自适应无网格伽辽金方法的研究

对基于位移或应变梯度的二阶导数和基于应变能密度的自适应无网格方法进行了研究。利用无网格方法结点排布灵活、结点添加删除方便等特点,给出两种自适应无网格伽辽金方法的原理、计算流程和程序实现。采用基于应变能密度的自适应方法,提出基于背景网格的误差估计和一种由结点组成的四边形局部结点加密技术,计算了悬臂梁拉伸和带小圆孔平板拉伸的算例。结果表明,自适应无网格伽辽金方法可以在较少结点数的情况下获得较准确的结果,具有一定的适用性。     

金属塑性成形过程无网格伽辽金法数值模拟技术研究

将无网格伽辽金法与刚(粘)塑性流动理论相结合,对无网格伽辽金法在金属塑性成形过程数值模拟中的应用技术进行了研究,分别采用混合变换法和罚函数法实现速度边界条件和体积不变条件,提出了基于刚(粘)塑性力学流动理论的无网格伽辽金法,给出了金属塑性成形无网格伽辽金法数值模拟的关键算法,拓展了无网格方法在金属成形领域的应用范围。典型算例的数值计算结果表明了方法的可行性。     

粗糙表面热弹塑性接触自适应无网格分析

为了在较真实地模拟接触状态的同时节省计算耗费,采用自适应无网格法求解粗糙表面热弹塑性接触问题。计算中考虑了屈服强度温度相关因素,将基于应变能梯度的自适应无网格法与线性规划-增量初应力法相结合,构建了热弹塑性接触自适应无网格分析模型,并给出相应的程序流程。通过粗糙表面与弹塑性平面热弹塑性接触算例进行验证,分别对两种不同工程材料考虑切向摩擦力、材料应变硬化和材料屈服强度温度相关等情况进行了讨论。结果表明,采用自适应无网格法能有效求解粗糙表面热弹塑性接触问题,在保证计算精度与整体加密相当的情况下,自适应加密的计算耗费约为整体加密计算耗费的10%。     

粗糙表面热弹塑性接触自适应无网格分析

为了在较真实地模拟接触状态的同时节省计算耗费,采用自适应无网格法求解粗糙表面热弹塑性接触问题.计算中考虑了屈服强度温度相关因素,将基于应变能梯度的自适应无网格法与线性规划一增量初应力法相结合,构建了热弹塑性接触自适应无网格分析模型,并给出相应的程序流程.通过粗糙表面与弹塑性平面热弹塑性接触算例进行验证,分别对两种不同工程材料考虑切向摩擦力、材料应变硬化和材料屈服强度温度相关等情况进行了讨论.结果表明,采用自适应无网格法能有效求解粗糙表面热弹塑性接触问题,在保证计算精度与整体加密相当的情况下,自适应加密的计算耗费约为整体加密计算耗费的10%.     

二维自适应粗糙表面弹塑性接触模型

提出一种基于自适应粗糙表面描述的弹塑性接触模型。根据某一给定阈值,去除粗糙表面上对接触力学行为影响较小的结点以减少计算时间。采用在计算区域划分及结点排布上较为灵活性的无网格伽辽金—有限元耦合方法求解自适应粗糙表面弹塑性接触模型。算例表明,当接触压力相对误差约为5%时,自适应接触模型的计算时间相对于非自适应接触计算可减少约50%。分析不同阈值对接触压力分布、接触间隙及接触体弹塑性应力分布的影响。     

基于刚塑性流动理论的无网格伽辽金方法

为了对金属体积成形过程进行数值模拟，给出基于刚塑性流动理论的无网格伽辽金方法。由于移动最小二乘近似的非插值特征给无网格伽辽金方法的应用带来了一定的困难，将再生核质点法中的完全变换法与无网格伽辽金方法相结合，通过对速度场的移动最小二乘近似进行修正，实现速度边界条件的精确施加；采用罚函数法引入体积不可压缩条件，运用刚塑性材料的不完全广义变分原理，给出基于刚塑性流动理论的无网格伽辽金方法的离散控制方程以及收敛判据和收敛控制方法。数值计算结果表明方法的可行性及有效性。     

刚(粘)塑性无网格伽辽金方法及其关键技术研究与应用

在国家杰出青年科学基金资助项目《塑性加工工艺及设备》(No.50425517)，以及国家自然科学基金资助项目《金属体积成形过程的刚(粘)塑性无网格伽辽金方法数值模拟理论及其关键技术研究》(No.50575125)支持下，开展刚(粘)塑性无网格伽辽金方法及其关键技术与应用的研究报告，对刚(粘)塑性无网格伽辽金方法的基础理论、数学模型建立方法、关键处理技术在金属塑性成形过程中应用研究的成果。 将无网格伽辽金方法引入塑性成形过程模拟，提出基于刚(粘)塑性理论的无网格伽辽金方法，推导刚度矩阵方程和求解列式。利用变换法施加本质边界条件，采用反正切摩擦模型描述摩擦力边界条件。对于模具边界任意的塑性成形过程，在局部坐标系下施加摩擦力边界条件，给出局部坐标系和整体坐标系的变换矩阵，解决了模具形状任意的二维塑性成形问题摩擦力边界条件的施加问题。采用直接迭代法获得初始速度场，利用Newton Raphson迭代方法求解刚度方程，给出模拟等温塑性成形问题的分析步骤。对于中高温条件下的塑性成形过程，推导出刚(粘)塑性无网格伽辽金方法热力耦合分析模型，给出热力     

用无网格法进行大变形柔性机构拓扑优化设

基于连续体结构的拓扑优化理论,将无网格伽辽金数值方法引入分布式大变形柔性机构拓扑优化设计,并解决了优化中的几何非线性问题。在优化问题中,基于各向同性固体材料惩罚模型（solidisotropic material with penalization,SIMP)和折衷规划法,同时考虑结构的柔性和刚度要求,建立了柔性机构拓扑优化的多准则优化模型,并利用优化准则法求解。采用无网格伽辽金法将求解域离散成节点,避免了有限元方法在处理大变形问题时由于使用网格而产生网格畸变等问题。求解经典算例,与基于线性理论的优化结果相比较分析,说明了该方法的正确性和有效性。     

无网格伽辽金方法在线弹性断裂力学中的应用研究

通过对移动最小二乘形函数进行局部修正,将混合变换法应用于无网格伽辽金方法,给出分析线弹性断裂力学问题的有效的无网格伽辽金方法。这一方法克服了无网格伽辽金方法中常用的拉格朗日乘子法和罚函数法的缺点,实现了本质边界条件在节点处的精确施加。运用线弹性断裂力学理论,采用基于t-分布的新型权函数和部分扩展基函数,对有限板单边裂纹的应力强度因子和拉剪复合型裂纹的扩展进行分析。由于该方法仅需节点信息,而不需要节点的连接信息,从而避免了有限元方法中的网格重构,大大简化了裂纹扩展的分析过程。数值计算结果表明了方法的有效性。     

刚塑性无网格方法中体积闭锁问题的缓解算法

将无网格伽辽金方法引入金属塑性成形过程分析,在给出基于刚塑性理论的无网格伽辽金方法及其基本理论方程的基础上,重点研究应用无网格方法分析金属塑性成形问题时所面临体积闭锁问题的缓解算法,采用体积应变率映射方法,对能量速率泛函中的体积应变率进行修正处理,通过将速度场计算的体积应变率映射到低维空间,降低独立约束方程数目,从而建立体积闭锁现象的缓解算法。应用所建立的方法对典型锻造过程进行了无网格分析,结果表明,该算法能够有效解决体积闭锁问题。     

Integrated adaptive element free Galerkin model for elastoplastic contact of engineering surfaces

Abstract

This paper presents an h-adaptive element free Galerkin (EFG) model and its formulation based on stain energy gradient for solving elastoplastic contact problems. In this model, the element free Galerkin finite element (EFG–FE) coupling method uses the initial stiffness method, and an error estimation approach based on strain energy gradient and a local adaptive refinement strategy are combined. Two-dimensional elastic contact problem between a rigid cylinder and an elastic plane is analysed to validate the adaptive elastoplastic EFG model. The results indicate that the adaptively refined solutions are accurate as compared to the Hertz theoretical solution or the uniformly refined solutions, while the cost of CPU time used by the adaptive model is less than that by the uniformly refined model under the same condition. Furthermore, the elastoplastic contacts involving a rough surface of elastic perfectly plastic and elastoplastic properties are investigated.     

Adaptive multiscale method for two-dimensional nanoscale adhesive contacts

There are two separate traditional approaches to model contact problems: continuum and atomistic theory. Continuum theory is successfully used in many domains, but when the scale of the model comes to nanometer, continuum approximation meets challenges. Atomistic theory can catch the detailed behaviors of an individual atom by using molecular dynamics (MD) or quantum mechanics, although accurately, it is usually time-consuming. A multiscale method coupled MD and finite element (FE) is presented. To mesh the FE region automatically, an adaptive method based on the strain energy gradient is introduced to the multiscale method to constitute an adaptive multiscale method. Utilizing the proposed method, adhesive contacts between a rigid cylinder and an elastic substrate are studied, and the results are compared with full MD simulations. The process of FE meshes refinement shows that adaptive multiscale method can make FE mesh generation more flexible. Comparison of the displacements of boundary atoms in the overlap region with the results from full MD simulations indicates that adaptive multiscale method can transfer displacements effectively. Displacements of atoms and FE nodes on the center line of the multiscale model agree well with that of atoms in full MD simulations, which shows the continuity in the overlap region. Furthermore, the Von Mises stress contours and contact force distributions in the contact region are almost same as full MD simulations. The method presented combines multiscale method and adaptive technique, and can provide a more effective way to multiscale method and to the investigation on nanoscale contact problems.     

Behaviour law for cutting process

To optimize the cutting conditions in machining, it is necessary to quantify energy parameters involved in the cutting process. These energy parameters are based on a hypothesis regarding the displacement of the material during the chip formation and the behaviour law of the material. However, a more refined model must be used to take into account complex phenomena in the cutting process. The strain gradient theory appears to be best in modelling rotation phenomena inside material. In order to solve this type of problem, the bases of the strain gradient theory are presented in this study.     

基于混沌映射与差分进化自适应教与学优化算法的太赫兹图像增强模型

为消除功率起伏效应引起的太赫兹(THz)图像局部伪影,构建了基于同态滤波的THz图像增强模型.然而,模型各参数取值差异大且耦合性强,给其参数确定带来了困难.为此,本文提出了混沌映射与差分进化自适应教与学优化算法以求解增强模型最优参数.首先,改进了标准Logistic混沌映射,提高了种群多样性.其次,引入适应度更新率,构...     

约束阻尼变密度材料属性合理近似插值模型全域灵敏度减振优化

针对圆柱壳约束阻尼拓扑减振优化问题,基于变密度材料属性合理近似插值模型,构建以模态损耗因子最小为优化目标的拓扑减振模型,引入振型控制因子MAC评价振型的阶跃状态,寻优迭代轻量化约束设计域内的减振优化解。推导目标灵敏度的插值模型,演算全域优化求解算法,建立∞-范数的全域灵敏度拓扑变量优化准则,避免非凸优化函数的部分目标灵敏度参与优化迭代的局部优化解和跳跃。编程实现了壳构件约束阻尼结构的减振优化分析。结果表明:全域优化算法寻优域广,迭代历程稳定时间短、敷设部分阻尼材料易获得有效的减振效果。     

基于Pareto最优原理的钻机钻进参数多目标优化

针对某双管定向钻机,提出了基于Pareto最优原理的钻进参数多目标优化方法。该方法根据钻机性能与工况,考虑水力对钻头比能影响,确定了钻进参数优化模型。针对罚函数处理约束条件的不足,引入了改进约束条件处理策略,提出了基于小生境思想拥挤度值计算方法及自适应交叉和变异算子。测试了改进算法的性能,并将改进算法用于求解基于某煤矿工程实际建立的钻机钻进参数优化模型。研究结果表明：与NSGA-Ⅱ和MOPSO算法相比,改进算法在求解测试问题时具有更好的收敛性与分布性。利用改进算法求解实际问题时得到的Pareto前端解集分布均匀,而且有效提高了机械钻速,延长了钻头寿命并降低了钻头比能。     

改进差分进化算法求解混合流水车间调度问题

对于求解混合流水车间调度问题,标准差分进化算法存在易陷入局部极值的缺点,为此,以最小化最大完工时间为目标函数建立了仿真优化模型,并提出了一种改进差分进化算法进行求解.将所提算法结合反向学习策略生成初始种群,在差分进化中进一步引入自适应差分因子,并在个体选择机制中引入模拟退火算法的Metropolis准则,有效提高了该算...     

A fast and efficient contact algorithm for fretting problems applied to fretting modes I,II and III

A computational contact algorithm is presented to solve both the normal and tangential contact problems that describe fretting contacts between two elastic half-spaces. The coupling between the normal and tangential contact problems can or not be taken into account. Nevertheless the coupling should be introduced when materials are dissimilar. Fast and efficient methods are used. The contact solver is based on a conjugate gradient method and acceleration techniques based on the Fast Fourier transforms (FFT) are employed. Very good agreements are found with analytical solutions of three fretting examples representing each fretting mode. However it is shown that these analytical solutions are based on approximations that can be too strong when materials are dissimilar.     

The superposition principle in the solution of contact friction stress problems

Comparative analyses of the solution of plane contact friction stress problems using the method of N. I. Muskhelishvili and using the superposition principle were carried out. It was established that the application of the superposition principle leads to an error whose value depends mainly upon the coefficient of friction and the Poisson coefficient. The nature of this dependence is different for the two cases of plane strain and a generalized plane stressed state. In solving contact stress problems considering frictional forces in the case of plane strain for an elastic half-space whose Poisson coefficient approximates $\text{1}\text{2}$, the superposition principle can also be applied at high friction coefficients.