课程文献中心 更多>>

关键词： Flow classification   Software-defined data center network   Deep learning   Deep residual learning   Gated recurrent unit  

摘要： in a data center network, accurately classifying flow is the key to optimal schedule flow. However, the existing classification methods cannot meet the demand of real networks in terms of classification performance, detection latency, and control overhead. Thus, by combining the ability of deep learning to describe the multi-dimensional features, and the advantage of software-defined networking (SDN) in centrally controlling the network from a global viewpoint, this paper proposes a fine-grained flow classification method. This paper uses random forest technology to select eight important features in three dimensions-time distribution feature of flow, real-time feature of flow and packet header feature-for the classification model. First, this paper proposes a 2-classification scheme with a two-level architecture of pre-classification and exact-classification to detect elephant/mice flows. The pre-classification model using deep residual learning + A-Softmax with cost-sensitive is deployed on a SDN switch at the network edge to filter out a large number of mice flows. The exact-classification model using deep residual learning + AM-Softmax is deployed on the SDN controller to accurately identify the elephant flows. Second, this paper proposes a 4-classification scheme based on gated recurrent unit (GRU) to detect elephant/cheetah/tortoise/porcupine flows. Finally, the experiment results show that, when the 5th packet of a flow arrives, the 2-classification scheme can achieve a recall of up to 97.31%, an accuracy of up to 93.6%, a control overhead of 0.1kbps, and a detection latency of 7 ms. At the same time, the 4-classification scheme can achieve a recall, an accuracy, a false positive rate (FPR), a Kappa, a control overhead, and a detection latency of 83.58%, 86.53%, 5.02%, 0.797, 0.61kbps and 7.5 ms on an average, respectively. Compared with the existing methods (FlowSeer, NELLY and ESCA), all performance measures are improved to different degrees. At the same ti

关键词： Dual fluidized bed   Gasification   Biomass   Segregation   Numerical simulation   Renewable energy  

摘要： In this work, computational fluid dynamics is adopted to simulate steam-biomass gasification in a dual fluidized bed reactor system. Under the Eulerian-Lagrangian multiphase framework, the large eddy simulation and multiphase particle-in-cell models are selected to track the gas and solid phases, respectively. The compositions of the gaseous products numerically obtained at the gasifier outlet are compared with the experimental data for model validation. Then, the radial and axial segregation of heat carrier with a wide particle size distribution together with the spatial variation of the microscale properties of solid phase are explored. The results show that the apparent size-induced segregation of heat carrier results in the distribution of large particles in the bottom region of both the combustor and gasifier, while the accumulation of fine particles in the freeboard of the gasifier. The mass of heat carrier distributed in the combustor is nearly 10% of that in the gasifier. The particles in the combustor have larger Reynolds number and dispersion coefficient as compared with that in the gasifier. For a specific particle size distribution, reducing the particle size enlarges the temperature, heat transfer coefficient, particle Reynolds number. Different solid residence behaviors appear in different components of the bed. A wider size distribution of heat carrier slightly enlarges the mean residence time in the combustor but obviously diminishes it in the gasifier, and also evidently reduces the solid circulation. (C) 2020 Elsevier Ltd. All rights reserved.

关键词： 柴达木盆地西北缘   上油砂山组   砂岩型铀矿   酸解烃  

摘要： 通过对柴达木盆地西北缘新近系上油砂山组18个样品的酸解烃分析,探讨了酸解烃中烃类气体组成特征、成因及来源,并结合该区铀矿地质特征初步分析了烃类流体与砂岩型铀成矿关系。酸解烃的分析结果表明,新近系上油砂山组砂岩层中烃类气体为有机成因的油型气和煤型气的混合气。烃类气主要处于成熟—高成熟阶段,是原油裂解气与原油伴生气的混合气,主要来源于路乐河组和下干柴沟组湖相腐泥型烃源岩。这些原油裂解气和原油伴生气的混合气主要沿断裂、微裂缝等通道向上运移到上油砂山组砂岩中,上油砂山组灰绿色砂岩即是烃类气体对红色砂岩还原的产物。研究区砂岩型铀矿成矿作用为原生红色—后生还原灰绿色的氧化还原作用过程,找矿标志为灰绿色含砾砂岩、砂砾岩,灰绿色砂岩层的古迎水面为该地区的找矿方向。

关键词： 页岩储层   扩散流动   扩散模型   动量方程   流动实验   气体扩散表征  

摘要： 为揭示页岩微纳米孔隙中气体的扩散机理,针对页岩储层气体扩散能力难以定量表征的问题,基于考虑流体黏性的微分形式动量方程,建立了考虑页岩孔隙度、迂曲度和流动Kn数(气体分子平均自由程与流动特征尺度之比)的气体扩散系数新模型,将模型与自主研发的页岩近平衡态实验进行验证,进而形成了页岩气扩散系数影响因素图版。研究表明:考虑了页岩孔隙度、迂曲度等多孔介质参数和流动Kn数,新建立的扩散系数更能准确表征页岩气扩散能力,与近平衡扩散流量吻合度达90%以上;扩散系数与压力负相关,而与孔隙直径正相关,在压力低于20 MPa、孔隙直径低于10 nm或Kn数高于0.2后必须考虑上述参数变化对扩散系数的影响。该研究实现了储层条件下页岩气扩散流量的定量计算,可用于页岩表观渗透率模型的建立,为不同生产阶段页岩气扩散对产量的贡献以及调整生产制度、提高单井产量提供科学依据。

关键词： 干旱型分支河流体系   沉积特征   演化过程   沉积构型   水槽沉积模拟实验  

摘要： 在干旱气候背景下,盆地边缘可形成一种发育于盆地边缘并顺源延伸至湖盆边缘终止的大型沉积体系,即分支河流体系,其主体往往被称为冲积扇。分支河流体系规模较大,沉积过程复杂,其沉积物可形成大规模的油气、水储层。为了明确干旱型分支河流体系沉积演化过程及其不同时期沉积特征的差异,通过水槽沉积实验模拟再现其沉积过程,并采用三维激光扫描、延时拍摄等技术手段精确监测。研究结果表明:1)分支河流体系的半径、边缘糙度与模拟期次(沉积物供给)呈对数相关关系,同时其增生规律呈现明显的阶段性差异,据此可将沉积演化过程划分为早期、中期及后期三个阶段。2)各阶段沉积特征存在较大差异,初期阶段以片流为主,片流单元在分支河流体系表面呈连片状多期次迁移叠置,中期阶段非限制性河道的侧向迁移及与其远端相连的限制性河道的形成与废弃共同主导了沉积过程,后期阶段非限制性河道发育程度大大降低而限制性河道发育程度大大增加并主导沉积过程。3)分支河流体系的三阶段演化导致其本身由三个沉积特征差异较大、空间结构迥异的沉积层构成,包括底层片流朵体复合体、中层非限制性河道与限制性河道切割叠覆体、顶层限制性河道切叠沉积体。

关键词： 助剂   吡虫啉   阿维菌素   苦参碱   表面张力   接触角  

摘要： 为达到农药减施增效的目的,将6种喷雾助剂(强力源、柔水通、有机硅、阿普顿、怀农特、倍创)与3种药剂(70%吡虫啉水分散粒剂、5%阿维菌素乳油、1.3%苦参碱水剂)混配,测定混配后药液的表面张力和接触角变化,分析药液表面张力和接触角的关系。结果表明,添加不同质量浓度的6种喷雾助剂均能降低药液的表面张力,其中,有机硅、强力源、怀农特降低3种药剂表面张力的效果最好,柔水通、阿普顿次之,倍创效果最差。有机硅、柔水通、强力源、怀农特降低3种药剂接触角的效果最好,阿普顿次之,倍创效果最差。分析了3种药剂在添加不同质量浓度的6种喷雾助剂后药液的表面张力与接触角余弦值的关系,结果显示,二者呈显著(P<0.05)负相关关系。综上,合理选用喷雾助剂有助于提高农药的利用率。研究结果可为田间施药和进一步研发喷雾助剂提供数据支持。

关键词： 旋涡提取方法   特征可视化   参考系不变性   流体特征   涡核  

摘要： 近年来,流体可视化已成为计算机图形学领域的一个研究热点,其最重要的目的之一是旋涡特征的提取与可视化。由于目前仍未有一个通用的定义描述旋涡,导致文献对旋涡是否存在的判断依据各不相同。为了对流体的旋涡特征提取方法进行较为系统的综述,首先对旋涡提取研究方向的相关概念进行解释,回顾流体旋涡特征提取方法的发展情况再进行总结,将常用的旋涡提取方法分为基于点、线、几何和基于机器学习的方法。对于新近提出的参考系不变性,将旋涡提取方法分为伽利略不变性、旋转不变性和拉格朗日不变性。为了比较不同方法的优势和缺陷,在综述每一类方法时分别给出若干经典方法,为研究者提供了一个清晰的研究思路。最后总结每类方法存在的难点和问题,并指出今后的研究重点。

摘要： Using the Structure of Observed Learning Outcomes (SOLO) taxonomy as the analytic framework, this study examined the impact of the reasoning flow scaffold (RFS) on students' written arguments. Two classes with a total of 88 10th grade students in a school participated in this study. One class, set as the experimental group, was taught scientific argumentation with RFS whereas the control class received conventional argumentation teaching. They all experienced three argument assignments of writing scientific arguments and the measurement task before and after the teaching intervention. The results of data analysis showed that after teaching intervention, students in the experimental group performed significantly better than those in the control group on evidence and rebuttal while there were no significant differences on claim or reason between the two groups. Some implications and suggestions are provided in the last part of this paper.