基于时变NARMA模型的非线性时变系统辨识

在非线性自回归滑动平均模型NARMA(NonlinearAutoRegressiveMovingAverage)中引入时间变量,将其扩展为时变NARMA模型,用Taylor展开将模型中的非线性函数展开为关于输入输出的多项式,得到关于参数线性时变的多项式形式的时变NARMA模型,再用基序列拟合模型的时变参数得到关于参数线性时不变的模型,最后用递推最小二乘法估计模型参数。仿真算例证明,与小波网络方法相比,辨识精度高,计算量小。     

基于Legendre正交多项式逼近法的结构可靠性分析

提出了结构可靠性分析的Legendre正交多项式逼近法。主要是基于数值逼近原理,以Legendre正交函数族做基,利用功能函数的高阶矩信息,通过计算功能函数概率密度函数的逼近表达式,然后根据工程结构可靠性的一般表达式来计算结构的失效概率,进行可靠性分析。通过数值检验,证明该方法可以很好地逼近各种经典理论分布曲线(正态分布、指数分布等6种经典分布)。文后给出了结构构件失效概率的实例计算,并和其它几种常用方法进行对比,进一步表明了Legendre正交多项式数值逼近法在结构可靠性分析理论上的正确性和实用性。     

关于Lagrange插值逼近中几个问题研究的新进展

本文将综述 L agrange插值逼近中几个问题研究的新进展 ,并提出几个新问题 .主要的论题是 Lagrange插值多项式序列的收敛与发散 ,用 L agrange插值多项式同时逼近可微分函数及其导数 ,以及修改的 L agrange插值多项式对函数的逼近     

基于优化组合核和Morlet小波核的LSSVM脉动风速预测方法

核函数是支持向量机的重要组成部分,直接影响预测模型的结果。根据Mercer定理,推导出了Morlet小波核函数,使其具有局部化、多层次、多分辨的优点。选择具有代表性的径向基(RBF)核函数和多项式(Poly)核函数构建出局部性和全局性相结合的线性组合核函数,使得预测模型保留RBF核函数所赋予的优越学习能力以及Poly核函数所拥有的强泛化能力;进一步,使用粒子群优化(PSO)算法,对惩罚参数、核参数、权重、尺度因子进行寻优,分别建立了基于Morlet小波核和组合核的PSO-LSSVM模型;使用建立的预测模型,对脉动风速进行了预测。通过比较预测性能评价指标,发现基于Morlet小波核和组合核PSO-LSSVM的预测精度优于常用的单核PSO-LSSVM模型。     

基于矩阵分式多项式时变结构模态参数最小二乘辨识

基于时间相关矩阵分式多项式传递函数模型,给出线性时变结构时频域参数化模型。以时频域参数化模型为基础,将现有广泛用于时不变结构模态参数辨识的最小二乘复指数法拓展到时频域,提出基于矩阵分式多项式模型的时频域线性时变结构模态参数最小二乘辨识方法;针对时频域最小二乘对计算资源庞大需求问题,给出基于缩减正则方程的最小二乘问题求解方法。通过对两质量连续变化三自由度时变结构仿真算例,说明最小二乘中待估参数约束对模态参数辨识影响,阐述所提线性时变结构模态参数辨识方法特点,说明方法的有效性及潜在实用性。     

基于区间B样条小波基函数时变多变量AR模型的时变结构参数识别

本文研究的是时变动力学结构的参数识别问题。仅利用线性时变结构的多个加速度响应测量数据,本文将区间B样条小波函数作为时变基函数,构造了基于区间B样条小波基函数的时变多变量自回归模型,并推导了时变结构瞬时频率的识别方法。文中利用该法对一个三自由度线性时变结构进行了仿真研究,针对结构瞬时频率周期变化,突变和线性变化三种不同的情况分别进行了瞬时频率识别,并在加速度响应数据中加入了不同信噪比的高斯白噪声以检验识别方法的抗噪能力。最后,文中对一个具有质量时变特性的悬臂梁结构进行了实验研究,实验结果充分说明了区间B样条小波基函数时变多变量自回归模型对时变结构参数识别的可行性,有效性和良好的抗噪能力。     

有高阶矩约束的最优投资组合模型及近似线性规划解法

本文讨论有包括偏度和峰度在内的高阶矩约束的最优投资组合模型。证明了最优投资组合决策的存在性并导出解析解的隐式表达式,然后利用线性逼近的方法得到近似解,并给出了具体算例,最后分析了模型中的权重参数对最优目标的影响。     

代数双曲空间中拟Legendre基的应用

鉴于Legendre基等正交基在代数多项式空间中的广泛应用,论文在深入研究代数双曲空间的拟Legendre基性质的基础上,给出了其在反函数逼近和等距曲线逼近上的应用。利用多项式和双曲函数的混合多项式序列来逼近反函数,并通过实例证明给出方法的有效性;对基曲线的法矢曲线进行逼近,构造H-Bézier曲线的等距曲线的最佳逼近,这种方法直接求得逼近曲线的控制顶点,计算简单,截断误差小。     

基于critic组合预测方法的共享单车故障预测

针对无桩式共享单车系统缺少故障预测方法的研究及预测准确率低不稳定等问题,在挖掘预测因素和故障单车之间的信息特征上,提出改进熵权法对预测因素分权,并采用BP神经网络、径向基函数和ELMAN神经网络3种单一预测模型建立基于critic权重的组合预测模型,在Matlab上进行实例求解。结果表明,相比单一预测方法,该组合预测方法提高了预测准确率5%左右,且能够降低预测风险,减少预测的系统误差,具有较好的实用价值。     

函数最佳平方逼近多项式是否唯一

本文对最佳一致逼近多项式与最佳平方逼近多项式相互比较,给出了相同的内积和基函数最佳平方逼近多项式是唯一,并且验证了勒让德展开部分和得到的最佳平方逼近多项式与Hn中的最佳平方逼近多项式是一致的。     

SAFE-FAIL LINE METHOD—AN ALGORITHM TO REDUCE THE CALCULATION OF COMPLEX CONSTRAINTS IN DESIGN OPTIMIZATION

A rather simple but quite effective algorithm is proposed for reducing the number of calculations of complex constraint functions in optimization problems. A type of monotonicity is used to set up a direct relationship between the design variables and the critical constraints. It can frequently be predicted whether or not the constraints are satisfied at a design point just by checking the variation in the design variables. It can be established in advance that a given constraint function varies with respect to a design variable in one of four ways, and rules can be defined, based on the previous history of function values, whether or not a constraint function need be evaluated. Trial calculations of a sample of problems suggest typical savings of 40-50%. The procedure is illustrated by stress constraints; but the same method can be used in other types.     

A TWO-LEVEL DECOMPOSITION METHOD FOR SHAPE OPTIMIZATION OF STRUCTURES

The paper proposes a two-level decomposition method for shape optimization of structures. The optimization problem is divided into two subproblems on the basis of the different effects on structural behaviour of different kinds of design variables. A minimum mass subproblem is solved to determine the sizing variables and a constraint evaluation function based on norm optimization is minimized to determine the shape variables. An efficient coupling technique is used between the subproblems to ensure very rapid and steady convergence. Numerical results are presented to demonstrate the effectiveness of the method. © 1997 by John Wiley & Sons, Ltd.     

R-CURVE TESTING AND ITS RELEVANCE TO STRUCTURAL ASSESSMENT

Crack growth resistance can be substantially affected by the constraint conditions of a structural member which in turn are mainly a function of geometrical variables and the degree of plasticity. Standardized test methods are restricted to high constraint conditions as represented by deeply cracked bend-type specimens and may hence lead to conservative structural assessments. It is demonstrated that adjusted testing can be used to reduce the degree of conservatism. Due to rapidly increasing computer capabilities, a combination of conventional R-curve testing with micromechanical models emerges as an accurate tool which may permit routine evaluations of practical situations in the near future.     

基于信息融合方法的切削颤振状态识别技术

讨论了切削颤振状态识别的信息融合方法．分别给出了基于基本概率分配函数和基于证据区间值的颤振状态识别方法。试验中，在同一个测量区内使用了功率传感器和加速度传感器，利用Dempster—Shafer证据论方法对两种传感器信息进行了分析融合。试验与理论分析表明：经过信息融合得到的基本概率分配函数可以作为一种颤振状态识别参数。如果同时考虑证据区间P1(A)-Bel(A)值进行识别会减小识别的不确定性．提高颤振状态识别的精度。     

概率及非概率不确定性条件下结构鲁棒设计方法

提出了在概率不确定性和非概率不确定性同时存在时的约束函数鲁棒性和目标函数鲁棒性的实现策略及结构鲁棒设计方法。将传统优化设计问题的约束条件改造成为能同时反映两类不确定性量波动变化影响的约束条件,以实现约束函数的鲁棒性;在传统优化设计问题目标函数中增加若干个关于目标函数灵敏度的新目标函数,构成一个多目标函数设计问题,以实现目标函数的鲁棒性。所提方法应用于一个10杆桁架结构设计,采用宽容排序法求解。计算结果表明,在相同的结构总质量限制条件下,目标函数鲁棒性程度随着变量不确定性程度的增加而降低;在相同的变量不确定性程度条件下,增加结构总质量能提高目标函数鲁棒性的程度。     

Solution to the topology optimization problem using a time-evolution equation

This article describes a numerical solution to the topology optimization problem using a time-evolution equation. The design variables of the topology optimization problem are defined as a mathematical scalar function in a given design domain. The scalar function is projected to the normalized density function. The adjoint variable method is used to determine the gradient defined as the ratio of the variation of the objective function or constraint function to the variation of the design variable. The variation of design variables is obtained using the solution of the time-evolution equation in which the source term and Neumann boundary condition are given as a negative gradient. The distribution of design variables yielding an optimal solution is obtained by time integration of the solution of the time-evolution equation. By solving the topology optimization problem using the proposed method, it is shown that the objective function decreases when the constraints are satisfied. Furthermore, we apply the proposed method to the thermal resistance minimization problem under the total volume constraint and the mean compliance minimization problem under the total volume constraint.     

Shape optimization using time evolution equations

In this study, we deal with a numerical solution based on time evolution equations to solve the optimization problem for finding the shape that minimizes the objective function under given constraints. The design variables of the shape optimization problem are defined on a given original domain on which the boundary value problems of partial differential equations are defined. The variations of the domain are obtained by the time integration of the solution to derive the time evolution equations defined in the original domain. The shape gradient with respect to the domain variations are given as the Neumann boundary condition defined on the original domain boundary. When the constraints are satisfied, the decreasing property of the objective function is guaranteed by the proposed method. Furthermore, the proposed method is used to minimize the heat resistance under a total volume constraint and to solve the minimization problem of mean compliance under a total volume constraint.     

基于SA算法的行星齿轮减速器离散变量优化设计

考虑制造工艺要求,将所有设计变量均视为离散变量,包括一般离散变量和伪离散变量,并就这两种情况下状态产生函数的设计原理进行深入研究,解决了将模拟退火算法用于离散变量函数优化的关键技术问题,介绍了一种基于模拟退火算法的离散变量函数优化的新方法。行星齿轮传动中各齿轮的齿数受传动比条件、同轴条件和装配条件的限制而不能任意取值,齿轮的模数也要受国家标准的制约只能取一些离散值,用以数学规划理论为基础的经典约束优化方法求解效果很差,用基于模拟退火算法的离散变量优化设计方法则可以方便快捷地获得满足各方面要求的最优设计方案。     

基于多峰分布的大尺度变形翼机构时变可靠性分析

针对机构可靠性的工程问题,提出了一种基于多峰分布的时变可靠性分析方法（iTRPD）,并应用于大尺度变形翼机构的可靠性分析。首先,将变形翼结构模型离散为几个瞬时功能函数,并将其转换为独立正态变量。然后,计算出不同时刻的瞬间可靠度与各向量间的自相关系数矩阵,得到对应的概率密度函数。最后,根据协方差特性与各向量间的相关性,利用1次高维高斯积分将独立标准正态空间的时变可靠度简化为大尺度变形翼机构整体的时变可靠度。结果表明：iTRPD在分析大尺度变形翼时变可靠性时,与蒙特卡洛仿真法（MCS）的相对误差仅为-2.842%,比常规方法TRPD好;对功能函数调用次数为415,远小于MCS的1×109次;对高维高斯积分的调用,常规时变可靠性方法为35次,iTRPD仅为1次。可见,iTRPD对涉及多模态分布的时变可靠性分析具有较高的计算精度和计算效率。     

Automatically deriving behavior constraints for performance variables in mechanical design

Generation of the behavioral constraint equations needed for analysis in the mechanical design synthesis-analysis iteration loop is currently based on approaches that model the behavior of an entire system or subsystem of the design. This paper presents a general approach to constraint derivation that is localized around design performance variables. The approach consists of an amenable design configuration representation, a standard format for representing behavior knowledge, and an algorithm that builds a constraint network that can be used to solve for the performance variables by applying behavioral knowledge to the configuration. The algorithm works outward from the initial performance variables, creating new performance variables as required, and processing only those behavior laws and components of the design needed to obtain a solution. It can take design variables (known conditions or variable parameters) into account, takes advantage of the part-subpart hierarchy in the configuration, can exploit previous solutions, and can detect indeterminacy that is sometimes caused by simplifying assumptions.