Modification of Boundary Fluctuations by LHCD in the HT—7 Tokamak

Measurements of boundary fluctuations and fluctuation driven electron fluxes have been performed in ohmic and lower hybrid current drive enhanced confinement plasma using a graphite Langmuir probe array on HT-7 tokamak. The fluctuations are significantly suppressed and the turbulent fluxes are remarkably depressed in the enhanced plasma. We characterized the statistical properties of fluctuations and the particle flux and found a non-Gaussian character in the whole scrape-off layer with minimum deviations from Gaussian in the proximity of the velocity shear layer in ohmic plasma. In the enhanced plasma the deviations in the boundary region are all reduces obviously. The fluctuations and induced electron fluxes show sporadic bursts asymmetric in time and the asymmetry is remarkably weakened in the lower hybrid current driving (LHCD) phase. The results suggest a coupling between the statistical behaviour of fluctuations and the turbulent flow.     

人工脑的三维模型及应用

将二维人工脑模型推广到三维，并给出了一个实际应用算例，还将二、三维模型的计算结果进行了对比，发现三维模型的计算效果明显优于二维模型。     

一种视频图像规则圆目标的超分辨率位置测量方法

介绍了一种用于视频图像规则圆目标的超分辨率位置测量方法。分析了超分辨率测量的原理 ,并提出和完成了测量方案 ,取得了满意的结果。首先在目标静止时 ,连续采集多帧视频图像 ,通过叠加平均进行降噪 ;其次通过罗伯特微分算子进行边界探测 ,求出目标的大致位置 ;最后通过边界重心拟合法、边界圆拟合法和最小二乘法拟合求出目标的精确位置。理论分析和实验表明 ,完整圆目标时 ,测量精度可达 1/4像素 ;而当目标图像有缺陷时 ,边界圆拟合法和最小二乘法拟合测量误差增加较少 ,边界重心法的测量误差增加较多。在具体情况下 ,算法及参数应该作相应的调整     

BOUNDARY LIMITS OF α-HARMONIC FUNCTIONS

In this paper, the existence of weighted boundary limits of α-harmonic functions is discussed.     

ABSTRACT FUNCTIONS OF BOUNDED VARIATION AND ABSOLUTE CONTINUITY

As well known that in 1938,I. M. Gelfand firstly introduced abstract functions of bounded variation from [a,b] to a Banach space. After Gelfand‘s work,many mathematicians investigated various properties and of this kind of abstract functions, and also paid attention to the abstract functions of absolute continuity In this paper, we summarize to explain our work [1-17] about this topic.     

一个扩散问题的自然边界元法与有限元法组合

本文讨论由Helmholtz方程描述的扩散问题的自然边界元法与有限元法的组合．取一个圆作为公共边界，用Fourier展开建立边界积分方程，将无界区域上的问题化为有界区域上的非局部边值问题．在变分方程中公共边界上的未知量只包含函数本身而不包含其法向导数，从而减少了未知数的数目，并且边界元剐度矩阵只有极少量不同的元素，有利于数值计算．这种组台方法优越于建立在直接边界元法基础上的组合方法．文中证明了变分解的唯一性，数值解的收敛性和误差估计．最后讨论了数值技术并给出一个算倒．     

基于BP算法神经网络的物理实验曲线拟合

介绍了人工神经元网络的基本原理和BP算法神经网络。将BP算法神经网络应用于物理实验中的幅频特性曲线拟合，在MATLAB下通过训练和仿真验证了应用BP算法神经网络拟合实验曲线的优越性。     

自组织人工神经元网络用于声图象识别的研究

本文通过实验研究了自组织人工神经元网络用于声图象识别的步骤和方法，讨论了在水声、超声医学等声图象识别中所遇到的一些关键性技术问题．