A novel application of the FPM to the buckling differential equation of non-uniform beams

In this study the buckling load (P) of non-uniform, deterministic and stochastically heterogeneous beams, is found by applying the Functional Perturbation Method (FPM) directly to the Buckling (eigenvalue) Differential Equation (BDE). The FPM is based on considering the unknown P and the transverse deflection (W) as functionals of heterogeneity, i.e., the elastic bending stiffness “K” (or the compliance S=1/K). The BDE is expanded functionally, yielding a set of successive differential equations for each order of the (Frèchet) functional derivatives of P and W. The obtained differential equations differ only in their RHS, and therefore a single modified Green function is needed for solving all orders. Consequently, an approximated value for the buckling load is obtained for any given morphology. Four examples of simply supported columns are solved and discussed. In the first three, deterministic realizations of K are considered, whereas in the fourth, K is assumed to be the stochastic field. The results are compared with solutions found in the literature for validation.     

特征值平移超分辨率算法

本文分析了逆幂迭代算法的收敛速度。为了克服逆幂迭代算法矩阵求逆的过程,提出了特征值平移幂迭代算法,并分析了它的收敛速度,给出了计算机模拟结果。     

一种通用二次旋转曲面面轮廓度算法

提出了一种通用求二次旋转曲面面轮廓度的算法。首先使用最小二乘法利用所有的测量点求出被测曲面的一般方程,根据被测曲面特征采用高精度算法从二次齐次变换矩阵求出一特定的特征值及其相应的特征矢量,则该矢量为二次旋转曲面的旋转轴线方向。然后以此方向作为被测曲面的基准轴线方向,确定出旋转及平移坐标变换矩阵。根据各参照曲面的特征,确定出被测曲面的面轮廓度计算公式。最后给出一个模拟实例以验证该算法的有效性。这一算法在数学模型建立过程中没有使用任何简化和近似的方法,所以不存在模型误差,仅有计算误差,具有精度高、算法稳定的特点,而且对被测物体的放置没有特殊的限制,在坐标类测量仪器中具有广阔的用途。     

Lurie时滞控制系统鲁棒绝对稳定的新判据

Ｌｕｒｉｅ型时滞控制系统是工程实践中经常遇到的重要的非线性控制系统,根据Ｌｙａｐｕｎｏｖ稳定性理论,提出了不确定性Ｌｕｒｉｅ时滞控制系统稳定的充分条件,并给出了一个应用实例．     

LMS算法自相关矩阵特征值表达式的计算分析

在研究线性滤波ＬＭＳ 算法的收敛条件及推导其稳态相对误差的上下限算式时, 都用到自相关矩阵的特征值的表达式。根据自相关矩阵的特点, 对ＬＭＳ 算法收敛步长的上限公式及其稳态相对误差的上下限公式中出现的特征值表达式估算作了分析, 使得有关的公式能用于具体指导ＬＭＳ 算法。     

柞树叶对麻织物的染色性能研究

柞树叶含有较高单宁,是较好的黑灰色染材。文中采用水煎法萃取柞树叶中的色素得到染液,探讨了不同染色方法对麻织物颜色特征值的影响,并优化了预媒染工艺条件。结果表明,直接染可染浅灰黄色调,应用范围有限;预媒染适合染较深的黑灰色调,同媒染适合染中灰色调,后媒染适合染浅灰黄色调。其中,预媒染应用价值最大。最佳预媒染色工艺为:浴比1∶50、染液质量浓度100.00%(pH值为6)、染色温度80℃、染色时间30 min、皂矾媒染剂质量浓度5.00 g/L、染色次数为7次。     

Some Refined Eigenvalue Perturbation Bounds for Two-by-Two Block Hermitian Matrices

We consider eigenvalue perturbation bounds for Hermitian matrices, which are associated with problems arising in various computational science and engineering applications. New bounds are discussed that are sharper than some existing ones, including the well-known Weyl bound. Two numerical examples are investigated, to illustrate our theoretical presentation.     

基于ANSYS的ITER重力支撑系统特征值屈曲分析

针对ITER重力支撑系统的结构特点,应用有限元软件ANSYS,建立了ITER重力支撑系统的有限元模型.采用Block Lanczos方法对ITER重力支撑系统在静载荷与地震载荷同时作用下的稳定性进行了特征值屈曲分析,求出了ITER重力支撑系统的前3阶屈曲特征值和特征值屈曲模态.结果表明ITER重力支撑系统不会发生屈曲,韧性板是整个支撑系统中最易发生屈曲的零件.屈曲分析的结果对ITER装置的设计与改进具有一定的指导意义.     

A new methodology for identification of critical lines using damping sensitivity analysis

When an inter-area mode dominates a low-frequency oscillation in a stressed condition, control of the active power flow of interface lines, with compensating devices, can effectively reduce the electromechanical power oscillations. In general, interface lines in which inter-area oscillations are large are considered to be good locations for installation of compensating devices. A sensitivity analysis with respect to change in active power flow can provide an important factor in electric power system operation. This paper proposes a new methodology to calculate the damping sensitivity with respect to change in active power flow, which can be useful for accurate selection of critical lines from the viewpoint of small-signal stability. In the proposed methodology, a damping sensitivity index is used to select the critical lines to damp power system oscillations. This paper describes how to derive the damping sensitivity for the selected mode and illustrates an example applying the proposed algorithm to a simple two-area system and the New England 39-bus test system.     

基于噪声子空间特征值重构的DOA估计算法

该文针对非等功率信号波达方向(DOA)估计问题,提出一种基于噪声子空间特征值重构(Eigenvalue Reconstruction of Noise Subspace, ERNS)的超分辨算法。算法对接收信号自相关矩阵进行特征值分解,通过重构噪声空间特征值以及引入虚拟信源来构造新的接收信号自相关矩阵,对该矩阵进行特征值分解得到新的噪声空间特征值。当虚拟信源与实际信源入射方向相同时,新噪声空间特征值与重构后噪声空间特征值保持不变,利用这一特性来估计信源入射方向。该文给出算法的原理及实现步骤,并通过仿真进行原理验证与性能分析,仿真结果表明与其他子空间算法和MUSIC 算法相比,ERNS算法能够提高弱信号估计成功的概率。