首页 | 本学科首页   官方微博 | 高级检索  
相似文献
 共查询到20条相似文献,搜索用时 109 毫秒
1.
基于LCF-Kriging模型的结构多失效模式可靠度计算   总被引:1,自引:0,他引:1  
针对多失效模式下结构体系可靠度计算中的代理模型构建成本与计算精度如何权衡的问题,本文以减小体系失效概率预测方差为出发点,推导出最大贡献函数(LCF-Largest Contribution Function)来识别对体系失效概率方差影响较大的样本。LCF函数可减少对体系失效概率方差影响较小区域内样本数量,进而提高代理模型的计算效率;通过置信水平和允许相对误差建立LCF函数的学习停止条件,能够保证已有样本信息不浪费。本文选取能够对多个功能函数联合构建的多输出Kriging模型作为代理模型,基于LCF-Kriging模型并结合MCS对体系可靠度进行计算,功能函数的相关性可通过各失效模式的逻辑关系予以考虑。数值算例表明,在适当的学习停止条件下,对于串联、并联和串并混联的结构体系可靠度评估,本文方法均能在计算精度和计算效率之间达到满意平衡。  相似文献   

2.
对于具有多失效模式的结构可靠度计算问题,利用多输出Kriging模型作为代理模型进行分析。该代理模型只需对所有功能函数进行一次建模,无需对每个功能函数建立各自的代理模型,且在建模过程中能够考虑各失效模式之间的相关性。本文方法设定的初始样本点不仅对随机变量均值附近区域给予足够重视,而且能够兼顾设计空间的边缘区域,进而确保初始代理模型在全局空间内具有较好精度,以减少后续利用学习函数更新代理模型的次数。数值算例表明,本文方法具有较好的计算精度和较高的计算效率,当失效模式较多时,计算效率大幅提升。  相似文献   

3.
进行复杂结构可靠度分析时,由于涉及隐式功能函数和耗时的数值计算,减少结构模型的调用次数在提高分析效率方面显得尤为重要。为此,本文基于贝叶斯支持向量回归机,提出了一种高效的自适应可靠度分析方法。该方法利用贝叶斯支持向量机提供的概率估计信息(均值和方差)构建学习函数,同时通过引入样本间的距离测度防止选取与现有样本过于临近的冗余点,进而能快速有效地选取极限状态曲面附近具有代表性的样本点,以提高代理模型的构建速度和预测精度。此外,在学习过程中引入了有效抽样域策略,有针对性地选取对失效概率估计误差贡献大的点,从而进一步提升结构可靠度分析的计算效率。最后,通过数值算例验证了本文方法对结构可靠度分析的适用性和有效性。  相似文献   

4.
对于具有多失效模式的结构,基于可靠性的结构优化计算成本是比较昂贵的。本文利用多输出高斯过程MOGP(Multiple Output Gaussian Process)代理模型以降低计算成本,首先利用Bucher方法生成初始样本,然后结合均匀训练样本和学习函数对MOGP代理模型进行构建。学习函数可在大范围内筛选出较为满意的训练样本,能够确保MOGP代理模型具有较好的全局精度,在整个优化过程中不再重新构建MOGP代理模型。利用协方差矩阵,MOGP代理模型能够考虑各失效模式的相关性,对多输入多输出系统具有良好的预测性能。数值算例表明,本文方法具有较好的计算结果,且计算效率较高,尤其是设计变量数目与失效模式数目较多时效率提升明显。  相似文献   

5.
苏永华  杨红波 《应用力学学报》2012,29(6):705-710,776
针对基于极限平衡模式而建立的边坡稳定可靠度功能函数的隐式特征及其导致的可靠度求解等问题,首先利用Kriging插值预测方法,以极限平衡理论的Janbu算法为例,建立了边坡稳定可靠度功能函数的新型显式代理模型;然后将拉丁超立方试验抽样设计方法、Janbu算法、几何可靠度指标计算三者结合起来,研制了该代理模型的求解程序;构建了基于Kriging插值预测的边坡稳定可靠度近似分析新方法。选用某著名边坡算例,将本文方法分析结果与基于蒙特卡洛法的精确解进行了对比,两者的相对误差为5.21%,但前者计算工作量不到后者的万分之一,表明了代理模型方法的实用性和有效性。最后用代理模型方法分析了衡桂高速某段路堑边坡的稳定可靠性,表明本文方法计算结果能够满足工程精度要求,具有一定的应用价值。  相似文献   

6.
采用重要抽样法的结构动力可靠度计算   总被引:2,自引:0,他引:2       下载免费PDF全文
首次对比分析了结构动力可靠度计算的三种重要抽样法,并对部分方法进行了补充修正.单元失效域法补充了依据随机教决定抽样区间的产生方法,根据单元失效域的条件概率和权重系数给出重要抽样密度函教.方差放大系数法直接通过激励过程的特性给出重要抽样密度函数的具体表达式.功率谱法的重要抽样密度函数仅为激励幅值的函数,根据结构反应的功率谱密度增大激励幅值的方差,建议幅值样本值的联合概率密度函数可表示为幅值样本值分量的概率密度函数的连乘形式.结果表明:对于线性体系三种方法的计算效率均比Monte-Carlo法有显著提高,而单元失效域法的计算效率又比另两种方法高.  相似文献   

7.
大型相关失效工程网络系统可靠度的近似算法   总被引:6,自引:0,他引:6  
何军  李杰 《计算力学学报》2003,20(3):261-266
为避开失效模式(或系统可靠路径)的NP难题(non—polynomial increase hard problem),提出递推分解算法,降低大型工程网络系统可靠度分析中的计算复杂性,同时,将改进的多维正态变量积分公式,引入递推分解算法之中,计算多失效模式(或系统可靠路径)的联合概率,并最终计算工程网络系统的失效概率和可靠度,对于特别复杂的工程网络系统,采用上下界的方法,给出具有很高精度的可靠度近似值。实例分析表明,本文提出的方法,具有较高的精度和计算效率。  相似文献   

8.
简要介绍了两种构造联合分布函数近似方法:基于Pearson相关系数的近似方法P和基于Spearman相关系数的近似方法S。推导了基于直接积分方法的串联系统失效概率计算公式,提出了两构件功能函数间负相关时串联结构系统失效概率上限值的计算公式。以理论联合概率分布函数是二维极值分布为例研究了两种近似方法在串联系统可靠度计算中的精度。结果表明,两种近似方法能够有效地计算串联结构系统可靠度,且精度很高,为不完备概率信息条件下串联结构系统可靠度分析提供了一条有效的途径。当两构件的功能函数正相关时,两种近似方法误差随串联系统失效概率的减小而增加,但近似方法与精确方法系统失效概率的差别最大也不会超过2倍;当两构件的功能函数负相关时,两种近似方法误差随系统失效概率的减小而减小,但两种近似方法的失效概率几乎与精确解一样。此外,两种近似方法误差不是随构件间相关性的增加而单调增加,而是存在一极大值。  相似文献   

9.
含概率与区间混合不确定性的系统可靠性分析方法   总被引:3,自引:0,他引:3  
系统可靠性问题中通常存在大量的不确定参数,传统方法一般是基于概率模型对系统进行可靠性分析,但是实际工程中由于数据缺乏或试验条件的限制往往难以得到参数的精确概率分布.本文将结构体系一部分样本信息充足的不确定变量用随机变量进行描述,而另一部分样本缺乏的用区间表示,并提出了一种新的含概率与区间混合不确定性的系统可靠性分析方法.首先,基于一个高效求解方法获得单失效模式下结构的最小可靠度指标;再针对多失效模式下含概率与区间混合不确定性问题建立了系统可靠性分析模型;考虑各失效模式之间的相关性,通过线性相关度计算方法求得相关系数矩阵;最后提出了串联体系和并联体系可靠度求解方法.3个数值算例表明,该方法可以实现含概率与区间混合的多个非线性失效模式下系统可靠度的计算.通过对比传统的概率可靠性分析方法,本文方法只需要少量的不确定信息便可确保系统更加安全,更适合复杂结构系统可靠性的分析和设计.  相似文献   

10.
多个模式联合失效的设计点及概率的计算   总被引:1,自引:0,他引:1  
提出了多个模式联合失效设计点的概念,并以此设计点作为计算联合失效概率的重要抽样函数的密度中心,算例表明本文方法可大大提高求多阶失效概率重要抽样法的投点效率和收敛速度,从而提高系统失效概率的计算精度。  相似文献   

11.
The series composed by beam mode function is used to approximate the displacement function of constrained damping of laminated cantilever plates, and the transverse deformation of the plate on which a concentrated force is acted is calculated using the principle of virtual work.By solving Lagrange's equation, the frequencies and model loss factors of free vibration of the plate are obtained, then the transient response of constrained damping of laminated cantilever plate is obtained, when the concentrated force is withdrawn suddenly.The theoretical calculations are compared with the experimental data, the results show:both the frequencies and the response time of theoretical calculation and its variational law with the parameters of the damping layer are identical with experimental results.Also, the response time of steel cantilever plate, unconstrained damping cantilever plate and constrained damping cantilever plate are brought into comparison, which shows that the constrained damping structure can effectively suppress the vibration.  相似文献   

12.
Using Reddy’s high-order shear theory for laminated plates and Hamilton’s principle, a nonlinear partial differential equation for the dynamics of a deploying cantilevered piezoelectric laminated composite plate, under the combined action of aerodynamic load and piezoelectric excitation, is introduced. Two-degree of freedom (DOF) nonlinear dynamic models for the time-varying coefficients describing the transverse vibration of the deploying laminate under the combined actions of a first-order aerodynamic force and piezoelectric excitation were obtained by selecting a suitable time-dependent modal function satisfying the displacement boundary conditions and applying second-order discretization using the Galerkin method. Using a numerical method, the time history curves of the deploying laminate were obtained, and its nonlinear dynamic characteristics, including extension speed and different piezoelectric excitations, were studied. The results suggest that the piezoelectric excitation has a clear effect on the change of the nonlinear dynamic characteristics of such piezoelectric laminated composite plates. The nonlinear vibration of the deploying cantilevered laminate can be effectively suppressed by choosing a suitable voltage and polarity.  相似文献   

13.
A nonlinear vibration analysis of a simply supported functionally graded rectangular plate with a through-width surface crack is presented in this paper. The plate is subjected to a transverse excitation force. Material properties are graded in the thickness direction according to exponential distributions. The cracked plate is treated as an assembly of two sub-plates connected by a rotational spring at the cracked section whose stiffness is calculated through stress intensity factor. Based on Reddy’s third-order shear deformation plate theory, the nonlinear governing equations of motion for the FGM plate are derived by using the Hamilton’s principle. The deflection of each sub-plate is assumed to be a combination of the first two mode shape functions with unknown constants to be determined from boundary and compatibility conditions. The Galerkin’s method is then utilized to convert the governing equations to a two-degree-of-freedom nonlinear system including quadratic and cubic nonlinear terms under the external excitation, which is numerically solved to obtain the nonlinear responses of cracked FGM rectangular plates. The influences of material property gradient, crack depth, crack location and plate thickness ratio on the vibration frequencies and transient response of the surface-racked FGM plate are discussed in detail through a parametric study.  相似文献   

14.
Based on Reddy's layerwise theory, the governing equations for dynamic response of viscoelastic laminated plate are derived by using the quadratic interpolation function for displacement in the direction of plate thickness. Vibration frequencies and loss factors are calculated for free vibration of simply supported viscoelastic sandwich plate, showing good agreement with the results in the literature. Harmonious transverse stresses can be obtained. The results show that the transverse shear stresses are the main factor to the delamination of viscoelastic laminated plate in lower-frequency free vibration, and the transverse normal stress is the main one in higher-frequency free vibration. Relationship between the modulus of viscoelastic materials and transverse stress is analyzed. Ratio between the transverse stress's maximum value and the in-plane stress's maximum-value is obtained. The results show that the proposed method, and the adopted equations and programs are reliable.  相似文献   

15.
This paper presents a study of non-linear response of a fluttered, cantilevered beam subjected to a random follower force at the free end. The random follower force is characterized as the sum of a post-critical static force and a stationary process with a zero mean. First, the Ritz-Galerkin method is applied to yield a set of discretized system equations. The system equations are then partially uncoupled by a special modal analysis based on normal modes of the corresponding linear, autonomous system at the onset of fluttering. Next, the stochastic averaging method is utilized to get Ito's differential equation governing the amplitude of the fluttered mode. Finally, the probability density function for the amplitude of the fluttered mode is obtained by solving the FPK equation. Numerical results show that the probability density function for the amplitude of the fluttered mode is determined by the sample behavior of the beam near the trivial equilibrium configuration.  相似文献   

16.
软铁磁薄板磁弹性屈曲的理论模型   总被引:5,自引:0,他引:5  
周又和  郑晓静 《力学学报》1996,28(6):651-660
铁磁弹性薄板的磁弹性屈曲问题一直作为电磁——弹性力学相互作用的一个基本模型进行研究,而作用在其磁介质上的磁力计算则是定量理论预测准确与否的关键.到目前为止,文献上已有的理论模型对悬臂铁磁梁式悬臂板在横向磁场中磁弹性屈曲的理论预测值始终高于实验值,有的甚至相差100%左右.本文基于电磁力计算的微观安培电流模型,严格给出了软铁磁薄板等效横向磁力的宏观计算表达式.在此基础上,建立了电磁——力学相互耦合作用的非线性理论模型.该模型能描述铁磁薄板结构在非均匀横向磁场环境中的磁弹性失稳(或屈曲).其定量分析采用了有限元法和有限差分法相结合.数值结果显示:本模型给出的磁弹性屈曲的临界磁场值与实验值符合良好.与此同时,文中还对文献中认为较成功的Moon-Pao模型的基本假设进行了分析.定量结果发现:Moon-Pao理论模型的基本假设仅在梁式板的长厚比L/h比较大时(约在200左右),是可以接受的,而当L/h较小时,该假设将导致理论值与实验值的较大误差.L/h比值越小,理论值与实验值的误差越大  相似文献   

17.
应用数值模拟方法研究磁场中旋转运动圆板的分叉与混沌问题。首先,基于薄板理论和麦克斯韦电磁场方程组,给出了动能、应变势能、外力虚功以及电磁力的表达式,再利用哈密顿原理,得到磁场中旋转运动圆板横向振动的非轴对称非线性磁弹性振动微分方程组。其次,采用贝塞尔函数作为圆板的振型函数进行伽辽金积分,得到了轴对称情况下横向振动的常微分方程组表达式。最后,针对主共振,取周边夹支边界条件的圆板作为算例,得到了当振型函数取一阶时,将磁感应强度、外激励振幅和激励频率作为控制参数的分叉图及庞加莱映射图等计算结果,并讨论了分叉参数对系统的分叉与混沌的影响。数值计算结果表明,这些控制参数的变化影响系统稳定性,在分叉参数逐渐变化的过程中,系统经历从混沌到多倍周期运动再到混沌的往复过程。  相似文献   

18.
The Green’s functions of a two-phase saturated medium subjected to a concentrated force are known to play an important role in seismology, earthquake engineering, soil dynamics, geophysics, and dynamic foundation theory. This paper presents a physical method for obtaining the dynamic Green’s functions of a two-phase saturated medium for materials considered to be isotropic and for low frequencies. First, the pore-fluid pressure in a two-phase saturated medium is divided into two parts: flow pressure and deformation pressure. Next, based on the compatibility condition of Biot’s equation and the property of the δ-function, the problem of coupled_fast and slow dilational waves is solved using the decomposition condition of the potential dilation field. The Green’s function for a concentrated force is then obtained by solving Biot’s complex modular equations, and their physical characteristics are discussed. The behavior of Green’s functions for the solid and fluid phases of a δ-impulsive force is investigated, from which the Green’s functions for a unit Heaviside force are also obtained by time integration. Finally, the present Green’s functions for a unit Heaviside force are compared with those obtained by a purely mathematical method; the two differ in form, but the numerical results are identical. The physical meaning of the expressions of Green’s functions obtained in this paper is evident. Therefore, the results may benefit future research on the dynamic responses of a two-phase saturated medium.  相似文献   

19.
董智法 《力学学报》1993,25(6):697-708
本文导出了由位移表示的带初始挠度的粘弹性大挠度矩形薄板的运动方程。在简支边界条件下,应用分歧理论和Melnikov方法,研究在周期外力作用下系统的动力行为。给出了出现次谐分枝和马蹄态的条件。  相似文献   

20.
An analysis and numerical results are presented for free transverse vibrations of non-homogeneous visco-elastic elliptic plate whose temperature and thickness spatial variations both are parabolic along a line through plate centre. The variation in density is assumed as parabolic along a line through plate centre, which raise non-homogeneity of the plate materials and make problem interesting as introducing variation in non-homogeneity of the material mass density reduce the problem practical importance in comparison to homogenous plate. For visco-elastic, the basic elastic and viscous elements are combined. We have taken Kelvin model for visco-elasticity that is the combination of the elastic and viscous elements in parallel. Here the elastic element means the spring and the viscous element means the dashpot. The governing differential equation of motion has been solved by Galerkin’s technique. Deflection, time period and logarithmic decrement corresponding to the first two modes of vibrations of a clamped non-homogeneous visco-elastic elliptic plate for various values of taper constant, thermal constants, non-homogeneity constant and aspect ratio are obtained and shown graphically.  相似文献   

设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号