一类连续时间广义边值系统的能控性和能观性

连续时间广义边值系统被描述为Ｅｘ（ｔ）＝Ａｘ（ｔ）＋Ｂｕ（ｔ），Ｖ０ｘ（０）＋Ｖｒｘ（Ｔ）＝ｖ，其中Ｅ是奇异方阵，ｕ（ｔ）满足任意次可微，本文讨论了一类连续时间广义边值系统的有关向内和向外边值过程的两个重要概念，并在此基础上给出了能控性和能观性定义和判别准则．     

Fundamental solutions for half plane with an oblique edge crack

An elastic half plane with an oblique edge crack is considered in this paper. A pair of concentrated forces or point dislocations is assumed to act at an arbitrary point in the half plane. The half plane with an edge crack is first mapped into a unit circle by a rational mapping function so that the following analysis can be carried out on the mapped plane analytically. Then the complex stress functions are derived by separating the whole problem into two parts; one is the principal part corresponding to the infinite plane acted on by concentrated forces or dislocations, the other is the holomorphic part, which can be determined by making use of the property of regularity of complex stress functions. The stress intensity factors of the crack can be calculated with different inclined angles of the crack, and the displacement and stress components at an arbitrary position in the half plane can be expressed explicitly.     

基于邻域统计特性的概率神经网络及其在自动目标识别中的应用

在概率神经的一种改进模型－ＦＤＯ网络的基础上，提出在设计网络收敛域时进一步考虑每一像素点周围８邻域的影响，对网络的作用函数加以修，使改进后的网络具有稳定性好且收敛速度快的优点。通过实验对改进前后网络的识别性能加以比较，证明改进后的网络特别适用于噪声图像的识别。     

挟沙水流中沙粒对底壁湍流猝发的影响

本文在常规的边界层方程中保留了法向压力梯度项和增加了固粒对流体的使用项，结合Ｂａｔｃｈｅｏｌｒ对交界面提供的分析方法和小扰动稳定性理论，由理论分析和计算得到了包括沙粒之类的悬浮体的尺寸，比重，流动Ｒｅ数，两相速度差以及扰动波数与水流底壁湍流猝发角的关系，计算结果与Ｒａｓｈｉｄｉ等的实验结果相符。文中所得结果有助于进一步了解江河底部水流的流动特性以及泥沙的沉积和输运机理。     

A BOUNDARY INTEGRAL EQUATION AND ITS NUMERICAL SOLUTION OF POTENTIAL FLOW AROUND SURFACE-IRREGULA-RITIES

In this paper,using complex functional theory,the authors turn the potentialflow around the surface irregularities in a pressure conduit and semi-infinite platforms intoDirichlet problem.Based on Schwarz formula and by the application of Plemelj's formula,theauthors change the problem into the integration of a Cauchy boundary integral equation in theflow plane through the substitution of variables.Using numerical integration,the authors obtainthe velocity distribution and pressure coefficient along surface irregularities and platforms.Thephysical concept of this method is clear,the convergent speed is rapid and the computative effi-ciency is high.The calculated values agree well with the measured results.It is an effective andsimple method in solving potential flow.     

人工智能方法用于熔喷过程在线检测的研究

熔喷生产最终产品质量的控制，对于过程优化具有重要的意义。产品质量的在线检测是过程监测与质量控制必不可少的。本文首次在熔喷系统中引入人工智能的研究方法，对熔喷非织造布的厚度与纤维直径进行在线预测。实测数据与人工神经网络的预测结果吻合较好，从而证实了人工神经网络用于模拟熔喷过程的可行性。网络的输入包括：挤出机温度、喷头温度、熔融体挤出速率、空气流率及接收距离。神经元计算的输出是纤维直径与熔喷非织布的厚度。文中进一步提出了将神经网络与专家系统相结合的研究路线，以实现熔喷过程的优化控制。     

Superplastic behaviour of Ti-48at.%Al two-phase titanium aluminide compacts made from mechanically alloyed powder

An HIP compact of MA-processed powder having a nominal composition of Ti-48at.% Al was produced. The compact consisted of a large amount of TiAl(λ) and a small amount of Ti₃Al (₂), in a completely ultra-fine equiaxed grain structure. This two-phase compact showed typical superplastic deformation behaviour. A maximum elongation of 550% was obtained. A strain exponent, n = 2, and grain size exponent, p = 2, were determined from the results of a strain-rate-change test and a creep test at constant initial stress using samples having various grain sizes, respectively. The activation energy for creep, Q_c at constant stress was calculated to be 350 kJ/mole. It is concluded that the superplastic deformation mechanism of the material under study is grain boundary sliding controlled by lattice diffusion in the TiAl phase.     

Microstructural Effects of Ion Implantation on Molybdenum

The effects of ion implantation and subsequent annealing on the microstructure of molybdenum have been investigated by transmission electron microscopy. The ions investigated were carbon, nitrogen, and tellurium. The ion-induced damage was found to give rise to grain boundary migration phenomena both during implantation and during subsequent annealing. Precipitation or ordering was found to occur for each ion on annealing.     

应用人工神经网络优化五点法水平井井网

水平井直井联合布井方式越来越多地应用到油田的实际开发中，但是井网的选择与优化方面的研究比较少，特别是没有根据不同开发阶段变化的生产动态对水平井井网进行优化。应用人工神经网络技术，在进行数值模拟计算的基础上，制作样本并训练网络，可预测任意面积、任意水平井长度下的五点法水平井井网在不同开发阶段的生产动态，优化出不同生产时间和不同井网面积条件下最优井网穿透比，即生产10a时最优井网穿透比为0．55～0．8。与数值模拟结果进行对比，最大误差在1％以内，表明这种方法既快速简便又具有较高的精确性，可以有效地指导五点法水平井井网优化设计。