细胞机械负载下的力学特性对于理解软骨细胞的正常和病理状态以及骨性关节炎的病因至关重要. 基于软骨细胞有限元计算模型的力学响应与其本构参数之间的高度复杂非线性, 本文提出了分别利用双向深度神经网络TW-Deepnets模型和随机森林RF模型并结合有限元方法来识别软骨细胞本构参数的两种反演方法. 首先, 建立了软骨细胞的无侧限压缩实验有限元模型, 收集MSnHS本构参数空间点与对应的有限元计算模型的压缩反作用力响应数据集. 其次, 结合贝叶斯超参数优化算法搭建了用于软骨细胞本构参数反求的TW-Deepnets模型和RF模型, 对有限元收集的数据进行训练, 并利用单个软骨细胞受到50%压缩程度下的实验数据对软骨细胞的MSnHS本构参数进行了反求. 最后, 通过与实验曲线的对比验证了所提出的反演方法的有效性, 并引入决定系数R2对两种模型的预测准确性进行了对比评估, 检验了模型对各本构参数的预测性能, 分析了MSnHS本构模型中各参数影响软骨细胞力学响应的重要性占比. 结果表明, 本研究提出的本构参数反演方法能够有效获取软骨细胞的本构参数值, 从而准确描述软骨细胞的时间依赖性力学特性, 该方法也可进一步推广到生物细胞在静态或动态负载条件下的复杂参数反演问题.      

热环境中黏弹性功能梯度材料及其结构的蠕变

基于经典的对应原理, 将 Mori-Tanaka 方法等细观力学结果推广于定常温度环境下的黏弹性情形. 根据泊松比与时间呈弱相关的特点, 给出 Laplace 象空间中功能梯度材料的松弛模量和热膨胀系数, 并直接建立耦合热应变的多维黏弹性本构关系. 在此基础上, 求解黏弹性功能梯度圆柱薄壳在热环境中的轴对称弯曲蠕变变形问题. 考虑材料热物参数的温度相关性, 首先确定稳态温度场, 导出相空间中轴对称弯曲变形的解析解, 采用数值反演得到蠕变变形. 算例表明, 蠕变初期, 热环境的影响明显, 随着时间增加, 热应力松弛, 影响逐渐消失. 当圆柱薄壳受轴压时, 相比于两端固支, 两端简支的端部变形更加明显. 通过圆柱薄壳的轴对称弯曲求解, 给出体积含量呈任意分布的黏弹性功能梯度结构在热机载荷下的蠕变分析途径.      

STUDY ON THE PARAMETER VALUE DOMAIN OF MACRO-MICRO CONSTITUTIVE MODEL

宏微观耦合本构模型的参数识别往往通过反分析方法进行,为了使参数识别结果具有高的置信度,需要确定合适的参数取值范围. 基于动态再结晶过程的微观机理以及相应本构方程的数学特征,提出一个确定参数取值范围的方法. 首先详细给出考虑动态再结晶的黏塑性本构模型,并根据模型构造物理机理,提出通过6 步确定该模型参数取值范围的方法;其次,对300M 低合金钢进行不同温度、应变速率下的热变形试验,测试宏观的流动应力-应变数据及微观的组织数据;然后应用提出的方法,依据试验数据,确定参数取值范围;最后,基于确定参数取值范围中获得的知识,对模型进行局部修改,使模型模拟结果更接近实验结果.     

非线性载荷作用下圆形薄膜的有限变形分析

利用修改后的应变能函数分析了作为细胞层载体的超弹性薄膜在非线性载荷作用下的有限变形问题.运用打靶法对控制方程进行了求解,讨论了本构参数n和α对薄膜变形的影响.计算结果表明:本构参数的增大会对材料产生强化作用;由于非线性载荷的影响,本构参数在n=1附近,薄膜曲率的变化趋势会相反;随着本构参数的减小,薄膜的不稳定点会在薄膜达到较小膨胀体积时出现.     

应用CBR确定面心立方金属的表面弹性参量

首先从能量变分出发基于同时考虑应变梯度效应和表面效应的跨尺度力学理论, 推导出表面能和表面弹性本构等基本关系, 然后基于简单的准连续Cauchy-Born法则(CBR)建立一种确定表面能密度以及表面弹性参量的方法.进一步以面心立方(face-centre-cubic,FCC)金属为例, 系统地获得了常用FCC金属表面弹性参量的数值, 结果与他人应用分子动力学计算得到的结果相吻合.     

考虑表面效应纳米压电双晶中波的频散分析

基于非局部应变梯度理论探究了考虑表面弹性和表面残余应力的纳米压电双晶中波的频散特性,压电双晶的上下压电层暴露在电场之中并且整体沉积在粘弹性基底之上.利用哈密顿原理和正弦剪切理论推导了控制方程,利用含非局部参数和长度尺度参数的尺度依赖本构关系得到了运动方程,带入谐波解求解相应的特征方程.数值揭示了表面弹性和表面残余应力、尺度参数和波数以及粘弹性基底对压电双晶的作用规律.研究表明,表面效应的存在对压电纳米双晶频率特性的研究至关重要,尺度参数和波数对频散特性具有耦合作用,弹性系数、阻尼系数和压电层厚度对频率的作用表现出区域性.     

Boussinesq–Flamant problem in gradient elasticity with surface energy

The strain gradient elasticity theory with surface energy is applied to Boussinesq–Flamant problem. The solution for the vertical displacements at the surface of half space due to the surface normal line load is presented. The theory includes both volumetric and surface energy terms. Boussinesq–Flamant problem in the strain gradient elasticity is solved by means of Fourier transform. The results obtained show that the vertical displacements of half space in the gradient elasticity are some different from that in the classical elasticity and the effects of the strain gradient parameters (material characteristic lengths) on the vertical displacements do exist.     

Gradient elasticity in statics and dynamics: An overview of formulations,length scale identification procedures,finite element implementations and new results

In this paper, we discuss various formats of gradient elasticity and their performance in static and dynamic applications. Gradient elasticity theories provide extensions of the classical equations of elasticity with additional higher-order spatial derivatives of strains, stresses and/or accelerations. We focus on the versatile class of gradient elasticity theories whereby the higher-order terms are the Laplacian of the corresponding lower-order terms. One of the challenges of formulating gradient elasticity theories is to keep the number of additional constitutive parameters to a minimum. We start with discussing the general Mindlin theory, that in its most general form has 903 constitutive elastic parameters but which were reduced by Mindlin to three independent material length scales. Further simplifications are often possible. In particular, the Aifantis theory has only one additional parameter in statics and opens up a whole new field of analytical and numerical solution procedures. We also address how this can be extended to dynamics. An overview of length scale identification and quantification procedures is given. Finite element implementations of the most commonly used versions of gradient elasticity are discussed together with the variationally consistent boundary conditions. Details are provided for particular formats of gradient elasticity that can be implemented with simple, linear finite element shape functions. New numerical results show the removal of singularities in statics and dynamics, as well as the size-dependent mechanical response predicted by gradient elasticity.     

Stress gradient versus strain gradient constitutive models within elasticity

A stress gradient elasticity theory is developed which is based on the Eringen method to address nonlocal elasticity by means of differential equations. By suitable thermodynamics arguments (involving the free enthalpy instead of the free internal energy), the restrictions on the related constitutive equations are determined, which include the well-known Eringen stress gradient constitutive equations, as well as the associated (so far uncertain) boundary conditions. The proposed theory exhibits complementary characters with respect to the analogous strain gradient elasticity theory. The associated boundary-value problem is shown to admit a unique solution characterized by a Hellinger–Reissner type variational principle. The main differences between the Eringen stress gradient model and the concomitant Aifantis strain gradient model are pointed out. A rigorous formulation of the stress gradient Euler–Bernoulli beam is provided; the response of this beam model is discussed as for its sensitivity to the stress gradient effects and compared with the analogous strain gradient beam model.     

Variational formulation of a simplified strain gradient elasticity theory and its application to a pressurized thick-walled cylinder problem

A detailed variational formulation is provided for a simplified strain gradient elasticity theory by using the principle of minimum total potential energy. This leads to the simultaneous determination of the equilibrium equations and the complete boundary conditions of the theory for the first time. To supplement the stress-based formulation, the coordinate-invariant displacement form of the simplified strain gradient elasticity theory is also derived anew. In view of the lack of a consistent and complete formulation, derivation details are included for the tutorial purpose. It is shown that both the stress and displacement forms of the simplified strain gradient elasticity theory obtained reduce to their counterparts in classical elasticity when the strain gradient effect (a measure of the underlying material microstructure) is not considered. As a direct application of the newly obtained displacement form of the theory, the problem of a pressurized thick-walled cylinder is analytically solved. The solution contains a material length scale parameter and can account for microstructural effects, which is qualitatively different from Lamé’s solution in classical elasticity. In the absence of the strain gradient effect, this strain gradient elasticity solution reduces to Lamé’s solution. The numerical results reveal that microstructural effects can be large and Lamé’s solution may not be accurate for materials exhibiting significant microstructure dependence.     

Gradient elasticity and nonstandard boundary conditions

Gradient elasticity for a second gradient model is addressed within a suitable thermodynamic framework apt to account for nonlocality. The pertinent thermodynamic restrictions upon the gradient constitutive equations are derived, which are shown to include, besides the field (differential) stress–strain laws, a set of nonstandard boundary conditions. Consistently with the latter thermodynamic requirements, a surface layer with membrane stresses is envisioned in the strained body, which together with the above nonstandard boundary conditions make the body constitutively insulated (i.e. no long distance energy flows out of the boundary surface due to nonlocality). The total strain energy is shown to include a bulk and surface strain energy. A minimum total potential energy principle is provided for the related structural boundary-value problem. The Toupin–Mindlin polar-type strain gradient material model is also addressed and compared with the above one, their substantial differences are pointed out, particularly for what regards the constitutive equations and the boundary conditions accompanying the solving displacement equilibrium equations. A gradient one-dimensional bar sample in tension is considered for a few applications of the proposed theory.     

Stress gradient continuum theory

A stress gradient continuum theory is presented that fundamentally differs from the well-established strain gradient model. It is based on the assumption that the deviatoric part of the gradient of the Cauchy stress tensor can contribute to the free energy density of solid materials. It requires the introduction of so-called micro-displacement degrees of freedom in addition to the usual displacement components. An isotropic linear elasticity theory is worked out for two-dimensional stress gradient media. The analytical solution of a simple boundary value problem illustrates the essential differences between stress and strain gradient models.     

Higher-order shear beam theories and enriched continuum

The buckling of higher-order shear beam-columns is studied in the light of enriched continuum. We show the equivalence between the enriched kinematics of usual higher-order shear beam theories with the nonlocal and gradient nature of the associated constitutive law. These equivalences are useful for a hierarchical classification of usual beam theories comprising Euler-Bernoulli beam theory, Timoshenko and third-order shear beam theories. A consistent variationnally presentation is derived for all generic theories, leading to meaningful buckling solutions. It is shown that Timoshenko or some other higher-order shear theories can be considered as nonlocal or gradient Euler-Bernoulli theories. The buckling problem of a third-order shear beam-column is analytically studied and treated in the framework of gradient elasticity Timoshenko theory. Some different gradient elasticity Timoshenko models are presented at the end of the paper with available buckling solutions for repetitive structures and microstructured beams.     

The Boussinesq problem in dipolar gradient elasticity

The three-dimensional axisymmetric Boussinesq problem of an isotropic half-space subjected to a concentrated normal quasi-static load is studied within the framework of dipolar gradient elasticity involving linear constitutive relations and small strains. Our main concern is to determine possible deviations from the predictions of classical linear elastostatics when a more refined theory is employed to attack the problem. Of special importance is the behavior of the new solution near to the point of application of the load where pathological singularities exist in the classical solution. The use of the theory of gradient elasticity is intended here to model the response of materials with microstructure in a manner that the classical theory cannot afford. A linear version of this theory (as regards both kinematics and constitutive response) results by considering a linear isotropic expression for the strain-energy density that depends on strain gradient terms, in addition to the standard strain terms appearing in classical elasticity and by considering small strains. Through this formulation, a microstructural material constant is introduced, in addition to the standard Lamé constants. The solution method is based on integral transforms and is exact. The present results show significant departure from the predictions of classical elasticity. Indeed, continuous and bounded displacements are predicted at the points of application of the concentrated load. Such a behavior of the displacement field is, of course, more natural than the singular behavior exhibited in the classical solution.     

优化方法用于模型校核

在工程实际中，常常会遇到需要选择本构模型和校准参数的模型校核问题。文中采用优化方法实现了复杂模型校核的自动化。文中，首先将各种本构模型映射为整数编号，并根据这些编号来激活相应的本构模型和本构模型参数，解决了本构模型的参数化问题；另外，根据模型校核问题的特点选择了试验结果和计算结果相对误差的平方和作为目标函数。通过这些步骤，用实现了优化方法模型校核自动化。     

Constitutive equations for porous plane-strain gradient elasticity obtained by homogenization

This paper addresses the problem of plane-strain gradient elasticity models derived by higher-order homogenization. A microstructure that consists of cylindrical voids surrounded by a linear elastic matrix material is considered. Both plane-stress and plane-strain conditions are assumed and the homogenization is performed by means of a cylindrical representative volume element (RVE) subjected to quadratic boundary displacements. The constitutive equations for the equivalent medium at the macroscale are obtained analytically by means of the Airy’s stress function in conjunction with Fourier series. Furthermore, a failure criterion based on the maximum hoop stress on the void surface is formulated. A mixed finite-element formulation has been implemented into the commercial finite-element program Abaqus. Using the constitutive relations derived, numerical simulations were performed in order to compute the stress concentration at a hole with varying parameters of the constitutive equations. The results predicted by the model are discussed in comparison with the results of the theory of simple materials.