拉格朗日反分析方法误差分析

系统地讨论了路径线法及曲面拟合法误差发展。反分析计算误差可分成截断误差及增殖误差两部分，截断误差及增殖误差受测量量计数目及量计间距影响，增加量计数，减小量计间距可以减小截断误差，但不利于控制增殖误差。反分析计算误差发展可分为三个过程:开始，计算误差主要来源于截断误差；其后，误差受截断误差和增殖误差共同影响；最后，增殖误差是主要的。无论是曲面拟合法还是路径线法，各量计线上增殖误差可用时间多项式函数来表示，且多项式最高幂次数与量计线数目有关。     

基于径向基函数的无单元法求解力学问题误差分析

径向基函数形状参数的选择在无单元法数值计算中一直是一个热门的问题,现在已总结出许多确定形状参数的经验公式. 但还没有相关研究表明这些形状参数是如何随着影响域尺寸而变化的. 本文研究了MQ(multi-quadrics) 径向基函数中形状参数对无单元法计算误差的影响. 首先,从理论上分析了形函数导数随着形状参数值的变化趋势,和以计算点为中心节点对称布置与不对称布置的形函数导数的变化规律;然后分析了影响域尺寸对误差的影响,得到了在不同影响域尺寸下,误差随形状参数值变化的规律;在此基础上,给出了影响域范围值.     

双重点最小二乘配点无网格法

配点类无网格法需要计算近似函数的二阶导数,因而在移动最小二乘(MLS)近似中至少要采用二次基函数。本文利用Voronoi图对双重点移动最小二乘近似法进行了改进,建立了基于Voronoi图的双重点移动最小二乘近似(VDG),并利用加权最小二乘法离散微分方程,导出了双重点最小二乘配点无网格法(MD GLS)。该方法将求解域用节点离散,并以节点为生成点建立Voronoi图,取Voronoi多边形的顶点为辅助点。近似函数及其二阶导数的计算过程可分解为两个步骤:首先用场函数节点值拟合辅助点处近似函数的一阶导数,再以辅助点处近似函数的一阶导数值拟合节点处近似函数的二阶导数。由于在每一步中只需计算MLS形函数及其一阶导数,这种近似方法需要较少的影响点和较小的影响域。同时借助于Voronoi结构的优良几何性质,可以快速地搜索影响点。研究表明,与基于MLS的加权最小二乘无网格法(MWLS)相比,这种方法可以显著提高计算效率,并且在精度和收敛性方面也有所改善。     

基于T样条的变网格等几何薄板动力学分析

具有大位移、大变形的薄板在接触碰撞等工况下, 其局部应变会产生剧烈变化. 为了保证对其进行动力学分析的精度和计算效率, 本文整合计算机辅助设计(CAD)与计算机辅助工程(CAE)系统, 提出了一种基于T样条曲面的变网格柔性系统等几何分析方法. 首先, 建立基于T样条曲面单元的基尔霍夫薄板运动学模型, 并根据非线性格林?拉格朗日应变建立由T样条曲面单元离散的薄板弹性模型. 其次, 通过在T网格中的局部区域插入节点的方式, 达到T样条曲面网格局部更新的目的. 利用T样条混合函数细化算法得到计算新广义变量的转换矩阵, 并结合广义α法创建了变自由度系统动力学方程的求解算法, 形成了系统的T样条单元局部细化算法. 最后, 静力学算例与柔性单摆模型分别验证了T样条薄板弹性模型的正确性, 以及T样条薄板单元在动力学分析上的精度和收敛性. 通过对受冲击柔性薄板的动力学分析表明, 本文所提出T样条单元及局部细化算法可以只在接触碰撞等应变剧烈变化的区域实现局部网格细化, 从而控制系统自由度数, 提高计算效率.      

反平面断裂问题的无单元伽辽金比例边界法

将比例边界法与无单元伽辽金法相结合,建立了反平面断裂分析的无单元伽辽金比例边界法。这是一种边界型无网格法,在环向方向上采用无单元伽辽金法进行离散,因此计算时仅需要边界上的节点信息,不需要边界元所要求的基本解。为了便于施加本质边界条件,通过建立节点值和虚拟节点值之间的关系给出了修正的移动最小二乘形函数。在径向方向上,该方法利用解析的方法求解,因此是一种半解析的数值方法。最后,给出了数值算例,并验证了所提方法后处理简单和计算精度高的特点,适合于求解反平面断裂问题。     

用自然三角形单元计算应力强度因子

用有限元法计算应力强度因子的各种奇异元中,由八节点或十二节点的四边形等参数单元所派生的奇异元最为简单,因为这种单元就是原有的等参数单元,无须程序变动,只要适当地改变裂纹顶点单元的边中点的坐标,就可使应变具有γ~(-1/2)的奇异性。这种单元是协调单元,满足收敛准则,当有人提出过渡单元概念以后,使计算结果更为精确,因此,这种单元是工程中广泛应用的一种计算应力强度因子的奇异元。与四边形等参数单元相应的一族单元是自然三角形单元。下面证明六节点的二次三角形单元和十节点的三次三角形单元,尽管它们的形状函数与上述四边形二次和三次单元通过一边收缩成一点而成的三角形单元的形状函数不同,但通过边上点取与上述等参数单元相同的位置,也可得到角点的应变具有γ~(-1/2)的奇异性。不同的是,对于六节点     

结点应力修正计算中边值问题的简便处理

在有限元分析中,对于利用数值积分的曲面(线)单元,我们发现在积分点上的应力精度远高于单元边界结点上的应力精度。本文以积分点的应力计算为基础,用带线性的约束的最小二乘法,简便地处理了曲面(线)单元结点应力修正计算中的边值问题。     

相关识别中的曲面拟合法

本文讨论了相关识别中的曲面拟合法。利用相关系数在真实匹配点附近的统计特性,通过对相关系数进行曲线（一维）或曲面（二维）拟合,根据其一阶导数确定最大相关系数所在的位置,快速得到亚像素匹配点。与逐步搜索法相比,该方法大大减少了运算次数,提高了数字图像相关运算的速度。     

基于参数曲面三维势问题的边界面法

提出了边界元法(BEM)的一种新的实现方法——边界面法(BFM)。在传统的边界元法中,单元不仅用来进行边界积分和函数插值,而且用来近似几何体。当离散网格较稀疏时,会引起较大几何误差,因而影响计算精度。本文基于参数曲面,将几何实体的边界曲面离散为参数空间里的曲面单元,边界积分和场变量的插值都是在曲面参数空间里进行。积分点...     

计算平面问题应力强度因子的任意高阶奇应变有限单元

W.K.Wilson提出的高阶奇应变圆单元(SSC)把有限单元法和裂纹尖端附近的线弹性解析解结合起来,能有效地计算应力强度因子K_Ⅰ或K_Ⅱ。A.Holston Jr.进一步把对称和反对称位移叠加起来,用以计算复合型应力强度因子K_Ⅰ和K_Ⅱ。但他们在对称和反对称的位移函数中最多只取到4项。胡海昌建议取更多的项,以便把圆单元的半径     

A new analytical definition of the dead material zone for forward extrusion of shaped sections

This article aims to provide a new formulation for the analysis of the extrusion process for non-axisymmetric sections. The upper bound theorem has been used to obtain a generalized kinematically admissible velocity field. The geometry of the deforming region has been formulated considering variation of the dead zone size at different angular positions and three-dimensional curved surfaces have been employed to define the entry and exit surfaces of the deformation zone. Using this analytical method, extrusion of square, rectangular and L-shaped sections were analyzed and the effect of shape complexity on material flow and dead material zone (DMZ) formation under different conditions has been investigated. Physical modelling experiments and finite element analysis were carried out to reveal the capability of the proposed theoretical method.     

压电智能结构荷载识别方法的研究

采用压电智能结构实测荷载的输出响应，基于BP神经网络与有限元逆分析提出一种识别荷 载位置及大小的方法. 首先在结构的不同位置施加单位荷载由有限元方法计算得到网络的学 习样本，经网络作逆分析识别荷载位置，继而通过有限元逆逼近方法确定荷载大小的最小二 乘解. 数值算例表明，该方法计算速度快、精度高，不受结构几何形状和边界条件的限制， 用于识别实际压电智能结构不确定荷载的位置及大小是可行的.     

高效的三维曲梁单元

三维井眼中延伸数千米的三维细长圆截面钢钻柱应力分析问题是一个复杂的力学问题，通常使用有限元数值分析方法对其进行受力分析。而在进行有限元分析时，现有的圆弧曲粱单元和空间直粱单元在几何上都不能很好地模拟三维曲线形状的钻柱。为了确保计算精度．其单元划分势必不能过大，结果是计算时间长，收敛性差。为了解决这一问题，显然必须构建一种新的较有效的曲梁单元。基于自然坐标系，依据圆截面空间曲粱单元节点有6个自由度——3个线位移和3个角位移，利用包含全部刚体位移模式和常应变的形函数，忽略剪切变形，假设变形后的梁轴线的弯曲曲率改变为线性变化，建立起了保证收敛性的具有12个自由度的有初始曲率和挠率的圆截面空间曲梁的有限元模型。为了证明给出的有限元模型的高效性，分析了几个静态问题，并与现有文献中的解析解或数值结果进行了比较。基于所给出的结果，可望该有限元模型可以作为分析三维空间曲粱结构的有效工具。     

Scaling investigation of plasma-induced flows over curved and flat surfaces: Comparison to the wall jet

The present investigation deals with Dielectric Barrier Discharge (DBD) induced jets flowing over curved surfaces and studied in the framework of a circulation control application, carried out by acting near the rounded trailing-edge of an airfoil. These jets are characterized experimentally via Particle Image Velocimetry (PIV) in quiescent air conditions. The study assesses the evolution of these flows in terms of self-similarity of the mean flow and of its turbulent components. DBD wall jets evolution in the streaming direction is also analyzed through the rate of spread and the maximum velocity decay evolution as commonly done for fluidic wall jets, and also through several normalized quantities deriving from different length and velocity scales. A comparison with a canonical flow, such as the classical wall jet flowing over plane or curved surfaces, is made in order to find out the similarities and the discrepancies between these two flows. Results reveal that DBD wall jets and canonical fluidic wall jets show comparable properties in the diffusion zone. Compared to the plane DBD wall jet, centrifugal forces are responsible of the greater spread of curved DBD wall jets and are likely the source of instabilities leading to their transitional state. The momentum flux of the induced jet and the radius of curvature of the surface are two relevant scales for DBD induced flows developing over curved surfaces.     

Out-of-plane responses of a circular curved Timoshenko beam due to a moving load

For the cases of using the finite curved beam elements and taking the effects of both the shear deformation and rotary inertias into consideration, the literature regarding either free or forced vibration analysis of the curved beams is rare. Thus, this paper tries to determine the dynamic responses of a circular curved Timoshenko beam due to a moving load using the curved beam elements. By taking account of the effect of shear deformation and that of rotary inertias due to bending and torsional vibrations, the stiffness matrix and the mass matrix of the curved beam element were obtained from the force–displacement relations and the kinetic energy equations, respectively. Since all the element property matrices for the curved beam element are derived based on the local polar coordinate system (rather than the local Cartesian one), their coefficients are invariant for any curved beam element with constant radius of curvature and subtended angle and one does not need to transform the property matrices of each curved beam element from the local coordinate system to the global one to achieve the overall property matrices for the entire curved beam structure before they are assembled. The availability of the presented approach has been verified by both the existing analytical solutions for the entire continuum curved beam and the numerical solutions for the entire discretized curved beam composed of the conventional straight beam elements based on either the consistent-mass model or the lumped-mass model. In addition to the typical circular curved beams, a hybrid curved beam composed of one curved-beam segment and two identical straight-beam segments subjected to a moving load was also studied. Influence on the dynamic responses of the curved beams of the slenderness ratio, moving-load speed, shear deformation and rotary inertias was investigated.     

对称变截面平面曲杆的单元刚度矩阵

本以一种新的思路和做法对变截面曲杆进行单元分析,直接把变截面曲杆作为单元,采用弹性中心法导出了单元刚度矩阵通用公式,确定了截面按余弦规律变化时,轴线分别为圆弧形、抛物线形、半椭圆形平面曲杆的单元刚度矩阵精确式,供结构设计参考使用。     

On Taylor weak statement finite element methods for computational fluid dynamics

A Taylor series augmentation of a weak statement (a ‘Taylor weak statement’ or ‘Taylor-Galerkin’ method) is used to systematically reduce the dispersion error in a finite element approximation of the one-dimensional transient advection equation. A frequency analysis is applied to determine the phase velocity of semi-implicit linear, quadratic and cubic basis one-dimensional finite element methods and of several comparative finite difference/finite volume algorithms. The finite element methods analysed include both Galerkin and Taylor weak statements. The frequency analysis is used to obtain an improved linear basis Taylor weak statement finite element algorithm. Solutions are reported for verification problems in one and two dimensions and are compared with finite volume solutions. The improved finite element algorithms have sufficient phase accuracy to achieve highly accurate linear transient solutions with little or no artificial diffusion.     

Linear viscoelastic analysis of straight and curved thin-walled laminated composite beams

This paper is devoted to study the behavior, in the range of linear viscoelasticity, of shear flexible thin-walled beam members constructed with composite laminated fiber-reinforced plastics. This work appeals to the correspondence principle in order to incorporate in unified model the motion equations of a curved or straight shear-flexible thin-walled beam member developed by the authors, together with the micromechanics and macromechanics of the reinforced plastic panels. Then, the analysis is performed in the Laplace or Carson domains. That is, the expressions describing the micromechanics and macromechanics of a plastic laminated composites and motion equations of the structural member are transformed into the Laplace or Carson domains where the relaxation components of the beam structure (straight or curved) are obtained. The resulting equations are numerically solved by means of finite element approaches defined in the Laplace or Carson domains. The finite element results are adjusted with a polynomial fitting. Then the creep behavior is obtained by means of a numerical technique for the inverse Laplace transform. Predictions of the present methodology are compared with experimental data and other approaches. New studies are performed focusing attention in the flexural–torsional behavior of shear flexible thin-walled straight composite beams as well as for thin-walled curved beams and frames.     

A non-horizontal multi-layer numerical model for the flow in natural rivers

A non-horizontal multi-layer element model is developed for the simulation for the flow in natural rivers. Either Cartesian coordinates or sigma coordinates will experience difficulties in dealing with the water surface and irregular bed topography of natural rivers. To create the surface-fitting and non-deformed cells, the newly developed model divides the water column into several layers with non-horizontal interfaces which are nearly parallel to the water surface. The irregular bed topography is also represented by the layered integration between non-horizontal interfaces. Two case studies for the flow in a straight channel and the flow in an S-shaped meander channel are conducted with good agreement between the numerical predictions and the analytical or experimental results. The model is further applied for the investigation of the flow in a 12-km-long and 3.46-m-drop reach of the Yangtze River with the water surface evaluation and the stream-wise velocity satisfactory accordance with the observed data.