接触碰撞动力学的多变量选取方法

在柔性多体的接触碰撞动力学问题中,多变量方法基于附加约束的接触模型,将柔性体的变形用不同变量来描述：接触局部区域的变形用有限元节点坐标描述,非接触局部区域的变形用模态坐标描述,兼顾了计算精度和效率. 将该方法推广到三维空间碰撞问题,对两柔性杆纵向碰撞过程进行动力学仿真,数值结果与实验结果吻合良好,验证了该方法的有效性. 针对柔性体各自区域的变量如何选取的问题,研究了节点取法、模态阶数以及材料参数对计算结果精度的影响,寻找到合理的多变量选取方法,保证精度的同时使自由度得到最大程度的缩减.     

薄板的多变量小波有限元分析

基于区间B样条小波和广义变分原理,提出了多变量小波有限元法,构造了一种新的薄板多变量小波有限单元.由广义变分原理推导结构的多变量有限元列式,区间B样条小波尺度函数作为插值函数构造的多变量小波有限元法中,广义应力和应变被作为独立变量进行插值,避免了传统方法中应力应变求解的微分运算,减小了计算误差.区间B样条小波良好的数值...     

基于模态缩减和模态实验数据的结构连接参数识别方法

利用在结构系统可测自由度上获得的不完备模态参数和子结构的有限元模型，根据模态缩减理论，建立了识别子结构间连接子结构参数的优化模型，采用逐次二次规划法求解，改善了测试噪声和模态截断误差的影响。该方法识别精度高、收敛速度快、计算量小，便于工程应用。     

热-动力学耦合多体系统建模与降阶

在轨运行的卫星天线受到太阳辐射热冲击后容易出现热致振动或指向不准确等问题, 严重时会导致航天器失效. 本文提出了一种基于改进模态综合法的刚柔热耦合多体系统建模与降阶方法. 采用绝对节点坐标法单元形函数对柔性天线的位移场与温度场进行统一离散插值, 避免了两种物理场网格不匹配带来的映射误差与效率问题, 并使用绝对节点坐标参考节点描述中心刚体. 在系统方程中考虑了热流输入和表面自热辐射. 针对绝对节点坐标法切线刚度阵高度非线性的特点, 利用一阶泰勒展开对系统动力学和传热学方程进行了分段线性化, 在线性化区间内切线刚度矩阵为常数矩阵, 避免了每个时间步上的弹性力及其雅克比矩阵的迭代计算, 并使得基于模态的降阶手段得以应用. 利用改进的模态综合方法划分子结构并缩减系统自由度. 相邻子结构之间通过约束方程保证连接精度和连续性. 通过纯导热半圆形薄板、薄板的热膨胀、柔性太阳能电池板和刚柔热耦合抛物线天线四个数值算例验证本文所提出方法的有效性. 结果表明, 本文提出的方法在保证仿真精度的前提下缩减了系统规模, 提高了仿真计算效率.      

考虑不确定因素的结构损伤定位方法

将应变模态作为实测资料应用概率方法进行结构损伤定位.采用的模态误差考虑了建模误差和测量误差等不确定因素,避免了测量自由度与模型自由度不匹配的问题.研究了损伤状态搜索策略,并对应变模态和位移模态分别作为实测资料的定位结果进行了比较,表明采用应变模态具有更高的定位精度.     

多自由度柔性结构非线性随机振动响应分析方法

随机等效线性化方法是多自由度结构非线性随机振动分析的有效方法,然而柔性结构隐式的系统方程限制了该方法的应用.本文首先提出了把多自由度柔性结构隐式系统方程显式化的等效非线性系统建立方法,该方法结合模态分析,将系统几何非线性作用等效为模态坐标的高次多项式组合,把多自由度物理系统转化为容易求解的模态系统;在此基础上运用随机等效线性化技术建立柔性结构非线性随机振动响应分析方法.该方法通过引入虚拟激励原理,大幅提高了计算效率,使对多自由柔性结构的非线性随机振动响应分析成为可能.通过算例分析,本文分析结果和精确解与数值模拟解的比较结果表明,本文方法具有较好的计算精度和较高的计算效率,能够应用于实际柔性工程结构的随机振动响应分析.     

非节点连接有限元及其在加筋结构中的应用

针对现有加筋结构有限元模型的不足,提出了自由度层次的非节点连接方法.加筋单元的各节点可位于一个或多个其它单元内部,内节点的自由度无需全部与母单元的位移场一致;通过在节点坐标系下对内节点设置独立自由度,可模拟加筋构件与基体材料之间的粘结滑移、无粘结和体外布置等位移不连续性.节点为内节点的单元的刚度矩阵和荷载向量利用虚功原理变换到对应于其广义自由度向量的形式,按照广义自由度的位置向结构整体刚度矩阵和荷载向量组装,以此实现单元问非节点位置的连接.利用开发的有限元软件计算了多个算例,验证了非节点连接方法用于加筋结构有限元建模的正确性和便利性.     

基于对称性的三维车辆轨道耦合系统随机振动虚拟激励方法

基于结构的对称性提出了用于三维车辆轨道耦合系统高效随机动力响应分析的虚拟激励方法.车辆采用刚体动力学模型,轨道结构利用三维轨道广义单元建模,车辆与轨道通过线性轮轨关系耦合.采用虚拟激励法将高低、方向和水平三类轨道不平顺转化为一系列筒谐的虚拟不平顺;考虑车辆及轨道结构的对称性,分别推导了耦合系统的对称和反对称凝聚矩阵,提出了用于车辆轨道耦合系统动力响应计算的自由度凝聚方法,将耦合系统的自由度缩减至原来的一半以下,并在此基础上实现了耦合系统随机振动的高效分析.数值算例将本文方法与传统有限元方法进行对比,验证了本文方法的正确性和有效性.     

多变量样条有限元法

本文提出了一种基于胡海昌－鹫津久一郎三类变量广义变分原理［１，２］的多变量样条有限元法，文中应用乘积型二元三次Ｂ样条插值函数来构造板壳的广义变量场函数，由三类变量广义变分原理来建立多变理样有限元模型。在计算种种场变量时，既不必求导，又无需用应力应变关系式，可直接算得其结果，因而对各种场变量均有足够的精度，文中，还解算了振动与稳定特征值问题，均得到了精度较好的数值结果。     

柔性多体碰撞问题的多变量方法

针对柔性多体系统碰撞问题的特点, 提出了柔性多体系统碰撞问题的多变量方法. 在未碰撞阶段和分离后的阶段, 多变量方法考虑了物体大范围运动与变形的耦合.在碰撞阶段, 考虑非碰撞区域的变形, 对非碰撞区域用浮动参考系方法求解, 对发生碰撞的局部区域用非线性有限元方法求解. 非碰撞区域采用浮动基位形坐标和柔性体相对浮动基的变形模态坐标描述, 大大减缩了变量的维数,提高了计算效率. 碰撞局部区域采用非线性有限元节点坐标描述, 可以得到碰撞局部区域高精度的应力、应变响应, 而且可以反映碰撞局部区域的大变形、塑性等非线性响应. 该方法既能体现碰撞对系统大范围运动和长期动力学仿真的影响, 又能精确反映出碰撞发生的微小时间段内物体碰撞局部区域的应力、应变响应, 表现出碰撞过程中物体接触区域的演化历程.介绍了两杆正碰撞与铅垂面内双摆撞击自由圆盘的算例, 设计了两柔性杆正碰撞实验,多变量方法的数值仿真结果与实验测量结果很好地吻合.      

基于应变能等效指标的结构损伤识别技术研究

为了解决结构的多损伤识别问题,提出了一种基于应变能等效指标的损伤识别方法。首先给出了损伤前后模态应变能变化的表达式以及能量耗散公式,然后根据结构的能量耗散与应变能的变化值等价的原理,建立了一个四阶等效方程,最后求出了该方程的四个根,并通过对该方程四个根的分析得到了一个应变能等效指标,通过该指标可以方便的求解损伤的位置和程度。数值仿真结果表明,基于应变能等效指标的损伤识别方法不仅可以精确的识别出损伤的位置和程度,而且其识别精度明显好于应变能耗散率方法。     

Nonlinear dynamics of doubly curved shallow microshells

The nonlinear dynamical characteristics of a doubly curved shallow microshell are investigated thoroughly. A consistent nonlinear model for the microshell is developed on the basis of the modified couple stress theory (MCST) in an orthogonal curvilinear coordinate system. In particular, based on Donnell’s nonlinear theory, the expressions for the strain and the symmetric rotation gradient tensors are obtained in the framework of MCST, which are then used to derive the potential energy of the microshell. The analytical geometrically nonlinear equations of motion of the doubly microshell are obtained for in-plane displacements as well as the out-of-plane one. These equations of partial differential type are reduced to a large set of ordinary differential equations making use of a two-dimensional Galerkin scheme. Extensive numerical simulations are conducted to obtain the nonlinear resonant response of the system for various principal radii of curvature and to examine the effect of modal interactions and the length-scale parameter.     

Strain Energy Functions for a Poisson Power Law Function in Simple Tension of Compressible Hyperelastic Materials

The most general strain energy function that yields a power law relationship between the principal stretches in the simple tension of nonlinear, elastic, homogeneous, compressible, isotropic materials is obtained. The approach taken generalises that used by Blatz and Ko. The strain energy function obtained depends on the choice of two stretch invariants. The forms of the strain energy function for a number of such choices are obtained. Finally, some consequences of the choice of strain energy function on the stress–strain relationship for uniaxial tension are investigated. This revised version was published online in July 2006 with corrections to the Cover Date.     

An improved Jones-Nelson nonlinear material model and rubber composite lamina analysis

A Jones-Nelson model has been applied to depict nonlinear stress-strain relations of composite laminae, where mechanical properties were expressed by strain energy density. The nonlinear material matrix is only a function of the strain energy density. Then a material model could be conveniently applied under complex stress condition. In this paper, by introducing large displacement stress-strain measurement and varying-Poisson's ratio idea, an improved Jones-Nelson material model is presented, where the expanding problem of material properties and convergence problems are overcome. Meanwhile a discuss of the reorientation of fiber and a material nonlinear analysis of rubber composite lamina under super large deformation conditions are made. The prediction results of improved material model are in fairly good agreement with those of the experiments.     

A direct approach to fiber and membrane reinforced bodies. Part II. Membrane reinforced bodies

The paper deals with membrane reinforced bodies with the membrane treated as a two-dimensional surface with concentrated material properties. The bulk response of the matrix is treated separately in two cases: (a) as a coercive nonlinear material with convex stored energy function expressed in the small strain tensor, and (b) as a no-tension material (where the coercivity assumption is not satisfied). The membrane response is assumed to be nonlinear in the surface strain tensor. For the nonlinear bulk response in Case (a), the existence of states of minimum energy is proved. Under suitable growth conditions, the equilibrium states are proved to be exactly states of minimum energy. Then, under appropriate invertibility condition of the stress function, the principle of minimum complementary energy is proved for equilibrium states. For the no-tension material in Case (b), the principle of minimum complementary energy (in the absence of the invertibility assumption) is proved. Also, a theorem is proved stating that the total energy of the system is bounded from below if and only if the loads can be equilibrated by a stress field that is statically admissible and the bulk stress is negative semidefinite. Two examples are given. In the first, we consider the elastic semi-infinite plate with attached stiffener on the boundary (Melan’s problem). In the second example, we present a stress solution for a rectangular panel with membrane occupying the main diagonal plane.     

Nonlinear formulation for flexible multibody system with large deformation

In this paper, nonlinear modeling for flexible multibody system with large deformation is investigated. Absolute nodal coordinates are employed to describe the displacement, and variational motion equations of a flexible body are derived on the basis of the geometric nonlinear theory, in which both the shear strain and the transverse normal strain are taken into account. By separating the inner and the boundary nodal coordinates, the motion equations of a flexible multibody system are assembled. The advantage of such formulation is that the constraint equations and the forward recursive equations become linear because the absolute nodal coordinates are used. A spatial double pendulum connected to the ground with a spherical joint is simulated to investigate the dynamic performance of flexible beams with large deformation. Finally, the resultant constant total energy validates the present formulation. The project supported by the National Natural Science Foundation of China (10472066, 10372057). The English text was polished by Yunming Chen.     

改进的Jones—Nelson物理非线性材料模型与橡胶复合材料层板分析

Ｊｏｎｅｓ－Ｎｅｌｓｏｎ模型是复合材料物理非线性应力应变关系的一种描述方法，它通过建立材料刚度与应变能密度的关系以及考虑纤维增强复合材料的物理非线性问题，非线性的材料矩阵只为应变能的子数，使得材料模型可以方便地应用地复杂应力状态下，通过 在大变形的应变应力变量和变泊松比概念，对Ｊｏｎｅｓ－Ｎｅｌｓｏｎ模型进行了改进，解决了材料特性的扩展问题和收敛问题，同时考虑了大变形中纤维铺设角度的重新取向，使其     

A thermodynamic approach to nonlinear ultrasonics for material state awareness and prognosis

We develop a thermodynamic framework for modeling nonlinear ultrasonic damage sensing and prognosis in materials undergoing progressive damage. The framework is based on the internal variable approach and relies on the construction of a pseudo-elastic strain energy function that captures the energetics associated with the damage progression. The pseudo-elastic strain energy function is composed of two energy functions—one that describes how a material stores energy in an elastic fashion and the other describes how material dissipates energy or stores it in an inelastic fashion. Experimental motivation for the choice of the above two functionals is discussed and some specific choices pertaining to damage progression during fatigue and creep are presented. The thermodynamic framework is employed to model the nonlinear response of material undergoing stress relaxation and creep-like degradation. For each of the above cases, evolution of the nonlinearity parameter with damage as well as with macroscopic measurables like accumulated plastic strain is obtained.     

Identification of Nonlinear Oscillator with Parametric White Noise Excitation

An identification technique is proposed for a nonlinear oscillator excited by response-dependent white noise. Stiffness, damping and excitation are estimated from records of the stationary stochastic response. The estimation of the stiffness is based on a nonparametric procedure in which the potential energy at the displacement extremes is obtained from the kinetic energy at the previous mean-level-crossing. A nonparametric estimate is obtained by an iterative averaging, in which the increased knowledge of the potential energy in each step is used to avoid bias. The second step in the procedure is to estimate the stationary probability potential in a nonparametric form from a histogram of the kinetic energy at mean-level-crossings. The damping is also estimated in a nonparametric way from approximate expressions of the covariance functions of a set of modified phase plane variables at a given energy level. Finally, the excitation is estimated from a relation between the stationary probability potential, the damping and the excitation. The separation of damping and excitation requires a parametric representation. The system identification technique is investigated by application to response records obtained by stochastic simulation. The stiffness estimation generally gives excellent results, while the damping and excitation estimation tend to be slightly biased for systems with strongly nonlinear stiffness.     

Growth estimates and lower bounds for solutions in nonlinear elastodynamics with indefinite strain energy

This paper investigates questions of nonexistence and growth of weak solutions of a system of equations of nonlinear elastodynamics under various hypotheses on the data and on the form of the strain energy function.