基于GAF-CNN的n/γ甄别方法研究

中子探测是核能开发领域的重要技术,由于中子闪烁体探测器往往会对中子和γ射线同时响应,因此有效分辨中子和γ射线是实现高精度中子探测的先决条件。为进一步提升n/γ甄别性能,本文结合脉冲形状甄别(PSD)技术和格拉姆角场(GAF)图像转换方法,将卷积神经网络(CNN)分类模型应用到n/γ甄别中。通过GAF将n/γ脉冲数据转化为二维图像,之后将其输入到CNN分类模型中达到样本辨别的目的。为验证GAF-CNN甄别的准确性,与传统CNN甄别法和电荷比较法进行了甄别效果对比。结果表明,GAF-CNN甄别法具有更低的辨别误差率和较短的处理时间,且n/γ甄别品质因子(FOM)有着数量级上的提升。同时其具备网络轻量化的特点,有助于实现CNN PSD算法的嵌入式部署,为研制高性能n/γ复合探测能谱仪提供了一种可行的PSD技术解决方案。

适用于级联型电力电子拓扑电磁暂态仿真的N端口网络通用等效建模方法

级联型电力电子设备复杂的拓扑结构给电磁暂态仿真带来巨大的计算压力。常见的模块化多电平换流器(modular multilevel converter,MMC)的等效建模方法相对成熟。然而,近年来出现大量复杂的新型级联型拓扑,目前仍缺乏通用的等效建模方法。为此,该文提出一种适用于任意拓扑N端口子模块、以任意级联方式组成的电力电子设备通用等效建模方法,利用状态空间表达式的形式将原始电路经数值积分代换后的端口特性提取出来,再根据得到的状态空间表达式反向构造等效电路模型。算例分析表明,所提方法不仅能消去子模块内部节点、提升仿真效率,还可自定义子模块拓扑结构和级联方式,相比传统等效建模方法更加灵活、通用,已有的MMC等效建模方法可以视为本文所提通用方法的一种特例。

基于N-BEATS的能源互联网短期负荷预测

短期负荷预测在能源互联网的规划中既占重要组成部分,又是能源系统可靠高效运行的基础。在能源互联网中能源的短期负荷预测精度问题是人们重点关注问题。N-BEATS的深度神经结构未使用时序特别组成成分,仅使用一种基于后向和前向残差链路以及非常深的全连接层堆栈的深度神经架构。该结构具有可解释性、适用于广泛的目标域、并且训练速度快等优点。实验使用N-BEATS模型对历史负荷数据进行训练,然后对未来负荷进行短期负荷预测,取得了较高的预测精度。测得平均绝对百分比误差(eMAPE)为1.26%,平均绝对误差(eMAE)为84.238 kW,决定系数(R^(2))为0.9955,实验结果表明采用该方法的预测精度高于传统的预测方法,如在eMAPE方面相比TCN降低了0.61%。

基于SSA_(n)-SSA_(l)-LSTM的短期空调负荷预测模型

本文提出了一种奇异谱分析(SSA_(n))和麻雀搜索算法(SSA_(l))优化的长短期记忆网络(LSTM)的组合空调负荷预测模型。使用皮尔逊相关系数和主成分分析法对输入特征进行挑选和处理,以消除特征之间的冗余性和相关性。针对空调负荷的波动性和随机性,采用SSA_(n)将空调负荷分解为多个分量。同时针对LSTM超参数设置的问题,采用SSA_(l)对模型进行优化,使用优化后的LSTM对各个分量进行预测,对预测结果进行重构。利用办公建筑和医疗建筑的空调负荷数据对模型进行了验证和分析。研究发现,与其他模型相比,SSA_(n)-SSA_(l)-LSTM模型表现最好,在预测办公建筑空调负荷时决定系数(R^(2))高达0.996 7,平均绝对百分比误差(MAPE)、平均绝对误差(MAE)和均方根误差(RMSE)分别为0.62%、14.42 kW和18.82 kW,在预测医疗建筑空调负荷时R^(2)高达0.992 7,MAPE、MAE和RMSE分别为0.50%、19.40 kW和25.71 kW。

含反馈电路的n阶电阻网络的新理论

研究了一类含有外接电阻的n阶电阻网络模型,给出了一个任意2节点间的等效电阻解析式,解决了一个电阻网络难题.文章研究所采用的方法主要是利用支路电流建立差分方程模型以及边界条件约束方程模型,然后求解差分方程组,由此导出等效电阻解析式.文章最后讨论了一些特殊情形的问题,并且将特殊情形的结论与其他相关结果进行了比较与验证.本文的研究工作对物理教育工作者开展实践创新研究具有比较好的示范作用.

含PNIPAM嵌段有机纳米管的合成及其对MB与ARS的吸附与释放性能探究

为了高效吸附染料废水中的染料分子,制备了一种具有温度可控性的新型吸附材料。首先,采用可控自由基聚合和开环聚合将聚N-异丙基丙烯酰胺(PNIPAM)引入多孔有机纳米管中,制备了具有核壳结构的多组分聚合物分子刷;其次,利用分子刷嵌段聚合物的傅克烷基化超交联过程生成了具有微孔、中孔和大孔结构的PNIPAM修饰的多级孔有机纳米管网络(MONNs)聚合物;最后,探究了MONNs对染料分子亚甲基蓝(MB)、茜素红S(ARS)的吸附与释放性能。结果表明,该方法制备的MONNs材料对染料分子具有良好的吸附效果。

基于N-K和SNA模型的物流仓储安全风险因素分析

为提高物流仓储的安全风险管理水平,降低事故发生率,收集2008—2022年的174起物流仓储安全事故,采用N-K模型进行风险因素耦合分析,借助社会网络模型实现风险因素关系可视化,并进行中心度和可达性分析。结果表明:多因素耦合导致物流仓储安全事故发生的概率最大;控制人—管双风险因素耦合可有效避免物流仓储安全事故发生;安全意识淡薄、安全投入不足、安全教育不到位是物流仓储安全事故的关键风险因素,需重点防范。

Biomass-Derived Nitrogen and Sulfur Co-Doped 3D Carbon Networks with Interconnected Meso-Microporous Structure for High-Performance Supercapacitors

Three-dimensional(3D)carbon networks have been explored as promising capacitive materials thanks to their unique structural features such as large ion-accessible surface area and interconnected porous networks,thus enhancing both ions and electrons transport.Here,sustainable bacterial cellulose(BC)is used both precursor and template for facile synthesis of free-standing N,S-codoped 3Dcarbon networks(a-NSC)by the pyrolysis and activation of polyrhodanine coated BC.The synthesized a-NSC shows highly conductive interconnected porous networks(24S·cm^(-1)),large surface area(1 420m^2·g^(-1))with hierarchical meso-microporosity,and high-level heteroatoms codoping(N:3.1%in atom,S:3.2%in atom).Benefitting from these,a-NSC as binder-free electrode exhibits an ultrahigh specific capacitance of 340F·g^(-1)(24μF·cm^(-2))at the current density of 0.5A·g^(-1)in 6MKOH electrolyte,high-rate capability(71%at 20A·g^(-1))and excellent cycle stability.Furthermore,the assembled symmetrical supercapacitor displays a much short time constant of 0.35sin 1MTEABF4/AN electrolyte,obtaining a maximum energy density of 32.1W·h·kg^(-1 )at power density of 637W·kg^(-1).The in situ multi-heteroatoms doping enables biocellulose-derived carbon networks to exploit its full potentials in energy storage applications,which can be extended to other dimensional carbon nanostructures.

Problem of Standardization of Configuration Data in Wireless Sensor Networks: The Case of Routing Protocols

The exploitation of Wireless Sensor Networks (WSN) is constrained by limited power, low computing power and storage and short-range radio transmission. Many routing protocols respecting these constraints were developed but, it still lacks formal and standardized solutions being able to help in their configuration. The configuration management that responds to this concern is very important in this type of network. It consists of the definition of data models to configure and is very necessary for the good network performance. Tangible results were obtained in traditional networks with the emergence of NETCONF and YANG standards, but on the best of our humble knowledge there are none yet in WSNs. We propose in this paper wsn-routing-protocol, a YANG data model for routing protocols configuration in WSNs. Following our model, we propose two YANG configuration data models based on the latter: they are respectively aodv for AODV and rpl for RPL.

Hierarchy property of traffic networks

The flourishing complex network theory has aroused increasing interest in studying the properties of real-world networks. Based on the traffic network of Chang-Zhu Tan urban agglomeration in central China, some basic network topological characteristics were computed with data collected from local traffic maps, which showed that the traffic networks were small-world networks with strong resilience against failure; more importantly, the investigations of as- sortativity coefficient and average nearestlneighbour degree implied the disassortativity of the traffic networks. Since traffic network hierarchy as an important basic property has been neither studied intensively nor proved quantitatively, the authors are inspired to analyse traffic network hierarchy with disassortativity and to finely characterize hierarchy in the traffic networks by using the n-degree-n-clustering coefficient relationship. Through numerical results and analyses an exciting conclusion is drawn that the traffic networks exhibit a significant hierarchy, that is, the traffic networks are proved to be hierarchically organized. The result provides important information and theoretical groundwork for optimal transport planning.

Research on Community Structure in Bus Transport Networks

We abstract the bus transport networks(BTNs)to two kinds of complex networks with space L and spaceP methods respectively.Using improved community detecting algorithm(PKM agglomerative algorithm),we analyzethe community property of two kinds of BTNs graphs.The results show that the BTNs graph described with space Lmethod have obvious community property,but the other kind of BTNs graph described with space P method have not.The reason is that the BTNs graph described with space P method have the intense overlapping community propertyand general community division algorithms can not identify this kind of community structure.To overcome this problem,we propose a novel community structure called N-depth community and present a corresponding community detectingalgorithm,which can detect overlapping community.Applying the novel community structure and detecting algorithmto a BTN evolution model described with space P,whose network property agrees well with real BTNs',we get obviouscommunity property.

Effects of Inter-Particle Frictional Coefficients on Evolution of Contact Networks in Landslide Process

During the process of landslide, its dynamic mechanism is important to understand and predict these kinds of natural hazard. In this paper, a new method, based on concepts of complex networks, has been proposed to investigate the evolution of contact networks in mesoscale during the sliding process of slope. A slope model was established using the discrete element method (DEM), and influences of inter-particle frictional coefficients with four different values on?dynamic landslides were studied. Both macroscopic analysis on slope?landslide?and mesoanalysis on structure evolution of contact networks, including the?average degree, clustering coefficient?and N-cycle, were done during the process?of landslide. The analysis results demonstrate that: 1) with increasing inter-particle?frictional coefficients, the displacement of slope decreases and the stable angle of slope post-failure increases, which is smaller than the peak internal frictional angle;2) the average degree decreases with the increase of inter-particle frictional coefficient. When the displacement at the toe of the slope is smaller,?the average degree there changes more greatly with increasing inter-particle?frictional coefficient;3) during the initial stage of landslide, the clustering coefficient?reduces sharply, which may leads to easily slide of slope. As the landslide?going?on, however, the clustering coefficient?increases denoting increasing stability?with?increasing inter-particle frictional coefficients. When the inter-particle?frictional coefficient is smaller than 0.3, its variation can affect the clustering coefficient?and stable inclination of slope post-failure greatly;and 4) the number of?3-cycle increases, but 4-cycle and 5-cycle decrease with increasing inter-particle frictional coefficients.

Spatial patterns nitrogen transfer models of ectomycorrhizal networks in a Mongolian scotch pine plantation

Ectomycorrhizal(EM)networks provide a variety of services to plants and ecosystems include nutrient uptake and transfer,seedling survival,internal cycling of nutrients,plant competition,and so on.To deeply their structure and function in ecosystems,we investigated the spatial patterns and nitrogen(N)transfer of EM networks usingN labelling technique in a Mongolian scotch pine(Pinus sylvestris var.mongolica Litv.)plantation in Northeastern China.In August 2011,four plots(20 × 20 m)were set up in the plantation.125 ml 5 at.%0.15 mol/LNHNOsolution was injected into soil at the center of each plot.Before and 2,6,30 and 215 days after theN application,needles(current year)of each pine were sampled along four 12 m sampling lines.Needle total N andN concentrations were analyzed.We observed needle N andN concentrations increased significantly over time afterN application,up to 31 and0.42%,respectively.There was no correlation between needle N concentration andN/N ratio(R2=0.40,n=5,P=0.156),while excess needle N concentration and excess needleN/N ratio were positively correlated across different time intervals(R~2=0.89,n=4,P\0.05),but deceased with time interval lengthening.NeedleN/N ratio increased with time,but it was not correlated with distance.NeedleN/N ratio was negative with distance before and 6th day and 30th day,positive with distance at 2nd day,but the trend was considerably weaker,their slop were close to zero.These results demonstrated that EM networks were ubiquitous and uniformly distributed in the Mongolian scotch pine plantation and a random network.We found N transfer efficiency was very high,absorbed N by EM network was transferred as wide as possible,we observed N uptake of plant had strong bias forN andN,namely N fractionation.Understanding the structure and function of EM networks in ecosystems may lead to a deeper understanding of ecological stability and evolution,and thus provide new theoretical approaches to improve conservation practices for the management of the Earth’s ecosystems.

Cavium Networks公司推出新型OCTEON^TMⅡ处理器

MIPS科技公司（MIPS Technologies,Inc）宣布,Cavium Networks公司推出的新型OCTEON^TMⅡ因特网应用处理器（IAP）中采用了MIPS科技的高性能MIPS64 架构。OCTEONⅡ系列采用多达32个MIPS64内核．是Cavium基于MIPS64架构的可扩展多核OCTEON处理器成功产品线的最新产品，并创下了在一个芯片中采用MIPS64处理器内核数目的新纪录。

融合N-K模型的复杂网络架桥机施工安全风险因素分析

为识别导致架桥机施工时发生安全事故的关键致因和链路,提出一种融合N-K模型的复杂网络架桥机施工安全事故风险因素分析法。首先,建立2000-2022年期间的架桥机事故案例库,分析事故致因,将架桥机施工安全事故风险因素归纳为4个一级风险因素和24个二级风险因素;然后,以事故类型和二级风险因素为节点,以致因关联为边,建立架桥机施工安全事故致因复杂网络模型,计算风险因素节点的亲近中心度和中介中心度;最后,建立N-K模型耦合分析一级风险因素,利用耦合交互信息值优化复杂网络风险因素节点的中心度,通过风险因素可达性分析和全局主路径分析,分别得到导致架桥机事故的7个关键致因和7条关键链路。结果表明:施工单位管理不到位、安全生产检查不充分2项管理因素的亲近中心度和中介中心度值较高。监理单位失职→施工单位管理不到位→安全生产检查不充分关键链路风险最大,针对这些关键风险因素和链路在架桥机施工过程中应重点管控。采用融合N-K模型的复杂网络分析得到的架桥机施工安全关键致因和链路与事故调查结论一致,准确实现了架桥机施工安全事故风险辨识。

高性能MIPS64架构推动Cavium Networks公司新型OCTEONⅡ处理器

MIPS科技公司宣布，Cevium Networks公司正式推出的新型OCTEONⅡ因特网应用处理器（IAP）中。采用了MIPS科技的高性能MIPS64架构。OCTEONⅡ系列采用多达32个MIPS64内核，是Cavium基于MIPS64架构的可扩展多核OCTEON处理器成功产品线的最新产品，并创下了在一个芯片中采用MIPS64处理器内核数目的新纪录。

基于N-BEATS神经网络的滚动轴承剩余寿命预测

针对轴承剩余使用寿命预测时采集到的信号序列复杂且不平稳,导致预测精度和性能较低的问题,采用经验模式分解对信号序列进行自适应分解,利用动态时间规整算法筛选主要退化特征,提取信号序列趋势特征,使用具有残差原理的深层神经网络N-BEATS进行预测。对于预测历史数据较少的问题,采用递归与直接相结合的预测结构对剩余寿命进行多步预测。将N-BEATS与长短时记忆神经网络、灰色预测模型进行比较,结果表明在不同工况下,所提出的方法预测结果的平均绝对误差相较于LSTM、灰色预测模型分别提升了3.2%以及3.3%,相对均方根误差分别提升了3.5%以及3.1%。

基于专利数据应用LDA和N-BEATS组合方法的技术主题预测研究

预测技术主题未来热点,有助于企业在技术层面判别现状、识别未来技术方向并提前规划战略布局。提出LDA和N-BEATS组合方法,运用LDA模型提取专利文献数据的技术主题,引入N-BEATS网络模型分析各技术主题专利数量的时间序列,发挥其分析可解释性时间序列的优势,在预测模型中加入技术研发活动周期性模块,并以芯片技术为例,运用该组合方法预测产业的技术主题和未来趋势。对比实验中LDA和N-BEATS组合方法的预测精度高于LDA-LSTM、IPC-N-BEATS和IPC-LSTM三种基准方法。案例结果表明,未来芯片产业研发热点是电子级树脂、蚀刻机、芯片封装、芯片键合、抛光液。

k元(n-1)方体子网络可靠性的近似评估方法

多处理器系统互连网络的拓扑性质对系统功能的实现起着重要的作用。k元n方体网络的子网络可靠性是以k元n方体为拓扑结构构建的多处理器系统处理计算任务时需要考虑的一个重要因素。为了精确高效地度量概率故障条件下k元n方体中k元(n-1)方体子网络的可靠性,提出基于反向传播(BP)神经网络的k元(n-1)方体子网络可靠性的近似评估方法。首先,利用蒙特卡洛仿真方法和k元(n-1)方体子网络可靠性的已有上下界给出用于训练BP神经网络的数据集的生成方法;其次,基于生成的训练数据集构造用于评估k元(n-1)方体子网络可靠性的BP神经网络模型;最后,对BP神经网络模型得出的k元(n-1)方体子网络可靠性的近似评估结果进行了分析,并与近似计算公式和基于蒙特卡洛的评估方法的结果进行了对比。与近似计算公式相比,所提方法得出的结果更为精确;与基于蒙特卡洛的评估方法相比,所提方法的评估耗时平均减少了约59%。实验结果表明,所提方法在兼顾精度和效率方面具有一定优势。