低对比度锥束CT图像缺陷分割算法

根据锥束CT图像的特点,通过获取切片图像中的灰度拓扑结构,将整幅图像中的低对比度分割问题转化为在局部拓扑结构中的较高对比度分割问题.然后针对每个拓扑结构,根据设置的2个阈值和4个检测模板,判断当前模板区域是缺陷区域、噪声区域或者背景材质区域,并将4个方向的检测结果相加得到该拓扑结构的检测结果,最终的检测结果为各拓扑结构的检测结果之和.对空心涡轮叶片蜡模锥束CT图像的实验表明,该算法可有效提取低对比度缺陷信息.     

衍射增强成像的若干基础问题研究Ⅰ.衍射增强成像的衬度与X射线能量的关系

为了探讨X射线衍射增强成像的衬度与X射线能量的关系,利用北京同步辐射光源4W1A光束线引出的硬X射线对大块正常和癌变的乳腺组织进行成像研究.在摇摆曲线顶部位置获得的图像(峰位图像)和表观吸收图像的衬度随X射线能量的变化关系是相似的,说明峰位图像和表观吸收图像包含的主要衬度相似,即吸收衬度,而折射图像的衬度随X射线能量的增大总体上呈现下降趋势.综合来看,对于乳腺类的软组织来说,DEI成像在低能量端有很好的衬度,反映了衍射增强成像更适合于主要由轻元素组成的物体的成像.     

X射线包裹图像颗粒状物品纹理分割

基于图像的纹理特征分割是图像分割的有效方法。X射线包裹图像中颗粒状物品具有比较明显的纹理特征。通过分析包裹中常见颗粒状和非颗粒状物品的纹理特性,采用滑动窗口对包裹图像进行模板扫描,提取局部图像的纹理特征,基于包裹内物品的纹理特征采用改进的K均值算法对其进行聚类,找出颗粒状物品区域并标记颗粒状物品边界,从而实现X射线包裹图像的颗粒状物品纹理分割。试验表明,该方法能够有效地分割出X射线包裹中颗粒状物品。该方法的特点是无监督、全自动、对纹理特征敏感。     

基于局部关联信息的视频监控图像中γ辐射噪斑的消除方法

为消除γ辐射环境视频监控图像内大量形状不规则、分布无规律、能量不均衡的噪斑,提出一种利用图像局部相邻像素关联信息去除噪斑的方法。首先根据γ辐射环境视频监控图像噪斑像素值的突变特性分割噪斑区域,再利用噪斑与当前帧噪斑相邻干净区域的局部关联性及噪斑与相邻帧相同坐标干净区域像素点的局部关联性,以两个干净区域像素值为基准、像素值差值为关联度,采用步进式方法修复噪斑区域。实验结果表明,与传统的图像噪斑去除算法相比,本方法在量化指标和视觉感知上均有较大提高。该噪斑消除方法可有效去除γ辐射环境视频监控图像的噪斑,保留图像细节纹理信息。     

基于局部关联信息的视频监控图像中γ辐射噪斑的消除方法

为消除γ辐射环境视频监控图像内大量形状不规则、分布无规律、能量不均衡的噪斑,提出一种利用图像局部相邻像素关联信息去除噪斑的方法。首先根据γ辐射环境视频监控图像噪斑像素值的突变特性分割噪斑区域,再利用噪斑与当前帧噪斑相邻干净区域的局部关联性及噪斑与相邻帧相同坐标干净区域像素点的局部关联性,以两个干净区域像素值为基准、像素值差值为关联度,采用步进式方法修复噪斑区域。实验结果表明,与传统的图像噪斑去除算法相比,本方法在量化指标和视觉感知上均有较大提高。该噪斑消除方法可有效去除γ辐射环境视频监控图像的噪斑,保留图像细节纹理信息。     

衍射增强成像的若干基础问题研究Ⅱ.衍射增强成像的边界效应和本底影响

在衍射增强成像原理的基础上,通过简单的实验模型分析了衍射增强成像(DEI)的过程与特点.在摇摆曲线顶部记录的峰位图像有很好的衬度,可以直接用于观察物体的内部结构.表观吸收图像的特点与传统X射线图像相似,但折射图像具有更好的边界衬度,而且腰位折射图像的成像效果比趾位折射图像的成像效果更好.同时也发现,分析晶体的本底信号会影响DEI成像的质量,但对折射图像没有任何影响,不会引起折射图像与实际物体之间的偏差.因而,折射图像用于辨别正常组织与病变组织、工业检测等方面具有很好的可信度.     

基于双树复小波变换的多聚焦图像融合

针对多聚焦图像的成像特点和变换后高低频系数的特性,提出了一种基于双树复小波变换的多聚焦图像融合方法。首先对源图像进行双树复小波变换,其次对低频系数采用基于改进的梯度算子的融合准则,而对高频系数采用基于局部区域能量匹配度的融合准则,然后进行一致性检验,最后进行重构得到融合结果。对融合结果采用交叉熵、均方根误差、扭曲程度等评价指标进行评估。实验结果表明,该方法能够获得清晰度更高的融合图像。     

小波重建中的小波选取

近年来发展的小波重建方法可以直接重建得到图像的小被系数，而且可以只利用通过感兴趣区域的局部投影数据进行重建，从而在进行局部重建时节省了计算时间。在进行小被重建时，不同的小波对重建效果有着不同的影响。主要讨论了Daubechies正交小波和双正交小波的重建效果．并总结了实际应用中小波选取的一些原则。     

用于堆芯换料优化的遗传算法的改进及检验

文章采用年龄代、相似度和价值函数等新技术对用于堆芯换料优化的遗传算法加以改进,其中年龄技术赋予了算法及时总结前一阶段方案搜索"经验"、引导算法更好地在局部最优邻域内进行搜索的能力;在对方案进行杂交之前首先评估方案的相似度,避免了对两个过于相似的方案进行杂交,从而防止算法早熟;价值函数的运用赋予算法依据较优方案共性特征的统计来产生新方案的能力.针对一个两环路堆芯换料优化基准题的数值检验说明,经改进的遗传算法可显著提高算法的搜索效率,同时也使优化解的质量得以提高.     

基于数学形态学BGA焊点分割和提取的改进方法

对经X射线采集到的BGA焊点缺陷图像进行预处理,传统的迭代阈值和灰度拉伸方法对灰度图像很难得到满意的结果,误分割很常见,针对BGA焊点图像的特性,提出了一种结合数学形态学的图像分割改进方法,分割后图像不仅保留了图像的重要信息,而且对于背景复杂、光照不均匀的图像也能取得良好的效果,该方法通过正确选取结构元素加快了图像分割的速度,在此基础上,采用连通域进行区域提取的方法能够准确无误的提取缺陷.实验结果表明,该方法很好实现了缺陷区域的分割和提取,且受噪声影响很小,为进一步的缺陷特征参量的处理莫定了良好的基础.     

Discussion of similarity principles for fluid-to-fluid scaling of heat transfer behaviour at supercritical pressures

In this paper, a rationale for comparing heat transfer behaviour of different fluids at supercritical pressures is discussed. The study is performed as an extension to heat transfer of a previous proposal of dimensionless groups suitable for assessing stability in heated channels with supercritical fluids. The additional difficulty to be overcome with respect to the case of stability of channels with imposed heat flux is related to the role played in the present case by thermo-physical properties, whose trends as a function of dimensionless enthalpy are different for different fluids, sharing anyway a general similarity across the pseudo-critical threshold.The proposed considerations are discussed with the aid of results obtained by computational fluid-dynamic codes, contributing to clarify the extent at which the reported considerations can be considered promising for developing a coherent similarity theory for heat transfer to fluids at supercritical pressures.     

Systematic and exact scaling analysis of the single-phase natural circulation flow: The hydraulic similarity

For the study of the hydraulic similarity in a single-phase natural circulation loop, the integral momentum equation is non-dimensionalized with respect to the initial flow kinematic energy of reference section, without intuitively specifying any reference parameters. By this mean, a unique hydraulic time scale, characterizing the system hydraulic response, is identified along with two dimensionless physical numbers: the dimensionless flow resistance number and the dimensionless gravitational force number. From the integral momentum equation, the mass flow rate at steady state is also obtained. The identified dimensionless parameters are then applied to derive a set of scaling criteria for the design of a full-pressure reduced-size similar model for a PWR (Pressurized Water Reactor). For exact hydraulic similarity, it was found for the first time that the cross sectional area scaling ratio should be related to the axial length scaling ratio. In addition, it is also found out that the relative cross-sectional area ratio should be preserved in order to preserve the flow resistances. Moreover, the scaling ratio for the number of the U-tubes was found to be unity if exact hydraulic similarity is pursued for the whole system. Three sets of scaling criteria for the design of a full-pressure model for a PWR are summarized in a table for different application. The accuracy and applicability of this proposed scaling method is demonstrated by proposing a simple loop and a PWR-like system, by scaling down the systems to get two corresponding models with this proposed scaling methodology, and by comparing the model results with their corresponding prototype results. Furthermore, the method for the evaluation of both system-level and local hydraulic scaling distortions are addressed.     

Numerical investigation on the similarity of developing laminar flows in helical pipes

This paper presents numerical investigations into the similarity of developing laminar flows at the entrance region of helical pipes. It is shown that the new characteristic parameters can be used to represent the flow development in helical pipes. The development of friction factor and flow pattern along with the characteristic parameters is not significantly affected by the variations of Reynolds number, dimensionless pitch, or curvature ratio at a given condition of modified Dean number and inlet velocity profile. The development of the friction factor is investigated in a modified Dean number range of 20–400, and a new correlation for the fully developed angles of laminar flows in helical pipes is proposed.     

A study on applicability of similarity rule to performances of centrifugal pumps driven in two-phase flow

It is studied in this report whether the similarity rule which is used for pumps in single-phase flow is applicable to those in two-phase flow using fundamental equations by which the hydraulic torques of pumps in two-phase flow are described pretty well.According to the equations, the similarity rule is applicable to pumps driven in two-phase flow if the force acting on the gas-liquid interface per unit length of the flow channel in the impeller is proportional to the pump size, and according to the study of the force it can be said that the force is proportional to the pump size unless the gas velocity is very small or liquid velocity is extremely large in the impeller.Therefore it is concluded that the similarity rule is applicable to pumps in two-phase flow excepting special operating conditions which are not usually realized.     

氟盐冷却球床堆与HTR-10和MSRE的定量相似性分析

氟盐冷却球床堆是当前国际上一种新的研究堆型,尚无已经建造完成的反应堆,因此,选择相似且具有运行经验的反应堆作为基准题有助于堆芯核设计软件适用性分析。利用国际上常采用的相似性分析软件,可对熔盐实验堆(Molten Salt Reactor Experiment,MSRE)及10 MW高温气冷堆(10 MW high-temperature gas-cooled test reactor,HTR-10)与氟盐冷却球床堆的相似性进行分析,定量判断它们作为基准题的合理性。分析结果表明,MSRE和氟盐冷却球床堆的能谱峰位能量接近且堆内元素种类相近,二者相似程度较高;常温临界HTR-10和氟盐冷却球床堆冷却剂不同,且能谱峰位能量差异较大,二者相似程度较低。因此,MSRE是氟盐冷却球床堆中子物理设计软件较理想的基准题。     

Experimental investigation on similarity between velocity and density profiles in density-stratified countercurrent flow in reactor horizontal leg

A density-stratified countercurrent flow was investigated to obtain data necessary to develop a physical model on a thermally stratified flow in a horizontal leg of a pressurized water reactor (PWR). The experiments were conducted at atmospheric pressure and temperature using fresh water and NaCl solution with a non-dimensional density ratio of up to 1.2. The emphasis was placed on measurements of velocity and concentration profiles near the interface between the two fluid layers. Measured mean velocity and concentration profiles were fitted consistently using the Monin–Obukhov similarity theory, which are well-known outcomes for stratified turbulent shear flow. The interfacial friction and entrainment coefficients obtained from the fitted profiles agreed well with existing results in literature, confirming the applicability of the Monin–Obukhov theory. Furthermore, a new empirical correlation was proposed for the prediction of a mixing layer thickness.     

On the applicability of Ishii's similarity parameters to integral system test facilities

This note evaluates the applicability range of Ishii's similarity parameters to integral system test facilities (IST) subjected to slow Small-Break Loss-Of-Coolant Accidents (SBLOCA). It is recognized that the development of Ishii's parameters is theoretically sound and rigorous. However, the range of applicability should be limited to conditions for which the original assumptions hold. For the IST facilities, the major out-of-range application is that Ishii's parameters, developed from the analysis of an open system, are applied to closed integral systems. An enthalpy transport analysis of a closed system demonstrates that such an extrapolation leads to erroneous conclusion. It is shown that two mechanisms which distinguish a closed system from an open system have significant effects on the scaling of system phenomena. One is the interaction/feedback mechanism, which is an inherent characteristics of an integral system. Another is the limited amount of available vapor space in closed systems.     

Quantitative prediction of natural circulation in an LMFR with a similarity law and a water test

This paper is devoted to logical derivation of a similarity law for single-phase natural circulation expected in a liquid metal fast reactor (LMFR). Though the way of the derivation is principally conventional, this paper shows that explicit definition of a representative velocity and a representative temperature difference, as used in previous studies, is generally inappropriate in formulating a similitude law. The paper also presents formulae which allow to directly convert water test results into actual plant values. Using the conversion formulae, it is demonstrated that data acquired in a small scale water experiment can be converted and the results are comparable with results of actual LMFR computation. The accuracy of the experimental prediction is discussed.     

Study of three—dimensional PET and MR image registration based on higher—order mutual information

Mutual information has currently been one of the most intensively reserached measures.It has been proven to be accurate and effective registration measure.Despite the general promising results,mutual information sometimes smight lead to misregistration because of neglecting spatial information and treating intensity variations with undue sensitivity.In this paper,an extension of mutual information framework was proposed in which higher-order spatial information regarding image structures was incorporated into the registration processing of PET and MR.The second-order estimate of mutual information algorithm was applied to the registration of seven patients.Evaluation from Vanderbilt University and our visual inspection showed that sub-voxel accuracy and robust results were achieved in all cases with second-order mutual information as the similarity measure and with Powell‘s multidimensional direction set method as optimization strategy.     

CFD-assisted scaling methodology and thermal-hydraulic experiment for a single spent fuel assembly

A computational fluid dynamics (CFDs)-assisted scaling methodology is proposed for a TN24P cask (Creer et al., 1987). In the proposed methodology, the height and the number of assemblies were scaled down based, respectively, on the integral scaling criteria by Ishii and Kataoka (1984) for a natural-circulation loop under single phase flow and on the concept of an insulator on the surface of the basket. The controlling scaling parameters for the TN24P cask (the scaling ratio of the heat flux, 1.6; the thickness and conductivity of the insulator, 9 mm and 0.05 W/m K, respectively) were estimated by comparing the results of a TN24P cask experiment (Creer et al., 1987) with those of a CFD simulation of a TN24P cask scaled down, using Fluent code. Based on the proposed scaling methodology and its scaling parameters, a thermal-hydraulic experiment with a half-height single assembly was carried out. The experiment was analyzed in comparison with a CFD simulation to validate the proposed CFD models in Fluent code. The results showed good agreement for the peak cladding temperature (215 °C from the experiment, 212 °C from the CFD). It is regarded that the proposed scaling methodology was reasonably validated as maintaining the similarity of the temperature gradient and the peak cladding temperature.