基于RBF神经网络的较低浓度下同步荧光光谱的溢油鉴别

针对海面溢油样品的含量难以确定,同时考虑到海水掺杂及风化等问题的影响,提出了在较低非线性浓度范围内采集溢油嫌疑样品的同步荧光光谱,获取其训练样本集,利用主成分分析法(Principal com-ponent analysis,PCA)提取其特征光谱,结合径向基函数(Radial basis function,RBF)神经网络对肇事样本和嫌疑样本进行模式识别的方法。通过对相近油源原油样品分类识别研究表明:该方法仅需单次对肇事样本同步光谱测量,再借助数据分析,就可以很好区分相近油源溢油样品,外扰对识别率影响也不大。RBF神经网络算法识别率在92%左右。该结论对海洋环境中溢油的实时检测及油指纹数据信息库的建立有重要意义。     

Analysis of film strain and stress in a film-substrate cantilever system

The bending problem of a magnetic film-nonmagnetic substrate cantilever system is studied by using the principle of energy minimization. Emphasis is placed on the analysis of geometrical and physical parameter dependence of the neutral plane, internal film stress and strain of the cantilever system, and then the influence of such a parameter on the bending characteristic is presented. The results indicate, owing to the anisotropic expanding feature of the magnetostriction, that the neutral plane is generally anisotropic, and moves downwards rapidly with the increasing thickness ratio. Meanwhile, the bounding rigidity of substrate on the film will decrease with the increasing thickness ratio, and thus release the film stress, i.e., it decreases, but the film strain increases. The effect of Poisson’s ratio of the materials on the film strain, the stress and the neutral plane in the direction transverse to the magnetization is prominent. For the strain and the stress in the magnetization, however, the role of Poisson’s ratio is inconspicuous. This property is due to the initiative elongating (or contracting) feature of the magnetic film along its magnetization. Supported by the National Natural Science Foundation of China (Grant No. 10762001), the Key Project of the Chinese Ministry of Education (Grant No. 206024), and the Program for New Century Excellent Talents in University of China (Grant No. NCET-2005-0272)     

不同算法原子电荷下的~1J_(CH)理论计算

用一种计算直接键连原子核自旋耦合常数的半经验公式,结合量子化学计算得到的32种有机分子稳定几何构型的7种不同算法下的原子电荷,探究原子电荷算法的不同对1JCH理论计算的影响,拟合出基于7种原子电荷的耦合常数计算公式,并利用拟合公式对5种分子进行了检验.计算结果表明拟合的32种分子及检验的5种分子的耦合常数的计算值均与实验值较好的符合,拟合得到的基于7种原子电荷的计算公式均可以对其他分子体系的耦合常数进行预测.另外,计算结果同样显示原子电荷算法的不同对1JCH理论计算值有一定的影响却不显著,其中基于电荷均衡方法电荷(QEq)得出的耦合常数计算值与实验值的偏差较其它6种原子电荷的小,结果更可靠.     

最小Higgs模型在ILC上可能的味改变物理迹象

在最小Higgs(LH)模型中, 由于引进了类矢量顶夸克T, 对弱作用流产生了新的影响. CKM矩阵从3×3变为4×3, 在树图下出现味改变中性流(FCNC). 本文研究了在国际高能直线对撞机(ILC)上能否通过e⁺e^－→tc+tc和e⁺e^－→WWν_eν_e→(tc+tc) ν_eν_e过程探测到LH模型所预言的FCNC信号.     

用振幅比值方法预言中性粲介子衰变到CP本征态的CP不对称性

用振幅比之方法计算了中性粲介子衰变到CP本征态的CP不对称性. 计算了时间相关和时间积分的CP不对称性. 结果表明, 时间积分的CP不对称参数约为千分之一的量级. 还讨论了在BESⅢ和B工厂上实验检验的可能性.     

非结构网格边粘性公式

利用Kuropatenko提出的人为粘性形式和Christiansen建议的TVD(Total Variation Diminishing)输运限制子构造非结构网格边粘性,用最大角原则选择邻边构造限制子,研制包括边粘性的相容流体程序,给出边粘性与体粘性数值模拟比较,对Saltzman活塞问题、Noh问题进行数值实验.数值实验结果比较说明,非结构网格边粘性准确检出强速度梯度不存在的情形,减弱人为粘性力的作用,有效捕捉流体的激波间断,提高模拟精度.     

Some fractal properties of the percolating backbone in two dimensions

A new algorithm is presented, based on elements of artificial intelligence theory, to determine the fractal properties of the backbone of the incipient infinite cluster. It is found that the fractal dimensionality of the backbone isd _f ^BB =1.61±0.01, the chemical dimensionality isd _t=1.40±0.01, and the fractal dimension of the minimum pathd _min=1.15 ± 0.02 for the two-dimensional triangular lattice.     

基于模型的制冷系统智能化仿真研究

为提高仿真方法对实际制冷空调装置多样性和复杂性的自适应性,将人工智能引入到制冷系统仿真研究中,构建基于数学模型与人工智能技术相结合方式的制冷系统智能仿真理论．本文介绍了上海交通大学在此方面所取得的研究成果,并提出了今后进一步发展的方向．本项研究有利于推进制冷装置设计方法现代化,研究思路对于一般热力系统仿真亦有参考价值．     

基于人工传输线结构的小型化低通滤波器

提出了一种基于人工传输线结构的小型化低通滤波器的设计方法。通过利用电磁仿真软件直接构建出微带电抗器件并对其逐一进行参数提取验证,将设计的微带电抗器件进行组合集成和调试后得到相应的低通滤波电路。对5阶巴特沃兹低通滤波电路和0.5 dB切比雪夫低通滤波电路的设计和测量结果对比表明,采用本方法所设计的滤波器结构紧凑,面积约为常规低通滤波电路的23%,测量结果与仿真吻合。     

Combustion mechanism of ultralean rotating counterflow twin premixed flame

In our previous numerical studies [Nishioka Makihito, Zhenyu Shen, and Akane Uemichi. “Ultra-lean combustion through the backflow of burned gas in rotating counterflow twin premixed flames.” Combustion and Flame 158.11 (2011): 2188–2198. Uemichi Akane, and Makihito Nishioka. “Numerical study on ultra-lean rotating counterflow twin premixed flame of hydrogen–air.” Proceedings of the Combustion Institute 34.1 (2013): 1135–1142]. we found that methane– and hydrogen–air rotating counterflow twin flames (RCTF) can achieve ultralean combustion when backward flow of burned gas occurs due to the centrifugal force created by rotation. In this study, we investigated the mechanisms of ultralean combustion in these flames by the detailed numerical analyses of the convective and diffusive transport of the main species. We found that, under ultralean conditions, the diffusive transport of fuel exceeds its backward convective transport in the flame zone, which is located on the burned-gas side of the stagnation point. In contrast, the relative magnitudes of diffusive and convective transport for oxygen are reversed compared to those for the fuel. The resulting flows for fuel and oxygen lead to what we call a ‘net flux imbalance’. This net flux imbalance increases the flame temperature and concentrations of active radicals. For hydrogen–air RCTF, a very large diffusivity of hydrogen enhances the net flux imbalance, significantly increasing the flame temperature. This behaviour is intrinsic to a very lean premixed flame in which the reaction zone is located in the backflow of its own burned gas.