Three-dimensional MHD high-resolution computations with CWENO employing adaptive mesh refinement

Until recently, numerical simulations of discontinuities in highly super-Alfvénic plasmas have been severely limited by comparatively crude resolution and accuracy. Significant progress in the numerical simulation of such plasmas was achieved with the recently implemented Central Weighted Essentially Non-Oscillatory (CWENO) scheme. Combining this technique with that of adaptive mesh refinement (AMR), we have developed a third-order numerical scheme, which is able to efficiently capture strong gradients on spatial scales being small compared to the overall scale of the plasma system considered. Here, we first describe important algorithmic aspects of the scheme as well as the physics included in it. Second, we present the results of various performance tests. And, third, we illustrate its application to ‘real world problems’ using the example of the dynamics of a Sedov-type explosion.     

High training set size reduction by space partitioning and prototype abstraction

Instance-based learning methods like the nearest neighbour classifier generally suffer from the indiscriminate storage of all training instances, resulting in large memory requirements and slow execution speed. In this paper, new training set size reduction methods based on prototype generation and space partitioning are proposed. Experimental results show that the new algorithms achieve a very high reduction rate with still an important classification accuracy.     

Low temperature synthesis of aluminium titanate by an aqueous sol-gel route

Formation of aluminium titanate (AT) has been achieved at low temperature through sol-gel process using boehmite and titanium hydroxide as precursors by controlling the particle size at nanoscale followed by in-situ peptisation. The formations of AT phase, particle size distributions, sintering and thermal expansion characteristics, and microstructural features have been reported. DTA and XRD analysis have been performed to confirm the formation of AT. A 94% relative density was obtained for aluminium titanate sintered at 1550 °C with controlled grain size in the range of 2-3 μm.     

一种新的变步长ICA自适应算法

结合互信息极小的优化判据，基于估计函数期望平方准则、外点法和最速下降法的思想，本文导出了一种新的变步长独立分量分析(ICA)自适应算法。该算法克服了固定步长在分离矩阵推导过程中出现的稳态失调问题，比基于模拟退火步长的ICA算法有更快的仞始收敛速度和较高的分离精度。同时，该算法还具有较好的时变系统跟踪能力。理论分析和仿真计算结果证实了其可以有效地提高ICA的自适虚性，更准确地完成盲信号分离。     

Anisotropic mesh adaption by metric‐driven optimization

We describe a Gauss–Seidel algorithm for optimizing a three‐dimensional unstructured grid so as to conform to a given metric. The objective function for the optimization process is based on the maximum value of an elemental residual measuring the distance of any simplex in the grid to the local target metric. We analyse different possible choices for the objective function, and we highlight their relative merits and deficiencies. Alternative strategies for conducting the optimization are compared and contrasted in terms of resulting grid quality and computational costs. Numerical simulations are used for demonstrating the features of the proposed methodology, and for studying some of its characteristics. Copyright © 2004 John Wiley & Sons, Ltd.     

影响铁精矿K2O、Na2O、SiO2含量的因素分析

用统计分析方法 ,分析了影响精矿K2 O、Na2 O、SiO2 含量的因素 ,并根据分析结论提出了对生产工艺进行定向控制的条件     

高可靠性烧结金属丝网防爆元件的应用研究

烧结金属多孔元件作为隔爆元件的应用，很好地满足了火焰过滤元件的性能要求，进一步拓宽了烧结金属多孔元件的应用领域，章在论述其性能特点、结构类型的同时，对其在微孔结构设计上的差异也进行了描述。     

高炉煤气净化系统新技术研究与应用

济钢第二炼铁厂针对原高炉煤气净化系统处理能力不足等问题，研制了新的煤气净化系统。该系统采用单根文氏管、串联花环、丝网脱水器湿法除尘，取代原串联文氏管电除尘净化工艺。应用表明，新系统运行安全，净煤气含尘量小于10^mg／m^3，含水量小于5g／m^3，高炉风温提高90℃。     

机械合金化过程中Fe70B30粉末晶粒尺寸和微观应变的研究

用Ｘ射线和电镜研究了Ｆｅ＿（７０）Ｂ＿（３０）粉末经不同时间高能球磨后晶粒尺寸和微观应力的变化。在机械合金化过程中，粉末的Ｘ射线衍射谱的宽度随球磨时间的增加逐渐加宽，这是晶粒细化和内部微观应力共同作用的结果。Ｘ射线衍射结果表明：随着机械合金化的进行，粉末的晶粒尺寸逐渐减小，球磨初期晶粒尺寸下降较快，经１５ｈ球磨，晶粒尺寸为２５ｎｍ，机械合金化进行到一定时间后晶粒尺寸下降缓慢，８０ｈ球磨后晶粒尺寸可达５ｎｍ。在机械合金化过程中球磨所造成的微观应变不大，球磨初期粉末的内应力随球磨时间增加而增加，当粉末粒子尺寸很小时，随球磨的进行粉末中的微观应变显著下降。     

Electrical/electronic effects of titanium and iron impurities in EFG and FZ solar cell silicon: SPV/EBIC analysis

In this paper results on surface photovoltage (SPV) and electron beam induced conductivity (EBIC) studies of edge-defined film-fed growth (EFG) and floating zone (FZ) silicon solar cell materials (both p-type) are presented. A systematic comparison based on minority carrier diffusion length and carrier recombination is made between: (i) samples contaminated with Ti and/or Fe, (ii) samples gettered by phosphorous diffusion, and (iii) as-received samples. Deep level transient spectroscopy (DLTS) measurements, together with the iron-boron (FeB) pairing kinetics [1] have successfully been used to detect the presence of Fe in the samples. Even though this process is effective in revealing Fe impurities in p-type FZ silicon it is evidently not suitable for Fe identification in p-type EFG silicon. Ti, like Fe, is found to be a prominent lifetime-limiting metallic impurity in both EFG and FZ silicon. Phosphorous diffusion is proven to be an effective external gettering technique for fast-diffusing impurities such as Fe, but not for Ti.