基于语义Choquet积分的知识服务能力测评方法

针对知识服务中考虑模糊语言关联性信息的知识服务能力评价与提升策略问题,提出了一种评价方法。为了解决该问题,首先构建了知识服务能力的评价指标体系和矩阵分析模型;然后,明晰了知识服务能力的语言评价指标会在不同程度上存在关联的问题特征及二元语义信息处理方法的特征,并进行了形式化描述;同时,提出并定义了二元语义Choquet积分算子,给出了一种基于二元语义Choquet积分算子评价知识服务能力的方法;最后,通过算例分析说明了本文给出方法的可行性和实用性。     

软件项目支付进度问题研究

软件项目的复杂性和特点使得其支付进度问题较具特殊性。针对软件质量的差异性,对软件项目支付进度问题进行研究。以承包商净现值最大化为优化目标,构建了具有惩罚结构的支付进度优化模型,该模型反映了不同软件质量对承包商收益的影响;针对该问题的特点设计了启发式求解算法;最后通过一个算例对模型进行了验证。可为承包商在签订外包合同及外包实施中的策略提供决策依据。     

Molecular Dynamics Studies on Ballistic Thermal Resistance of Graphene Nano-Junctions

Ballistic thermal resistance of graphene nano-junctions is investigated using non-equilibrium molecular dynamics simulation. The simulation system is consisted of two symmetrical trapezoidal or rectangular graphene nanoribbons(GNRs) and a connecting nanoscale constriction in between. From the simulated temperature profile, a big temperature jump resulted from the constriction is found, which is proportional to the heat current and corresponds to a local ballistic thermal resistance. Fixing the constriction width and the length of GNRs, this ballistic thermal resistance is independent of the width of the GNRs bottom layer, i.e., the convex angle. But interestingly, this thermal resistance has obvious size effect. It is inversely proportional to the constriction width and will disappear with the constriction being wider. Moreover, based on the phonon dynamics theory, a theoretical model of the ballistic thermal resistance in two-dimensional nano-systems is developed, which gives a good explanation on microcosmic level and agrees well with the simulation result quantitatively and qualitatively.     

红外无损检测在面向等离子体部件检测的应用

建立了主动式内热源激励的红外无损检测平台,并进行了不同缺陷大小的EAST W/Cu 面向等离子体部件(PFC)的无损检测实验(NDT),得到试件表面的红外热图。通过最大温差值融合法消除由试件表面发射率不均匀引起的温度误差,通过快速离散傅里叶变换法提高了图像的信噪比,实现了对W/Cu 面向等离子体部件缺陷的分辨。通过数值模拟的手段对可能影响该检测技术的关键因素进行了定量分析。     

水冷包层模块局部模型流量分配数值模拟

用JAEA 水冷包层(WCSB)模块侧壁的流量分配实验验证了SST 湍流模型,建立了水冷包层模块第一壁的数值分析模型,研究了第一壁各冷却管中的流量分配情况。分析结果发现,第一壁模型入口集管中存在复杂的流动行为,冷却剂可以明显的区分为主流和逆流两大部分,其中主流进入冷却管的位置是影响流量分配的主要因素,逆流区形成了一系列次级涡流和沿管壁的环向流动。分析结果表明,在第一壁模型中,在所以条件下,各冷却剂通道中存在流量分配不均匀现象,流速最大值与最小值偏差均小于2%。     

傅里叶变换红外光谱在胃癌淋巴结转移诊断中的应用研究

观察傅里叶变换红外光谱技术无创、原位、快速诊断胃癌淋巴结转移的可行性。联合使用衰减全反射探头及傅里叶变换红外光谱仪测量新鲜离体胃周淋巴结红外光谱,发现每条光谱在吸收波长3 000～1 000 cm-1之间循序出现13条谱带,依据病理检测结果将淋巴结分为转移组及非转移组,比较两组淋巴结红外光谱的峰位和相对峰强等指标结果,最后进行标准统计学分析。36例胃癌患者,共检测淋巴结720枚,其中转移性淋巴结180枚,未转移540枚;与非转移淋巴结相比,转移淋巴结红外光谱有如下特征：(1)与核酸相关的峰强比I_{1 240}/I_{1 460}(p=0.015)和I_{1 080}/I_{1 460}(p=0.034)显著升高,提示转移淋巴结细胞的核酸含量增多;(2)与蛋白相关的I_{1 640}/I_{1 460}(p=0.001)和I_{1 546}/I_{1 460}(p=0.027)峰强比值升高,表明转移淋巴结组织的蛋白质含量明显升高;(3)与脂类相关的I_{2 855}/I_{1 460}和I_{1 740}/I_{1 460}显著降低(p＜0.001),提示癌组织脂类含量相对减少;(4)I_{1 160}/I_{1 460}(p=0.023)显著降低,表明恶性细胞糖类物质的减少。研究结果显示,傅里叶变换红外光谱分析技术有望成为术中原位、在体和快速诊断胃癌淋巴结转移的新方法。     

花生种子品质的可见-近红外光谱分析

利用600～1 100 nm波段研究花生品种的可见-近红外反射光谱,对选取的三种具有代表性的花生种子进行实验。使用近红外光纤光谱仪采集光谱数据,对原始光谱进行小波分析以提取光谱特征,再用主成分分析方法进行聚类分析,最后把每一个样品的前4个主成分得分作为识别模型的输入,品种类别作为模型的输出,以马氏距离作为判别函数,建立了线性判别分析模型。对于每个品种的50个样品,随机挑选30个样本作为训练集,剩余的20个样本作为预测集。该识别模型对3个花生品种的平均正确识别率为95%。表明该方法能有效的识别花生种子,得到较好的分类效果,为花生种子品质的区分和鉴别提供了一种新方法。     

基于光谱分析的重载车辆变速箱可靠性评估方法研究

润滑油光谱分析是变速箱磨损状态监测的重要手段。以某重载车辆变速箱台架试验中润滑油原子发射光谱分析得到的元素成分含量检测数据为基础,通过相关性分析和磨损机理分析确定了能够反映整机磨损状态和失效特点的金属元素,并对补油、换油后的光谱检测数据进行了修正,将变速箱试验过程中特征金属元素的含量随试验时间的变化规律表示为线性函数的形式。然后,综合考虑检测数据的时序规律与离散性,利用不同时刻光谱检测数据的正态分布均值和标准差进行变速箱可靠性评估,进一步讨论了失效阈值对可靠性评估结果的影响。研究表明,重载车辆变速箱润滑油中的Cu元素浓度与其他金属元素浓度的相关性较高、绝对值较大,既能够反映变速箱整体磨损状态,又便于检测,利用多个试验样本的润滑油Cu元素浓度光谱检测时序数据,可以对重载车辆变速箱进行可靠性评估;若失效阈值取定值,则失效阈值越大,同一时刻变速箱的可靠度越高;若失效阈值取随机分布,则失效阈值的离散度越大,可靠度随时间延长而降低的趋势越缓慢。当可靠度R大于0.9时,失效阈值的均值对可靠性评估结果的影响十分显著,同一时刻变速箱的可靠度随失效阈值标准差的增大而降低。将光谱分析与概率统计学结合进行变速箱可靠性评估方法研究,扩展了光谱分析技术的应用范围,具有创新性。     

用于空间激光通信的InGaAsP 激光二极管辐射损伤研究(英文)

对1:55 μm波长DFB 结构的InGaAsP 多量子阱激光二极管开展电子和60Co- 射线辐照试验。试验结果表明,激光二极管的斜度效率主要受带电粒子沉积的电离总剂量影响,而阈值电流和光功率主要受位移损伤剂量的影响。利用位移损伤剂量方法评价激光二极管的辐射损伤特征,并且预测其在空间辐射环境中的光功率衰退情况。模拟计算结果表明,MEO轨道辐射环境对激光二极管光功率辐射损伤远大于GEO轨道的影响,这主要是由于MEO轨道辐射环境的高能电子通量密度远大于GEO轨道的通量密度。The 1.55 μm InGaAsP multi-quantum-well laser diodes with distributed feedback structures were irradiated by electrons and 60Co- rays. The experimental results show the slope efficiency of laser diode is mostly affected by the total ionizing dose produced by charging particles, and the threshold current and the optical power mainly by displacement damage dose. The displacement damage dose methodology was employed to evaluate radiation damage of the laser diodes, and to predict the power degradations of these diodes in space. The calculated results indicate that the optical powers of the diodes will have more serious degradation for medium Earth orbit than for geosynchronous Earth orbit,due to higher fluence density of high energy electrons in GEO orbits.     

A data transmission method for particle physics experiments based on Ethernet physical layer

Due to its advantages of universality, flexibility and high performance, fast Ethernet is widely used in readout system design for modern particle physics experiments. However, Ethernet is usually used together with the TCP/IP protocol stack, which makes it difficult to implement readout systems because designers have to use the operating system to process this protocol. Furthermore, TCP/IP degrades the transmission efficiency and real-time performance. To maximize the performance of Ethernet in physics experiment applications, a data readout method based on the physical layer(PHY) is proposed. In this method, TCP/IP is replaced with a customized and simple protocol, which makes it easier to implement. On each readout module, data from the front-end electronics is first fed into an FPGA for protocol processing and then sent out to a PHY chip controlled by this FPGA for transmission.This kind of data path is fully implemented by hardware. From the side of the data acquisition system(DAQ),however, the absence of a standard protocol causes problems for the network related applications. To solve this problem, in the operating system kernel space, data received by the network interface card is redirected from the traditional flow to a specified memory space by a customized program. This memory space can easily be accessed by applications in user space. For the purpose of verification, a prototype system has been designed and implemented.Preliminary test results show that this method can meet the requirements of data transmission from the readout module to the DAQ with an efficient and simple manner.