NONPARALLEL BOUNDARY LAYER STABILITY IN HIGH SPEED FLOWS

The parabolized stability equations （PSEs） for high speed flows, especially supersonic and hypersonic flows, are derived and used to analyze the nonparallel boundary layer stability. The proposed numerical techniques for solving PSE include the following contents： introducing the efficiently normal transformation of the boundary layer, improving the computational accuracy by using a high-order differential scheme near the wall, employing the predictor-corrector and iterative approach to satisfy the important normalization condition, and implementing the stable spatial marching. Since the second mode dominates the growth of the disturbance in high Mach number flows, it is used in the computation. The evolution and characteristics of the boundary layer stability in the high speed flow are demonstrated in the examples. The effects of the nonparallelizm, the compressibility and the cooling wall on the stability are analyzed. And computational results are in good agreement with the relevant data.

On Nonlinear Evolution of C-type Instability in Nonparallel Boundary Layers

The process of evolution, especially that of nonlinear evolution, of C-type instability of laminar-turbulent flow transition in nonparallel boundary layers are studied by means of a newly developed method called parabolic stability equations （PSE）. Initial conditions, which are very important for the nonlinear problem, are investigated by computing initial solution of the harmonic waves, modifying the mean-flow-distortion, and giving initial value of TS wave and its subharmonic waves at initial station by solving linear PSE. A numerical method with high-order accuracy are developed in the text, the key normalization conditions in the PSE are satisfied, and nonlinear PSE are solved efficiently and implemented stably by the spatial marching. It has been shown that the computed process of nonlinear evolution of C-type instability in Blasius boundary layer is in good agreement with the experimental results.

New Method of Reconstruction from Nonparallel Stereo and Application to Surgical Navigator

A new method to reconstruct 3D scene points from nonparallel stereo is proposed. From a pair of conjugate images in an arbitrarily configured stereo system that has been calibrated, coordinates of 3D scene points can be computed directly using the method, bypassing the process of rectifying images or iterative solution involved in existing methods. Experiment results from both simulated data and real images validate the method. Practical application to surgical navigator shows that the method has advantages to improve efficiency and accuracy of 3D reconstruction from nonparallel stereo system in comparison with the conventional method that employs algorithm for standard parallel axes stereo geometry.

NONLINEAR EVOLUTION ANALYSIS OF T-S DISTURBANCE WAVE AT FINITE AMPLITUDE IN NONPARALLEL BOUNDARY LAYERS

The nonlinear evolution problem in nonparallel boundary layer stability was studied. The relative parabolized stability equations of nonlinear nonparallel boundary layer were derived. The developed numerical method, which is very effective, was used to study the nonlinear evolution. of T-S disturbance wave at finite amplitudes. Solving nonlinear equations of different modes by using predictor-corrector and iterative approach, which is uncoupled between modes, improving computational accuracy by using high order compact differential scheme, satisfying normalization condition I determining tables of nonlinear terms at different modes, and implementing stably the spatial marching, were included in this method. With different initial amplitudes, the nonlinear evolution of T-S wave was studied. The nonlinear nonparallel results of examples compare with data of direct numerical simulations (DNS) using full Navier-Stokes equations.

Nonlinear Stability of Supersonic Nonparallel Boundary Layer Flows

This article studies the nonlinear evolution of disturbance waves in supersonic nonparallel boundary layer flows by using nonlinear parabolic stability equations （NPSE）. An accurate numerical method is developed to solve the equations and march the NPSE in a stable manner. Through computation,are obtained the curves of amplitude and disturbance shape function of harmonic waves. Especially are demonstrated the physical characteristics of nonlinear stability of various harmonic waves,including instantaneous stream wise vortices,spanwise vortices and Λ structure etc,and are used to study and analyze the mechanism of the transition process. The calculated results have evidenced the effectiveness of the proposed NPSE method to research the nonlinear stability of the supersonic boundary layers.

ON NONLINEAR STABILITY IN NONPARALLEL BOUNDARY LAYER FLOW

The nonlinear stability problem in nonparallel boundary layer flow fortwo-dimensional disturbances was studied by using a newly presented method called ParabolicStability Equations (PSE). A series of new modes generated by the nonlinear interaction ofdisturbance waves were tabu-lately analyzed, and the Mean Flow Distortion (MFD) was numericallygiven. The computational techniques developed, including the higher-order spectral method and themore effective algebraic mapping, increased greatly the numerical accuracy and the rate ofconvergence. With the predictor-corrector approach in the marching procedure, the normalizationcondition was satisfied, and the stability of numerical calculation could be ensured. With differentinitial amplitudes, the nonlinear stability of disturbance wave was studied. The results ofexamples show good agreement with the data given by the DNS using the full Navier-Stokes equations.

EFFECTS OF NONPARALLELISM ON THE BOUNDARY LAYER STABILITY

The nonparallel effects on the stability of the boundary layer flow was investigated using the Parabolie Stability Equations (PSE). In order to improve the accuracy of the calculation which is very important for the investigation of stability, higher order expansions in orthogonal functions in normal direction and the effective algebraic mapping to deal with the problem of infinite region were used and the way to collocate the boundary point based on its characteristics was adopted. With the effective control of step size in the marching procedure, the special condition was satisfied, and the stability of calculation was assured. From the curves of the neutral stability, the growth rate, the amplitude variation and disturbed velocity profile, the effects of the nonparallelism were given accurately and analyzed detailedly. It is found that the nonparallelism of the flow amplifies the amplitude and growth rate of disturbances, especially for three-dimensional disturbances, even can change the sign of flow stability from stability to instability for some cases. Computed results are in good agreement with the classical experimental results.

Nonparallel Support Vector Machine Based on One Optimization Problem for Pattern Recognition

In this paper,we present a novel nonparallel support vector machine based on one optimization problem(NSVMOOP)for binary classification.Our NSVMOOP is formulated aiming to separate classes from the largest possible angle between the normal vectors and the decision hyperplanes in the feature space,at the same time implementing the structural risk minimization principle.Different from other nonparallel classifiers,such as the representative twin support vector machine,it constructs two nonparallel hyperplanes simultaneously by solving a single quadratic programming problem,on which a modified sequential minimization optimization algorithm is explored.The NSVMOOP is analyzed theoretically and implemented experimentally.Experimental results on both artificial and publicly available benchmark datasets show its feasibility and effectiveness.

Narrow-spectral-span spectral beam combining with a nonparallel double-grating structure

We propose a nonparallel double-grating structure in a spectral-beam combining technique, where two gratings are placed nonparallel satisfying the Littrow mount in the focal region of the convergent lens. The most attractive advantage of this approach is that it will compress the spectral span into half of its original spectrum, which means the number of combined elements can be doubled in the gain range of diode lasers. Experimental results demonstrate that the CW output power of the combined beam is 30.9 W with a spectral span of 7.0 nm, compared with its original spectrum span of 13.6 nm, and the spectral beam combining efficiency is 70.5%. In consideration that a single grating could have a high efficiency of 〉97% in a bandwidth of over ten nanometers, the efficiency loss of the grating pair should be less than 6%, which is acceptable for most applications, so this method of using double gratings should be highly interesting for practical applications when a nearly doubled number of diode lasers could be combined into one single laser compared with the previous single-grating methods.

Measurement errors of ultrasonic velocity and attenuation due to nonparallelness of specimen faces

The measurement errors due to nonparallelness of specimen faces in the ultrasonic velocity and attenuation coefficient measured by the pulse reflection method in VHF range have been investigated theoretically, and the theoretical expressions for estimating these measurement errors are given. It has also been shown that, the attenuation coefficient error is depending on both the nonparallelness angle of specimen faces and the ultrasonic frequency, but the velocity error is only depending on the former. Furthermore, for pure silica glass specimen it is estimated that, in VHF range in order to insure that the attenuation coefficient and velocity errors due to nonparallelness of specimen faces are less than 10% and 0.01%, respectively, the nonparallelness angle of specimen faces must be less than 10 s and 40 s correspondingly.

基于最小二乘孪生支持向量机的不确定数据学习算法

孪生支持向量机通过计算2个二次规划问题,得到2个不平行的超平面,用于解决二分类问题。然而在实际的应用中,数据通常包含不确定信息,这将会对构建模型带来困难。对此,提出了一种用于求解带有不确定数据的最小二乘孪生支持向量机模型。首先,对于每个实例,该方法都分配一个噪声向量来构建噪声信息。其次,将噪声向量结合到最小二乘孪生支持向量机,并在训练阶段得到优化。最后,采用一个2步循环迭代的启发式框架求解得到分类器和更新噪声向量。实验表明,跟其他对比方法比较,本方法采用噪声向量对不确定信息进行建模,并将孪生支持向量机的二次规划问题转化为线性方程,具有更好的分类精度和更高的训练效率。

改进的湿式离合器带排转矩模型

基于表面张力和表面波度以及摩擦副不平行度的影响,建立了改进的湿式离合器带排转矩模型。推导了考虑表面张力和表面湿润的摩擦副径向压力分布方程,同时推导了摩擦副不平行度和表面波度影响下的油膜厚度。通过计算研究发现:表面张力、摩擦副不平行度和表面波度均使离合器带排转矩增大,而且不平行程度越大,带排转矩增加越快,而表面波度对带排转矩的影响较小。将改进的模型与原模型和试验数据进行了对比,验证了改进模型的有效性。

结构化支持向量机研究综述

结构化支持向量机(Structural Support Vector Machine,SSVM)是支持向量机(Support Vector Machine,SVM)的变体算法,被广泛应用于多个领域。阐述了SSVM的发展过程,详细分析了SSVM各种具体实现算法的思想及表现上的优劣;并通过实验的对比讨论,发现了SSVM的各种具体实现算法在分类性能和分类效率上优于其他SVM算法,而在稳定性上则逊于后者;基于此,给出了SSVM的后续研究方向。

有限振幅T-S波在非平行边界层中的非线性演化研究

研究对非平行边界层稳定性有重要影响的非线性演化问题 ,导出与其相应的抛物化稳定性方程组 ,发展了求解有限振幅T_S波的非线性演化的高效数值方法· 这一数值方法包括预估_校正迭代求解各模态非线性方程并避免模态间的耦合 ,采用高阶紧致差分格式 ,满足正规化条件 ,确定不同模态非线性项表和数值稳定地作空间推进· 通过给出T_S波不同的初始幅值 ,研究其非线性演化· 算例与全Navier_Stokes方程的直接数值模拟 (DNS)的结果作了比较·

改进的投影孪生支持向量机

针对投影孪生支持向量机(Projection Twin Support VectorMachine,PTSVM)在训练和求解过程中存在的问题,提出了一类改进的投影孪生支持向量机(Improved PTSVM),简称为IPTSVM.该文首先构造了改进的线性投影孪生支持向量机,然后利用核技巧轻松将其推广到了非线性形式.本文的主要贡献有:(1)提出了投影孪生支持向量机的新模型,克服了原始PTSVM在训练之前需要求解两个逆矩阵的问题;(2)继承了传统SVM(Support VectorMachine)的精髓,利用核技巧直接将线性IPTSVM推广到非线性形式;(3)引入了一个新的参数,可以调节模型的性能,提高了IPTSVM的分类精度.实验结果表明,与PTSVM算法相比较,IPTSVM不仅提高了分类精度,而且克服了PTSVM的一些不足.

可压缩非平行流边界层稳定性研究

采用线性稳定性理论和多重尺度方法 ,研究三维可压缩的非平行流边界层稳定性问题。分析了可解条件的特征 ,导出精确计算所需的伴随问题方程渐近外边界条件的矩阵表达式 ,给出有控制的重正化方法 ,以有效地克服刚性方程在积分求解中的困难。探讨与非平行性作用相关的方程和影响因素 ,特别是新的特征函数畸变对扰动增加率的作用。通过算例 ,清楚地显示了流动的非平行性对边界层稳定性的影响。

非平移收益曲线的风险免疫策略

在债券收益曲线呈刚体运动的假设条件下,引入Fisher&Weil久期的概念.从收益曲线的运动分析出发,提出了非平移收益曲线的风险免疫模型,基于该模型研究了风险最小化债券组合的对冲技术和方法.通过数值模拟实验,采用风险值VaR次序统计量估计技术对不同免疫策略下债券组合的风险敞口进行了分析.结果表明,所提出的风险免疫策略能有效地防范和控制无违约债券的利率风险.

三维扰动波的非平行边界层稳定性研究

导出了三维扰动波的原始变量形式的抛物化稳定性方程（PSE）,研究了三维空间模态TS波的非平行边界层稳定性问题.采用了法向四阶紧致格式,以提高计算精度.通过给出不会导致奇性的坐标变换、修改外边界条件以及克服平行流初始值的瞬态影响和推进步长的限制,保证了计算的数值稳定.用补全元素带状矩阵法求解块三对角矩阵,大大提高了速度.计算结果清楚地显示了三维扰动波的演化过程和非平行性对边界层稳定性的影响,特别是,观察到非平行性对三维扰动波的影响,有时会使其稳定性出现逆转的现象.还研究了逆压梯度的作用.算例的结果与其他结果符合良好.

F-P干涉仪腔镜的不平行度对高斯光束透射特性的影响

基于多光束干涉原理,推导出了高斯光束经过两腔镜不严格平行的F P干涉仪后透射光强度的表达式。在此基础上,数值研究了F P干涉仪腔镜的不平行度对高斯光束透射特性的影响。结果表明:随着F P两腔镜不平行度的增大,透射光的峰值强度总体呈现下降的趋势,且在下降的过程中会出现波动;峰值强度位置将发生移动;光斑的均方根宽度将发生变化。

基于流形正则化的半监督投影双子支持向量机

投影双子支持向量机(PTSVM)是一种有监督学习方法,其性能极大依赖于有监督信息量的大小.受流形正则化框架启发,文中提出半监督投影双子支持向量机(SPTSVM).该方法可同时利用有监督(有标签样本)信息和无监督(无标签样本)信息构造一个更合理的半监督学习器.SPTSVM不仅继承PTSVM有监督分类性能,而且使用流形正则项捕获蕴含在无标签数据中的潜在几何信息.通过选择合理的参数,SPTSVM退化为有监督PTSVM或正则化PTSVM.在人工数据集和实际数据集上的对比实验验证文中方法的有效性.