基于Isomap的核爆地震模式识别

本文利用Isomap算法对核爆地震和天然地震的特征空间进行特征的降维,并用距离判别、Fisher判别和Logistic回归对降维后的结果进行了分类测试.实验结果表明:在核爆地震模式识别中,Isomap算法能在保持原数据绝大部分信息的前提下,很好地降低特征空间的维数,从而得到较高的正确识别率,并且优于PCA算法.     

基于高分辨率时频分析的核爆地震信号能量统计初至检测算法

核爆地震波检测是禁核试核查的关键环节之一。常用的短时长时能量均值比、赤池信息准则、傅里叶变换等方法虚警率都比较高,为此,提出一种基于高分辨率时频分析的核爆地震信号能量统计初至检测算法。其基本原理是根据核爆地震事件信号与噪声的频域分布差异,采用窄带分频技术将信号时域波形变换到时频域,在各频带计算局部能量均值与全局能量均值之比并进行二值化处理,当结果大于设定的阈值时,认为此频带该时刻属于地震事件信号的一部分,如果在某一阈值下出现二值化结果相同的情况,只需继续增大阈值进行检测,直到最终只有一个时间点检测到频段的数量最多即可,最大值的跳跃点即认为是初值时刻。利用地下核爆产生的地震信号进行实验研究,结果表明本文方法能够在提高检测准确率的同时解决高虚警问题,在低信噪比信号条件下具有较好的适用性与鲁棒性。     

基于序优化方法的特征选择研究

针对模式识别中特征子集的选择存在组合爆炸的情况．以核爆地震识别中的特征选择为例,将序优化中的BP(Blind Picking)规则进行用于特征选择研究。实验结果表明,此优化方法能迅速地减小搜索空间,提高搜索效率,从而为特征选择问题提供了一种新的处理方法。     

基于模糊C-均值的原型模式选择及其在核爆地震识别中的应用

协同模式识别是一种全新的有着抗噪声、抗缺损等诸多优良特性的模式识别方法,但将其用于核爆地震和天然地震的分类时,采用现有的原型模式选择方法识别效果并不理想.本文提出了一种基于模糊C-均值的原型模式选择方法,该方法首先对每一类训练样本采用模糊C-均值聚类的方法聚为c类,然后选取这c类的c个重心或c个聚类中心作为该类的原型模式进行核爆地震的协同模式识别.实验结果表明,同现有的原型模式选择方法相比,该方法使识别率有了较大提高.     

核爆地震的模式识别研究

首先简要介绍了核爆地震自动识别研究的历史和现状,重点介绍了笔者所在的研究集体的研究进展;然后．针对核爆地震自动识别研究的方法理论——模式识别中的几个关键问题进行了讨论,给出了初步研究结果,并提出了进一步研究设想。     

基于相匹配滤波的面波自动检测算法的实现

准确和可靠的面波测量对于监控全面禁止核试验条约(CTBT)的实施是十分重要的,相匹配滤波可有效检测和测量小当量核爆事件的面波信号,提高天然地震与核爆的识别能力。本文首先介绍了相匹配滤波的理论基础,然后设计了面波自动检测流程及方法,并将其应用到禁核试国家数据中心地震数据处理流水线中。测试结果表明:改进后的面波检测和测量程序提高了面波检测的精度,降低了小震级事件的检测阈值。     

电磁脉冲电场探头校准装置及不确定度评定

研制了一种核爆电磁脉冲电场探头校准标准装置,该标准装置利用TEM小室产生标准电场,可完成核爆电磁脉冲电场传感器频域及时域校准工作,为核爆电磁脉冲电场传感器的研制及性能考核提供了技术途径。文章还对频域及时域校准方法的不确定度进行了评定。     

核电厂地震风险量化方法

核电厂地震风险评价应该采用合适的数据处理方法和分析技巧。自主开发地震量化软件,采用蒙特卡洛抽样方法,对地震发生频率及设备失效条件概率进行模拟,并结合地震事故序列对电厂地震风险水平进行评价。该方法弥补了传统概率安全评价(PSA)建模软件在处理地震风险评价方面的不足。与国外同类型软件相比,在不确定性方法的处理上更合理,功能上更完善。     

地下核爆炸的自动识别研究

本文简要叙述了远区核爆探测的研究历史和现状，从研究的目的与任务，系统结构与研究思路，程序设计与实现，信号分析理论与模式分类器设计和模式识别结果分析等几方面介绍了“核爆地震模式识别系统”并对下一步研究工作提出了设想。     

基于相关性分析的特征选择方法研究

讨论了特征相关性的大小与分类能力之间的联系,提出了一种以相关系数和为判据的特征自动选择方法,并将它应用到地下核爆炸与天然地震的自动识别中,实验结果表明该方法所提取的特征子集具有很好的分类能力。     

Elucidation of Metallic Plume and Spatter Characteristics Based on SVM During High-Power Disk Laser Welding

During deep penetration laser welding,there exist plume(weak plasma) and spatters,which are the results of weld material ejection due to strong laser heating.The characteristics of plume and spatters are related to welding stability and quality.Characteristics of metallic plume and spatters were investigated during high-power disk laser bead-on-plate welding of Type 304 austenitic stainless steel plates at a continuous wave laser power of 10 kW.An ultraviolet and visible sensitive high-speed camera was used to capture the metallic plume and spatter images.Plume area,laser beam path through the plume,swing angle,distance between laser beam focus and plume image centroid,abscissa of plume centroid and spatter numbers are defined as eigenvalues,and the weld bead width was used as a characteristic parameter that reflected welding stability.Welding status was distinguished by SVM(support vector machine) after data normalization and characteristic analysis.Also,PCA(principal components analysis) feature extraction was used to reduce the dimensions of feature space,and PSO(particle swarm optimization) was used to optimize the parameters of SVM.Finally a classification model based on SVM was established to estimate the weld bead width and welding stability.Experimental results show that the established algorithm based on SVM could effectively distinguish the variation of weld bead width,thus providing an experimental example of monitoring high-power disk laser welding quality.     

Non-invasive on-line two-phase flow regime identification employing artificial neural networks

A novel non-invasive approach to the on-line identification of BWR two-phase flow regimes is investigated. The proposed approach receives neutron radiography images of coolant flow recordings as its input and performs feature extraction on each image via simple and directly computable statistical operators. The extracted features are subsequently used as inputs to an ensemble of self-organizing maps whose outputs demonstrate swift and accurate classification of each image into its corresponding flow regime. The novelty of the approach lies in the use of the self-organizing map which generates the different classes by itself, according to feature similarity of the corresponding images; this contrasts traditional artificial neural networks where the user has to define both the number of distinct classes as well as to supply separate training vectors for each class.     

基于PCA-GA-SVM的竖直下降两相流流型预测

为提高小样本条件下的流型识别精度和时效性,提出了一种融合小波包分解（WPD）、主元分析（PCA）、遗传算法（GA）和支持向量机（SVM）的优化识别模型,并成功应用在竖直下降两相流流型辨识工作中。利用WPD对非平稳电导波动信号进行分解、重构,提取小波包能量构造特征向量;通过PCA对特征向量进行降维,降低特征输入的复杂性;同时采取GA全局迭代寻优的方式确定SVM的关键参数惩罚因子（C）和核函数参数（g）。对PCA-GA-SVM识别效果进行验证后与SVM、PCA-SVM、GA-SVM网络进行对比。结果表明,经过PCA和GA优化后的SVM网络在流型识别精度和时效性方面均提升显著,对泡状流、弹状流、搅拌流和环状流的总体预测精度达到了94.87%,耗时仅3.95 s,可满足在线识别需求。      

基于KNN改进算法的闪烁体探测器故障诊断

为研究核探测器的可靠性,本文提出了一种基于K近邻（KNN）算法的闪烁体探测器故障诊断方法。首先通过提取不同工况下的核脉冲信号的下降沿时间、信号幅值及能谱信号的能峰位置和低道址计数等特征参数,建立故障核信号统计特征信息库。通过修正权重因子,改变邻点距离计算方式等方法改进KNN算法建立闪烁体探测器故障诊断模型,并搭建故障数据采集验证系统,提取探测器输出信号的特征信息放入到模型中进行诊断实验。实验结果表明,该方法不仅能实现对探测器故障类别的智能诊断,而且能对不同故障的严重程度做出良好的判别。     

Laser-induced breakdown spectroscopy for the classification of wood materials using machine learning methods combined with feature selection

In this paper, we explore whether a feature selection method can improve model performance by using some classical machine learning models, artificial neural network, k-nearest neighbor,partial least squares-discrimination analysis, random forest, and support vector machine(SVM),combined with the feature selection methods, distance correlation coefficient(DCC), important weight of linear discriminant analysis(IW-LDA), and Relief-F algorithms, to discriminate eight species of wood(African rosewood, Brazilian bubinga, elm, larch, Myanmar padauk,Pterocarpus erinaceus, poplar, and sycamore) based on the laser-induced breakdown spectroscopy(LIBS) technique. The spectral data are normalized by the maximum of line intensity and principal component analysis is applied to the exploratory data analysis. The feature spectral lines are selected out based on the important weight assessed by DCC, IW-LDA,and Relief-F. All models are built by using the different number of feature lines(sorted by their important weight) as input. The relationship between the number of feature lines and the correct classification rate(CCR) of the model is analyzed. The CCRs of all models are improved by using a suitable feature selection. The highest CCR achieves(98.55...0.39)% when the SVM model is established from 86 feature lines selected by the IW-LDA method. The result demonstrates that a suitable feature selection method can improve model recognition ability and reduce modeling time in the application of wood materials classification using LIBS.     

我国高放废液中铯分离研究进展

由于高放废液的放射性强、毒性大、组成复杂，从高放废液中分离铯是一个世界性难题。多年来国内外研究者一直在探索研究从高放废液中分离铯的方法，开发适合工业应用的铯分离技术，以解决从高放废液中分离铯的难题。一方面，我国现存的生产堆高放废液，浓缩倍数大、盐分高、放射性强，长期贮存风险大，需要进行妥善处理；另一方面，随着我国核电的快速发展和民用核燃料后处理的工业化，动力堆高放废液的处理问题也日益突出。针对这些需求，我国科技工作者们开展了大量从高放废液中分离铯的研究工作，取得了系列研究成果。近几十年来我国主要开展了离子交换、萃取色层和溶剂萃取分离高放废液中铯的研究，先后开发了亚铁氰化钛钾离子交换分离工艺以及杯芳烃冠醚萃取分离工艺，并进行了热实验验证以及台架实验。杯芳烃冠醚从高放废液中萃取分离铯的工作不但具备了工程应用的技术条件，也走在了世界前列。     

Fuzzy methodology applied to Probabilistic Safety Assessment for digital system in nuclear power plants

A fuzzy inference system (FIS) modeling technique to treat a nuclear reliability engineering problem is presented. Recently, many nuclear power plants (NPPs) have performed a shift in technology to digital systems due to analog obsolescence and digital advantages. The fuzzy inference engine uses these fuzzy IF-THEN rules to determine a mapping of the input universe of discourse over the output universe of discourse based on fuzzy logic principles. The risk priority number (RPN) (typical of a traditional failure mode and effects analysis - FMEA) is calculated and compared to fuzzy risk priority number (FRPN), obtained by the use of the scores from expert opinions. It was adopted the digital feedwater control system as a practical example in the case study. The results demonstrated the potential of the inference system to this class of problem.     

拌摩擦焊焊缝缺陷超声检测信号特征分析与神经网络模式识别

以搅拌摩擦焊（FSW）焊缝的包铝层伸入、未焊透、隧道孔缺陷为对象，将小波分析理论应用于缺陷超声检测信号特征提取问题的研究，使用小波包分解重构节点能量、小波包分解节点系数、缺陷信号的功率谱密度小波分解这三种方法对缺陷的超声检测信号进行特征提取。利用类别可分离性判据和BP神经网络分别对提取的特征量进行评估和识别。结果表明，缺陷信号的功率谱密度小波分解这一特征提取方式具有最好的类别可分性，并且以该特征量为网络输入的BP神经网络具有85.71%缺陷识别率。      

SSE在中子脉冲序列核信号实时频谱分析中的应用

为解决中子核信号频谱分析中快速傅立叶变换(FFT)高速计算的实时性关键技术问题,本论文针对用于频谱分析的中子脉冲序列信号本身所具有特殊"0、1"结构的特点,构造旋转因子表,采用SSE指令对算法优化设计,研究了一种仅需加法和寻址运算的FFT快速算法,实现了基于PC平台,在1 GHz采样率下中子脉冲序列信号的FFT实时频谱分析处理工作.性能测试结果表明,在中子脉冲计数率为3×106 s-1时,与相应的FFTW算法相比较,该算法所需的计算时间(效率)提高了517%.其中,SSE指令的优化设计可使算法的计算效率提升209%,满足了中子脉冲序列核信号频谱分析时对实时性的要求.     

Hybrid classification of coal and biomass by laser-induced breakdown spectroscopy combined with K-means and SVM

Laser-induced breakdown spectroscopy(LIBS) is a new technology suitable for classification of various materials. This paper proposes a hybrid classification scheme for coal, municipal sludge and biomass by using LIBS combined with K-means and support vector machine(SVM)algorithm. In the study, 10 samples were classified in 3 groups without supervision by K-means clustering, then a further supervised classification of 6 kinds of biomass samples by SVM was carried out. The results show that the comprehensive accuracy of the hybrid classification model is over 98%. In comparison with the single SVM classification model, the hybrid classification model can save 58.92% of operation time while guaranteeing the accuracy. The results demonstrate that the hybrid classification model is able to make an efficient, fast and accurate classification of coal, municipal sludge and biomass, furthermore, it is precise for the detection of various kinds of biomass fuel.