平面弹性力学问题的离散元法

根据离散元的基本原理,基于变形体的理论提出了适用于平面弹性力学问题的界面位移、应变和应力模式,建立了求解平面弹性力学问题的离散元方程和相应的迭代求解方法.通过界面位移可以简洁地将位移和力的边界条件引入离散系统的控制方程,也可以方便地求解节点位移.数值算例表明,与具有相同网格的有限元结果相比,离散元能同时给出精度相对较高的应力解和精度相当的位移解.     

不可压缩平面问题的位移-压力混合重心插值配点法

引入人工压力变量,将弹性本构方程以应力、应变和压力表达,建立求解不可压缩平面弹性问题的位移-压力方程和不可压缩条件方程的耦合偏微分方程组。利用张量积型重心Lagrange插值近似二元函数,得到计算插值节点处偏导数的偏微分矩阵。采用配点法离散不可压缩弹性控制方程,利用偏微分矩阵直接离散弹性力学控制方程为矩阵形式方程组。利用插值公式离散位移和应力边界条件,将离散边界条件与离散控制方程组合为新的方程组,得到求解弹性问题的过约束线性代数方程组;利用最小二乘法求解线性方程组,得到弹性力学问题位移数值解。数值算例验证了所提方法的数值计算精度为10-14~10-10。     

基于辛弹性力学解析本征函数的有限元应力磨平方法

在实际工程结构的结构强度与优化等力学数值分析中,应力计算结果的精度是非常重要的。有限元法是得到最广泛应用的一类数值方法,并形成了众多通用的有限元程序系统。这些程序系统采用的几乎都是基于最小总势能的位移法,虽然其分析给出的有限元位移场具有较高的精度,但所得到的有限元应力场的精度较位移场大大降低。基于极坐标辛对偶体系所提供的平面弹性力学的解析辛本征展开解,并借用有限元程序系统所得到的节点位移,本文提出了一个应力分析的改进方法。数值结果表明,本方法给出的应力分析精度得到大幅提高,并具有良好的数值稳定性,可用于有限元程序系统的后处理,以提高应力尤其是关键区域应力的分析精度。     

平面弹性问题的位移-应力混合重心插值配点法

提出数值分析平面弹性问题的位移-应力混合重心插值配点法。将弹性力学控制方程表达为位移和应力的耦合偏微分方程组,采用重心插值近似未知量,利用重心插值微分矩阵得到平面问题控制方程的矩阵形式离散表达式。使用重心插值离散位移和应力边界条件,采用附加法施加边界条件,得到求解平面弹性问题的过约束线性代数方程组,应用最小二乘法求解过约束方程组,得到平面弹性问题位移和应力数值解。数值算例结果表明,重心Lagrange插值方法的计算精度可达到10~(-10)量级。位移-应力混合重心插值配点法的计算公式简单、程序实施方便,是一种高精度的无网格数值分析方法。     

薄壁弯曲结构有限元计算精度研究

基于h-p型有限元精度计算法,以薄壁弯曲结构为研究对象,系统地介绍了实体单元常见的分类方法及优缺点;通过理论公式推导了薄壁弯曲结构发生弹性和弹塑性变形时的位移和应力理论解;采用有限元法计算数值解,研究了影响有限元计算精度的因素和规律,并用算例证实了研究结果的合理性。研究结果表明:当单元类型、积分方式、阶次、长高比相同时,只有1层实体单元情况下得到的计算误差总是大于多层单元;只要严格控制单元长高比为1左右,单元层数不小于4层,采用一阶全积分六面体单元就可以控制位移及应力误差在5%以内;当采用一阶减缩积分六面体单元,只需2层单元就可以控制弹性位移误差在1%左右,但此时应力误差达30%以上,对于塑性变形,单元层数达6层时其位移误差仍达8%以上;对于二阶六面体及二阶四面体单元,只需2层单元,且不需严格控制单元长高比为1左右就可以使位移及应力计算误差在5%以内。     

分区混合有限元法计算应力强度因子

本文应用分区混合能量原理,提出分区混合有限元法,用以计算应力强度因子,方法的特点是:在裂纹尖端附近采用应力型奇异单元,在外部采用位移型常规单元。由于针对问题的受力特点,合理地把应力型与位移型、奇异元与常规元、解析解与数值解加以结合,各自发挥所长,从而能以较疏的网格取得较高的精度。 本文不仅为计算应力强度因子提供了一种有特点的有效解法,而且为分区混合有限元法的广泛应用提供了最初的例证。     

弹性力学位移构造解待定参数的优化算法研究

从弹性力学平面问题位移解析构造通解的基本原理出发,针对含未知参量的位移函数确定问题,分析了应力边界、位移边界、混合边界的离散节点需要满足的函数关系,构建了以位移解析构造解中未知参量为设计变量,以边界离散节点满足的代数关系为目标函数的优化问题,提出了获得任意边界平面问题的位移构造解中未知参量的优化求解算法,编制了任意节点边界条件的未知参量通用求解程序,给定误差计算的判定方法。求解了平面应力问题的具体实例,通过本文算法与有限元计算结果的误差对比,表明所研究算法的正确性,为任意边界的复杂工程问题求解提供依据。     

正交各向异性板平面剪切型裂纹应力强度因子的复变—变分解法

1 引言为了改善计算的精度和效率并消除离散化所带来的力学模型不确定性,本文提供了求解具有内部裂纹的有限宽板平面剪切型应力强度因子的复变-变分解法.2 各向异性边缘裂纹板的应力与位移场由二维各向异性弹性理论,满足所有基本方程的应力与位移分量可以表达为如下形式     

四节点四边形数值流形方法及其改进

本文对四节点四边形流形元提出了改进措施,将覆盖位移函数用自然坐标表示,使得在一般非规则有限数学覆盖网格下,数值积分变得比较容易,克服了现有四节点四边形流形单元数值积分困难的缺点。数值算例将其应用于复合材料数值模拟,计算结果表明,当覆盖位移函数采用完全一阶等参多项式时,计算精度较传统有限元法有很大改进。在力应集中或应力突变的区域,无需网格加密,只需提高覆盖位移函数的阶次即可。     

求解弹性力学问题的有理单元法

传统的位移有限元法采用多项式形式的位移试函数,对于边数大于4的多边形单元,构造满足单元间协调性要求的多项式形式位移插值函数是一件困难的工作。本文利用逆距离权插值的思想并考虑到单元节点的分布,建立了边数大于4多边形单元上的有理函数形式的形函数。利用有理试函数,采用Galerkin法推导出求解平面弹性力学问题的有理单元法。采用有理单元法求解弹性力学问题,求解区域根据需要可以划分为任意多边形单元,极大地提高了网格划分的灵活性。有理单元法不依赖等参变换,不同单元的形函数表达形式统一,方便计算程序的编写。     

平面广义四节点等参元GQ4及其性能探讨

广义节点有限元是将传统有限元方法中的节点广义化,在不增加节点个数的前提下,仅通过提高广义节点的插值函数的阶次,从而达到提高有限元解精度的目的.与现有的p型和hp型有限元不同,在这种新的有限元中,节点自由度全部定义在节点处,在理论与程序实现上与传统有限元方法具有很好的相容性,传统有限元方法是这种新方法的广义节点退化为0阶时的特殊情形.文中主要讨论了这一新方法的四节点等参元（记为GQ4）的形式.对GQ4进行的各种数值试验表明,所发展的广义四节点等参单元具有精度高且无剪切自锁与体积自锁等的特点.     

Mixed-mode thermal stress intensity factors from the finite element discretized symplectic method

A finite element discretized symplectic method is introduced to find the thermal stress intensity factors (TSIFs) under steady-state thermal loading by symplectic expansion. The cracked body is modeled by the conventional finite elements and divided into two regions: near and far fields. In the near field, Hamiltonian systems are established for the heat conduction and thermoelasticity problems respectively. Closed form temperature and displacement functions are expressed by symplectic eigen-solutions in polar coordinates. Combined with the analytic symplectic series and the classical finite elements for arbitrary boundary conditions, the main unknowns are no longer the nodal temperature and displacements but are the coefficients of the symplectic series after matrix transformation. The TSIFs, temperatures, displacements and stresses at the singular region are obtained simultaneously without any post-processing. A number of numerical examples as well as convergence studies are given and are found to be in good agreement with the existing solutions.     

Solution of the connection problems between a finite holed plate and a stiffener by using the partitioning concept of the generalized variational method

In this paper a new finite element method is presented, in which complex functions are chosen to be the finite element model and the partitioning concept of the generalized variational method is utilized. The stress concentration factors for a finite holed plate welded by a stiffener are calculated and the analytical solutions in series form are obtained. From some computer trials it is demonstrated that the problem of displacement compatibility and continuity of tractions between the holed plate and the stiffener is successfully analysed by using this method. Since only three elements need to be formulated, relatively less storage is required than the usual finite element methods. Furthermore, the accuracy of solutions is improved and the computer time requirements are considerably reduced. Numerical results of stress concentration factors and stresses along the welded-line which may be referential to engineers are shown in tables.     

A point assembly method for stress analysis for two-dimensional solids

A novel point assembly method (PAM) is presented for stress analysis for two-dimensional solids. In the present method, the boundaries of the problem domain are represented by a set of discrete points, and the domain itself is represented by properly scattered points. The displacement in the influence triangular areas of a point is interpolated by the displacements at the point and pairs of surrounding points using shape functions. The shape functions used in this work are obtained in the same way as those of a triangular element in the conventional finite element method (FEM). A variational (weak) form of the equilibrium equation is used to produce a set of system equations. These equations are assembled for all the points in the domain, and solved for the displacement field. Stresses and strains at a point are then computed using the displacements obtained for the point and pairs of the surrounding points. A PAM program with an automatic point-searching algorithm has been developed in fortran. Patch tests and convergence studies have been carried out to verify the convergence of the present method and program. Examples are also presented to demonstrate the efficiency and accuracy of the present method compared with analytical solutions as well as the conventional FEM solutions.     

二阶非自伴两点边值问题Garlerkin有限元的超收敛算法

提出了基于改进位移模式的二阶非自伴两点边值问题Garlerkin有限元的超收敛算法. 用常规有限元解的位移模式与高阶有限元解的位移模式之和构造新的位移模式,基于Garlerkin 方法,采用积分形式推导了单元平衡方程. 对于线性单元,本文给出了有代表性的算例,结点和单元的位移、导数都达到了h⁴阶的超收敛精度.     

影响函数与有限元应力计算

用有限元法得到位移场后,总要计算应力场。通常的做法是对位移进行微商计算应变,再根据应力-应变关系计算应力。有限元位移计算的精度比较高,但通过用位移微商来计算应力,精度会大大降低。本文利用Hamilton对偶体系的已有成果,解析求解位移和应力的影响函数,利用有限元法计算得到的位移和节点力,通过功的互等定理,可以求得一点的应力值。因影响函数是分析解,而且计算应力时不必进行微商,应力精度大幅提高。数值结果表明该方法是可行的和有效的。由该方法编制成的计算程序,可作为有限元通用程序应力计算的一个模块,将较大地提高有限元应力计算的精度和稳定性。     

Application of the extended Kantorovich method to the bending of clamped cylindrical panels

Applicability and performance of the extended Kantorovich method (EKM) to obtain highly accurate approximate closed form solution for bending analysis of a cylindrical panel is studied. Fully clamped panel subjected to both uniform and non-uniform loadings is considered. Based on the Love–Kirchhoff first approximation for thin shallow cylindrical panels, the governing equations of the problem in terms of three displacement components include a system of two second order and one forth order partial differential equations. The governing PDE system is converted to a double set of ODE systems by assuming separable functions for displacements together with utilization of the extended Kantorovich method. The resulted ODE systems are solved iteratively. In each iteration, exact closed form solutions are presented for both ODE systems. Rapid convergence and high accuracy of the method is shown for various examples. Both displacement and stress predictions show close agreement with other analytical and finite element analysis.     

从矩阵位移法看有限元应力精度的损失与恢复

矩阵位移法在计算杆端力时须叠加一个“固端力”项，而在有限元法中结点（应）力是直接对位移求导获得的，丢失了“固端力”一项，致使应力的精度大为下降．其实，对于一维有限元，同样可以对结点力叠加一个“固端力”项，使结点内力的精度与位移不相上下，而且这一做法几乎可以直接推广到半解析的有限元线法的二维问题中．本文简要介绍这一最新研究的思路、做法和一些初步的数值结果．     

High-order triangular finite element for shell analysis

A curved-shell finite element of triangular shape is described which is based on conventional shell theory expressed in terms of surface coordinates and displacements Each of the three surface displacement components is independently represented by a two-dimensional polynomial of constrained-quintic order giving the element a total of 54 degrees of freedom. Two particular geometric forms of the element are considered, viz. doubly-curved shallow and circular cylindrical. The high level of accuracy which can be achieved using few elements is demonstrated in a range of problems where comparison is made with previous finite element solutions.     

An efficient method for computing local quantities of interest in elasticity based on finite element error estimation

We present an efficient finite element method for computing the engineering quantities of interest that are linear functionals of displacement in elasticity based on a posteriori error estimate. The accuracy of quantities is greatly improved by adding the approximate cross inner product of errors in the primal and dual problems, which is calculated with an inexpensive gradient recovery type error estimate, to the quantities obtained from the finite element solution. With less CPU time, the accuracy of the improved quantities obtained with the proposed method on the coarse finite element mesh is similar to that of the quantities obtained from the finite element solutions on the finer mesh. Three quantities related to the local displacement, local stress and stress intensity factor are computed with the proposed method to verify its efficiency.