一种新的减基法中的采样方法及误差估计

针对减基法中求解精度受减基空间完备性影响的问题,提出了一种基于向量空间逼近原理的采样方法及相应的误差估计。该方法使每一步采样得到的特征向量与之前已得到的特征向量所张成的向量空间角度最大,从而保证了每步采样所对应的特征向量都具有最弱相关性,进而使最终得到的特征向量基空间更具完备性。并且由该方法产生了一种先验特征值误差界,...     

Karhunen–Loève Galerkin method with decimated sampling technique for the simulation of complex fluids defined in the phase space

In complex fluids, solute molecules with structural length scales much larger than atomic are dispersed in solvents of simple fluids such as water. The rheological properties of complex fluids are determined by dynamics of solute molecules which can be modeled by the Fokker–Planck equation defined in a six-dimensional phase space. In the present investigation, we devise a method of efficient simulation of complex fluid flows employing the Karhunen–Loève Galerkin (KLG) method. Adopting the decimated sampling of solvent flow fields, a reduced-order model for the Fokker–Planck equation is obtained, which can be employed for the the simulation of complex fluids with a decent computer time. As a specific example, we consider a flow of dilute polymeric liquids over a cylinder, whose constitutive equation is the FENE (finitely extensible nonlinear elastic) model. It is found that the KLG method with the decimated sampling technique yields accurate results at a computational cost less than a hundredth of that for the numerical simulation using the Fokker–Planck equation. The KLG method supplemented by the decimated sampling technique is an efficient method of coarse-graining for equations of complex fluids defined in the phase space.     

非正态变量可靠性分析的鞍点线抽样方法

结合鞍点概率分布估计和传统线抽样方法的优点.提出了非正态变量可靠性分析的鞍点线抽样方法.传统的线抽样方法对非正态变量可靠性问题进行分析时需将非正态变量等价转换为标准正态变量,这种非线性转换将增加响应功能函数的非线性程度,进而加大了转换后响应函数失效概率估计的难度.所提鞍点线抽样方法则无需将非正态变量转化为标准正态变量,它利用鞍点概率分布估计方法可以直接估计非正态变量空间中线性响应函数概率分布的特点,并利用线抽样方法可以将非线性功能函数的失效概率转化为一系列线性功能函数失效概率平均值进行估计的优点,实现了非正态变量空间非线性功能函数失效概率的高精度估计.鞍点线抽样方法使用前需将变量进行标准化变换,这种变换是线性的,通过对变量的标准化变换可以消除变量的量纲,从而使得标准化变量空间概率分布更具规律性.理论推导可以证明:鞍点线抽样方法在基本变量服从正态分布时将退化为传统的线抽样方法.算例验证结果表明:针对非线性功能函数的可靠性问题,鞍点线抽样方法比传统的直接鞍点估计具有更高的精度,比直接Monte Carlo模拟有更高的效率.     

Free-form deformation,mesh morphing and reduced-order methods: enablers for efficient aerodynamic shape optimisation

ABSTRACT

In this work, we provide an integrated pipeline for the model-order reduction of turbulent flows around parametrised geometries in aerodynamics. In particular, free-form deformation is applied for geometry parametrisation, whereas two different reduced-order models based on proper orthogonal decomposition (POD) are employed in order to speed-up the full-order simulations: the first method exploits POD with interpolation, while the second one is based on domain decomposition. For the sampling of the parameter space, we adopt a Greedy strategy coupled with Constrained Centroidal Voronoi Tessellations, in order to guarantee a good compromise between space exploration and exploitation. The proposed framework is tested on an industrially relevant application, i.e. the front-bumper morphing of the DrivAer car model, using the finite-volume method for the full-order resolution of the Reynolds-Averaged Navier–Stokes equations.     

子集模拟的混合采样算法

子集模拟是当前可靠度计算领域常用的估计算法,相比于直接蒙特卡罗积分法,极大减少了函数调用的次数。子集模拟法一般使用单一采样器,然而不同采样器适用范围不同。如使用单一椭圆切片采样器,其遍历性较好但函数调用次数较多;而使用单一自适应条件采样时,其采样效率较高但样本容易陷入局部极值。单一采样器由于本身特性面对不同问题时失效概率的积分结果可能出现偏差,模拟效果不稳定。本文首次提出了一种混合采样子集模拟法,在子集模拟的前几层使用椭圆切片采样,此时失效区域收缩程度有限,函数调用次数在可接受的范围内,样本经过采样扩充后能充分探索参数空间,更有效地检测出所有失效区域。当失效区域收缩至一定限度后,使用自适应条件采样,此时种子样本继承了前几层样本较低的相关性,并在此基础上通过自适应条件采样更高效地增殖样本。本文通过四个模拟算例多种维度下的数值积分验证了该算法具有椭圆切片采样器较好的遍历性,同时采样效率位于椭圆切片采样与自适应条件采样之间,对于不同问题拥有良好的通用性。     

正态变量相关情况下可靠性灵敏度分析的新方法

基于独立正态变量情况下可靠性灵敏度分析的线抽样法,提出了一种求解正态相关变量情况下可靠性灵敏度的新方法。在所提方法中,首先将正态相关变量等效变换为正态独立变量,然后利用线抽样方法独立完成等效独立变量情况下失效概率对独立变量的所有分布参数的灵敏度分析,最后依据等效变换前后变量分布参数之间的解析关系和复合函数求导公式,求得...     

COORDINATED PATH PLANNING BY INTEGRATING IMPROVED RRT$^{\small*}$ AND QUARTIC SPLINE$^{\bf 1)}$

针对空间机器人抓捕空间非合作目标的在轨服务任务,同时考虑机器人运动学约束和动力学约束,提出一种分层式的自由漂浮双臂空间机器人协调路径规划方法. 首先,在路径规划层面上基于 RRT* 算法分别规划双臂末端执行器在笛卡尔空间下的初始可行路径,为双臂设置独立的采样空间,保证路径规划过程中双臂系统不发生自身碰撞. 然后,在轨迹规划层面上利用四次样条曲线平滑 RRT* 算法生成的初始路径,设计满足样条曲线的一阶、二阶及三阶微分连续约束,同时考虑机械臂末端执行器的初末速度约束条件、初始加速度约束条件,得到适合于空间机器人执行的动力学可行的平滑 轨迹.最后,计算所规划路径的最大速度、最大加速度与机械臂末端执行器物理极限值的比值,取最小上限,即为最少路径规划时间. 所提路径规划方法能够设计出满足特定路径点约束的协调路径,且所设计的路径考虑了机械臂的物理限制条件,通过对自由漂浮双臂空间机器人进行仿真试验,验证了所提路径规划算法的有效性.     

一种改进 RRT* 结合四次样条的协调路径规划方法

针对空间机器人抓捕空间非合作目标的在轨服务任务,同时考虑机器人运动学约束和动力学约束,提出一种分层式的自由漂浮双臂空间机器人协调路径规划方法. 首先,在路径规划层面上基于 RRT* 算法分别规划双臂末端执行器在笛卡尔空间下的初始可行路径,为双臂设置独立的采样空间,保证路径规划过程中双臂系统不发生自身碰撞. 然后,在轨迹规划层面上利用四次样条曲线平滑 RRT* 算法生成的初始路径,设计满足样条曲线的一阶、二阶及三阶微分连续约束,同时考虑机械臂末端执行器的初末速度约束条件、初始加速度约束条件,得到适合于空间机器人执行的动力学可行的平滑 轨迹.最后,计算所规划路径的最大速度、最大加速度与机械臂末端执行器物理极限值的比值,取最小上限,即为最少路径规划时间. 所提路径规划方法能够设计出满足特定路径点约束的协调路径,且所设计的路径考虑了机械臂的物理限制条件,通过对自由漂浮双臂空间机器人进行仿真试验,验证了所提路径规划算法的有效性.      

Dynamics of Digital Force Control Applied in Rehabilitation Robotics

Controlled structures are often required to keep desired contact forces between some of its elements. A classical example is the controlled interaction of a robot with its environment when the control of the contact force between the robotic actuator and the workpiece is prescribed. Experiments often call the attention to the destabilising digital effects, like sampling, in these systems. In this paper the stability of a newly developed force based teaching-in method is analysed. The method is applied in rehabilitation robotics. The stability limits are presented in the parameter space of the sampling time, control gains and mechanical parameters of the robot. The least force error and the fastest settling force signal are calculated. The influence of the elasticity of the force sensor is analysed as well as the possible bifurcations. Real parameter case study confirms the analytical predictions.     

基于减基概念的凸集结构可靠性分析方法

针对椭球凸集参数域结构的可靠性分析问题,提出了一种基于减基概念的快速求解方法。首先,将椭球参数域进行坐标正交变换,获得标准的椭球域及其相应的矩形域,在矩形域采样且通过坐标逆向变换获得原椭球参数域的样本参数点集,并以此构建结构的减基空间及其相应的减基算法;随后,在标准椭球域产生均匀的等概率抽样点,并通过坐标逆向变换和相应的减基算法进行蒙特卡洛减基模拟来分析结构的可靠度及其可靠域。由于是在低维的逼近空间中进行椭球参数域结构的位移向量解分析,故而较之有限元法能够获得较高的计算效率。算例测试验证了本文方法的有效性。     

壁湍流相干结构的辨识

本文评价了辨识湍流边界层近壁区域相干结构的条件采样法、自相关函数法、数字滤波法和子波分析法。结果表明：条件采样方法对门限值和平均窗宽等检测参数具有较强的依赖性，可以通过调整这些参数来选择最佳的参数值使得相干结构的误判概率达到最小。自相关函数法不需要任何参数就能够得出确定的结论，因此是一种比较客观的方法。但是，这种方法只能从统计上确定平均猝发周期，并不能实时地辨识相干结构。数字滤波法能够将湍流脉动信号分解为大尺度涡和近似各向同性的小尺度涡两部分，但是如何去掉大尺度信号中的非相干部分是仍然需要解决的问题。子波分析是分解湍流信号辨识湍流相干结构的有效手段，它能够将湍流信号在时域空间和频域空间进行双局部化分解，用能量最大准则确定相干结构对应的时间尺度参数，用子波逆变换提取相干结构对应的速度信号波形     

一种随机裂纹结构的裂纹扩展路径分析方法

裂纹结构中存在大量不确定性因素,如裂纹长度、材料性质、外部载荷等,裂纹扩展路径的不确定性分析对研究随机裂纹结构损伤和断裂的力学特性并预测其性能及可靠性具有重要意义。本文提出了一种适应于混合载荷模式下随机裂纹结构的裂纹扩展路径分析方法。该方法考虑了裂纹长度、材料性质和外部载荷等的随机性,并通过蒙特卡洛方法对随机参数空间进行采样。采用比例边界有限元方法计算结构应力强度因子,进而模拟单次裂纹扩展路径。在此基础上,通过概率分析方法获得随机裂纹结构中裂纹扩展路径的统计特性。最后给出了两个数值算例验证了本文方法的有效性。     

A high-order accurate wavelet method for solving Schrdinger equations with general nonlinearity

A sampling approximation for a function defined on a bounded interval is proposed by combining the Coiflet-type wavelet expansion and the boundary extension technique. Based on such a wavelet approximation scheme, a Galerkin procedure is developed for the spatial discretization of the generalized nonlinear Schr¨odinger(NLS) equations, and a system of ordinary differential equations for the time dependent unknowns is obtained. Then, the classical fourth-order explicit Runge-Kutta method is used to solve this semi-discretization system. To justify the present method, several widely considered problems are solved as the test examples, and the results demonstrate that the proposed wavelet algorithm has much better accuracy and a faster convergence rate in space than many existing numerical methods.     

Signal reconstruction by random sampling in chirp space

Reconstruction by irregular sampling and sampling at rates below the Nyquist rate are important issues for the experimental characterization of dynamical systems. For the space of functions that can be approximated by chirps, we prove a reconstruction theorem by random sampling at arbitrary rates.     

板片空间结构初始缺陷的动力稳定性影响分析

对板片空间结构体系的研究大都局限于静力载荷工况,结构在动力载荷作用下的稳定性及失效特征研究仍相对缺乏．板片空间钢结构先天存在各样缺陷,而缺陷对结构稳定性的影响不可忽略．本文引入拉丁超立方法,结合Budiansky-Roth 准则,提出适用于求解含初始缺陷结构动力临界荷载的方法,对板片空间结构初始缺陷的动力稳定性开展研究．研究表明,相比于其它抽样方法,拉丁超立方法能显著减少样本数,有效提高计算效率．针对典型的板片空间结构,施加不同方向的阶跃荷载,分析了节点位置偏差缺陷、杆件截面尺寸缺陷以及板片厚度缺陷对板片空间结构临界承载力的影响．结果显示,节点位置偏差缺陷对该结构临界承载力影响最为显著,其它两类缺陷影响甚微．     

Generalized Wall Functions for Rough Walls Based on Data Assimilation

We present a systematic identification method for the development of equivalent boundary conditions for turbulent flow simulation over rough walls in engineering applications where the typical roughness involved is known for a given application. After parameterizing the roughness and sampling of the parameter space, numerical simulations or experiences are used to generate a database which is used to identify the coefficients of a polynomial reduced-order model. These laws are parameterized by an a priori length scale for which we propose a new a posteriori criterion.     

Measurements and treatment of LDA signals, comparison with hot-wire signals

 An experimental investigation was carried out in the wind tunnel F2 of the ONERA Fauga centre for the measurement of the characteristics of a turbulent wake behind a wing. As these measurements require the calculation of time and space correlations, two different types of acquisition means are used : a pair of crossed hot wires and a one-dimensional laser-Doppler anemometer. The non-intrusive characteristics of laser anemometry allow the measurement of fluid velocity upstream of a hot wire probe without disrupting the flow. It is well known that LDA generates individual realisations of randomly sampled velocity data because the random arrival of seeding particles in the measurement volume is nonperiodic. A detailed study of this random sampling quantifies the deviations from the theoretical lows, shows the limiting factors of this sampling, and gives a characterisation of the particles arrival law. The simultaneous acquisition of the two velocity signals at very close points allows a good comparison between the signals. A statistical analysis of the two signals enables us to precisely measure the error value of the velocity estimation made by the anemometer. The spectrum analysis of the laser signal coupled with one of the hot wire signals requires resampling the signal at constant steps. Two different methods of interpolation are analysed: the sample and hold interpolation method and the linear interpolation method. The influence of these interpolation methods on the spectrum of LDA signals is studied. Different estimators are then calculated to evaluate the convection velocities and the coherence length of the turbulence. Received: 9 December 1997/Accepted: 19 March 1999     

The effects of the irregular sample and missing data in time series analysis

Human self-report time series data are typically marked by irregularities in sampling rates; furthermore, these irregularities are typically natural outcomes of the data generation process. Relatively little has been published to assist the analysis of irregularly sampled data. We report the results of a series of computational experiments on synthetic data sets designed to assess the utility of techniques for handling irregular time series data. The behavior of a conservative quasiperiodic, a dissipative chaotic, and a self-organized critical dynamics were sampled regularly in time and the regular sampling was disrupted by data point removal or by stochastic shifts in time. Missing data segments were then patched by means of segment concatenation, by segment filling with average data values, or by local interpolation in phase space. We compared results of nonlinear analytical tools such as autocorrelations and correlation dimensions using complete and patched sets, as well as power spectra with Lomb periodograms of the decimated sets. Local interpolation in phase space was particularly successful at preserving key features of the original data, but required potentially impractical quantities of intact data as a primer. While the other patching methods are not limited by the need for intact data, they distort results relative to the intact series. We conclude that irregularly sampled data sets with as much as 15 percent missing data can potentially be re-sampled or repaired for analysis with techniques that assume regular sampling without introducing substantial errors.     

Modeling of metallic foam morphology using the Representative Volume Element approach: Development and experimental validation

This paper focuses on the development of an algorithm capable of generating morphologically-representative foam structures using the Representative Volume Element (RVE) approach. Stereology, a sampling method based on direct observations of the foam cross-sections, is used to characterize the pore size and shape distributions. Using the morphology generation algorithm, the smallest RVEs corresponding to the numerically-convergent foam morphologies are calculated for different foam porosities. To validate the foam generation algorithm, the pore size and shape distributions of the numerically-generated foams are compared to those of the titanium foams manufactured by the space holder method.     

Using the Sequential Self-calibration Method and Genetic Algorithm Method to Optimally Design Tracer Test to Estimate Conductivity Distribution

A gradient-based inverse method – the sequential self-calibrated method – is combined with a genetic algorithm method to search the optimal design scheme for a tracer test. The sequential self-calibrated method is developed for estimating conductivity distribution in a study domain conditioning on tracer test data. To improve the calculation efficiency, a fast streamline-based approach is used to compute the derivative of concentration with respect to the changes of hydraulic conductivity. Performance of the sequential self-calibrated method has been studied using a synthetic aquifer having a sandwich-like geologic structure where hypothetical tracer tests are conducted. The study results indicate that the locations and number of sampling wells will significantly affect accuracy in the estimates. To maximize estimating accuracy in the sequential self-calibrated method for a fixed number of sampling wells, a genetic algorithm method is applied to search the optimal locations for sampling wells. The results indicate that the optimal sampling well locations depend on the apparent geologic structure and the difference in conductivity values for the various regions. For the sandwich-like structure, when the difference between conductivity values in the two separate regions is large enough, the optimal locations for the sampling wells will be fixed, regardless of conductivity values. The study results also show that based on the optimal sampling-well scheme, estimating accuracy will increase as the number of sampling wells increases, even though the rate of increasing accuracy slows as the number of wells increases.