Flow patterns past two circular cylinders in proximity
多相流基础 chapter 2-two-phase flow patterns and flow pattern maps
Tip:the types of flow patterns increase in heating channel when heat flux increases.
6、Flow patterns in horizontal tubes, heating up flow(水平管内两相流型,加热流 )
(1)Bubbly flow(泡状流)
Liquid phase is the continuous phase Gas or vapor phase is the dispersed phase Gas or vapor bubbles are of approximately uniform size.
5、Flow patterns in vertical upward tubes ——gas and liquid phase both flowing upwards,heating up flow(垂直上升管内的流型 —— 气液两相均向上流 ,加热流)
Heat addition Phase change More complicated than adiabatic flow The evolvement of flow patterns are greatly affected by the heat flux(热流密度)
(4)Annular flow(环状流)
Gas flows in the center of the tube
Liquid partially flows as an annular film on the walls of the tube Liquid partially flows as small droplets distributed in the gas flowing in the center of the tube
帕金森转棒实验的英文
帕金森转棒实验的英文The Parkinson's Turning Stick ExperimentParkinson's disease is a neurodegenerative disorder that affects the central nervous system. It primarily affects movement, causing tremors, stiffness, and difficulty with balance and coordination. One of the challenges faced by individuals with Parkinson's disease is turning while walking. This experiment aims to investigate the effectiveness of a specialized turning stick in improving turning abilities for Parkinson's patients.The experiment will involve a group of participants diagnosed with Parkinson's disease. They will be divided into two groups: an experimental group and a control group. The experimental group will use the turning stick, while the control group will not. The participants will undergo a series of assessments and measurements before and after the intervention.The turning stick is designed to provide support and stability during turning movements. It has a unique handle that allows for a secure grip and a rotating mechanism thatassists in turning. The stick will be adjusted to each participant's height and handedness for optimal comfort and ease of use.The assessments will include various tests to evaluate the participants' turning abilities, such as the Timed Up and Go test, which measures the time taken to stand up from a chair, walk a short distance, turn around, and sit back down. Other tests may include the Functional Gait Assessment and the Freezing of Gait questionnaire, which assesses freezing episodes during walking.The participants in the experimental group will receive training on how to use the turning stick effectively. They will practice walking and turning with the stick under the supervision of a physiotherapist. The control group will not receive any specific intervention or training related to turning.After a designated period, the participants will undergo the same assessments and measurements as before the intervention. The data collected will be analyzed to determine if there are any significant differences betweenthe experimental and control groups in terms of turning abilities.It is hypothesized that the participants in the experimental group, who use the turning stick, will show improvements in turning abilities compared to the control group. The specialized handle and rotating mechanism of the turning stick are expected to provide additional support and stability during turning movements, thereby reducing the risk of falls and improving overall mobility.If the results of the experiment support the hypothesis, it could lead to the development of more effective interventions and assistive devices for individuals with Parkinson's disease. The turning stick could potentially be incorporated into rehabilitation programs and daily routines to enhance mobility and quality of life for Parkinson's patients.帕金森转棒实验帕金森病是一种神经退行性疾病,影响中枢神经系统。
浙江大学基础医学导论题库各论I题库
各论I题库第一篇运动系统 (2)第一章骨与骨的连接 (2)第二章骨骼肌解剖结构 (29)第三章骨骼肌生理 (45)第二篇血液与造血系统 (58)第一章血液的形态学与血细胞的发生 (58)第二章血液生理62 第三章淋巴和造血系统疾病 (77)第三篇循环系统 (88)第一章循环系统解剖 (88)第二章循环系统的组织结构与发生93 第三章血液循环生理 (108)第四章循环系统病理153 第五章循环系统功能障碍156 第六章心血管系统药理 (178)鸣谢:初菁菁、郑佳佳、董军乐、胡慧美、王靖、王懿、范欣、秦祯蔚、彭德清、王灿锋、姜治伟等人的辛苦整理~医学院0707&07032009.3.10第一篇运动系统第一章骨与骨的连接第一节骨学总论一、骨的分类1. 下列哪项不是青枝骨折的特征A. 多发生在儿童B. 是一种不完全骨折C. 无明显功能障碍D. 无局部压痛及纵形叩击痛E. 畸形不严重2. 骨损伤后能参与修复的结构是A.骨质B.骨骺C.骨膜D.骨髓E.关节面的软骨3. 黄骨髓存在于A. 所有骨的内部B.幼儿长骨骨干内部C.成人长骨骨干内部D. 幼儿长骨骨骺内部E. 成人扁骨内部4. 有关骨髓腔正确的是A. 位于骨骺内B. 位于长骨的骨干内C. 成人骨髓腔内含红骨髓D. 小儿骨髓腔内含黄骨髓E. 以上全不对5. 有关红骨髓正确的是A. 成人存在于髓腔内B. 不存在于板障内C. 胎儿期造血,成年期不造血D. 髂骨、胸骨、椎骨内终生保存红骨髓E. 以上全不对6. 有关骨的构造,正确的说法是A. 骨干由松质构成B. 骨骺由密质构成C. 骨膜有血管无神经D. 骨髓有神经无血管E. 以上全不对答案:DCCBDE二、骨的构造(共6题)1. 下列哪项不是青枝骨折的特征A. 多发生在儿童B. 是一种不完全骨折C. 无明显功能障碍D. 无局部压痛及纵形叩击痛E. 畸形不严重2. 骨损伤后能参与修复的结构是A.骨质B.骨骺C.骨膜D.骨髓E.关节面的软骨3. 黄骨髓存在于A. 所有骨的内部B.幼儿长骨骨干内部C.成人长骨骨干内部D. 幼儿长骨骨骺内部E. 成人扁骨内部4. 有关骨髓腔正确的是A. 位于骨骺内B. 位于长骨的骨干内C. 成人骨髓腔内含红骨髓D. 小儿骨髓腔内含黄骨髓E. 以上全不对5. 有关红骨髓正确的是A. 成人存在于髓腔内B. 不存在于板障内C. 胎儿期造血,成年期不造血D. 髂骨、胸骨、椎骨内终生保存红骨髓E. 以上全不对6. 有关骨的构造,正确的说法是A. 骨干由松质构成B. 骨骺由密质构成C. 骨膜有血管无神经D. 骨髓有神经无血管E. 以上全不对答案:DCCBDE三、骨的化学成分和物理性质及年龄差异无第二节中枢骨一、躯干骨1. 男性,35岁,头痛、恶心、呕吐、发热,诊断为脑脊髓膜炎,要行腰椎穿刺抽取脑脊液化验,其进针部位应在A.第12胸椎与第1腰椎棘突间隙B.第1腰椎与第2腰椎棘突间隙C.第3腰椎与第4腰椎棘突间隙D.第5腰椎与骶椎间隙E.第2对骶后孔处2. 关于椎弓的说法,错误的是A.位于椎体的后方B.呈半环状C.与椎体共同围成椎孔D.由一对椎弓板构成E.相邻椎弓间有黄韧带3. 关于椎体的说法,错误的是A.呈短圆柱状B.位于椎骨前部C.主要由松质构成D.表面有一层密质E.椎体中央有椎孔4. 躯干骨在体表易摸到的骨性标志是A.全部颈椎棘突B.全部肋骨C.胸椎横突D.胸骨角E.骶岬5. 骶骨的正确描述是A.有5对骶前孔B.由4块骶椎融合而成C.与第4腰椎相关节D.骶管内有脊髓通过E.于骶角处可寻骶管裂孔进行神经阻滞麻醉6. 骶管麻醉的穿刺部位正对A.骶角B.骶管裂孔C.骶前孔D.骶后孔E.骶岬7. 关于骶骨正确的是A. 由3块骶椎组成B. 骶前孔、骶后孔均与骶管相通C. 骶骨的耳状面与髋骨的耳状面相连,属直接连接D. 骶管与椎管不通E.女性骶岬较男性突出8. 关于肋骨正确的是A. 为长骨B. 分体和前、后两端C. 肋头与胸椎横突相关节D. 肋沟位于各肋内面上缘E.以上都不对9. 关于肋正确的是A. 上6对肋直接与胸骨相连,故称真肋B. 呈长条形,属长骨C. 第8~12肋组成肋弓D. 肋结节与椎体肋凹相关节E. 第2肋平对胸骨角10. 关于腰椎的正确说法是A. 椎体粗壮,横切面呈三角形B. 椎孔呈圆形C. 各棘突的间隙较宽D. 上关节面是冠状位E. 上关节面呈水平位11. 颈椎特有的结构是A. 椎孔呈三角形B. 椎弓C. 关节突D. 横突E. 横突孔12. 关于胸椎的特点是A. 横突上有横突孔B. 棘突分叉C. 上、下关节突不明显D. 棘突水平伸向后方E. 椎体侧面后部有肋凹13. 椎骨正确的是A. 所有颈椎棘突分叉B. 第6颈椎称隆椎C. 腰椎关节呈冠状位D. 腰椎棘突宽而短呈板状E. 胸椎椎体都有一完整肋凹14. 椎骨正确的是A. 是短骨B. 椎体之间有椎间关节相连C. 相邻椎弓之间构成椎间孔D. 椎体与椎弓共同围成椎孔E. 以上全不对15. 关于颈椎,正确的是A. 均有椎体及椎弓B. 1 ~2颈椎无横突孔C. 棘突末端都分叉D. 第6颈椎棘突末端膨大成颈动脉结节E. 第7颈椎又名隆椎答案:CDEDE BBBEC EEDDE1. 一年轻人被车撞倒,头部摔伤造成颅底骨折及脑膜和鼻旁窦损伤,有血液和脑脊液从鼻腔流出,此情况提示最可能损伤哪个鼻旁窦A. 额窦B. 上颌窦C. 额窦和上颌窦D. 蝶窦E. 筛窦2. 某建筑工人自高处跌下引起颅底骨折,患侧额纹消失,眼周围肌和口周围肌瘫痪,口角偏向对侧,舌运动灵活,患侧舌前2/3味觉消失,听觉过敏,无眼角干燥的现象,其他感觉正常。
Boundary-Layer Separation
拇指血流灌注指数试验与改良Allen试验的比较
硕士学位论文论 文 题 目: 拇指血流灌注指数试验与改良Allen试验的比较Evaluation of the patency of the hand collateralarteries with thumb Perfusion Index test:Comparison with the modified Allen’s test研 究 生 姓 名: 吴阳指导教师: 刘松学科专业: 麻醉学研究方向: 麻醉学临床技能训练与研究论文工作时间: 2015年6月至2016年12月目录中文摘要 (1)英文摘要 (2)正 文 (3)前 言 (3)资料与方法 (7)结 果 (10)讨 论 (15)结 论 (22)参考文献 (23)致 谢 (33)附录A (34)附录B (44)拇指血流灌注指数试验与改良Allen试验的比较中文摘要目的:探讨拇指血流灌注指数(Perfusion Index,PI)试验替代改良Allen试验(modified Allen's test,MAT)评价掌部组织侧支循环血流灌注的可行性。
方法:选择1108例拟行择期手术并需要经桡动脉入路进行有创动脉压力监测的患者,在桡动脉穿刺前先后用MAT和拇指PI值试验分别评价患者试验侧掌部组织侧支循环血流灌注的情况,并将两种试验方法结果进行统计学比较和分析。
结果:在1108例患者中MAT阴性患者1035例(93.41%),阳性患者73例(6.59%);拇指PI值试验阴性患者1090例(98.38%),其中包括57例MAT阳性患者,阳性患者18例(1.62%)。
拇指PI值试验阴性患者行经该侧桡动脉入路进行有创动脉压力监测,两种试验方法结果进行卡方检验,差异有统计学意义(x2=51.27, P<0.05)。
两种试验方法影响因素进行logistic回归分析发现两种试验方法结果阳性率均与年龄和性别有相关性(P<0.05)。
结论:在本研究中用拇指PI值试验筛选出1.62%的患者不宜行经桡动脉入路进行有创动脉压力监测。
毕业设计论文塑料注射成型
Modeling of morphology evolution in the injection moldingprocess of thermoplastic polymersR.Pantani,I.Coccorullo,V.Speranza,G.Titomanlio* Department of Chemical and Food Engineering,University of Salerno,via Ponte don Melillo,I-84084Fisciano(Salerno),Italy Received13May2005;received in revised form30August2005;accepted12September2005AbstractA thorough analysis of the effect of operative conditions of injection molding process on the morphology distribution inside the obtained moldings is performed,with particular reference to semi-crystalline polymers.The paper is divided into two parts:in the first part,the state of the art on the subject is outlined and discussed;in the second part,an example of the characterization required for a satisfactorily understanding and description of the phenomena is presented,starting from material characterization,passing through the monitoring of the process cycle and arriving to a deep analysis of morphology distribution inside the moldings.In particular,fully characterized injection molding tests are presented using an isotactic polypropylene,previously carefully characterized as far as most of properties of interest.The effects of both injectionflow rate and mold temperature are analyzed.The resulting moldings morphology(in terms of distribution of crystallinity degree,molecular orientation and crystals structure and dimensions)are analyzed by adopting different experimental techniques(optical,electronic and atomic force microscopy,IR and WAXS analysis).Final morphological characteristics of the samples are compared with the predictions of a simulation code developed at University of Salerno for the simulation of the injection molding process.q2005Elsevier Ltd.All rights reserved.Keywords:Injection molding;Crystallization kinetics;Morphology;Modeling;Isotactic polypropyleneContents1.Introduction (1186)1.1.Morphology distribution in injection molded iPP parts:state of the art (1189)1.1.1.Modeling of the injection molding process (1190)1.1.2.Modeling of the crystallization kinetics (1190)1.1.3.Modeling of the morphology evolution (1191)1.1.4.Modeling of the effect of crystallinity on rheology (1192)1.1.5.Modeling of the molecular orientation (1193)1.1.6.Modeling of theflow-induced crystallization (1195)ments on the state of the art (1197)2.Material and characterization (1198)2.1.PVT description (1198)*Corresponding author.Tel.:C39089964152;fax:C39089964057.E-mail address:gtitomanlio@unisa.it(G.Titomanlio).2.2.Quiescent crystallization kinetics (1198)2.3.Viscosity (1199)2.4.Viscoelastic behavior (1200)3.Injection molding tests and analysis of the moldings (1200)3.1.Injection molding tests and sample preparation (1200)3.2.Microscopy (1202)3.2.1.Optical microscopy (1202)3.2.2.SEM and AFM analysis (1202)3.3.Distribution of crystallinity (1202)3.3.1.IR analysis (1202)3.3.2.X-ray analysis (1203)3.4.Distribution of molecular orientation (1203)4.Analysis of experimental results (1203)4.1.Injection molding tests (1203)4.2.Morphology distribution along thickness direction (1204)4.2.1.Optical microscopy (1204)4.2.2.SEM and AFM analysis (1204)4.3.Morphology distribution alongflow direction (1208)4.4.Distribution of crystallinity (1210)4.4.1.Distribution of crystallinity along thickness direction (1210)4.4.2.Crystallinity distribution alongflow direction (1212)4.5.Distribution of molecular orientation (1212)4.5.1.Orientation along thickness direction (1212)4.5.2.Orientation alongflow direction (1213)4.5.3.Direction of orientation (1214)5.Simulation (1214)5.1.Pressure curves (1215)5.2.Morphology distribution (1215)5.3.Molecular orientation (1216)5.3.1.Molecular orientation distribution along thickness direction (1216)5.3.2.Molecular orientation distribution alongflow direction (1216)5.3.3.Direction of orientation (1217)5.4.Crystallinity distribution (1217)6.Conclusions (1217)References (1219)1.IntroductionInjection molding is one of the most widely employed methods for manufacturing polymeric products.Three main steps are recognized in the molding:filling,packing/holding and cooling.During thefilling stage,a hot polymer melt rapidlyfills a cold mold reproducing a cavity of the desired product shape. During the packing/holding stage,the pressure is raised and extra material is forced into the mold to compensate for the effects that both temperature decrease and crystallinity development determine on density during solidification.The cooling stage starts at the solidification of a thin section at cavity entrance (gate),starting from that instant no more material can enter or exit from the mold impression and holding pressure can be released.When the solid layer on the mold surface reaches a thickness sufficient to assure required rigidity,the product is ejected from the mold.Due to the thermomechanical history experienced by the polymer during processing,macromolecules in injection-molded objects present a local order.This order is referred to as‘morphology’which literally means‘the study of the form’where form stands for the shape and arrangement of parts of the object.When referred to polymers,the word morphology is adopted to indicate:–crystallinity,which is the relative volume occupied by each of the crystalline phases,including mesophases;–dimensions,shape,distribution and orientation of the crystallites;–orientation of amorphous phase.R.Pantani et al./Prog.Polym.Sci.30(2005)1185–1222 1186R.Pantani et al./Prog.Polym.Sci.30(2005)1185–12221187Apart from the scientific interest in understandingthe mechanisms leading to different order levels inside a polymer,the great technological importance of morphology relies on the fact that polymer character-istics (above all mechanical,but also optical,electrical,transport and chemical)are to a great extent affected by morphology.For instance,crystallinity has a pro-nounced effect on the mechanical properties of the bulk material since crystals are generally stiffer than amorphous material,and also orientation induces anisotropy and other changes in mechanical properties.In this work,a thorough analysis of the effect of injection molding operative conditions on morphology distribution in moldings with particular reference to crystalline materials is performed.The aim of the paper is twofold:first,to outline the state of the art on the subject;second,to present an example of the characterization required for asatisfactorilyR.Pantani et al./Prog.Polym.Sci.30(2005)1185–12221188understanding and description of the phenomena, starting from material description,passing through the monitoring of the process cycle and arriving to a deep analysis of morphology distribution inside the mold-ings.To these purposes,fully characterized injection molding tests were performed using an isotactic polypropylene,previously carefully characterized as far as most of properties of interest,in particular quiescent nucleation density,spherulitic growth rate and rheological properties(viscosity and relaxation time)were determined.The resulting moldings mor-phology(in terms of distribution of crystallinity degree, molecular orientation and crystals structure and dimensions)was analyzed by adopting different experimental techniques(optical,electronic and atomic force microscopy,IR and WAXS analysis).Final morphological characteristics of the samples were compared with the predictions of a simulation code developed at University of Salerno for the simulation of the injection molding process.The effects of both injectionflow rate and mold temperature were analyzed.1.1.Morphology distribution in injection molded iPP parts:state of the artFrom many experimental observations,it is shown that a highly oriented lamellar crystallite microstructure, usually referred to as‘skin layer’forms close to the surface of injection molded articles of semi-crystalline polymers.Far from the wall,the melt is allowed to crystallize three dimensionally to form spherulitic structures.Relative dimensions and morphology of both skin and core layers are dependent on local thermo-mechanical history,which is characterized on the surface by high stress levels,decreasing to very small values toward the core region.As a result,the skin and the core reveal distinct characteristics across the thickness and also along theflow path[1].Structural and morphological characterization of the injection molded polypropylene has attracted the interest of researchers in the past three decades.In the early seventies,Kantz et al.[2]studied the morphology of injection molded iPP tensile bars by using optical microscopy and X-ray diffraction.The microscopic results revealed the presence of three distinct crystalline zones on the cross-section:a highly oriented non-spherulitic skin;a shear zone with molecular chains oriented essentially parallel to the injection direction;a spherulitic core with essentially no preferred orientation.The X-ray diffraction studies indicated that the skin layer contains biaxially oriented crystallites due to the biaxial extensionalflow at theflow front.A similar multilayered morphology was also reported by Menges et al.[3].Later on,Fujiyama et al.[4] investigated the skin–core morphology of injection molded iPP samples using X-ray Small and Wide Angle Scattering techniques,and suggested that the shear region contains shish–kebab structures.The same shish–kebab structure was observed by Wenig and Herzog in the shear region of their molded samples[5].A similar investigation was conducted by Titomanlio and co-workers[6],who analyzed the morphology distribution in injection moldings of iPP. They observed a skin–core morphology distribution with an isotropic spherulitic core,a skin layer characterized by afine crystalline structure and an intermediate layer appearing as a dark band in crossed polarized light,this layer being characterized by high crystallinity.Kalay and Bevis[7]pointed out that,although iPP crystallizes essentially in the a-form,a small amount of b-form can be found in the skin layer and in the shear region.The amount of b-form was found to increase by effect of high shear rates[8].A wide analysis on the effect of processing conditions on the morphology of injection molded iPP was conducted by Viana et al.[9]and,more recently, by Mendoza et al.[10].In particular,Mendoza et al. report that the highest level of crystallinity orientation is found inside the shear zone and that a high level of orientation was also found in the skin layer,with an orientation angle tilted toward the core.It is rather difficult to theoretically establish the relationship between the observed microstructure and processing conditions.Indeed,a model of the injection molding process able to predict morphology distribution in thefinal samples is not yet available,even if it would be of enormous strategic importance.This is mainly because a complete understanding of crystallization kinetics in processing conditions(high cooling rates and pressures,strong and complexflowfields)has not yet been reached.In this section,the most relevant aspects for process modeling and morphology development are identified. In particular,a successful path leading to a reliable description of morphology evolution during polymer processing should necessarily pass through:–a good description of morphology evolution under quiescent conditions(accounting all competing crystallization processes),including the range of cooling rates characteristic of processing operations (from1to10008C/s);R.Pantani et al./Prog.Polym.Sci.30(2005)1185–12221189–a description capturing the main features of melt morphology(orientation and stretch)evolution under processing conditions;–a good coupling of the two(quiescent crystallization and orientation)in order to capture the effect of crystallinity on viscosity and the effect offlow on crystallization kinetics.The points listed above outline the strategy to be followed in order to achieve the basic understanding for a satisfactory description of morphology evolution during all polymer processing operations.In the following,the state of art for each of those points will be analyzed in a dedicated section.1.1.1.Modeling of the injection molding processThefirst step in the prediction of the morphology distribution within injection moldings is obviously the thermo-mechanical simulation of the process.Much of the efforts in the past were focused on the prediction of pressure and temperature evolution during the process and on the prediction of the melt front advancement [11–15].The simulation of injection molding involves the simultaneous solution of the mass,energy and momentum balance equations.Thefluid is non-New-tonian(and viscoelastic)with all parameters dependent upon temperature,pressure,crystallinity,which are all function of pressibility cannot be neglected as theflow during the packing/holding step is determined by density changes due to temperature, pressure and crystallinity evolution.Indeed,apart from some attempts to introduce a full 3D approach[16–19],the analysis is currently still often restricted to the Hele–Shaw(or thinfilm) approximation,which is warranted by the fact that most injection molded parts have the characteristic of being thin.Furthermore,it is recognized that the viscoelastic behavior of the polymer only marginally influences theflow kinematics[20–22]thus the melt is normally considered as a non-Newtonian viscousfluid for the description of pressure and velocity gradients evolution.Some examples of adopting a viscoelastic constitutive equation in the momentum balance equations are found in the literature[23],but the improvements in accuracy do not justify a considerable extension of computational effort.It has to be mentioned that the analysis of some features of kinematics and temperature gradients affecting the description of morphology need a more accurate description with respect to the analysis of pressure distributions.Some aspects of the process which were often neglected and may have a critical importance are the description of the heat transfer at polymer–mold interface[24–26]and of the effect of mold deformation[24,27,28].Another aspect of particular interest to the develop-ment of morphology is the fountainflow[29–32], which is often neglected being restricted to a rather small region at theflow front and close to the mold walls.1.1.2.Modeling of the crystallization kineticsIt is obvious that the description of crystallization kinetics is necessary if thefinal morphology of the molded object wants to be described.Also,the development of a crystalline degree during the process influences the evolution of all material properties like density and,above all,viscosity(see below).Further-more,crystallization kinetics enters explicitly in the generation term of the energy balance,through the latent heat of crystallization[26,33].It is therefore clear that the crystallinity degree is not only a result of simulation but also(and above all)a phenomenon to be kept into account in each step of process modeling.In spite of its dramatic influence on the process,the efforts to simulate the injection molding of semi-crystalline polymers are crude in most of the commercial software for processing simulation and rather scarce in the fleur and Kamal[34],Papatanasiu[35], Titomanlio et al.[15],Han and Wang[36],Ito et al.[37],Manzione[38],Guo and Isayev[26],and Hieber [25]adopted the following equation(Kolmogoroff–Avrami–Evans,KAE)to predict the development of crystallinityd xd tZð1K xÞd d cd t(1)where x is the relative degree of crystallization;d c is the undisturbed volume fraction of the crystals(if no impingement would occur).A significant improvement in the prediction of crystallinity development was introduced by Titoman-lio and co-workers[39]who kept into account the possibility of the formation of different crystalline phases.This was done by assuming a parallel of several non-interacting kinetic processes competing for the available amorphous volume.The evolution of each phase can thus be described byd x id tZð1K xÞd d c id t(2)where the subscript i stands for a particular phase,x i is the relative degree of crystallization,x ZPix i and d c iR.Pantani et al./Prog.Polym.Sci.30(2005)1185–1222 1190is the expectancy of volume fraction of each phase if no impingement would occur.Eq.(2)assumes that,for each phase,the probability of the fraction increase of a single crystalline phase is simply the product of the rate of growth of the corresponding undisturbed volume fraction and of the amount of available amorphous fraction.By summing up the phase evolution equations of all phases(Eq.(2))over the index i,and solving the resulting differential equation,one simply obtainsxðtÞZ1K exp½K d cðtÞ (3)where d c Z Pid c i and Eq.(1)is recovered.It was shown by Coccorullo et al.[40]with reference to an iPP,that the description of the kinetic competition between phases is crucial to a reliable prediction of solidified structures:indeed,it is not possible to describe iPP crystallization kinetics in the range of cooling rates of interest for processing(i.e.up to several hundreds of8C/s)if the mesomorphic phase is neglected:in the cooling rate range10–1008C/s, spherulite crystals in the a-phase are overcome by the formation of the mesophase.Furthermore,it has been found that in some conditions(mainly at pressures higher than100MPa,and low cooling rates),the g-phase can also form[41].In spite of this,the presence of different crystalline phases is usually neglected in the literature,essentially because the range of cooling rates investigated for characterization falls in the DSC range (well lower than typical cooling rates of interest for the process)and only one crystalline phase is formed for iPP at low cooling rates.It has to be noticed that for iPP,which presents a T g well lower than ambient temperature,high values of crystallinity degree are always found in solids which passed through ambient temperature,and the cooling rate can only determine which crystalline phase forms, roughly a-phase at low cooling rates(below about 508C/s)and mesomorphic phase at higher cooling rates.The most widespread approach to the description of kinetic constant is the isokinetic approach introduced by Nakamura et al.According to this model,d c in Eq.(1)is calculated asd cðtÞZ ln2ðt0KðTðsÞÞd s2 435n(4)where K is the kinetic constant and n is the so-called Avrami index.When introduced as in Eq.(4),the reciprocal of the kinetic constant is a characteristic time for crystallization,namely the crystallization half-time, t05.If a polymer is cooled through the crystallization temperature,crystallization takes place at the tempera-ture at which crystallization half-time is of the order of characteristic cooling time t q defined ast q Z D T=q(5) where q is the cooling rate and D T is a temperature interval over which the crystallization kinetic constant changes of at least one order of magnitude.The temperature dependence of the kinetic constant is modeled using some analytical function which,in the simplest approach,is described by a Gaussian shaped curve:KðTÞZ K0exp K4ln2ðT K T maxÞ2D2(6)The following Hoffman–Lauritzen expression[42] is also commonly adopted:K½TðtÞ Z K0exp KUÃR$ðTðtÞK T NÞ!exp KKÃ$ðTðtÞC T mÞ2TðtÞ2$ðT m K TðtÞÞð7ÞBoth equations describe a bell shaped curve with a maximum which for Eq.(6)is located at T Z T max and for Eq.(7)lies at a temperature between T m(the melting temperature)and T N(which is classically assumed to be 308C below the glass transition temperature).Accord-ing to Eq.(7),the kinetic constant is exactly zero at T Z T m and at T Z T N,whereas Eq.(6)describes a reduction of several orders of magnitude when the temperature departs from T max of a value higher than2D.It is worth mentioning that only three parameters are needed for Eq.(6),whereas Eq.(7)needs the definition offive parameters.Some authors[43,44]couple the above equations with the so-called‘induction time’,which can be defined as the time the crystallization process starts, when the temperature is below the equilibrium melting temperature.It is normally described as[45]Dt indDtZðT0m K TÞat m(8)where t m,T0m and a are material constants.It should be mentioned that it has been found[46,47]that there is no need to explicitly incorporate an induction time when the modeling is based upon the KAE equation(Eq.(1)).1.1.3.Modeling of the morphology evolutionDespite of the fact that the approaches based on Eq.(4)do represent a significant step toward the descriptionR.Pantani et al./Prog.Polym.Sci.30(2005)1185–12221191of morphology,it has often been pointed out in the literature that the isokinetic approach on which Nakamura’s equation (Eq.(4))is based does not describe details of structure formation [48].For instance,the well-known experience that,with many polymers,the number of spherulites in the final solid sample increases strongly with increasing cooling rate,is indeed not taken into account by this approach.Furthermore,Eq.(4)describes an increase of crystal-linity (at constant temperature)depending only on the current value of crystallinity degree itself,whereas it is expected that the crystallization rate should depend also on the number of crystalline entities present in the material.These limits are overcome by considering the crystallization phenomenon as the consequence of nucleation and growth.Kolmogoroff’s model [49],which describes crystallinity evolution accounting of the number of nuclei per unit volume and spherulitic growth rate can then be applied.In this case,d c in Eq.(1)is described asd ðt ÞZ C m ðt 0d N ðs Þd s$ðt sG ðu Þd u 2435nd s (9)where C m is a shape factor (C 3Z 4/3p ,for spherical growth),G (T (t ))is the linear growth rate,and N (T (t ))is the nucleation density.The following Hoffman–Lauritzen expression is normally adopted for the growth rateG ½T ðt Þ Z G 0exp KUR $ðT ðt ÞK T N Þ!exp K K g $ðT ðt ÞC T m Þ2T ðt Þ2$ðT m K T ðt ÞÞð10ÞEqs.(7)and (10)have the same form,however the values of the constants are different.The nucleation mechanism can be either homo-geneous or heterogeneous.In the case of heterogeneous nucleation,two equations are reported in the literature,both describing the nucleation density as a function of temperature [37,50]:N ðT ðt ÞÞZ N 0exp ½j $ðT m K T ðt ÞÞ (11)N ðT ðt ÞÞZ N 0exp K 3$T mT ðt ÞðT m K T ðt ÞÞ(12)In the case of homogeneous nucleation,the nucleation rate rather than the nucleation density is function of temperature,and a Hoffman–Lauritzen expression isadoptedd N ðT ðt ÞÞd t Z N 0exp K C 1ðT ðt ÞK T N Þ!exp KC 2$ðT ðt ÞC T m ÞT ðt Þ$ðT m K T ðt ÞÞð13ÞConcentration of nucleating particles is usually quite significant in commercial polymers,and thus hetero-geneous nucleation becomes the dominant mechanism.When Kolmogoroff’s approach is followed,the number N a of active nuclei at the end of the crystal-lization process can be calculated as [48]N a ;final Zðt final 0d N ½T ðs Þd sð1K x ðs ÞÞd s (14)and the average dimension of crystalline structures can be attained by geometrical considerations.Pantani et al.[51]and Zuidema et al.[22]exploited this method to describe the distribution of crystallinity and the final average radius of the spherulites in injection moldings of polypropylene;in particular,they adopted the following equationR Z ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi3x a ;final 4p N a ;final 3s (15)A different approach is also present in the literature,somehow halfway between Nakamura’s and Kolmo-goroff’s models:the growth rate (G )and the kinetic constant (K )are described independently,and the number of active nuclei (and consequently the average dimensions of crystalline entities)can be obtained by coupling Eqs.(4)and (9)asN a ðT ÞZ 3ln 24p K ðT ÞG ðT Þ 3(16)where heterogeneous nucleation and spherical growth is assumed (Avrami’s index Z 3).Guo et al.[43]adopted this approach to describe the dimensions of spherulites in injection moldings of polypropylene.1.1.4.Modeling of the effect of crystallinity on rheology As mentioned above,crystallization has a dramatic influence on material viscosity.This phenomenon must obviously be taken into account and,indeed,the solidification of a semi-crystalline material is essen-tially caused by crystallization rather than by tempera-ture in normal processing conditions.Despite of the importance of the subject,the relevant literature on the effect of crystallinity on viscosity isR.Pantani et al./Prog.Polym.Sci.30(2005)1185–12221192rather scarce.This might be due to the difficulties in measuring simultaneously rheological properties and crystallinity evolution during the same tests.Apart from some attempts to obtain simultaneous measure-ments of crystallinity and viscosity by special setups [52,53],more often viscosity and crystallinity are measured during separate tests having the same thermal history,thus greatly simplifying the experimental approach.Nevertheless,very few works can be retrieved in the literature in which(shear or complex) viscosity can be somehow linked to a crystallinity development.This is the case of Winter and co-workers [54],Vleeshouwers and Meijer[55](crystallinity evolution can be drawn from Swartjes[56]),Boutahar et al.[57],Titomanlio et al.[15],Han and Wang[36], Floudas et al.[58],Wassner and Maier[59],Pantani et al.[60],Pogodina et al.[61],Acierno and Grizzuti[62].All the authors essentially agree that melt viscosity experiences an abrupt increase when crystallinity degree reaches a certain‘critical’value,x c[15]. However,little agreement is found in the literature on the value of this critical crystallinity degree:assuming that x c is reached when the viscosity increases of one order of magnitude with respect to the molten state,it is found in the literature that,for iPP,x c ranges from a value of a few percent[15,62,60,58]up to values of20–30%[58,61]or even higher than40%[59,54,57].Some studies are also reported on the secondary effects of relevant variables such as temperature or shear rate(or frequency)on the dependence of crystallinity on viscosity.As for the effect of temperature,Titomanlio[15]found for an iPP that the increase of viscosity for the same crystallinity degree was higher at lower temperatures,whereas Winter[63] reports the opposite trend for a thermoplastic elasto-meric polypropylene.As for the effect of shear rate,a general agreement is found in the literature that the increase of viscosity for the same crystallinity degree is lower at higher deformation rates[62,61,57].Essentially,the equations adopted to describe the effect of crystallinity on viscosity of polymers can be grouped into two main categories:–equations based on suspensions theories(for a review,see[64]or[65]);–empirical equations.Some of the equations adopted in the literature with regard to polymer processing are summarized in Table1.Apart from Eq.(17)adopted by Katayama and Yoon [66],all equations predict a sharp increase of viscosity on increasing crystallinity,sometimes reaching infinite (Eqs.(18)and(21)).All authors consider that the relevant variable is the volume occupied by crystalline entities(i.e.x),even if the dimensions of the crystals should reasonably have an effect.1.1.5.Modeling of the molecular orientationOne of the most challenging problems to present day polymer science regards the reliable prediction of molecular orientation during transformation processes. Indeed,although pressure and velocity distribution during injection molding can be satisfactorily described by viscous models,details of the viscoelastic nature of the polymer need to be accounted for in the descriptionTable1List of the most used equations to describe the effect of crystallinity on viscosityEquation Author Derivation Parameters h=h0Z1C a0x(17)Katayama[66]Suspensions a Z99h=h0Z1=ðx K x cÞa0(18)Ziabicki[67]Empirical x c Z0.1h=h0Z1C a1expðK a2=x a3Þ(19)Titomanlio[15],also adopted byGuo[68]and Hieber[25]Empiricalh=h0Z expða1x a2Þ(20)Shimizu[69],also adopted byZuidema[22]and Hieber[25]Empiricalh=h0Z1Cðx=a1Þa2=ð1Kðx=a1Þa2Þ(21)Tanner[70]Empirical,basedon suspensionsa1Z0.44for compact crystallitesa1Z0.68for spherical crystallitesh=h0Z expða1x C a2x2Þ(22)Han[36]Empiricalh=h0Z1C a1x C a2x2(23)Tanner[71]Empirical a1Z0.54,a2Z4,x!0.4h=h0Zð1K x=a0ÞK2(24)Metzner[65],also adopted byTanner[70]Suspensions a Z0.68for smooth spheresR.Pantani et al./Prog.Polym.Sci.30(2005)1185–12221193。
ARVC2010诊断标准
SPECIAL REPORT Diagnosis of arrhythmogenic right ventricularcardiomyopathy/dysplasiaProposed Modification of the Task Force CriteriaFrank I.Marcus1*Chair,William J.McKenna2Co-Chair,Duane Sherrill1,Cristina Basso3,Barbara Bauce3,David A.Bluemke4,Hugh Calkins5,Domenico Corrado3,Moniek G.P.J.Cox6,James P.Daubert7,Guy Fontaine10,Kathleen Gear1,Richard Hauer6,Andrea Nava3,Michael H.Picard11,Nikos Protonotarios13,Jeffrey E.Saffitz12,Danita M.Yoerger Sanborn11,Jonathan S.Steinberg9,Harikrishna Tandri5,Gaetano Thiene3,Jeffrey A.Towbin14, Adalena Tsatsopoulou13,Thomas Wichter15,and Wojciech Zareba81University of Arizona,Tucson,AZ;2The Heart Hospital,London,United Kingdom;3University of Padua Medical School,Padua,Italy;4National Institutes of Health,Clinical Center, Bethesda;5Johns Hopkins Hospital,Baltimore,MD;6University Medical Center Utrecht,Utrecht,The Netherlands;7Strong Memorial Hospital,Rochester,NY;8University ofRochester Medical Center,Rochester,NY;9St.Luke’s-Roosevelt Hospital Center,New York,NY;10Hopital La Salpetriere,Paris,France;11Massachusetts General Hospital,Boston,MA;12Beth Israel Deaconess Medical Center,Boston,MA;13Yannis Protonotarios Medical Centre,Hora Naxos,Greece;14Cincinnati Children’s Hospital,Cincinnati,OH;and15Marienhospital Osnabru¨ck,Osnabru¨ck,GermanyOnline publish-ahead-of-print19February2010This paper was guest edited by Douglas P.Zipes*Correspondence to Dr Frank I.Marcus,Sarver Heart Center,1501N Campbell,Rm5153,Box245037,Tucson,AZ.Email fmarcus@This article has been co-published in the April2010issue of Circulation(Vol.121,Issue13)&2010American Heart Association,Inc.and European Society of Cardiology.European Heart Journal(2010)31,806–814doi:10.1093/eurheartj/ehq025Downloaded fromplasia’.2Progression to more diffuse RV disease and left ventricular (LV)involvement,typically affecting the posterior lateral wall,is common.3Predominant LV disease is also recognized.4Postmor-tem diagnosis may require extensive sampling and transillumina-tion.5Disease expression is variable.In the early‘concealed phase’,individuals are often asymptomatic but may nonetheless be at risk of sudden cardiac death,notably during exertion.6In the overt‘electrical phase’,individuals present with symptomatic arrhythmias,and RV morphological abnormalities are readily dis-cernible by conventional ter,diffuse disease may result in biventricular heart failure,whereas ventricular arrhythmias may or may not be present.The ultimate phenotype may resemble dilated cardiomyopathy.Clinical manifestations vary with age and stage of disease.7ARVC/D is considered to be familial with autosomal dominant inheritance,although there are recessive forms(eg,Naxos disease,Carvajal syndrome)that are associated with a cutaneous phenotype.8,9Genetic variations have been found in the desmo-somes that are responsible for cell-to-cell binding10,11(Figure1). Seven genes have been identified that are associated with ARVC/ D:plakoglobin(JUP),12desmoplakin(DSP),13plakophilin-2 (PKP2),14desmoglein-2(DSG2),15,16desmocollin-2(DSC2),17,18 transforming growth factor beta-3(TGFß3),19and TMEM43.20 Mutations in RYR2coding the ryanodine receptor have been reported in ARVC/D in patients with an arrhythmic presentation (stress-induced bidirectional ventricular tachycardia)in the absence of significant electrocardiographic or structural abnormal-ities.At present,catecholaminergic polymorphic ventricular tachycardia is considered a disorder distinct from ARVC/D.11Pre-liminary observations suggest that the mechanical defect of the desmosomes alters function of the gap junction.Electrocardio-graphic(ECG)changes and arrhythmias may develop before histo-logical evidence of myocyte loss or clinical evidence of RV dysfunction.21,22It has been proposed that similar clinical pheno-types occur that are based on disruption of a‘final common pathway’by mutations in genes encoding proteins in the defined desmosomal pathway.23Recognition of the genetic basis of ARVC/D facilitates examination of the pathogenesis in relation to arrhythmogenesis and disease progression.24It has been suggested that patients with ARVC/D may be predis-posed or susceptible to viral myocarditis,which could lead to a decrease in cardiac function and accelerate progression of the disease.25–27The link between ARVC/D and myocarditis is still undefined.BackgroundThe original1994International Task Force criteria for the clinical diagnosis of ARVC/D were based on structural,histological, ECG,arrhythmic,and familial features of the disease28(Table1).phy.Arrhythmias of RV origin,another cardinal feature of ARVC/D,was designated a minor criterion because of its occur-rence in other diseases,particularly idiopathic RV outflow tract tachycardia.Furthermore,the1994criteria focused on RV disease manifestations and stipulated the absence of or only mildLV involvement because of the need to exclude common disorderssuch as ischemic heart disease and dilated cardiomyopathy.At the time of the publication of the original Task Force guide-lines,clinical experience with ARVC/D was dominated by sympto-matic index cases and sudden cardiac death victims–the overt or severe end of the disease spectrum.Consequently,the1994cri-teria were highly specific,but they lacked sensitivity for early and familial disease.29–31Over the past15years,additional ECG markers have been pro-posed.32–34In addition,the genetic basis of the disease has been recognized,with the potential for mutation analysis.Experiencewith quantification of imaging criteria of ARVC/D has increased,and newer imaging techniques have been introduced,such as contrast-enhanced echocardiography,3-dimensional echocardio-graphy,cardiovascular magnetic resonance with late enhancement,and electroanatomic voltage mapping.35–40Figure1The cardiac desmosome and proposed roles of the desmosome in(A)supporting structural stability through cell–cell adhesion,(B)regulating transcription of genes involved in adi-pogenesis and apoptosis,and maintaining proper electrical con-ductivity through regulation of(C)gap junctions and(D)calcium homeostasis.Dsc2indicates desmocollin-2;Dsg2, desmoglein-2;Dsp,desmoplakin;Pkg,plakoglobin;Pkp2, plakophilin-2;and PM,plasma membrane.Reprinted by per-mission from Macmillan Publishers Ltd:Nat Clin Pract CardiovascMed11,&2008.at Peking University on March 6, 2011Downloaded from.............................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................Table 1Comparison of original and revised task force criteriaOriginal task force criteriaRevised task force criteriaI.Global or regional dysfunction and structural alterations*Major†Severe dilatation and reduction of RV ejection fraction with no (or only mild)LV impairment†Localized RV aneurysms (akinetic or dyskinetic areas with diastolic bulging)†Severe segmental dilatation of the RVBy 2D echo:†Regional RV akinesia,dyskinesia,or aneurysm †and 1of the following (end diastole):—PLAX RVOT 32mm (corrected for body size [PLAX/BSA] 19mm/m 2)—PSAX RVOT 36mm (corrected for body size [PSAX/BSA] 21mm/m 2)—or fractional area change 33%By MRI:†Regional RV akinesia or dyskinesia or dyssynchronous RV contraction †and 1of the following:—Ratio of RV end-diastolic volume to BSA 110mL/m 2(male)or 100mL/m 2(female)—or RV ejection fraction 40%By RV angiography:†Regional RV akinesia,dyskinesia,or aneurysmMinor†Mild global RV dilatation and/or ejection fraction reduction with normal LV†Mild segmental dilatation of the RV †Regional RV hypokinesiaBy 2D echo:†Regional RV akinesia or dyskinesia †and 1of the following (end diastole):—PLAX RVOT 29to ,32mm (corrected for body size [PLAX/BSA] 16to,19mm/m 2)—PSAX RVOT 32to ,36mm (corrected for body size [PSAX/BSA] 18to,21mm/m 2)—or fractional area change .33%to 40%By MRI:†Regional RV akinesia or dyskinesia or dyssynchronous RV contraction †and 1of the following:—Ratio of RV end-diastolic volume to BSA 100to ,110mL/m 2(male)or90to ,100mL/m 2(female)—or RV ejection fraction .40%to 45%II.Tissue characterization of wall Major†Fibrofatty replacement of myocardium on endomyocardial biopsy †Residual myocytes ,60%by morphometric analysis (or ,50%if estimated),with fibrous replacement of the RV free wall myocardium in 1sample,with or without fatty replacement of tissue on endomyocardial biopsyMinor†Residual myocytes 60%to 75%by morphometric analysis (or 50%to 65%if estimated),with fibrous replacement of the RV free wall myocardium in 1sample,with or without fatty replacement of tissue on endomyocardial biopsyIII.Repolarization abnormalities Major†Inverted T waves in right precordial leads (V 1,V 2,and V 3)or beyond in individuals .14years of age (in the absence of complete right bundle-branch block QRS 120ms)Minor†Inverted T waves in right precordial leads (V 2and V 3)(people age .12years,in absence of right bundle-branch block)†Inverted T waves in leads V 1and V 2in individuals .14years of age (in the absence of complete right bundle-branch block)or in V 4,V 5,or V 6†Inverted T waves in leads V 1,V 2,V 3,and V 4in individuals .14years of age in the presence of complete right bundle-branch blockContinuedF.I.Marcus et al .808at Peking University on March 6, 2011 Downloaded fromSince publication of the 1994Task Force guidelines,cardiovascu-lar evaluation of the relatives of ARVC/D index cases and,more recently,genotype–phenotype association studies have also high-lighted the shortcomings of the criteria.It is now recognized that LV involvement may occur early in the course of the diseasewith some frequency.4,41The criteria also lack sensitivity for the diagnosis of familial disease.Modifications of the original criteria have been proposed to facilitate clinical diagnosis in first-degree relatives who often have incomplete expression of the disease.42According to these recommendations,in the context of proven.............................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................Table 1ContinuedOriginal task force criteriaRevised task force criteriaIV.Depolarization/conduction abnormalities Major†Epsilon waves or localized prolongation (.110ms)of the QRS complex in right precordial leads (V 1to V 3)†Epsilon wave (reproducible low-amplitude signals between end of QRS complex to onset of the T wave)in the right precordial leads (V 1to V 3)Minor†Late potentials (SAECG)†Late potentials by SAECG in 1of 3parameters in the absence of a QRS duration of 110ms on the standard ECG†Filtered QRS duration (fQRS) 114ms†Duration of terminal QRS ,40m V (low-amplitude signal duration) 38ms †Root-mean-square voltage of terminal 40ms 20m V†Terminal activation duration of QRS 55ms measured from the nadir of the S wave to the end of the QRS,including R 0,in V 1,V 2,or V 3,in the absence of complete right bundle-branch blockV.Arrhythmias Major†Nonsustained or sustained ventricular tachycardia of left bundle-branchmorphology with superior axis (negative or indeterminate QRS in leads II,III,and aVF and positive in lead aVL)Minor†Left bundle-branch block-type ventricular tachycardia (sustained and nonsustained)(ECG,Holter,exercise)†Frequent ventricular extrasystoles (.1000per 24hours)(Holter)†Nonsustained or sustained ventricular tachycardia of RV outflow configuration,left bundle-branch block morphology with inferior axis (positive QRS in leads II,III,and aVF and negative in lead aVL)or of unknown axis †.500ventricular extrasystoles per 24hours (Holter)VI.Family history Major†Familial disease confirmed at necropsy or surgery†ARVC/D confirmed in a first-degree relative who meets current Task Force criteria†ARVC/D confirmed pathologically at autopsy or surgery in a first-degree relative †Identification of a pathogenic mutation †categorized as associated or probably associated with ARVC/D in the patient under evaluation Minor†Family history of premature sudden death (,35years of age)due to suspected ARVC/D†Familial history (clinical diagnosis based on present criteria)†History of ARVC/D in a first-degree relative in whom it is not possible or practical to determine whether the family member meets current Task Force criteria †Premature sudden death (,35years of age)due to suspected ARVC/D in a first-degree relative†ARVC/D confirmed pathologically or by current Task Force Criteria in second-degree relativePLAX indicates parasternal long-axis view;RVOT,RV outflow tract;BSA,body surface area;PSAX,parasternal short-axis view;aVF,augmented voltage unipolar left foot lead;and aVL,augmented voltage unipolar left arm lead.Diagnostic terminology for original criteria:This diagnosis is fulfilled by the presence of 2major,or 1major plus 2minor criteria or 4minor criteria from different groups.Diagnostic terminology for revised criteria:definite diagnosis:2major or 1major and 2minor criteria or 4minor from different categories;borderline:1major and 1minor or 3minor criteria from different categories;possible:1major or 2minor criteria from different categories.*Hypokinesis is not included in this or subsequent definitions of RV regional wall motion abnormalities for the proposed modified criteria.†A pathogenic mutation is a DNA alteration associated with ARVC/D that alters or is expected to alter the encoded protein,is unobserved or rare in a large non-ARVC/D control population,and either alters or is predicted to alter the structure or function of the protein or has demonstrated linkage to the disease phenotype in a conclusive pedigree.Diagnosis of arrhythmogenic right ventricular cardiomyopathy/dysplasia809at Peking University on March 6, 2011 Downloaded fromis based on the documentation of one of the following in a family member:(1)T-wave inversion in right precordial leads V1,V2,and V3inindividuals over the age of14years.(2)Late potentials by signal-averaged ECG(SAECG).(3)Ventricular tachycardia of left bundle-branch block mor-phology on ECG,Holter monitor,or during exercise testing or.200premature ventricular contractions in24hours. (4)Either mild global dilatation or reduction in RV ejection frac-tion with normal LV or mild segmental dilatation of the RV or regional RV hypokinesis.Revision of the diagnostic criteria is important to provide guidance on the role of emerging diagnostic modalities and to recognize advances in the genetics of ARVC/D.The criteria have been modi-fied to incorporate new knowledge and technology to improve diagnostic sensitivity,but with the important requisite of maintain-ing diagnostic specificity,and they include quantitative parameters for Task Force criteria,particularly for the imaging studies (Table1).The approach of classifying structural,histological, ECG,arrhythmic,and genetic features of the disease as major and minor criteria has been maintained.MethodsA limitation of the previous Task Force criteria was the reliance on subjective criteria for assessing ventricular structure and function and for evaluation of myocardial histology.In this modification of the Task Force criteria,quantitative criteria are proposed and abnormalities are defined on the basis of comparison with normal subject data(Table1).The data from108probands with newly diagnosed ARVC/D,age 12years,who were enrolled in the National Institutes of Health-supported Multidisciplinary Study of Right Ventricular Dysplasia,43were compared with those of normal subjects(online-only Data Supplement).The cri-teria were selected on the basis of analysis of sensitivity and speci-ficity from receiver operating characteristic curves.For analysis of each test[e.g.echocardiogram,magnetic resonance imaging (MRI)],proband data were excluded if that test was crucial for the diagnosis of the individual patient.This was done to eliminate bias in estimating the sensitivity and specificity of that particular test.In general,when determining the sensitivity and specificity of a new screening test,it is recommended that none of the screening test elements be used in making the primary diagnosis; this principle also holds when establishing diagnostic criteria. ResultsThere were44proband MRIs compared with462MRIs of normal subjects,69proband echocardiograms compared with450echo-cardiograms of normal subjects,69proband SAECGs compared with103SAECGs of normal subjects,and68proband Holters compared with398Holters of normal subjects.The minor criteria for echocardiography were selected where specificity and sensi-tivity are equal(sensitivity equals specificity)(Table2).The major criteria were selected as the value that yielded95%specificity.Sen-sitivity and specificity for the MRI criteria were made RV end-diastolic volume indexed to body surface area(size)andRV ejection fraction(function)simultaneously by using the OR logistical function.If either RV size or function was positive in con-junction with RV wall motion abnormality,then the subject wouldbe classified as having a major criterion for the MRI.The sensitivityof RV size alone or function alone ranged from41%to50%formajor criteria and31%to41%for minor criteria,with specificityof96%to100%.Using the OR logistical function improved the sensitivity of the MRI to79%to89%for major criteria and68%to78%for minor criteria.The original Task Force criteria list late potentials as a minor cri-terion.It has become common practice,though not based on evi-dence,to state that the SAECG is positive if2of the following3 parameters are abnormal:filtered QRS duration(fQRS),root-mean square voltage of the terminal40ms of the QRS,or durationof the terminal QRS signal,40m V.Analysis of each of the single parameters of the SAECG with late potentials by using a40-to250-Hzfilter had a sensitivity ranging from58%to60%,with aspecificity of94%to96%.Two of three parameters had a sensi-tivity of66%and specificity of95%,adding little advantage withregard to sensitivity and specifiing any one of the3 SAECG parameters had a sensitivity of74%and specificity of92%.A definitive diagnosis of ARVC/D is based on histological dem-onstration of transmuralfibrofatty replacement of RV myocardiumat biopsy(Figure2),necropsy,or surgery.5,44In most patients, however,assessment of transmural myocardium is not possible.In addition,diagnosis based on RV endomyocardial biopsy speci-mens is limited because the segmental nature of the disease causes false e of electroanatomic voltage mappingto identify pathological areas for biopsy sampling may improvethe yield.45RV free wall biopsy has a slight risk of perforation,but the more accessible interventricular septum rarely exhibits his-tological changes.Nevertheless,endomyocardial biopsy may ident-ify other conditions(e.g.myocarditis,sarcoidosis,endomyocardialfibrosis),and the recognition of myocyte loss withfibrous orfibro-fatty replacement can be a valuable diagnostic feature.46The identification of disease-causing genes has led to the recognition of a broader spectrum of disease expression within families,including individuals who have predominantly LV disease, manifest clinically by inferolateral T-wave changes,ventricular ectopy,or ventricular tachycardia with right bundle-branch block morphology and epicardial or midmyocardial late enhancementby MRI.4,7,38,39,41The importance of familial disease highlights arole for mutation analysis of probands with cascade screening of relatives that offers an alternative strategy to serial noninvasive car-diovascular evaluation of families.A positive diagnosis in a family member changes the probability of disease in an individual sus-pected of the disease to1:2from1:1000to1:5000.Thus,con-firmed disease in afirst degree relative is a major criterion for diagnosis.42DiscussionThe diagnosis of ARVC/D relies on the demonstration of struc-tural,functional,and electrophysiological abnormalities that are caused by or reflect the underlying histological changes.Technical advances in MRI and2-dimensional echocardiography haveat Peking University on March 6, 2011Downloaded fromimproved the capability to image the RV with reproducible measure-ments of volume and systolic function,which permits classification of severity and differentiation from normality 47(Table 2).Previous diag-nostic reliance on subjective assessment of RV wall thinning,wall motion abnormalities,and fatty infiltration of the myocardium byMRI has proven problematic.48,49Recognition of significant fatty involvement without concomitant fibrosis of the RV in normal indi-viduals renders this unique MRI capability of limited te enhancement on MRI permits myocardial tissue characterization in the LV.It can be difficult to be certain of late enhancement for..............................................................................................................................................................................................................................................................................................................................................................Table 2Sensitivity and specificity of proposed RV imaging criteria*ValueSensitivity,%Specificity,%EchocardiogramMajorPLAX RVOT (diastole)32mm 7595Corrected for body size (PLAX/BSA) 19mm/m2PSAX RVOT (diastole)36mm 6295Corrected for body size (PSAX/BSA) 21mm/m 2Fractional area change 33%5595MinorPLAX RVOT (diastole)29mm8787Corrected for body size (PLAX/BSA) 16to 18mm/m 2iPSAX RVOT (diastole)32mm8080Corrected for body size (PSAX/BSA) 18to 20mm/m 2Fractional area change 40%7676MRI †MajorRatio of RV end-diastolic volume to BSAMales 110mL/m 2Females 100mL/m 27690F or6898CRV ejection fraction 40%MinorRatio of RV end-diastolic volume to BSAMales 100mL/m 2Females 90mL/m 27985F or8997CRV ejection fraction45%Abbreviations as in Table 1.*All the major and minor criteria listed in this table are in addition to the requirement that regional wall motion abnormalities must also be present.†The sensitivity and specificity for males and females are the same as listed if,in addition to the stated wall motion criteria,there is either abnormal RV size or function or both.Figure 2Endomyocardial biopsy findings in a proband affected by a diffuse form of ARVC/D.All 3biopsy samples are from different regionsof the RV free wall.There is extensive fibrofatty tissue replacement with myocardial atrophy,which is a major criterion (i.e.residual myocytes ,60%by morphometric analysis or ,50%if estimated).Contributed by C.Basso,Padua,Italy.Diagnosis of arrhythmogenic right ventricular cardiomyopathy/dysplasia811at Peking University on March 6, 2011 Downloaded fromcharacterization of RV myocardium because of the thin wall of the RV and possible confusion with fat.50There also have been recent developments to quantify the extent of RV wall motion abnormalities by angiography with computer-based analysis,as well as to determine RV volumes.51,52In addition,commercial software is available to deter-mine RV volumes and ejection fraction.53The RV angiogram obtained in multiple views is considered to be a reliable imaging test to assess wall motion abnormalities but requires considerable experience.Standardized protocols for performance of these diag-nostic studies (ECG,SAECG,echocardiogram,RV angiogram,and MRI)are available on .Repolarization abnormalities are early and sensitive markers of disease expression in ARVC/D.T-wave inversion in V1,V2,and V3and beyond in individuals .14years of age who are otherwise healthy is observed in only 4%of healthy women and 1%of men.Therefore,it is reasonably specific in this population and con-sidered a major diagnostic abnormality in ARVC/D.54Depolariz-ation delay in right precordial leads is also common in ARVC/D.33,34Evaluation of the duration of terminal QRS activation (Figure 3)incorporates slurring of the S wave,as well as R 0,into a single measure of terminal activation duration.34Depolarization abnormalities cannot be evaluated in the presence of typical com-plete right bundle-branch block with terminal delay in leads I and V 6.However,T-wave inversion in V 1,V 2,V 3,and V 4is uncommon in patients with right bundle-branch block who do not have ARVC/D and is seen frequently in those who do have the disease.Con-ventional definitions are used for ventricular arrhythmias.An abnormal SAECG is based on time domain criteria with cutoffs generated from receiver operating characteristic curves.55,56The sensitivity and specificity of any one of the time domain criteria is similar to that of any 2or 3of these criteria;therefore,any one of the criteria is proposed as a criterion for this modality.The presence of left bundle-branch block ventricular tachycardia with an inferior axis (R wave positive in leads II and III and negative in lead aVL)is typical of focal RV outflow tract tachycardia.57Similar features may be seen in patients with ARVC/D but usually coexist with anterior T-wave inversion and ventricular arrhythmias of varying morphologies.The presence of ventricular ectopy increases with age,but .200ventricular premature beats in 24hours in an adult ,50years of age suggests underlying myo-cardial disease.58The revised criteria were applied post hoc to 108newly diag-nosed probands enrolled in the Multidisciplinary Study of Right Ventricular Dysplasia,a study supported by the National Institutes of Health.They had been carefully evaluated,including assessment of diagnostic tests by expert core laboratories.43Of the 73pro-bands with final classification as ‘affected’,71remain affected and 2were reclassified as borderline.The change from affected to bor-derline in the 2was due to the echocardiogram’s fulfilling only minor criteria in one and only mild hypokinesis in the angiogram of the other.Of the 28probands classified as borderline (met some but not all of the original Task Force criteria–i.e.1major and 1minor or 3minor),5remain borderline and 16were reclas-sified by the new criteria as affected.Seven became unaffected (did not meet the proposed modified Task Force criteria).Of 7pro-bands previously classified as unaffected,4remained unaffected,1became affected,and 2became borderline.Therefore,the effect of the revised criteria is to increase the sensitivity of the classification,primarily in probands previously classified as borderline.Nine of 28probands classified as borderline by original criteria had gene variants consistent with ARVC/D.The sensitivity of the revised criteria is not perfect,as exemplified by the observation that if the genetic criteria are ignored,the proposed criteria classi-fied 2as unaffected and 3remained borderline,and 4became affected.Including the proposed genetic criteria resulted in all 9being classified asaffected.Figure 3ECG from proband with T-wave inversion in V 1through V 4and prolongation of the terminal activation duration 55ms measuredfrom the nadir of the S wave to the end of the QRS complex in V 1.Contributed by M.G.P.J.Cox,Utrecht,The Netherlands.F.I.Marcus et al .812at Peking University on March 6, 2011 Downloaded from。
第11届水科学发展论坛将在大连举办
International Journal for Numerical in Fluids, 2011,67(9):1160-1174.[2]王兴勇,索丽生,程永光,等.用Lattice Boltzmann方法模拟方柱绕流[J].河海大学学报(自然科学版),2003, 31 ( 3 ):259-263. (WANG Xingyong,SUOLisheng,CHENG Yongguang,et al. Simulation of flowfields around a square cylinder with Lattice BoltzmannMethod [ J ]. Journal of Hohai University ( NaturalScience),2003,31(3) :259-263. (in Chinese))[3]张兰丁.桥墩群体绕流的复变函数理论解[J].水利水电科技进展,2005,25 (1 ):14-16. (ZHANG Landing.Theoretical solution of complex function about flow aroundbridge piers[J]. Advances in Science and Technology ofWater Resources,2005,25(1) :14-16. (in Chinese))[4] CAO S Y,GEY J,TAMURA Y. Shear effects on flow pasta square cylinder at moderate Reynolds numbers [ J].Journal of Engineering Mechanics,2012,138 (1): 116123.[5] LANKADASU A,VENGADESAN S. Onset of vortexshedding in planar shear flow past a square cylinder[J].International Journal of Heat and Fluid Flow,2008,29(4) :1054-1059.[ 6 ] HWANG R R, SUE Y C. Numerical simulation of shear effect on vortex shedding behind a square cylinder [J].International Journal for Numerical Methods in Fluids,1998,25(12) :1409-1420.[7] AYUKAWA K,OCHI J,HIRAO T. Effects of shear rate onthe flow around a square cylinder in a uniform shear flow[J ]. Journal of Wind Engineering and IndustrialAerodynamics,1993,50(93) :97-106.[8]陈明杰,施卫平.用格子Boltzmann方法计算剪切流的方柱绕流问题[J].吉林大学学报(理学版),2007,45(I): 11-14. ( CHENG Mingjie,SHI Weiping. Numericalsimulation of linear shear flow past a square cylinder withLattice Boltzmann method [ J] . Journal of Jilin University(Science Edition),2007,45(1) :11-14. (in Chinese)) [ 9 ] SAHA A K, BISWAS G, MURALIDHAR K. Influence of inlet shear on structure of wake behind square cylinder[J]. Journal of Engineering Mechanics,1999,125 : 359363.[10] WANG D G,WANG H J,XIONG J H,et al.Characteristic-based operator-splitting finite elementmethod for Navier-Stokes equations [ J ] . Science ChinaTechnological Sciences,2011,54(8):2157-2166. [11] BREUER M,BERNSDORF J,ZEISER T,et al. Accuratecomputations of the laminar flow past a square cylinderbased on two different methods:Lattice-Boltzmann andfinite-volume[J]. International Journal of Heat and FluidFlow,2000,21(2):186-196.[12] BREUER M,RODI W. Large eddy simulation for complexturbulent flows of practical interest [J]. Flow Simulationswith High-Performance Computer II,1996,52:258-274. [13] DAVIS R W,MOORE E F. A numerical study of vortexshedding from rectangles[J]. Journal of Fluid Mechanics,1982,116:475-506.[14] TREIDLER E B. An experimental and numericalinvestigation of flow past ribs in a channel[ R]. Berkeley:University of California,1991.[15] LI G,HUMPHREY J A C. Numerical modelling ofconfined flow past a cylinder of square cross-section atvarious orientations [ J ] . International Journal forNumerical Methods in Fluids,1995,20(21) :1215-1236.[16] ARNAL M P,GOERING D J,HUMPHREY J A C. Vortexshedding from a bluff body adjacent to a pane siding wall[ J ] . Transaction of American Society of MechanicalEngineering,1991,113(3) :384-398.[17] OKAJIMA A. Strouhal numbers of rectangular cylinders[J]. Journal of Fluid Mechanics,1982,123 :379-398. [18] CHENG M,TAN S H N,HUNG K C. Linear shear flowover a square cylinder at low Reynolds number [ J ] .Physics of Fluids,2005,17(7) :078103-1-78103^. [19] KWON T S,SUNG H J,HYUN J M,et al. Experimentalinvestigation of uniform-shear flow past a circular cylinder[J]. Journal of Fluids Engineering,1992,114(3) : 457 -460.(收稿日期:2015 - 11 -07编辑:熊水斌)•简讯•第11届水科学发展论坛将在大连举办随着自然条件的变化和人类活动的加剧,水科 学问题越来越复杂。
髋膝关节假体周围感染
李树源 1,陈雷雷 2,黄林峰 1,赵赫然 1,吴素雯 1,蒋元康 1
文章快速阅读:
髋膝假体感染
病因和机制 流行病学 风险因素 诊断和分型
Hydrocepnalus2
Four main intracranial components : the brain, the CSF, the arterial and the venous blood Any volume increase in one component causes a matching decrease in the other components Fluids are incompressible The systolic expansion of the intracranial arteries is balanced by a matching expulsion of CSF
无流动补赏标准双回声序列来诊断是否有脑室内梗阻 交通性脑积水:a,中脑导水管明显扩张(质子密度像)。 b,心脏收缩期导水管、三四脑室的CSF流空信号(T2像)
青岛大学医学院附属医院 1898
脑积水
流体动力学理论 病因与临床症状 治疗
典型病例
青岛大学医学院附属医院 1898
脑积水
病因学:蛛网膜下腔出血、脑膜炎、颅脑损 伤以及静脉栓塞和颅内肿瘤等.
头部平扫MRI:胼胝体膝部左侧、压部多发梗死灶, 脑积水
青岛大学医学院附属医院 1898 诊断:病毒性脑膜炎
青岛大学医学院附属医院 1898
8月后复查
青岛大学医学院附属医院 1898
患者因“头痛、头晕伴恶心、呕吐7天” 。于 2017-4-22再次入院。
查体:颈部略抵抗,双侧克氏征(+)。 颅脑CT示脑积水并轻度脑脊液外渗,胼胝体压部软
through the foramen magnum and expulsion of blood from the veins into the dural venous
On the origin and excitation of the extended nebula surrounding NGC 1275
a r X i v :a s t r o -p h /0512331v 1 13 D e c 2005Mon.Not.R.Astron.Soc.000,000–000(0000)Printed 5February 2008(MN L A T E X style file v2.2)On the origin and excitation of the extended nebula surroundingNGC 1275N.A.Hatch,⋆C.S.Crawford,R.M.Johnstone and A.C.FabianInstitute of Astronomy,Madingley Road,Cambridge,CB30HA5February 2008ABSTRACTWe use line-of-sight velocity information on the filamentary emission-line nebula of NGC 1275to infer a dynamical model of the nebula’s flow through the surrounding intra-cluster gas.We detect outflowing gas and flow patterns that match simulations of buoyantly rising bubbles from which we deduce that some of the nebula filaments have been drawn out of NGC 1275.We find a radial gradient of the ratio [N II ]λ6584/H αwhich may be due to a variation in metallicity,interactions with the surrounding intracluster medium or a hardening of the excitation mechanism.We find no preferred spatial correlation of stellar clusters within the filaments and there is a notable lack of [O III ]λ5007emission,therefore it is unlikely that the filaments are ionized by stellar UV .Key words:galaxies:clusters:individual:Perseus -cooling flows -galaxies:individual:NGC 1275-intergalactic medium.1INTRODUCTIONNGC 1275is the central galaxy of the X-ray luminous Perseus clus-ter (A426),which has bright centrally peaked X-ray emission,and a cool core with a central temperature of a third the virial tem-perature (Schmidt et al.2002;Sanders et al.2004).Cavities in the X-ray emission are observed in a number of locations surround-ing the central galaxy (Fabian et al.2003a).Where these coincide with GHz radio emission they are interpreted as bubbles filled with relativistic plasma,which has been injected into the intracluster medium (ICM)by the central engine.Cavities with no observed radio emission have been described as ‘ghost bubbles’,and are thought to have originated from an earlier epoch of activity in the central engine.Minkowski (1957)discovered that the nebula of NGC 1275comprises of two distinct emission-line systems:a high-velocity system (8200kms −1),identified as a disrupted foreground galaxy (Boroson 1990)at least 60kpc in front of NGC 1275(Gillmon et al.2004),and a low-velocity system (5265kms −1).Although the high-velocity system lies directly in front of NGC 1275,the emis-sion lines are easily distinguished in wavelength from those of the low-velocity system,and clearly indicate photoionization by hot,young stars (Kent &Sargent 1979).The low-velocity system associated with the central galaxy NGC 1275is the focus of this work.It is known to extend over 100kpc in a large array of thin filaments (Lynds 1970;Conselice et al.2001).Whilst the nebula is extremely luminous–4.1×1042ergs −1in H αand [N II ](Heckman et al.1989),with a total line luminosity probably 20times that in H α–the power⋆E-mail:nah@source remains unknown.Ionization by the central active galac-tic nucleus (AGN)residing in NGC 1275can be ruled out as the dominant source of power for the extended nebula on the grounds that the H αluminosity does not decrease with distance from the nucleus (Johnstone &Fabian 1988),although it may be important in the luminous inner regions.Ionization by hot young stars is an attractive option as it is a local mechanism,but the line ratios are drastically different to those seen in H II regions (Kent &Sargent 1979).Models of heating by X-rays from the ICM have been put forward (Donahue &V oit 1991),as well as conduction from the ICM (Donahue et al.2000),shocks (Sabra et al.2000)and turbu-lent mixing layers (Crawford &Fabian 1992).Soft X-ray emission is associated with some of the optical fil-aments of NGC 1275(Fabian et al.2003b).The filaments are less luminous in the X-ray than the optical/UV by up to two orders of magnitude implying they are not excited by X-radiation.The soft X-ray emission indicates an interaction between the warm fil-aments and the hot ICM,possibly via heat conduction.Large deposits of molecular hydrogen have been discov-ered in the central regions of NGC 1275(Krabbe et al.2000;Donahue et al.2000)similar to other central cluster galaxies with emission-line nebulae (Edge et al.2002).Recently,molecu-lar hydrogen was observed in the outer filaments of NGC 1275(Hatch et al.2005),indicating gas at 2000K exists within the hot ICM at radii of over 25kpc.The origin of the filaments remain a mystery.Current theories include condensing gas from the ICM in the form of a cooling flow (Fabian et al.1984;Heckman et al.1989;Donahue &V oit 1991),gas accreting from previous mergers (Braine et al.1995),the explo-sive expulsion of gas from NGC 1275(Burbidge &Burbidge 1965)or gas drawn out (Fabian et al.2003b).The filaments are very thin,c0000RAS2N.A.Hatch,C.S.Crawford,A.C.Fabian&R.M.Johnstone long and the majority are radial.Submerged within the ICM,theyenable us to constrain the level of turbulence in the ICM and arguefor a laminarflow.As the ICM moves it may drag the warm opti-cally emitting gas,thus thefilaments can act as streamlines tracing theflow direction(Fabian et al.2003b).None of these problems are exclusive to NGC1275as ex-tended emission-line nebulae are commonly found surrounding other massive galaxies in the centre of X-ray bright‘cool cores’, where the X-ray emission is centrally peaked(Crawford et al. 1999).This work presents new spectroscopic data that explore the kinematic and line-emission properties of the nebula that surrounds NGC1275.After analysing and interpreting the kinematics we put forward a dynamical model of the nebula and discuss the origin of thefilaments.The redshift of NGC1275is0.0176,which using H0=70kms−1Mpc−1,gives1kpc≃2.7arcsec.2OBSERV ATIONS AND DATA REDUCTIONThe data were obtained on the nights of2004Sept23and2004 Oct06using the GMOS North instrument on the Gemini North telescope on Mauna Kea,Hawaii.The sky on both nights was pho-tometric and the seeing was less than0.8arcsec.Six slit positions were chosen using the map of Conselice et al.(2001;see Fig1), selected to feature particular structures of interest.Two bright stars were aligned on each slit so its exact position was known.The slit width was0.5arcsec,filter R831+G590was used and the exposure time for each slit was900seconds(3×300s);all exposures were binned2×2before readout.This setup allows us to explore the observed wavelength range4850–6945˚A.The spectroscopic stan-dards BD+28d4211and G1912B2were observed in order toflux-calibrate the data.Aflat was taken with the GCAL instrument after every slit position observed.CuAr arcs and bias frames were taken with the GCAL instrument on the nights of the2004Sept23and 2004Oct05.A summary of all science exposures taken is given in Table1.The data were reduced using the IRAF Gemini package(ver-sion1.7).The bias frames were combined,flatfields were nor-malised and mosaiced together.All science and standard star frames were reduced by bias subtraction,mosaicing the individ-ual chips together,flatfielding and interpolating across the detector gaps.The arcs were calibrated and checked manually against a line list before the wavelength solution was transferred to the data.The science observations wereflux calibrated using the standard star G1912B2(BD+28d4211was not used as it was observed35◦from the parallactic angle and showed signs of strong atmospheric dif-ferential refraction),andfinally sky-subtracted.The data were de-reddened for Galactic extinction using E(B–V)=0.315.The spec-tra were extracted,converted to ASCII format and analysis was performed using QDP(Tennant1990).The red section covering [O I]λ6300to[S II]λ6731wasfitted separately to a blue section covering the lines of Hβto[N I]λ5199.The He Iλ5876wasfit-ted separately.The lines were assumed to be Gaussian in profile and share the same redshift and velocity width.[N II]λ6548was assumed to have a third of the intensity of the[N II]λ6584line.Acquisition images were taken at all slit positions through the r G0303filter.These images were reduced by bias subtraction,flatfielding,then normalised and combined using the IRAF IM-AGES package to form an R-band image with a scale of0.1454 arcsec/pixel.The image was used to accurately align the slits with the Hαimage from Conselice et al.(2001).Slit1Spectral R831+G590900 Slit1Image rG0303100 Slit2Spectral R831+G590900 Slit3Spectral R831+G590900 Slit3Image rG030360 Slit4Spectral R831+G590900 Slit4Image rG0303300 Slit5Spectral R831+G590900 Slit5Image rG030380 Slit6Spectral R831+G590900 Slit6Image rG0303300On the origin and excitation of the extended nebula surrounding NGC12753Figure1.Position of the six longslit observations on the Hαnebula surrounding NGC1275.The background image is taken from the data of Conselice et al. (2001)within the low velocity system.The spatial connection and similar-ity in colour to the structures within the high-velocity system sug-gest that many of these blue star clusters may be associated with the disrupted foreground galaxy.The high-resolution HST image shows the Hαfilaments themselves to be highly smooth(Figure4) and continuous,consisting of several strands which are individually unresolved on scales of0.25arcsec(corresponding to90parsecs).Excluding the central region where continuum from the galaxy is prominent,our spectra show signs of continuum only at9re-gions.Two of these regions exhibit no line emission so the con-tinuum may have originated from stellar clusters which typically do not have Hαemission(Holtzman et al.1992).The other7re-gions exhibited line emission with very similar line intensity ratios to regions without continuum and only one exhibited[O III]λ5007 emission(see Figure16in section7.3.1).The Hαluminosity from these continuum regions is fairly typical.Although the continuum regions are distinct in the R-band acquisition image as stellar clus-ters,the HST images show that these bright knots are ubiquitous throughout the region stretching beyond the emission-line nebula. These are likely to be chance associations of alignment.It does not appear that the stellar clusters are formed or located within thefil-aments.5KINEMATICS OF THE NEBULAThe morphology of thefilaments suggest they act as streamlines tracing the gasflow in the ICM.Therefore the Doppler shifts of thefilaments may reveal the velocityfield in the core of the clus-ter,near NGC1275.The forces that could drive theflow offil-aments are gravity(from NGC1275),which would draw thefil-aments inwards,or the outward pull following a buoyantly ris-ing radio/ghost bubble as proposed by Bohringer et al.(1995);c 0000RAS,MNRAS000,000–0004N.A.Hatch,C.S.Crawford,A.C.Fabian&R.M.Johnstonebined colour image of the North and West environment of NGC1275as seen through the HST F450W(shown as blue),F702W(green)and F814W(red)broad-bandfilters.The F702Wfilter encompasses Hα+[N II]line emission at the redshift of NGC1275.The image is120arcsec by73arcsec in size.The Northwest and Northernfilaments including the‘horseshoe’loop are visible.A1-arcminute-long chain of blue stellar clusters runs from the top left(above the cluster galaxy)toward the bottom right where it meets the Western edge of the infalling galaxy(high-velocity system)near the bright star.The white arrow points to a region where the detected Hαemission is redshifted by5538km s−1placing it in the low-velocity system.Figure3.A zoom to the Northwest of the previousfigures,showing detail of the loops in the region of the‘horseshoe’filament.The three colour images have each been unsharp-masked to remove the light from the underlying central galaxy.The image is66by38arcsec in size.c 0000RAS,MNRAS000,000–000On the origin and excitation of the extended nebula surrounding NGC12755Figure4.A detailed image of the Hα‘horseshoe’filament as seen through the HST F702W broad-bandfilter,showing thefine structure and smooth-ness of the individual strands.The image is24arcsec on a side,and has been unsharp-masked(i.e.it has had a highly smoothed version of itself subtracted)to remove the underlying gradient due to the light from the cen-tral galaxy continuum.Churazov et al.(2001);Reynolds et al.(2005).A three dimensional flow pattern can differentiate between galactic outflow and inflow models,and thus constrain the origin of thefilaments.The ve-locities are determined from binning the spatial dimension of the longslit spectra in bins of4pixels(0.58arcsecond).All veloci-ties presented are heliocentric and the line-of-sight zero point is defined as the velocity of NGC1275,assumed to be5265kms−1 (Ferruit et al.1997).All distances referred to are projected dis-tances.5.1NorthernfilamentThe Northernfilament is the dominant long(∼60kpc),thin (<1kpc)structure stretching radially North-South,situated North of NGC1275;slit6was positioned along thisfilament.The veloc-ity structure of the Hαand[N II]lines is shown in Figure5.The filament appears extremely radial for the majority of its length sug-gesting that the dominant direction offlow is also radial.As this filament is the only structure detected so far out from the galaxy it is extremely unlikely to be in projection with anotherfilament,and therefore we assume it to be a single structure.It is improbable that we should be viewing an intrinsically-curvedfilament as a linear one,so we assume that it is intrinsically straight.Thefilament has a kinematic North-South divide:the North displaying a velocity blueshift by up to−180km s−1,whereas the South is erratically redshifted.The Northern half(above37kpc) is clumpy on scales of up to5kpc in length.Each clump exhibits smooth velocity gradients,although there are velocity discontinu-ities between the clumps.It is possible that some of the velocity could be due to thefilament twisting as it falls(note the helical na-ture of the lower parts of thefilament,Fig.5).The Southern part of thisfilament is split into two vertical segments:slit6covers only the western dimmer segment,whilst slits1and2slice across both. The Western segment is very thin and redshifted.The Eastern seg-ment appears thicker and clumpier,and slit2shows that the emis-sion is blueshifted.Slit1shows that the very bottom of the Eastern segment is redshifted.The kinematic North-south divide indicates the lower part of thefilament is moving in the opposite direction to the upper part of thefilament:thefilament is either being stretched or is collaps-ing,depending on whether it is orientated toward or away from the observer.Half of thisfilament must be falling into the galaxy,and the other half must beflowing away from the galaxy.Thus we can immediately rule out a model in which thefilaments are smoothly falling onto the galaxy below.The Doppler shifts alone do not en-able us to determine which end of thefilament is inflowing or out-flowing since we then need to know the inclination.However,as part of thefilament must beflowing away from the galaxy,there must be a mechanism for drawing gas away from the galaxy.We note that the Southern end of the three radial Northern filaments coincides with a shock front seen in the X-ray images (Fabian et al.2003a).This front is due to the formation of the inner Northern bubble around the radio source which is a cyclical pro-cess taking place every107yr or so(as indicated by the presence of the outer ghost bubbles).If an expanding shock front destroys the emission-linefilaments,then the lower part of the Northernfila-ments must previously have been at a larger radial distance in order to have survived the shock emitted from the Northwest ghost bub-ble when it was forming.Therefore the lower half of the Northern filament is probably moving inward,whilst the upper segment is moving outward,i.e.thefilament is likely to be stretching.There is a depression in the thermal pressure just above the Northernfilament(Fabian et al.2005),that could be a remnant of a ghost bubble that has buoyantly risen from the central region. We can now interpret thefilament in the context of the rising bubble models of Bohringer et al.(1995);Churazov et al.(2001); Fabian et al.(2003b);Reynolds et al.(2005).The radialfilament morphology traces the primary direction offlow therefore it acts as a streamline.Part of thefilament isflowing away from the galaxy due to the uplift caused by the ghost bubble’s buoyant rise through the ICM,whilst the other half has been overcome by the galaxy’s gravity and is nowflowing back.The pull from the bubbles com-petes with gravity.For a totalfilament length of25kpc and a range in velocity of 400km s−1the dispersion time is6×107years;if thefilament is at a small angle from the plane of the sky(as is likely due to its large projected length)the velocity range may be much larger,reducing the dispersion time.5.2Northwestfilaments and‘horseshoe’featureTo the Northwest of the galaxy lies an array of radialfilaments, one of which extends30kpc from the nucleus and ends in a curved loop that Conselice et al.(2001)refer to as the‘horseshoe’(detail in Fig.4).The loop is positioned underneath a ghost bubble visible as a prominent depression in the X-ray image(Fabian et al.2003a).The morphology of thesefilaments has previously been noted to resemble theflow underneath an air bubble rising in water (Fabian et al.2003b).Figure6details the line-of-sight velocities of thesefilaments.The long radialfilament(Western part of the ‘horseshoe’loop)begins with a redshifted line-of-sight velocity of95kms−1which remains fairly constant for5kpc until the line emission shifts Southwest beyond the slit for3kpc,to reappear at ac 0000RAS,MNRAS000,000–0006N.A.Hatch,C.S.Crawford,A.C.Fabian &R.M.JohnstoneFigure 5.Line-of-sight velocities of the Northern filaments.Purple-blue indicate blueshifted emission,yellow-red indicate redshifted emission,whilst green has zero velocity relative to the central galaxy.Velocities from slit 1(bottom)and slit 2(top)which cut across the Northern filament are displayed in the image but not shown in graph.Background images are from the data of Conselice et al.(2001).distance of 18kpc from the galaxy with a velocity of 60kms −1.The difference of 35kms −1between the two sides of this gap cannot be unambiguously attributed to a change in speed as a small change in orientation to the plane of our line of sight could also produce the observed velocity deviations.From 18kpc upward,this filament divides into two velocity structures before curving into the loop which starts at 22kpc (see right panel of Fig.6).These velocity structures have smooth gra-dients with no small scale random deviations in excess of the error (1-10km s −1depending on the line strength).The low velocities,morphological structure of the filaments,and the spherical cap ap-pearance of the ghost bubble,indicate the filament may be close to being in the plane of the sky.The curved part of the ‘horseshoe’starts at a projected dis-tance of 22kpc from the nucleus where the velocity increases rapidly to 200kms −1,and remains steady over 4kpc.The top of the loop has the highest velocity,peaking at 300kms −1,then slow-ing down to 200kms −1as the loop turns over.On the short side of the loop the emission is still redshifted but the velocity dies to 60kms −1in under 1kpc and remains steady for the rest of the loop.Above the loop is gas with a blueshifted line-of-sight velocity of –230kms −1,a jump in velocity space of more than 480kms −1over a projected distance of 3kpc.Further above the central axis of the bubble slit 5cuts across some dim emission which is also blueshifted.The gas in the loop and above the loop surrounds theghost bubble suggesting the gas above the loop is part of the same structure as that in the loop and is not just a projection effect.The flow pattern qualitatively matches the simulations of Reynolds et al.(2005)of a bubble rising through a viscous ICM:gas above the bubble flows in the opposite direction to the gas be-low the bubble,and the largest velocities occur near the central axis and close to the bubble.If the flow of cool gas starts from the galaxy,to stream up the long straight side,around the curve and down the short straight side,we would expect the material in the short straight side to be flowing in the opposite direction to the long straight side,similar to that in an eddy,with blueshifted emission underneath the bubble.Indeed the emission from the short straight side is blueshifted relative to the rest of the loop:at a height of 25kpc from the nucleus,one side of the loop has a velocity of 200kms −1whereas the other side is at 60kms −1.The agreement of the velocities,morphology and the clear ghost bubble visible in the X-ray images of the Perseus clus-ter (Fabian et al.2003a),suggest the most likely dynamical model is one in which the loop and radial filament is flow-ing out of the galaxy,with the filament slightly orientated away from us.As the bubble has risen (possibly with velocities of 700kms −1Fabian et al.2003b),it has dragged up cool material from the galaxy producing the filamentary structure we observe.c0000RAS,MNRAS 000,000–000On the origin and excitation of the extended nebula surrounding NGC 12757-200-1000100200300Line-of-sight velocity (km/s)101520253035P r o j e c t e d d i s t a n c e f r o m g a l a x y c e n t r e (k p c )050100150200250300Figure 6.Left:Line-of-sight velocities along the ‘horseshoe’loop.Positive velocities are red,blue indicates negative velocities relative to galaxy.Right:Left panel shows velocities on the short straight of the loop covered by slit 3(triangles)and along the top of the loop,covered by slit 2(stars).Right panel shows data from the long straight on the right-hand side covered by slit 4(crosses)and data from slit 5that crossed through the loop (squares).Only data above dashed line is presented in image.Background image is from the data of (Conselice et al.2001).Figure 7.Line-of-sight velocities of the tangential filament running along the Northeast of NGC 1275.Purple-blue indicate blueshifted emission,yellow-red is redshifted emission and green has zero velocity relative to the central galaxy.Data from slit 5is presented in the graph with crosses.Slit 3also covered some nearby regions which have been marked by the square symbols.Luminous radial filaments extending from the galaxy to the tangential filament appear at the bottom right of the image.Background image is from the data of Conselice et al.(2001).c0000RAS,MNRAS 000,000–0008N.A.Hatch,C.S.Crawford,A.C.Fabian &R.M.Johnstone5.3Tangential filamentFig.7shows the line-of-sight velocities of the filament that appears to run tangentially along the North of the galaxy.Slit 3also covered line emitting regions situated between the tangential filament and the galaxy whose velocities have been added into Figure 7as square symbols.The velocity structure along 14.5kpc varies without large ve-locity jumps so the Eastern section of the filament appears to be a coherent structure.The emission is blueshifted to a similar velocity along ∼10kpc of its length then smoothly decreases in speed at the most Eastern end.Small scale deviation from the large scale trends are seen in excess of the error (∼1–10km s −1depending on the line intensity),suggesting some small scale random variations in velocity.Beyond 14.5kpc the emission first jumps by –80kms −1in velocity then jumps again by almost +400kms −1.The Western section of this filament is not coherent in velocity space,and the morphology suggests that slit 5is slicing across radial filaments that extend to the Northwest,in the same direction as the ‘horse-shoe’feature.The Eastern section is a puzzling structure since it is tangential,unlike the majority of the filaments.Interpretation of this region is complicated by the presence of the Northern radio lobe and complex X-ray emission.6FILAMENT LINE-WIDTHS AND VELOCITIES The FWHM (full width at half maximum)of the instrumental profile is 85km s −1(2.85pixels)at H αdetermined from nearby sky lines.Most of the material has FWHM line-widths of 50-160kms −1after correcting for the instrumental broadening (Fig.8),much greater than the thermal width of hydrogen gas at 10,000K (∼20kms −1).If the filaments represent an inflow of gas as predicted by early models of cooling flows (Fabian et al.1984)we would expect an anti-correlation between line-widths and radial distance from the nucleus and extremely large (∼500–1000kms −1)central line-widths (Heckman et al.1989).We observe a few points within 10kpc that have large (>200kms −1)line-widths.It is within 10kpc of the galactic centre that the density of the line-emitting filaments increases greatly and there are many regions where the spectra display double peaked lines indicating that the line-of-sight crosses at least two clumps of line-emitting material which have different kinematics.It is not necessary that these clumps be physi-cally close,therefore they do not imply small-scale velocity devia-tions along a single filament as observed in slit 5.Examples of such regions are shown in Fig.9,and were either resolved into two sets of lines or removed from the dataset if the result was ambiguous.However,it is likely that some of these central regions would have clumps with similar line-of-sight velocities and result in spectra with a single wide peak.Most regions with large line-widths also have a large H αsurface brightness (Fig.8),therefore it is probable that the spectra from these inner regions are caused by filaments overlapping in the same line-of-sight with slightly different veloc-ities.Beyond the inner 10kpc,the line-widths are uniformly 2–8times the thermal width of gas at 10,000K.Some radially extend-ing filaments exhibit similar line-widths along their whole length (see Fig.14in section 7).Therefore the line-widths provide no ev-idence to suggest the nebula flows into the galaxy.No line-of-sight velocity greater than 350kms −1was de-tected.This work primarily probes the outer filaments,which are likely to have small angles from the plane of the sky due to their large projected distance from the galaxy,and therefore are not ex-pected to have large line-of-sight velocities.Cigan et al.(2004)0102030405060Projected radial distance from galaxy centre (kpc)50100150200250300L i n e w i d t h (k m s -1)0102030405060H α (10-15 erg s -1cm -2arcsec -2)Figure 8.Left :Radial projection of emission line widths in kms −1based on the H αand [N II ]lines.Right :Line width verses H αsurface brightness.Most points with large line-widths (>200km s −1)are within the central few kpc and have large H αsurface brightness suggesting they may be overlap-ping filaments.Figure 9.Examples of central clumps which show double-peaked line emis-sion.These regions are all centrally located.HV denotes emission from the high-velocity system which lies infront of NGC 1275.who probe the central regions as well as the outer filaments find no velocities greater than 450kms −1.In section 5.2we argue that the ‘horseshoe’feature and the Northwest filaments covered by slit 4are very close to being in the plane of the sky.Therefore the observed line-of-sight velocity of 200kms −1,at the top of the ‘horseshoe’loop could transform to a velocity much greater than 700kms −1(expected if tilted by a conservative 75◦from the line-of-sight).This is far beyond what is observed in the rest of the dataset.It is possible that the two inner radio lobes have pushed or destroyed the filaments pointing toward our line-of-sight.7SPECTRAL FEATURESOur data probe the outer nebula that extends beyond 10kpc in de-tail for the first time.All intensity line ratios and line intensities are determined from binning the spatial dimension of the longslit spec-tra in bins of 6pixels (0.87arcsecond).A typical spectrum is shown in Fig.11and line intensity ratios are summarised in Table 2.Allc0000RAS,MNRAS 000,000–000。
NONLINEAR TIME SERIES ANALYSIS
More informationNONLINEAR TIME SERIES ANALYSISThis book represents a modern approach to time series analysis which is based onthe theory of dynamical systems.It starts from a sound outline of the underlyingtheory to arrive at very practical issues,which are illustrated using a large number ofempirical data sets taken from variousfields.This book will hence be highly usefulfor scientists and engineers from all disciplines who study time variable signals,including the earth,life and social sciences.The paradigm of deterministic chaos has influenced thinking in manyfields ofscience.Chaotic systems show rich and surprising mathematical structures.In theapplied sciences,deterministic chaos provides a striking explanation for irregulartemporal behaviour and anomalies in systems which do not seem to be inherentlystochastic.The most direct link between chaos theory and the real world is the anal-ysis of time series from real systems in terms of nonlinear dynamics.Experimentaltechnique and data analysis have seen such dramatic progress that,by now,mostfundamental properties of nonlinear dynamical systems have been observed in thelaboratory.Great efforts are being made to exploit ideas from chaos theory where-ver the data display more structure than can be captured by traditional methods.Problems of this kind are typical in biology and physiology but also in geophysics,economics and many other sciences.This revised edition has been significantly rewritten an expanded,includingseveral new chapters.In view of applications,the most relevant novelties will be thetreatment of non-stationary data sets and of nonlinear stochastic processes insidethe framework of a state space reconstruction by the method of delays.Hence,non-linear time series analysis has left the rather narrow niche of strictly deterministicsystems.Moreover,the analysis of multivariate data sets has gained more atten-tion.For a direct application of the methods of this book to the reader’s own datasets,this book closely refers to the publicly available software package TISEAN.The availability of this software will facilitate the solution of the exercises,so thatreaders now can easily gain their own experience with the analysis of data sets.Holger Kantz,born in November1960,received his diploma in physics fromthe University of Wuppertal in January1986with a thesis on transient chaos.InJanuary1989he obtained his Ph.D.in theoretical physics from the same place,having worked under the supervision of Peter Grassberger on Hamiltonian many-particle dynamics.During his postdoctoral time,he spent one year on a Marie Curiefellowship of the European Union at the physics department of the University ofMore informationFlorence in Italy.In January1995he took up an appointment at the newly foundedMax Planck Institute for the Physics of Complex Systems in Dresden,where heestablished the research group‘Nonlinear Dynamics and Time Series Analysis’.In1996he received his venia legendi and in2002he became adjunct professorin theoretical physics at Wuppertal University.In addition to time series analysis,he works on low-and high-dimensional nonlinear dynamics and its applications.More recently,he has been trying to bridge the gap between dynamics and statis-tical physics.He has(co-)authored more than75peer-reviewed articles in scien-tific journals and holds two international patents.For up-to-date information seehttp://www.mpipks-dresden.mpg.de/mpi-doc/kantzgruppe.html.Thomas Schreiber,born1963,did his diploma work with Peter Grassberger atWuppertal University on phase transitions and information transport in spatio-temporal chaos.He joined the chaos group of Predrag Cvitanovi´c at the Niels BohrInstitute in Copenhagen to study periodic orbit theory of diffusion and anomaloustransport.There he also developed a strong interest in real-world applications ofchaos theory,leading to his Ph.D.thesis on nonlinear time series analysis(Univer-sity of Wuppertal,1994).As a research assistant at Wuppertal University and duringseveral extended appointments at the Max Planck Institute for the Physics of Com-plex Systems in Dresden he published numerous research articles on time seriesmethods and applications ranging from physiology to the stock market.His habil-itation thesis(University of Wuppertal)appeared as a review in Physics Reportsin1999.Thomas Schreiber has extensive experience teaching nonlinear dynamicsto students and experts from variousfields and at all levels.Recently,he has leftacademia to undertake industrial research.NONLINEAR TIME SERIES ANALYSIS HOLGER KANTZ AND THOMAS SCHREIBERMax Planck Institute for the Physics of Complex Systems,DresdenMore informationMore informationpublished by the press syndicate of the university of cambridgeThe Pitt Building,Trumpington Street,Cambridge,United Kingdomcambridge university pressThe Edinburgh Building,Cambridge CB22RU,UK40West20th Street,New York,NY10011–4211,USA477Williamstown Road,Port Melbourne,VIC3207,AustraliaRuiz de Alarc´o n13,28014Madrid,SpainDock House,The Waterfront,Cape Town8001,South AfricaC Holger Kantz and Thomas Schreiber,2000,2003This book is in copyright.Subject to statutory exceptionand to the provisions of relevant collective licensing agreements,no reproduction of any part may take place withoutthe written permission of Cambridge University Press.First published2000Second edition published2003Printed in the United Kingdom at the University Press,CambridgeTypeface Times11/14pt.System L A T E X2ε[tb]A catalogue record for this book is available from the British LibraryLibrary of Congress Cataloguing in Publication dataKantz,Holger,1960–Nonlinear time series analysis/Holger Kantz and Thomas Schreiber.–[2nd ed.].p.cm.Includes bibliographical references and index.ISBN0521821509–ISBN0521529026(paperback)1.Time-series analysis.2.Nonlinear theories.I.Schreiber,Thomas,1963–II.TitleQA280.K3552003519.5 5–dc212003044031ISBN0521821509hardbackISBN0521529026paperbackThe publisher has used its best endeavours to ensure that the URLs for external websites referred to in this bookare correct and active at the time of going to press.However,the publisher has no responsibility for the websites and can make no guarantee that a site will remain live or that the content is or will remain appropriate.More informationContentsPreface to thefirst edition page xiPreface to the second edition xiiiAcknowledgements xvI Basic topics11Introduction:why nonlinear methods?32Linear tools and general considerations132.1Stationarity and sampling132.2Testing for stationarity152.3Linear correlations and the power spectrum182.3.1Stationarity and the low-frequency component in thepower spectrum232.4Linearfilters242.5Linear predictions273Phase space methods303.1Determinism:uniqueness in phase space303.2Delay reconstruction353.3Finding a good embedding363.3.1False neighbours373.3.2The time lag393.4Visual inspection of data393.5Poincar´e surface of section413.6Recurrence plots434Determinism and predictability484.1Sources of predictability484.2Simple nonlinear prediction algorithm504.3Verification of successful prediction534.4Cross-prediction errors:probing stationarity564.5Simple nonlinear noise reduction58vMore informationvi Contents5Instability:Lyapunov exponents655.1Sensitive dependence on initial conditions655.2Exponential divergence665.3Measuring the maximal exponent from data696Self-similarity:dimensions756.1Attractor geometry and fractals756.2Correlation dimension776.3Correlation sum from a time series786.4Interpretation and pitfalls826.5Temporal correlations,non-stationarity,and space timeseparation plots876.6Practical considerations916.7A useful application:determination of the noise level using thecorrelation integral926.8Multi-scale or self-similar signals956.8.1Scaling laws966.8.2Detrendedfluctuation analysis1007Using nonlinear methods when determinism is weak1057.1Testing for nonlinearity with surrogate data1077.1.1The null hypothesis1097.1.2How to make surrogate data sets1107.1.3Which statistics to use1137.1.4What can go wrong1157.1.5What we have learned1177.2Nonlinear statistics for system discrimination1187.3Extracting qualitative information from a time series1218Selected nonlinear phenomena1268.1Robustness and limit cycles1268.2Coexistence of attractors1288.3Transients1288.4Intermittency1298.5Structural stability1338.6Bifurcations1358.7Quasi-periodicity139II Advanced topics1419Advanced embedding methods1439.1Embedding theorems1439.1.1Whitney’s embedding theorem1449.1.2Takens’s delay embedding theorem1469.2The time lag148More informationContents vii9.3Filtered delay embeddings1529.3.1Derivative coordinates1529.3.2Principal component analysis1549.4Fluctuating time intervals1589.5Multichannel measurements1599.5.1Equivalent variables at different positions1609.5.2Variables with different physical meanings1619.5.3Distributed systems1619.6Embedding of interspike intervals1629.7High dimensional chaos and the limitations of the time delayembedding1659.8Embedding for systems with time delayed feedback17110Chaotic data and noise17410.1Measurement noise and dynamical noise17410.2Effects of noise17510.3Nonlinear noise reduction17810.3.1Noise reduction by gradient descent17910.3.2Local projective noise reduction18010.3.3Implementation of locally projective noise reduction18310.3.4How much noise is taken out?18610.3.5Consistency tests19110.4An application:foetal ECG extraction19311More about invariant quantities19711.1Ergodicity and strange attractors19711.2Lyapunov exponents II19911.2.1The spectrum of Lyapunov exponents and invariantmanifolds20011.2.2Flows versus maps20211.2.3Tangent space method20311.2.4Spurious exponents20511.2.5Almost two dimensionalflows21111.3Dimensions II21211.3.1Generalised dimensions,multi-fractals21311.3.2Information dimension from a time series21511.4Entropies21711.4.1Chaos and theflow of information21711.4.2Entropies of a static distribution21811.4.3The Kolmogorov–Sinai entropy22011.4.4The -entropy per unit time22211.4.5Entropies from time series data226More informationviii Contents11.5How things are related22911.5.1Pesin’s identity22911.5.2Kaplan–Yorke conjecture23112Modelling and forecasting23412.1Linear stochastic models andfilters23612.1.1Linearfilters23712.1.2Nonlinearfilters23912.2Deterministic dynamics24012.3Local methods in phase space24112.3.1Almost model free methods24112.3.2Local linearfits24212.4Global nonlinear models24412.4.1Polynomials24412.4.2Radial basis functions24512.4.3Neural networks24612.4.4What to do in practice24812.5Improved cost functions24912.5.1Overfitting and model costs24912.5.2The errors-in-variables problem25112.5.3Modelling versus prediction25312.6Model verification25312.7Nonlinear stochastic processes from data25612.7.1Fokker–Planck equations from data25712.7.2Markov chains in embedding space25912.7.3No embedding theorem for Markov chains26012.7.4Predictions for Markov chain data26112.7.5Modelling Markov chain data26212.7.6Choosing embedding parameters for Markov chains26312.7.7Application:prediction of surface wind velocities26412.8Predicting prediction errors26712.8.1Predictability map26712.8.2Individual error prediction26812.9Multi-step predictions versus iterated one-step predictions27113Non-stationary signals27513.1Detecting non-stationarity27613.1.1Making non-stationary data stationary27913.2Over-embedding28013.2.1Deterministic systems with parameter drift28013.2.2Markov chain with parameter drift28113.2.3Data analysis in over-embedding spaces283More informationContents ix13.2.4Application:noise reduction for human voice28613.3Parameter spaces from data28814Coupling and synchronisation of nonlinear systems29214.1Measures for interdependence29214.2Transfer entropy29714.3Synchronisation29915Chaos control30415.1Unstable periodic orbits and their invariant manifolds30615.1.1Locating periodic orbits30615.1.2Stable/unstable manifolds from data31215.2OGY-control and derivates31315.3Variants of OGY-control31615.4Delayed feedback31715.5Tracking31815.6Related aspects319A Using the TISEAN programs321A.1Information relevant to most of the routines322A.1.1Efficient neighbour searching322A.1.2Re-occurring command options325A.2Second-order statistics and linear models326A.3Phase space tools327A.4Prediction and modelling329A.4.1Locally constant predictor329A.4.2Locally linear prediction329A.4.3Global nonlinear models330A.5Lyapunov exponents331A.6Dimensions and entropies331A.6.1The correlation sum331A.6.2Information dimension,fixed mass algorithm332A.6.3Entropies333A.7Surrogate data and test statistics334A.8Noise reduction335A.9Finding unstable periodic orbits336A.10Multivariate data336B Description of the experimental data sets338B.1Lorenz-like chaos in an NH3laser338B.2Chaos in a periodically modulated NMR laser340B.3Vibrating string342B.4Taylor–Couetteflow342B.5Multichannel physiological data343More informationx ContentsB.6Heart rate during atrialfibrillation343B.7Human electrocardiogram(ECG)344B.8Phonation data345B.9Postural control data345B.10Autonomous CO2laser with feedback345B.11Nonlinear electric resonance circuit346B.12Frequency doubling solid state laser348B.13Surface wind velocities349References350Index365More informationPreface to thefirst editionThe paradigm of deterministic chaos has influenced thinking in manyfields of sci-ence.As mathematical objects,chaotic systems show rich and surprising structures.Most appealing for researchers in the applied sciences is the fact that determinis-tic chaos provides a striking explanation for irregular behaviour and anomalies insystems which do not seem to be inherently stochastic.The most direct link between chaos theory and the real world is the analysis oftime series from real systems in terms of nonlinear dynamics.On the one hand,experimental technique and data analysis have seen such dramatic progress that,by now,most fundamental properties of nonlinear dynamical systems have beenobserved in the laboratory.On the other hand,great efforts are being made to exploitideas from chaos theory in cases where the system is not necessarily deterministicbut the data displays more structure than can be captured by traditional methods.Problems of this kind are typical in biology and physiology but also in geophysics,economics,and many other sciences.In all thesefields,even simple models,be they microscopic or phenomenological,can create extremely complicated dynamics.How can one verify that one’s model isa good counterpart to the equally complicated signal that one receives from nature?Very often,good models are lacking and one has to study the system just from theobservations made in a single time series,which is the case for most non-laboratorysystems in particular.The theory of nonlinear dynamical systems provides new toolsand quantities for the characterisation of irregular time series data.The scope ofthese methods ranges from invariants such as Lyapunov exponents and dimensionswhich yield an accurate description of the structure of a system(provided thedata are of high quality)to statistical techniques which allow for classification anddiagnosis even in situations where determinism is almost lacking.This book provides the experimental researcher in nonlinear dynamics with meth-ods for processing,enhancing,and analysing the measured signals.The theorist willbe offered discussions about the practical applicability of mathematical results.ThexiMore informationxii Preface to thefirst editiontime series analyst in economics,meteorology,and otherfields willfind inspira-tion for the development of new prediction algorithms.Some of the techniquespresented here have also been considered as possible diagnostic tools in clinical re-search.We will adopt a critical but constructive point of view,pointing out ways ofobtaining more meaningful results with limited data.We hope that everybody whohas a time series problem which cannot be solved by traditional,linear methodswillfind inspiring material in this book.Dresden and WuppertalNovember1996More informationPreface to the second editionIn afield as dynamic as nonlinear science,new ideas,methods and experimentsemerge constantly and the focus of interest shifts accordingly.There is a continuousstream of new results,and existing knowledge is seen from a different angle aftervery few years.Five years after thefirst edition of“Nonlinear Time Series Analysis”we feel that thefield has matured in a way that deserves being reflected in a secondedition.The modification that is most immediately visible is that the program listingshave been be replaced by a thorough discussion of the publicly available softwareTISEAN.Already a few months after thefirst edition appeared,it became clearthat most users would need something more convenient to use than the bare libraryroutines printed in the book.Thus,together with Rainer Hegger we prepared stand-alone routines based on the book but with input/output functionality and advancedfeatures.Thefirst public release was made available in1998and subsequent releasesare in widespread use now.Today,TISEAN is a mature piece of software thatcovers much more than the programs we gave in thefirst edition.Now,readerscan immediately apply most methods studied in the book on their own data usingTISEAN programs.By replacing the somewhat terse program listings by minuteinstructions of the proper use of the TISEAN routines,the link between book andsoftware is strengthened,supposedly to the benefit of the readers and users.Hencewe recommend a download and installation of the package,such that the exercisescan be readily done by help of these ready-to-use routines.The current edition has be extended in view of enlarging the class of data sets to betreated.The core idea of phase space reconstruction was inspired by the analysis ofdeterministic chaotic data.In contrast to many expectations,purely deterministicand low-dimensional data are rare,and most data fromfield measurements areevidently of different nature.Hence,it was an effort of our scientific work over thepast years,and it was a guiding concept for the revision of this book,to explore thepossibilities to treat other than purely deterministic data sets.xiiiMore informationxiv Preface to the second editionThere is a whole new chapter on non-stationary time series.While detectingnon-stationarity is still briefly discussed early on in the book,methods to deal withmanifestly non-stationary sequences are described in some detail in the secondpart.As an illustration,a data source of lasting interest,human speech,is used.Also,a new chapter deals with concepts of synchrony between systems,linear andnonlinear correlations,information transfer,and phase synchronisation.Recent attempts on modelling nonlinear stochastic processes are discussed inChapter12.The theoretical framework forfitting Fokker–Planck equations to datawill be reviewed and evaluated.While Chapter9presents some progress that hasbeen made in modelling input–output systems with stochastic but observed inputand on the embedding of time delayed feedback systems,the chapter on mod-elling considers a data driven phase space approach towards Markov chains.Windspeed measurements are used as data which are best considered to be of nonlinearstochastic nature despite the fact that a physically adequate mathematical model isthe deterministic Navier–Stokes equation.In the chapter on invariant quantities,new material on entropy has been included,mainly on the -and continuous entropies.Estimation problems for stochastic ver-sus deterministic data and data with multiple length and time scales are discussed.Since more and more experiments now yield good multivariate data,alternativesto time delay embedding using multiple probe measurements are considered at var-ious places in the text.This new development is also reflected in the functionalityof the TISEAN programs.A new multivariate data set from a nonlinear semicon-ductor electronic circuit is introduced and used in several places.In particular,adifferential equation has been successfully established for this system by analysingthe data set.Among other smaller rearrangements,the material from the former chapter“Other selected topics”,has been relocated to places in the text where a connectioncan be made more naturally.High dimensional and spatio-temporal data is now dis-cussed in the context of embedding.We discuss multi-scale and self-similar signalsnow in a more appropriate way right after fractal sets,and include recent techniquesto analyse power law correlations,for example detrendedfluctuation analysis.Of course,many new publications have appeared since1997which are potentiallyrelevant to the scope of this book.At least two new monographs are concerned withthe same topic and a number of review articles.The bibliography has been updatedbut remains a selection not unaffected by personal preferences.We hope that the extended book will prove its usefulness in many applicationsof the methods and further stimulate thefield of time series analysis.DresdenDecember2002More informationAcknowledgementsIf there is any feature of this book that we are proud of,it is the fact that almost allthe methods are illustrated with real,experimental data.However,this is anythingbut our own achievement–we exploited other people’s work.Thus we are deeplyindebted to the experimental groups who supplied data sets and granted permissionto use them in this book.The production of every one of these data sets requiredskills,experience,and equipment that we ourselves do not have,not forgetting thehours and hours of work spent in the laboratory.We appreciate the generosity ofthe following experimental groups:NMR laser.Our contact persons at the Institute for Physics at Z¨u rich University were Leci Flepp and Joe Simonet;the head of the experimental group is E.Brun.(See AppendixB.2.)Vibrating string.Data were provided by Tim Molteno and Nick Tufillaro,Otago University, Dunedin,New Zealand.(See Appendix B.3.)Taylor–Couetteflow.The experiment was carried out at the Institute for Applied Physics at Kiel University by Thorsten Buzug and Gerd Pfister.(See Appendix B.4.) Atrialfibrillation.This data set is taken from the MIT-BIH Arrhythmia Database,collected by G.B.Moody and R.Mark at Beth Israel Hospital in Boston.(See Appendix B.6.) Human ECG.The ECG recordings we used were taken by Petr Saparin at Saratov State University.(See Appendix B.7.)Foetal ECG.We used noninvasively recorded(human)foetal ECGs taken by John F.Hofmeister as the Department of Obstetrics and Gynecology,University of Colorado,Denver CO.(See Appendix B.7.)Phonation data.This data set was made available by Hanspeter Herzel at the Technical University in Berlin.(See Appendix B.8.)Human posture data.The time series was provided by Steven Boker and Bennett Bertenthal at the Department of Psychology,University of Virginia,Charlottesville V A.(SeeAppendix B.9.)xvMore informationxvi AcknowledgementsAutonomous CO2laser with feedback.The data were taken by Riccardo Meucci and Marco Ciofini at the INO in Firenze,Italy.(See Appendix B.10.)Nonlinear electric resonance circuit.The experiment was designed and operated by M.Diestelhorst at the University of Halle,Germany.(See Appendix B.11.)Nd:YAG laser.The data we use were recorded in the University of Oldenburg,where we wish to thank Achim Kittel,Falk Lange,Tobias Letz,and J¨u rgen Parisi.(See AppendixB.12.)We used the following data sets published for the Santa Fe Institute Time SeriesContest,which was organised by Neil Gershenfeld and Andreas Weigend in1991:NH3laser.We used data set A and its continuation,which was published after the contest was closed.The data was supplied by U.H¨u bner,N.B.Abraham,and C.O.Weiss.(SeeAppendix B.1.)Human breath rate.The data we used is part of data set B of the contest.It was submitted by Ari Goldberger and coworkers.(See Appendix B.5.)During the composition of the text we asked various people to read all or part of themanuscript.The responses ranged from general encouragement to detailed technicalcomments.In particular we thank Peter Grassberger,James Theiler,Daniel Kaplan,Ulrich Parlitz,and Martin Wiesenfeld for their helpful remarks.Members of ourresearch groups who either contributed by joint work to our experience and knowl-edge or who volunteered to check the correctness of the text are Rainer Hegger,Andreas Schmitz,Marcus Richter,Mario Ragwitz,Frank Schm¨u ser,RathinaswamyBhavanan Govindan,and Sharon Sessions.We have also considerably profited fromcomments and remarks of the readers of thefirst edition of the book.Their effortin writing to us is gratefully appreciated.Last but not least we acknowledge the encouragement and support by SimonCapelin from Cambridge University Press and the excellent help in questions ofstyle and English grammar by Sheila Shepherd.。
斯仑贝谢所有测井曲线英文名称解释
˹Âر´Ð»ËùÓв⾮ÇúÏßÓ¢ÎÄÃû³Æ½âÊÍ OCEAN DRILLING PROGRAMACRONYMS USED FOR WIRELINE SCHLUMBERGER TOOLSACT Aluminum Clay ToolAMS Auxiliary Measurement SondeAPS Accelerator Porosity SondeARI Azimuthal Resistivity ImagerASI Array Sonic ImagerBGKT Vertical Seismic Profile ToolBHC Borehole Compensated Sonic ToolBHTV Borehole TeleviewerCBL Casing Bond LogCNT Compensated Neutron ToolDIT Dual Induction ToolDLL Dual LaterologDSI Dipole Sonic ImagerFMS Formation MicroScannerGHMT Geologic High Resolution Magnetic Tool GPIT General Purpose Inclinometer ToolGR Natural Gamma RayGST Induced Gamma Ray Spectrometry ToolHLDS Hostile Environment Lithodensity Sonde HLDT Hostile Environment Lithodensity Tool HNGS Hostile Environment Gamma Ray Sonde LDT Lithodensity ToolLSS Long Spacing Sonic ToolMCD Mechanical Caliper DeviceNGT Natural Gamma Ray Spectrometry Tool NMRT Nuclear Resonance Magnetic ToolQSST Inline Checkshot ToolSDT Digital Sonic ToolSGT Scintillation Gamma Ray ToolSUMT Susceptibility Magnetic ToolUBI Ultrasonic Borehole ImagerVSI Vertical Seismic ImagerWST Well Seismic ToolWST-3 3-Components Well Seismic ToolOCEAN DRILLING PROGRAMACRONYMS USED FOR LWD SCHLUMBERGER TOOLS ADN Azimuthal Density-NeutronCDN Compensated Density-NeutronCDR Compensated Dual ResistivityISONIC Ideal Sonic-While-DrillingNMR Nuclear Magnetic ResonanceRAB Resistivity-at-the-BitOCEAN DRILLING PROGRAMACRONYMS USED FOR NON-SCHLUMBERGER SPECIALTY TOOLSMCS Multichannel Sonic ToolMGT Multisensor Gamma ToolSST Shear Sonic ToolTAP Temperature-Acceleration-Pressure ToolTLT Temperature Logging ToolOCEAN DRILLING PROGRAMACRONYMS AND UNITS USED FOR WIRELINE SCHLUMBERGER LOGSAFEC APS Far Detector Counts (cps)ANEC APS Near Detector Counts (cps)AX Acceleration X Axis (ft/s2)AY Acceleration Y Axis (ft/s2)AZ Acceleration Z Axis (ft/s2)AZIM Constant Azimuth for Deviation Correction (deg) APLC APS Near/Array Limestone Porosity Corrected (%)C1 FMS Caliper 1 (in)C2 FMS Caliper 2 (in)CALI Caliper (in)CFEC Corrected Far Epithermal Counts (cps)CFTC Corrected Far Thermal Counts (cps)CGR Computed (Th+K) Gamma Ray (API units)CHR2 Peak Coherence, Receiver Array, Upper DipoleCHRP Compressional Peak Coherence, Receiver Array, P&S CHRS Shear Peak Coherence, Receiver Array, P&SCHTP Compressional Peak Coherence, Transmitter Array, P&S CHTS Shear Peak Coherence, Transmitter Array, P&SCNEC Corrected Near Epithermal Counts (cps)CNTC Corrected Near Thermal Counts (cps)CS Cable Speed (m/hr)CVEL Compressional Velocity (km/s)DATN Discriminated Attenuation (db/m)DBI Discriminated Bond IndexDEVI Hole Deviation (degrees)DF Drilling Force (lbf)DIFF Difference Between MEAN and MEDIAN in Delta-Time Proc. (microsec/ft) DRH HLDS Bulk Density Correction (g/cm3)DRHO Bulk Density Correction (g/cm3)DT Short Spacing Delta-Time (10'-8' spacing; microsec/ft)DT1 Delta-Time Shear, Lower Dipole (microsec/ft)DT2 Delta-Time Shear, Upper Dipole (microsec/ft)DT4P Delta- Time Compressional, P&S (microsec/ft)DT4S Delta- Time Shear, P&S (microsec/ft))DT1R Delta- Time Shear, Receiver Array, Lower Dipole (microsec/ft)DT2R Delta- Time Shear, Receiver Array, Upper Dipole (microsec/ft)DT1T Delta-Time Shear, Transmitter Array, Lower Dipole (microsec/ft)DT2T Delta-Time Shear, Transmitter Array, Upper Dipole (microsec/ft) DTCO Delta- Time Compressional (microsec/ft)DTL Long Spacing Delta-Time (12'-10' spacing; microsec/ft)DTLF Long Spacing Delta-Time (12'-10' spacing; microsec/ft)DTLN Short Spacing Delta-Time (10'-8' spacing; microsec/ftDTRP Delta-Time Compressional, Receiver Array, P&S (microsec/ft)DTRS Delta-Time Shear, Receiver Array, P&S (microsec/ft)DTSM Delta-Time Shear (microsec/ft)DTST Delta-Time Stoneley (microsec/ft)DTTP Delta-Time Compressional, Transmitter Array, P&S (microsec/ft)DTTS Delta-Time Shear, Transmitter Array, P&S (microsec/ft)ECGR Environmentally Corrected Gamma Ray (API units)EHGR Environmentally Corrected High Resolution Gamma Ray (API units) ENPH Epithermal Neutron Porosity (%)ENRA Epithermal Neutron RatioETIM Elapsed Time (sec)FINC Magnetic Field Inclination (degrees)FNOR Magnetic Field Total Moment (oersted)FX Magnetic Field on X Axis (oersted)FY Magnetic Field on Y Axis (oersted)FZ Magnetic Field on Z Axis (oersted)GR Natural Gamma Ray (API units)HALC High Res. Near/Array Limestone Porosity Corrected (%)HAZI Hole Azimuth (degrees)HBDC High Res. Bulk Density Correction (g/cm3)HBHK HNGS Borehole Potassium (%)HCFT High Resolution Corrected Far Thermal Counts (cps)HCGR HNGS Computed Gamma Ray (API units)HCNT High Resolution Corrected Near Thermal Counts (cps)HDEB High Res. Enhanced Bulk Density (g/cm3)HDRH High Resolution Density Correction (g/cm3)HFEC High Res. Far Detector Counts (cps)HFK HNGS Formation Potassium (%)HFLC High Res. Near/Far Limestone Porosity Corrected (%)HEGR Environmentally Corrected High Resolution Natural Gamma Ray (API units) HGR High Resolution Natural Gamma Ray (API units)HLCA High Res. Caliper (inHLEF High Res. Long-spaced Photoelectric Effect (barns/e-)HNEC High Res. Near Detector Counts (cps)HNPO High Resolution Enhanced Thermal Nutron Porosity (%)HNRH High Resolution Bulk Density (g/cm3)HPEF High Resolution Photoelectric Effect (barns/e-)HRHO High Resolution Bulk Density (g/cm3)HROM High Res. Corrected Bulk Density (g/cm3)HSGR HNGS Standard (total) Gamma Ray (API units)HSIG High Res. Formation Capture Cross Section (capture units)HSTO High Res. Computed Standoff (in)HTHO HNGS Thorium (ppm)HTNP High Resolution Thermal Neutron Porosity (%)HURA HNGS Uranium (ppm)IDPH Phasor Deep Induction (ohmm)IIR Iron Indicator Ratio [CFE/(CCA+CSI)]ILD Deep Resistivity (ohmm)ILM Medium Resistivity (ohmm)IMPH Phasor Medium Induction (ohmm)ITT Integrated Transit Time (s)LCAL HLDS Caliper (in)LIR Lithology Indicator Ratio [CSI/(CCA+CSI)]LLD Laterolog Deep (ohmm)LLS Laterolog Shallow (ohmm)LTT1 Transit Time (10'; microsec)LTT2 Transit Time (8'; microsec)LTT3 Transit Time (12'; microsec)LTT4 Transit Time (10'; microsec)MAGB Earth's Magnetic Field (nTes)MAGC Earth Conductivity (ppm)MAGS Magnetic Susceptibility (ppm)MEDIAN Median Delta-T Recomputed (microsec/ft)MEAN Mean Delta-T Recomputed (microsec/ft)NATN Near Pseudo-Attenuation (db/m)NMST Magnetometer Temperature (degC)NMSV Magnetometer Signal Level (V)NPHI Neutron Porosity (%)NRHB LDS Bulk Density (g/cm3)P1AZ Pad 1 Azimuth (degrees)PEF Photoelectric Effect (barns/e-)PEFL LDS Long-spaced Photoelectric Effect (barns/e-)PIR Porosity Indicator Ratio [CHY/(CCA+CSI)]POTA Potassium (%)RB Pad 1 Relative Bearing (degrees)RHL LDS Long-spaced Bulk Density (g/cm3)RHOB Bulk Density (g/cm3)RHOM HLDS Corrected Bulk Density (g/cm3)RMGS Low Resolution Susceptibility (ppm)SFLU Spherically Focused Log (ohmm)SGR Total Gamma Ray (API units)SIGF APS Formation Capture Cross Section (capture units) SP Spontaneous Potential (mV)STOF APS Computed Standoff (in)SURT Receiver Coil Temperature (degC)SVEL Shear Velocity (km/s)SXRT NMRS differential Temperature (degC)TENS Tension (lb)THOR Thorium (ppm)TNRA Thermal Neutron RatioTT1 Transit Time (10' spacing; microsec)TT2 Transit Time (8' spacing; microsec)TT3 Transit Time (12' spacing; microsec)TT4 Transit Time (10' spacing; microsec)URAN Uranium (ppm)V4P Compressional Velocity, from DT4P (P&S; km/s)V4S Shear Velocity, from DT4S (P&S; km/s)VELP Compressional Velocity (processed from waveforms; km/s)VELS Shear Velocity (processed from waveforms; km/s)VP1 Compressional Velocity, from DT, DTLN, or MEAN (km/s)VP2 Compressional Velocity, from DTL, DTLF, or MEDIAN (km/s)VCO Compressional Velocity, from DTCO (km/s)VS Shear Velocity, from DTSM (km/s)VST Stonely Velocity, from DTST km/s)VS1 Shear Velocity, from DT1 (Lower Dipole; km/s)VS2 Shear Velocity, from DT2 (Upper Dipole; km/s)VRP Compressional Velocity, from DTRP (Receiver Array, P&S; km/s)VRS Shear Velocity, from DTRS (Receiver Array, P&S; km/s)VS1R Shear Velocity, from DT1R (Receiver Array, Lower Dipole; km/s)VS2R Shear Velocity, from DT2R (Receiver Array, Upper Dipole; km/s)VS1T Shear Velocity, from DT1T (Transmitter Array, Lower Dipole; km/s) VS2T Shear Velocity, from DT2T (Transmitter Array, Upper Dipole; km/s) VTP Compressional Velocity, from DTTP (Transmitter Array, P&S; km/s) VTS Shear Velocity, from DTTS (Transmitter Array, P&S; km/s)#POINTS Number of Transmitter-Receiver Pairs Used in Sonic Processing W1NG NGT Window 1 counts (cps)W2NG NGT Window 2 counts (cps)W3NG NGT Window 3 counts (cps)W4NG NGT Window 4 counts (cps)W5NG NGT Window 5 counts (cps)OCEAN DRILLING PROGRAMACRONYMS AND UNITS USED FOR LWD SCHLUMBERGER LOGSAT1F Attenuation Resistivity (1 ft resolution; ohmm) AT2F Attenuation Resistivity (2 ft resolution; ohmm) AT3F Attenuation Resistivity (3 ft resolution; ohmm) AT4F Attenuation Resistivity (4 ft resolution; ohmm) AT5F Attenuation Resistivity (5 ft resolution; ohmm) ATR Attenuation Resistivity (deep; ohmm)BFV Bound Fluid Volume (%)B1TM RAB Shallow Resistivity Time after Bit (s)B2TM RAB Medium Resistivity Time after Bit (s)B3TM RAB Deep Resistivity Time after Bit (s)BDAV Deep Resistivity Average (ohmm)BMAV Medium Resistivity Average (ohmm)BSAV Shallow Resistivity Average (ohmm)CGR Computed (Th+K) Gamma Ray (API units)DCAL Differential Caliper (in)DROR Correction for CDN rotational density (g/cm3).DRRT Correction for ADN rotational density (g/cm3).DTAB AND or CDN Density Time after Bit (hr)FFV Free Fluid Volume (%)GR Gamma Ray (API Units)GR7 Sum Gamma Ray Windows GRW7+GRW8+GRW9-Equivalent to Wireline NGT window 5 (cps) GRW3 Gamma Ray Window 3 counts (cps)-Equivalent to Wireline NGT window 1GRW4 Gamma Ray Window 4 counts (cps)-Equivalent to Wireline NGT window 2GRW5 Gamma Ray Window 5 counts (cps)-Equivalent to Wireline NGT window 3GRW6 Gamma Ray Window 6 counts (cps)-Equivalent to Wireline NGT window 4GRW7 Gamma Ray Window 7 counts (cps)GRW8 Gamma Ray Window 8 counts (cps)GRW9 Gamma Ray Window 9 counts (cps)GTIM CDR Gamma Ray Time after Bit (s)GRTK RAB Gamma Ray Time after Bit (s)HEF1 Far He Bank 1 counts (cps)HEF2 Far He Bank 2 counts (cps)HEF3 Far He Bank 3 counts (cps)HEF4 Far He Bank 4 counts (cps)HEN1 Near He Bank 1 counts (cps)HEN2 Near He Bank 2 counts (cps)HEN3 Near He Bank 3 counts (cps)HEN4 Near He Bank 4 counts (cps)MRP Magnetic Resonance PorosityNTAB ADN or CDN Neutron Time after Bit (hr)PEF Photoelectric Effect (barns/e-)POTA Potassium (%) ROPE Rate of Penetration (ft/hr)PS1F Phase Shift Resistivity (1 ft resolution; ohmm)PS2F Phase Shift Resistivity (2 ft resolution; ohmm)PS3F Phase Shift Resistivity (3 ft resolution; ohmm)PS4F Phase Shift Resistivity (4 ft resolution; ohmm)PS5F Phase Shift Resistivity (5 ft resolution; ohmm)PSR Phase Shift Resistivity (shallow; ohmm)RBIT Bit Resistivity (ohmm)RBTM RAB Resistivity Time After Bit (s)RING Ring Resistivity (ohmm)ROMT Max. Density Total (g/cm3) from rotational processing ROP Rate of Penetration (m/hr)ROP1 Rate of Penetration, average over last 1 ft (m/hr). ROP5 Rate of Penetration, average over last 5 ft (m/hr) ROPE Rate of Penetration, averaged over last 5 ft (ft/hr) RPM RAB Tool Rotation Speed (rpm)RTIM CDR or RAB Resistivity Time after Bit (hr)SGR Total Gamma Ray (API units)T2 T2 Distribution (%)T2LM T2 Logarithmic Mean (ms)THOR Thorium (ppm)TNPH Thermal Neutron Porosity (%)TNRA Thermal RatioURAN Uranium (ppm)OCEAN DRILLING PROGRAMADDITIONAL ACRONYMS AND UNITS(PROCESSED LOGS FROM GEOCHEMICAL TOOL STRING)AL2O3 Computed Al2O3 (dry weight %)AL2O3MIN Computed Al2O3 Standard Deviation (dry weight %) AL2O3MAX Computed Al2O3 Standard Deviation (dry weight %) CAO Computed CaO (dry weight %)CAOMIN Computed CaO Standard Deviation (dry weight %) CAOMAX Computed CaO Standard Deviation (dry weight %) CACO3 Computed CaCO3 (dry weight %)CACO3MIN Computed CaCO3 Standard Deviation (dry weight %)CACO3MAX Computed CaCO3 Standard Deviation (dry weight %) CCA Calcium Yield (decimal fraction)CCHL Chlorine Yield (decimal fraction)CFE Iron Yield (decimal fraction)CGD Gadolinium Yield (decimal fraction)CHY Hydrogen Yield (decimal fraction)CK Potassium Yield (decimal fraction)CSI Silicon Yield (decimal fraction)CSIG Capture Cross Section (capture units)CSUL Sulfur Yield (decimal fraction)CTB Background Yield (decimal fraction)CTI Titanium Yield (decimal fraction)FACT Quality Control CurveFEO Computed FeO (dry weight %)FEOMIN Computed FeO Standard Deviation (dry weight %) FEOMAX Computed FeO Standard Deviation (dry weight %) FEO* Computed FeO* (dry weight %)FEO*MIN Computed FeO* Standard Deviation (dry weight %) FEO*MAX Computed FeO* Standard Deviation (dry weight %) FE2O3 Computed Fe2O3 (dry weight %)FE2O3MIN Computed Fe2O3 Standard Deviation (dry weight %) FE2O3MAX Computed Fe2O3 Standard Deviation (dry weight %)GD Computed Gadolinium (dry weight %)GDMIN Computed Gadolinium Standard Deviation (dry weight %) GDMAX Computed Gadolinium Standard Deviation (dry weight %) K2O Computed K2O (dry weight %)K2OMIN Computed K2O Standard Deviation (dry weight %)K2OMAX Computed K2O Standard Deviation (dry weight %)MGO Computed MgO (dry weight %)MGOMIN Computed MgO Standard Deviation (dry weight %) MGOMAX Computed MgO Standard Deviation (dry weight %)S Computed Sulfur (dry weight %)SMIN Computed Sulfur Standard Deviation (dry weight %) SMAX Computed Sulfur Standard Deviation (dry weight %)SIO2 Computed SiO2 (dry weight %)SIO2MIN Computed SiO2 Standard Deviation (dry weight %)SIO2MAX Computed SiO2 Standard Deviation (dry weight %) THORMIN Computed Thorium Standard Deviation (ppm)THORMAX Computed Thorium Standard Deviation (ppm)TIO2 Computed TiO2 (dry weight %)TIO2MIN Computed TiO2 Standard Deviation (dry weight %)TIO2MAX Computed TiO2 Standard Deviation (dry weight %) URANMIN Computed Uranium Standard Deviation (ppm)URANMAX Computed Uranium Standard Deviation (ppm)VARCA Variable CaCO3/CaO calcium carbonate/oxide factor。
二尖瓣反流量简化评估法英语
二尖瓣反流量简化评估法英语Simplified Echocardiographic Assessment of Mitral Regurgitation.Mitral regurgitation (MR) is the backward flow of blood from the left ventricle into the left atrium during systole, which can be caused by various structural or functional abnormalities of the mitral valve. Echocardiography is a widely used imaging modality for the assessment of MR, and the simplified echocardiographic assessment of MR aims to provide a quick and reliable method to estimate theseverity of MR.Methods for Assessing Mitral Regurgitation.There are several methods for assessing MR by echocardiography, including:Doppler echocardiography: Measures the velocity of blood flow across the mitral valve during systole andregurgitation. The degree of MR is estimated based on the regurgitant jet velocity and regurgitant volume.Color Doppler echocardiography: Visualizes the direction and extent of the regurgitant jet.Pulsed-wave Doppler: Measures the velocity profile of the regurgitant jet at a specific point in time, which can be used to estimate the regurgitant volume.Continuous-wave Doppler: Measures the instantaneous velocity of the regurgitant jet throughout the cardiac cycle, which can provide additional information about the timing and severity of MR.Simplified Echocardiographic Assessment of Mitral Regurgitation.The simplified echocardiographic assessment of MR is based on the regurgitant jet area method, which estimates the severity of MR by measuring the area of the regurgitant jet in a parasternal long-axis view. This method isrelatively simple and can be performed quickly, making it suitable for routine clinical use.To perform the simplified echocardiographic assessment of MR, the following steps are typically followed:1. Obtain a parasternal long-axis view of the heart.2. Identify the mitral valve and visualize the regurgitant jet.3. Measure the area of the regurgitant jet using the planimetry tool of the echocardiography machine.4. Calculate the regurgitant fraction (%) by dividing the area of the regurgitant jet by the area of the left ventricular outflow tract (LVOT).Grading of Mitral Regurgitation Severity.The severity of MR is typically graded based on the regurgitant fraction as follows:Mild MR: Regurgitant fraction < 20%。
Flow Pattern and Pressure Drop of Upward Two-Phase Flow in Vertical
Flow Pattern and Pressure Drop of Upward Two-Phase Flow in Vertical CapillariesDingsheng Liu†,‡and Shudong Wang*,†Dalian Institute of Chemical Physics,Chinese Academy of Sciences,Dalian116023,People’s Republic ofChina,and Graduate School of Chinese Academy of Sciences,Beijing100039,People’s Republic of ChinaGas-liquid two-phase flow patterns were investigated visually with a digital camera for upward air-waterflow in capillaries with diameters of1.47,2.37,and3.04mm.Several distinctive flow patterns were observedand described.The flow pattern maps developed in this work were compared with the conventional flow-pattern transition criteria for larger-diameter tubes and the experimental data that have been published.Thetotal pressure drop in capillaries was measured with a differential pressure transducer for Taylor flow,whichexisted broadly in multiphase monolith reactors and microchannel reactors.A new correlation was presentedto calculate the total pressure gradient(TPG)of Taylor flow.In addition,the other three models,which wereproposed previously by other researchers,were also used to predict the TPG.The four theoretical results allwere in overall agreement with the experimental data.Finally,the effects of method used to calculate thevoid fraction on the TPG were checked.It was observed that the void fraction influenced the TPG re-markably,and the drift flux model was not suitable to evaluate the void fraction under the present experimentalconditions.1.IntroductionMonolith honeycomb reactors are now widely considered to be a promising alternative to conventional gas-liquid-solid three-phase reactors,such as trickle-bed and slurry reactors.1-3 Generally,the monolith structure consists of an array of parallel, straight,uniform channels with square or circular geometry, typically having hydraulic diameters in the range of1-5mm.1-3 Fundamental knowledge of the two-phase flow in small chan-nels,such as the flow pattern and the pressure drop,is,thus, demanded urgently for optimizing engineering design as well as evaluating practical performance.A pioneer systematic study on the characteristics of the gas-liquid flow in monolith reactors was performed by Satterfield and Ozel.4They showed that the two-phase flow in a single channel of monolith reactors could be characterized by the two-phase flow in a capillary.1.1.Flow Pattern.It is well-known that the morphology of gas-liquid flow greatly influences the rates of mass and heat transfer in small tubes.To this day,many papers related to flow patterns of multiphase flow in capillaries or microchannels have been published.Barnea et al.5compared the experimental flow patterns from the capillaries with diameters on the order of millimeters,with the theory6applied to the tubes with diameters on the centimeter scale.Barnea et al.considered surface tension to account for the deviation between the experimental data and the theoretical results.For capillaries,surface tension was a dominant factor,whereas,for larger diameter tubes,Kelvin-Helmholtz instability was a key factor.5Mishima and Hibiki7 experimentally observed some flow regimes that were peculiar to capillaries,such as intermittent strings of bubbly flow.In addition,the boundaries between flow regimes were well-predicted by the model of Mishima and Ishii.8Coleman and Garimella9considered hydraulic diameter and surface tension to be important factors in determining the locations of flow-pattern transitions.Decreasing the capillary diameters shifted the transition to a dispersed flow pattern to a higher superficial liquid velocity,because of the combination of surface tensions and tube diameters.The transition to purely annular flow almost occurred at a constant superficial gas velocity and approached a limiting value as the capillary diameters decreased.Triplett et al.10conducted an experimental investigation of two-phase flow patterns in horizontal circular capillaries.In their experi-ments,five major flow patterns were observed:bubbly,Taylor, churn,slug-annular,and annular flow.The data agreed well with some existing experimental data but poor consistency was observed with the predictions based on the relevant flow-pattern transition models.For vertical triangular capillaries with hy-draulic diameters of2.866and1.443mm,Zhao et al.11observed flow patterns similar to that encountered in the conventional large-sized vertical circular tubes.It was also determined that the transition boundaries from slug flow to churn flow and from churn flow to annular flow in the flow pattern maps shifted to the right as the capillary hydraulic diameters decreased.Chen et al.12noticed,experimentally,a peculiar flow pattern that was called bubble-train slug flow.The flow pattern was characterized by the presence of a bubble train that was formed by several bubbles connected together with a membrane between two neighboring bubbles.The number of bubbles in a bubble-train slug appeared in an irregular manner,which was indicative of its inherent turbulent nature.Liu et al.13studied two-phase flow using air and different liquids in circular and square capillaries. Four flow patterns,including bubbly,slug-bubbly,Taylor,and churn flow,were observed,under their experimental conditions. In addition another typical flow pattern,annular flow,occurred only at excessive high gas and liquid velocities.Chen et al.14 studied,experimentally,the effects of capillary diameters on vertical two-phase flow patterns.The flow-pattern maps showed that the transition boundaries of slug-churn flow and churn-annular flow were dependent strongly on the capillary diameters. In contrast,the boundaries of dispersed bubble flow to churn flow and from bubbly flow to slug flow were less affected.*To whom correspondence should be addressed.Tel.:+86-411-84662365.Fax:+86-411-84662365.E-mail address:Wangsd@.†Dalian Institute of Chemical Physics,Chinese Academy of Sci-ences.‡Graduate School of Chinese Academy of Sciences.243Ind.Eng.Chem.Res.2008,47,243-25510.1021/ie070901h CCC:$40.75©2008American Chemical SocietyPublished on Web12/01/2007Moreover,the experimental transition boundaries showed poor agreement with that predicted by existing models applied to larger tubes.1.2.Pressure Drop.For two-phase flow in capillaries or microchannels,pressure drop is a very important parameter and is often used to design reactors and optimize operational conditions.Mishima et al.7studied the frictional pressure drop of a gas-liquid two-phase flow in vertical capillaries using the Lockhart-Martinelli correlation15and Chisholm’s equation.16 It was found that the value of the original Chisholm’s parameter (C)was dependent on not only the flow regimes but also the capillary diameters.The two-phase frictional pressure drop could be corrected well by Chisholm’s equation,in association with Chisholm’s parameter C,which is a function of the capillary diameter.Triplett et al.17calculated the two-phase pressure drop in transparent,long,horizontal microchannels with circular and semitriangular cross-sections,using various two-phase friction models.For bubbly and slug flow patterns,the two-phase friction factor based on homogeneous mixture assumption provided the best agreement with the experimental data.For annular flow, the homogeneous mixture model and other correlations signifi-cantly overpredicted the frictional pressure drop.Zhao et al.18 reported the experimental data for the pressure drop of upward air-water flow in vertical miniature triangular channels.The results showed that two-phase frictional pressure drop could be well-predicted by the Lockhart-Martinelli method if their friction factor correlation for single-phase flow was adopted. Kawahara et al.19measured and analyzed two-phase frictional pressure drop data in circular microchannels.In their experi-ments,the two-phase flow patterns were much less homoge-neous,as indicated by video images and very large slip ratios. Thus,the agreements between the experimental data and the homogeneous flow model were generally poor,except for the good predictions(within(20%)obtained Dukler et al.’s function,20which was used to calculate the mixture viscosity. In contrast,two-phase friction multiplier values were correlated well(within(10%)with the separated flow model of Lockhart and Martinelli15when the coefficient C,which is used in Chisholm’s equation,was given by the model by Lee and Lee.21 Later,a conclusion similar to that of Kawahara et al.19was also drawn by Yue et al.22by investigating the pressure drop of two-phase flow through T-type microchannel mixers.Recently,Liu et al.13specifically studied the pressure drop of Taylor flow, one interesting and practical two-phase flow pattern in vertical capillaries.They correlated the dimensionless two-phase pres-sure factors and the total pressure drop using gra V ity-equi V alent V elocities.To calculate the pressure drop of the homogeneous flow and nonhomogeneous flow,two different functions were developed to evaluate the pressure factors.In addition,the experimental data were predicted well by their theory,especially at low and moderate liquid velocities.Kreutzer et al.23experi-mentally measured the total pressure drop of Taylor flow in a 2.3-mm vertical capillary,using a technique that independently controlled the lengths of liquid slugs and gas bubbles.They proposed a new method to calculate the total pressure drop by means of the lengths of liquid slugs.More recently,Akbar et al.24simulated Taylor flow in capillaries based on the Volume of Fluid technique.Similar to Kreutzer et al.,23they obtained a new function,which was used to calculate the frictional pressure gradient.The agreements between the simulation and the experimental data were very good.This review shows that many works about the two-phase flow patterns in capillaries have been conducted.However,rather inconsistent results and conclusions have also been obtained by various researchers.This confusion may partially stem from varieties of nomenclatures defining flow-pattern transition boundaries or even flow patterns themselves.Generally,there is one undoubted fact:the behavior of fluid in capillaries is different from that in conventional larger tubes,because of size effects.It is significant to understand the physical mechanisms of flow patterns further,to enhance the performances of multiphase monolith reactors and microchannel reactors. The aforementioned review also indicates that many methods have been proposed to calculate the two-phase pressure drop in capillaries or microchannels.Most of them are independent of two-phase flow patterns,which make the methods inap-plicable under certain circumstances.Therefore,it is not puzzling to observe significant deviations from experimental data without knowing specific flow regimes.Taylor flow is a very common and important flow pattern in multiphase monolith reactors. Although some researchers,including Liu et al.,13Kreutzer et al.,23and Akbar et al.,24have reported some new methods and data about pressure drop of Taylor flow in capillaries,there is no universal theory yet and more data on the pressure drop in Taylor flow must be accumulated.Therefore,the work performed here was designed to sys-tematically study the flow patterns and pressure drop of the gas-liquid flow in vertical capillaries,which were similar to the channels of the monolith reactors.The experimental results were compared to the available relevant data and theories.2.ExperimentThe experimental setup of the upward gas-liquid flow in vertical capillaries was clearly demonstrated in Figure1.As shown in Figure2,the inlet part of the capillary was connected to a T-shaped mixer with the same internal diameter to the capillary.The outlet part of the capillary was connected to a straight junction with the same structure as the T-mixer,but without the side liquid inlet.Deionized water and air were chosen for the liquid phase and the gas phase,respectively.To force liquid through the capillaries at a constant velocity,a tank with invariable pressure was designed to supply water.The water velocity was adjusted and measured by a liquid rotameter between the water bank and the T-mixer.The air velocity was adjusted by a mass flow controller.Water and air were fed into the capillaries from the side and bottom of the T-mixer, respectively.A separator was used to drain water and discharge air at the exit of the system,where the pressure was atmospheric. Three different circular quartz-glass capillaries were used to investigate flow patterns and pressure drop.The capillary internal diameters were determined by imaging the capillary cross-sections using a microscope.The physical parameters of the capillaries are given in Table1.The flow patterns were captured by a digital charge-coupled device(CCD)camera with the shutter speed of1/1000-1/4000s and then were saved onto the hard drive of a personal computer(PC)for later analysis. The pressure drop along the capillaries was measured with a differential pressure transducer(DPT).As shown in Figure1, the two sampling ports of the DPT were connected to the side ports of the T-mixer and the straight junction,respectively. Figure2shows that the distance from the T-mixer’s side port to the capillary inlets was only2mm,so it was negligible,in regard to the local effect of the T-mixer on the total pressure drop,with respect to the total lengths of the capillaries.The same conclusion was suited for the capillary outlets.3.Results and Discussion3.1.Two-Phase Flow Pattern.3.1.1.Flow Pattern.As shown in Figures3and4,different flow patterns were observed244Ind.Eng.Chem.Res.,Vol.47,No.1,2008at different superficial gas and liquid velocities for upward gas -liquid flow in the vertical capillaries.To determine flow patterns clearly and eliminate confusion of terminologies used by various researchers,the specific definitions for different flow patterns are given as follows:Bubbly flow :Bubbly flow often occurs when superficial gas velocities are low while superficial liquid velocities are high.It is characterized by distinct or distorted sphere bubbles;generally,the diameters of the bubbles are less than or equal to the capillary inner diameters.Taylor flow :Taylor flow (which is also called slug flow,bubble train flow or segmented flow)is characterized by long bubbles separated by liquid slugs.The lengths of the bubbles are greater than the capillary inner diameters.Thin liquid films exist between the bubbles and the capillary walls.Slug -bubbly flow :This flow pattern simultaneously has the characteristics of Taylor flow and bubbly flow.The flowpatternFigure 1.Schematic diagram of the experimentalsetup.Figure 2.Detailed structure of the T-mixer.Table 1.Specifications of the Capillaries Used in the Experimentinner diameter(mm)length (m)capillary number 1.470.76112.370.91723.040.8343Figure 3.Representative flow patterns in the 1.47-mm capillary at a low superficial liquid velocity:(a)U G )0.0113m/s,U L )0.0786m/s,bubbly flow;(b)U G )0.3048m/s,U L )0.0786m/s,Taylor flow;(c)U G )1.2203m/s,U L )0.0786m/s,bubble-train slug flow;(d)U G )4.4745m/s,U L )0.0786m/s,churn flow;and (e)U G )11.0351m/s,U L )0.0786m/s,annular flow.Ind.Eng.Chem.Res.,Vol.47,No.1,2008245is more disordered and is often considered to be a transition flow pattern.Bubble-train slug flow :This flow pattern typically consists of series of bubbles,similar to trains,with a clear interface between the connecting bubbles.12The number of bubbles in a “train”seems to appear in a random manner.In addition,the shape and size of bubbles are not always uniform,under certain circumstances.Churn flow :This flow occurs when superficial gas velocities are large enough for the successive several bubbles to coalesce to one bubble after breaking through the liquid slugs between them.A wave or ripple motion is often observed at the bubble tail with tiny gas bubbles entrained in the liquid slug because of the high superficial gas velocity.To some extent,churn flow is similar to Taylor flow,but the former is more chaotic and disordered.Annular flow :Annular flow is observed at excessively high superficial gas velocities and very low superficial liquid velocities.This flow is comprised of a continuous gas phase in the central core and a continuous liquid phase that is deposited on the circumference of the capillary walls.One noteworthy feature is that the major flow patterns,classified as given previously,can be further subdivided into more-specific flow patterns.In other words,some slight difference in appearance can be observed for the flow pattern with the same name.For example,Figure 4a -c showed the diversity of appearance of bubbly flow due to different superficial gas velocities.Figures 3and 4show typical images of the two-phase flow observed in the 1.47-mm capillary at low and high superficial liquid velocities,respectively.With increasing superficial gas velocities,five distinct flow patterns s including bubbly flow,Taylor flow,bubble-train slug flow,churn flow,and annular flow s appeared one after another at a constant low superficial liquid velocity in Figure 3.Moreover,these flow patterns were comparative ordered or regular.In contrast,at a high superficial liquid velocity in Figure 4,every flow pattern appeared much more disordered and chaotic.In addition,bubbly flow,slug -bubbly flow,bubble-train flow,and churn flow were recorded in a sequence with increasing superficial gas velocities.Fur-thermore,the comparison of the flow patterns between the low and high liquid velocities were performed.Bubbles of the bubblyflow in Figure 3were spherical or spheroidal,and the distances between the bubbles were uniform.Bubbles confined in the capillary rose straight regularly.Figures 4a -c show that the number of bubbles in the bubbly flow increased as the gas velocities were enhanced.Also,the bubbles moving up along the capillary oscillated from one side of the wall to another side,which was similar to the phenomenon noticed by Krishna et al.25In addition,slug -bubbly flow was not observed at the low superficial liquid velocity,and Taylor flow and annular flow were not noticed at the high superficial liquid velocity.3.1.2.Flow Pattern Map.Figures 5a,5b,and 5c show the overall flow pattern maps for the three circular capillaries (with diameters of 1.47,2.37,and 3.04mm),where the superficial liquid and gas velocities were used as the ordinate and abscissa,respectively.The solid lines represented the boundaries at which flow pattern transitions occurred.Four major flow patterns s including bubbly flow,Taylor flow,churn flow,and annular flow s divided the flow pattern maps into four regions,where the flow patterns possessed definite characteristics that were easily observed.The two transition flow patterns,slug -bubbly flow and bubble-train slug flow,were situated in the small central transition regions of the flow pattern maps.To predict the pressure drop of Taylor flow exactly,which will be described in section 3.2,quasi-Taylor flow was presented as an indepen-dent flow pattern only in Figure 5.Here,quasi-Taylor flow was defined as the Taylor flow of which the lengths of liquid slugs were smaller than or equal to capillary diameters.The flow patterns in the transition regions represented ambiguity char-acteristics and were judged more subjectively.Because these measurements were on the same order of size for the capillary diameters in this study (millimeters),the effects of diameter on flow patterns were not remarkable.Moreover,the three flow pattern maps in Figure 5were similar (overall).It is well-known that flow patterns are dependent on not only liquid velocities but also gas velocities.In other words,if the gas (or liquid)velocity is not appropriate,some specific flow patterns will not appear at any liquid (or gas)velocity in Figure 5.Therefore,the flow pattern maps are helpful in determining the operational conditions in industry.parison with Existing Transition Correlations.The present experimental data were compared with two famous models that have been developed to predict flow pattern transitions during steady gas -liquid flow in vertical circular tubes:one by Taitel et al.6and another by Mishima et al.8To predict the operational conditions under which the transitions will occur,it is very important to understand the physical mechanisms of flow transitions.However,considerable dis-agreements among various transition mechanisms still exist.Therefore,the theories with different physical mechanisms sometimes predicted different flow patterns under the same operational conditions.In 1980,Taitel et al.6presented a physical model and developed theoretically based transition equations as follows.For the transition from bubbly flow to Taylor flow,Taitel et al.6gave an expression for the flow at which dispersed bubbles were observed:This function was plotted in Figure 6as curve A;abovethis curve,bubbly flow existed.However,regardless of the amount of liquid turbulent energy for breaking and dispersing the gas phase,bubbly flow could not exist at a gas voidfractionFigure 4.Representative flow patterns in the 1.47-mm capillary at a high superficial liquid velocity:(a)U G )0.0113m/s,U L )1.6365m/s,bubbly flow;(b)U G )0.0727m/s,U L )1.6365m/s,bubbly flow;(c)U G )1.2203m/s,U L )1.6365m/s,bubbly flow;(d)U G )2.3259m/s,U L )1.6365m/s,slug-bubbly flow;(e)U G )4.4745m/s,U L )1.6365m/s,bubble-train slug flow;and (f)U G )11.0351m/s,U L )1.6365m/s,churn flow.U L )-U G +4.0{d 0.429(σ/F L )0.089νL0.072[g (F L -F G )F L ]0.446}(1)246Ind.Eng.Chem.Res.,Vol.47,No.1,2008of R )0.52,which was represented as beeline B in Figure 6.Here,Thus,curve A,which delimited bubbly flow,terminated at beeline B.Bubbly flow existed in the regions to the left of beeline B and on curve A.Below curve A,Taylor flow dominated.Increasing the gas velocities further would shiftTaylor flow to churn flow.The criteria of the transition between Taylor flow and churn flow was expressed aswhere L E was the entrance length of the tube.In this work,L Erepresented the length from the entrance to the midpoint of the capillary that was used.Equation 3was denoted by curveD.Figure 5.Flow pattern maps for upward gas -liquid flow in vertical capillaries:(a)d )1.47mm,(b)d )2.37mm,and (c)d )3.04mm.R )U G U G +U L(2)Figure parison of the experimental flow pattern maps with Taitel et al.’s model:6(a)d )1.47mm,(b)d )2.37mm,and (c)d )3.04mm.L E d )40.6(U G +U Lgd+0.22)(3)Ind.Eng.Chem.Res.,Vol.47,No.1,2008247Moreover,Taitel et al.6considered the fact that churn flow could be expected only at R>0.25.Thus,the locus of R)0.25is shown as dashed beeline E in Figure6and represents the terminus of curve D.For higher gas velocities,the flow became annular.The transition boundary between churn flow and annular flow was given by The transition criteria were plotted as vertical beeline C in Figure 6.The regions to the right of beeline C were dominated by annular flow.(Note that the flow pattern names shown in the brackets in Figures6-8are those given by the respective authors.)As seen from Figure6,the theoretical predictions are not consistent with the experimental results completely.As mentioned previously,bubble-train slug and slug-bubbly are two transition flow patterns.In Figure6,beeline B and curve D were within or near the transition regions,which meant the theory predicted the transition from bubbly flow to churn flow and the transition from Taylor flow to churn and annular flow well.For the capillaries with diameters of2.37and3.04mm, the correct locus of beeline E also meant good predictions. Discrepancies between the theory results and the experimental data existed,but the theoretical boundaries between bubbly flow and Taylor flow were still helpful to identify bubbly flow above curve A correctly.From Figure6,the theoretical beeline C poorly predicted the transition from churn flow to annular flow, which might result from the fact that the simple criteria(eq4) did not take into account the effects of liquid velocities and capillary diameters on flow patterns.In1984,Mishima et al.8considered the void fraction(R),as a simple and reliable parameter to predict flow pattern.From this point of view,one new flow pattern transition criterion for upward gas-liquid flow in vertical tubes was developed.The transition boundary of bubbly flow to Taylor flow was given by R)0.3,which was converted to a conventional form based on superficial velocities:whereEquation5was plotted as curve A in Figure7.The regions to the left of the curves marked A were dominated by bubbly flow,and the proximate zones to the right of the curves marked A were controlled by Taylor flow.The transition from Taylor flow to churn flow was postulated to occur whenHere,R was calculated based on the drift-flux modeland R m represents the mean void fraction of Taylor bubbles and was expressed as follows:The transition for the Taylor flow to churn flow occurred at the curves marked B,based on eq7.The curves marked B intersected the beelines that were marked C and formed the churn-flow regions,which were below curve B and to the left of beeline C.At higher liquid and gas velocities,the Taylorparison of the experimental flow pattern maps with Mishima et al.’s model:8(a)d)1.47mm,(b)d)2.37mm,and(c)d)3.04mm.UG FG1/2[σg(FL -FG)]1/4)3.1(4)UL)(3.33C0-1)U G-0.76C0[σg(F L-F G)F L2]1/4(5)C)1.2-0.2 F G F L(6)R g Rm(7)R)UGC(UG+UL)+0.35 ∆F gd F L(8)Rm)1-0.813×{(C0-1)(U G+U L)+0.35 ∆F gd/F L(UG+UL)+0.75 ∆F gd/F L[∆F gd3/(F LνL2)]1/18}0.75(9) 248Ind.Eng.Chem.Res.,Vol.47,No.1,2008flow was limited by beeline D.At gas velocities above beeline D,liquid slugs would be dispersed into liquid droplets by entrainment;thus,annular flow should be observed.For two different mechanisms for occurrence of annular flow,Mishima et al.derived two transition equations as follows:8where R should satisfy the condition given by eq7: Equations10and11were designated as beelines C and D, respectively.Beelines C and D intersected curve B at two different points.The three curves also consisted of a zone that was dominated by annular flow.Generally,wide discrepancies between the Mishima et al.8 theory and this study are observed in Figure7.The transition boundaries from bubbly flow to Taylor flow was dependent only on the properties of both phases;they were not affected by the capillary diameters.In other words,the loci of the three curves marked A in Figure7were completely the same.From the experimental data,the transition boundaries from bubbly flow to Taylor flow were inclined to be parallel to the abscissas,while the theoretical curves marked A were vertical to the abscissa in the mass.Another experimental fact was that bubbly flow existed infrequently in the regions to the right of the curves marked A,except for the1.47-mm capillary.The Taylor-flow regions predicted by the theory actually included several different flow patterns observed in the experiment.The agree-ment between the theory and the experiment was good only at the lower liquid velocities.With further increases in the gas velocities,the theoretical transition boundaries showed that churn flow and annular flow occurred at low and high liquid velocities,respectively.The prediction zones for churn flow just fell in the experimental transition region,including mainlybubble-train slug flow,partial Taylor flow,annular flow,and churn flow.The theoretical annular flow region was apparently the combination of the experimental churn and annular flow zones.That meant that,for any liquid velocity,the flow pattern would always be annular flow when the gas velocity was larger than the critical value.Obviously,the theoretical results for annular flow were proved incorrect by the experimental data in Figure7.parison with Existing Data.The flow pattern map for the1.47-mm capillary was also compared with the existing experimental data of Triplett et al.10and Zhao et al.11 in Figure8.Triplett et al.10distinguished five flow patterns in a horizontal circular capillary that had a diameter of1.45mm: bubbly,Taylor,slug-annular,churn,and annular flow.Zhao et al.11observed bubbly,Taylor,churn,and annular flow in the vertical triangular capillary with a diameter of1.44mm.Figure 8a shows that Triplett et al.’s experimental data agreed well with the overall results determined in this work,except for the Taylor-flow region.The slug-annular flow pattern defined by Triplett et al.10was assigned to annular flow here.The Taylor flow observed by Triplett et al.10had a wider distribution than that in this work.The bubbly-and Taylor-flow regions observed by Zhao et al.11in Figure8b were similar to that recorded by Triplett et al.10in Figure8a.That made the bubbly-flow region smaller and the Taylor-flow region bigger than that in this work. The churn-flow region of Zhao et al.11extended to lower liquid velocities but disappeared in trends at higher gas velocities.This distribution of churn flow was different from the results of the present work and Triplett et al.10In Figure8b,for any liquid velocity,the annular flow always occurred at higher gas velocities,which was inconsistent with the experimental data.3.2.Two-Phase Pressure Drop.It is well-known that the total pressure gradient(TPG)of two-phase flow consists of three components,as follows:where[d P/d X]T,[d P/d X]F,[d P/d X]A,and[d P/d X]S represent TPG,the frictional pressure gradient(FPG),the acceleration pressure gradient(APG),and the static pressure gradient(SPG), respectively.During the present experiment,the pressure drop along the capillaries is very small,in comparison to atmosphere pressure,and no phase changing occurs,so the acceleration pressure drop can be negligible.Then,eq12becomesAs discussed previously,the pressure drop of only Taylor flow was evaluated.The long bubbles and liquid slugs are distributed alternately and orderly in Taylor flow.The liquidUG) ∆F gd F G(R-0.11)(10)UG g(σg∆F F G2)1/4[µL(FLσ σ/(∆F g))1/2]-1/5(11)Figure8.Flow-regime map comparison of the present experimental datain1.47-mm capillary with published experimental data:(a)Triplett et al.10and(b)Zhao et al.11[d Pd X]T)[d Pd X]F+[d Pd X]A+[d Pd X]S(12)[d Pd X]T)[d Pd X]F+[d Pd X]S(13)Ind.Eng.Chem.Res.,Vol.47,No.1,2008249。
Experiments on two-phase flow distribution inside parallel channels of compact heat exchangers
Experiments on two-phase flow distribution inside parallel channels of compact heat exchangersA.Marchitto,F.Devia,M.Fossa,G.Guglielmini *,C.SchenoneDiptem,University of Genoa,Via all’Opera Pia 15A,16145Genoa,Italy Received 14February 2007;received in revised form 16July 2007AbstractUneven distribution in heat exchangers is a cause of reduction in both thermal and fluid-dynamic performances.Many papers have dealt with single-phase flow and both flow distribution data and analytical or numerical models are available for header design.With regard to two-phase flow,phase separation in manifolds with several outlets is so complicated that,to date,there is no general way to predict the distribution of two-phase mixtures at header-channel junctions.The design of headers for new generation compact heat exchangers and multi-microchannel evaporators is still based on an empirical approach,as a number of variables act together:geometrical parameters and orientation of the manifolds and of the chan-nels,operating conditions,fluid physical properties.In the present paper measurements of the two-phase air–water distributions occurring in a cylindrical horizontal header supplying 16vertical channels are reported for upward flow.The effects of the operating conditions,of the header-channel distribution area ratios and of the inlet port orifice plates were investigated.The flow rates of each phase flowing in the different channels were measured.Time varying,void fraction data were also analysed to characterise the two-phase flow patterns.Video records were taken in order to infer different flow patterns (from intermittent to annular)inside the header-channel system.Ó2007Elsevier Ltd.All rights reserved.Keywords:Flow distribution;Air–water mixture;Parallel channels;Plate heat exchangers;Orifice plates1.IntroductionOne of the factors that most strongly influence the performance of compact heat exchangers is the degree of flow rate uniformity in the various parallel channels where the heat transfer occurs.Plate heat exchangers (PHE)have long been used in a wide range of industrial applications regarding sin-gle-phase flow.Recently,two-phase gas/liquid mixtures have been utilized in such exchangers in processes involving vaporisation and condensation.Typical examples include refrigerating cycles,in which the use of such components favours compactness and improves heat transfer performance.Multi-microchannel tube0301-9322/$-see front matter Ó2007Elsevier Ltd.All rights reserved.doi:10.1016/j.ijmultiphaseflow.2007.08.005*Corresponding author.Tel.:+390103532877;fax:+39010311870.E-mail address:guglielm@ditec.unige.it (G.Guglielmini).Available online at International Journal of Multiphase Flow 34(2008)128–144A.Marchitto et al./International Journal of Multiphase Flow34(2008)128–144129 heat exchangers are also relatively common and especially in mobile air-conditioning applications.Interest in micro-channel heat exchangers is keen,owing to their compactness,which saves space,weight and refrigerant charge.Both evaporators and condensers raise the problem of ensuring uniform distribution of the two-phaseflow from the distributor to the multi-channel system that makes up the exchanger.Uneven two-phase distribution can occur both inside each channel,owing to the asymmetrical parallel and diagonalflow,and inside the header,owing to the separation of the two-phase mixture in the header-tube junctions.Flow mal-distribution in heat exchangers may be spatial,temporal or both,its effect being to reduce both thermal andfluid-dynamic performances.Many papers have dealt with the effects offlow mal-distribution on the performance of heat exchangers(Mueller,1987;Mueller and Chiou,1988;Probhakara Rao et al.,2005). Some mal-distributions are the result of fabrication conditions,such as mechanical design or manufacturing tolerances;others are caused by the heat transfer andfluidflow process itself,by fouling and/or corrosion,or by the typical non-uniformities of two-phaseflows.The effects on heat transfer efficiency and operating con-trol vary.While some cases of bad distribution have little effect on heat exchanger performances,others result in significant loss of performance and/or mechanical failure of the devices(Mueller and Chiou,1988).For these reasons,several recent studies have analysed the effects of geometrical configuration and operating con-ditions on the distribution of two-phaseflows.The present paper reports the results of several experiments carried out on a cylindrical horizontal two-phaseflow header supplying upwardly oriented sixteen vertical channels.The aim of this work is to analyse the effects on two-phase distribution of orifice plates placed upstream of the header and perforated plates con-necting it to the channels.Although these configurations are widely used in real distributors,they have not been extensively studied in the literature.Measurements of air–waterflow rate distributions were taken for a number of operating conditions and for different geometrical configurations.Time varying,cross-sectional void fraction data were analysed to characterise the two-phaseflow patterns in the header section.Video records were taken in order to infer theflow patterns inside the distributor during intermittent and almost annularflows.2.Previous worksWith regard to single-phaseflow,many papers have been published onflow distribution data and analytical and numerical models for manifold design.Among the early works,Acrivos et al.(1959)obtained an iterative solution for viscous,single-phaseflow in headers.Bajura and Jones(1976)studiedflow distribution in lateral branches of different manifolds,both analytically and experimentally.They showed that uniformflow distri-bution in the lateral branches is attained only when the headers act as infinite reservoirs.Bassiouny and Mar-tin(1984)attempted tofind a correlation between the header geometry and theflow distribution.Kim et al. (1995)investigated the effects of header shapes and the Reynolds number in a parallel-flow manifold to be used in a liquid module for electronic packaging.They found that theflow distribution depended sensibly on the header shape and on the Reynolds number.Yin et al.(2002)experimentally studied the pressure drops inside the complex headers and parallel circuits of a micro-channel heat exchanger.A pressure drop model for the whole heat exchanger was developed and the pressure and massflow rate distribution inside the heat exchanger were predicted.Two-phaseflow distribution from a header to parallel channels gained great attention to predict the heat transfer performance of compact heat exchangers,evaporators and condensers.The distribution of two-phase mixtures through the channels is often non-uniform,and in extreme cases there is almost no liquidflow through some channels.In evaporators,uniform distribution is essential in order to avoid dry-out phenomena and the resulting poor heat exchange performance.In condensers,uneven distribution of liquid could create zones of reduced heat transfer due to high liquid loading.Thus,in designing compact heat exchangers,under-standing separation phenomena in the manifolds is of great importance.The two-phaseflow structure in compact heat exchangers is very complex and to date there is no general way to predict the distribution of the two-phase mixtures at the header-channel junctions.In fact many vari-ables act together,such as the geometric factors(manifold shape,channel junctions,flow orientation)and the130 A.Marchitto et al./International Journal of Multiphase Flow34(2008)128–144operating conditions(massflow rate and vapour quality at the inlet of the distributors and heat load on the tubes).Several authors have investigated two-phaseflow division in T-junctions.However,phase separation in manifolds is so complicated that T-junctionfindings cannot be directly applied to the study of a multi-channel system(Watanabe et al.,1995).Only a limited number of studies have investigated the complex phenomena that occur in two-phaseflow distributors.A few studies have attempted to identify an appropriate CFD model for refrigerant distributors(Li et al., 2002a,b).Experiments were performed that aimed to validate the general trends associated with the CFD results.However,it is important to note that the systems studied were characterised by only a few branches and by operating conditions that could be well represented by the homogeneousflow model.Most studies on two-phase distribution in compact heat exchangers have been experimental.Our literature search is summarized in Table1,which includes the description of test geometry,experimental conditions and two-phase mixtures used.Generally speaking,experiments have been performed under adiabatic conditions.Only Vist and Pettersen (2004)investigated the influence of changing the heatflux to the evaporator tubes on the two-phase distribu-tion in the manifold.Changing the heat load on the evaporator test section had little influence on the two-phaseflow distribution,while the two-phaseflow distribution markedly influenced the heat exchanged between the refrigerant and the counter-flowing water.All found studies used air–water mixtures or halocar-bon refrigerants.Rong et al.(1996)used air–water mixtures to simulate an R-134a refrigerant system at the same inlet volumetricflow rates.Webb and Chung(2005)simulated the actual refrigerantflow conditions by means of air–water mixtures with the same liquid–vapour density ratio and the same Martinelli parameters (Xtt).Operating conditions varied from bubbly and slugflows(Osakabe et al.,1999;Horiki and Osakabe,1999) to annularflow(Lee and Lee,2004)at the header inlet.Lee recently presented an extensive review(2006)of research on two-phase distribution in dividing tubes and parallel channels.A review of experimental and theoretical studies was also presented by Guglielmini (2006).The effects of tube outlet direction,tube protrusion depth,massflow rate and quality were experimen-tally investigated by Kim and Sin(2006)for a horizontal round header and30verticalflat tubes simulating a parallel-flow heat exchanger;both upward and downwardflow were considered.Forflush-mounted configu-rations,their results confirmed what other researchers had found:downwardflow and upwardflow led water toflow mostly through the front and rear parts of the header,respectively.The effect of tube protrusion depth was also studied:for the downwardflow configuration,as protrusion depth was increased,more water was forced to the rear part of the header,while for the upwardflow configuration,theflow distribution was not significantly altered.A negligible difference in two-phaseflow distribution was observed between the par-allel and the reverse-flow configurations.In recent years,several studies have been devoted to the comprehension of theflowfield in manifolds con-necting a series of micro-channels(Beaver et al.,2000;Tomkins et al.,2002;Hrnjak,2004;Fei et al.,2002;Cho et al.,2003;Fei and Hrnjak,2004).These studies have been performed by using either air–water mixtures or various refrigerants(e.g.R22,R134a,R744).The main parameters examined have been the orientation of the header(horizontal or vertical)theflow direction of the refrigerant into the inlet header(in-line,parallel and cross-flow)and operating conditions.Most studies have confirmed the difficulty of obtaining uniform two-phaseflow distribution.The distribution of thefluid into each parallel micro-channel tube is a function of theflow regime in the header.However,it is also significantly affected by the type of connection between the pipes and the header,the shape of the inlet port,the protrusion depth,the orientation,etc.Experiments on two-phaseflow distribution have demonstrated that,among the main factors affecting the splitting of the phases,a very important role is played by the various geometric factors involved:distributor shape and size,channel junctions,header and channel orientation,the size and length of the inlet pipe,and the intrusion depth of the channels into the header wall.All these parameters can strongly affect the distribution of gas and liquidflow rates among the channel pipes.However,the behaviour can be completely changed by par-ticular upstream conditions,such as the presence offittings(e.g.nozzles or short inlet tubes),which are able to modify the two-phaseflow pattern at the inlet of the manifold.As Webb and Chung(2005)concluded,the design of devices to improve the distribution is nowadays‘‘highly empirical’’.A.Marchitto et al./International Journal of Multiphase Flow34(2008)128–144133A further confirmation of this condition is the point that,whereas orifices and nozzles are commonly used by heat exchanger manufacturers to empirically adjust theflow rate distribution on the basis of the operating conditions and of the unit dimension,these aspects have not yet been sufficiently investigated,and no system-atic study is reported in the literature.The aim of the experimental campaign is therefore to deeply study some phenomenological aspects of the two-phaseflow separation produced either by varying the area restriction caused by thin orifice plates(between the straight pipe and the distributor)or by varying the header-channel area ratios at the connection of the header to the parallel channels by inserting perforated plates of different sizes.3.ExperimentsExperiments were carried out on a simple test section in order to investigate some phenomenological aspects of the two-phase distribution in compact heat exchanger manifolds.Theflow inside the vertical chan-nels was upward.The test section was designed to allow the visualisation offlow structure in the inlet port and inside the header.The instrumentation was designed to record the pressure and void fraction evolution inside the header and to measure the liquid and gasflow rates insides the parallel channels,downstream of the header.3.1.Experimental set-up and procedureThis experimental apparatus consists of two supply lines of air and water that merge into a horizontal pipe (Fig.1).Phase mixing(through a Tfitting)allows intermittent and annularflow regimes to be generated. Downstream of the mixer,the mixtureflows horizontally inside a2.0m long acrylic pipe with an inner diam-eter of26mm.The pipe is connected by aflange to the inlet port of the test section.An overall sketch of the experimental apparatus is given is Fig.1.Downstream of the test section,an array of valves can operate in order to extract theflow rate coming from a pair of vertical channels and to divert it toward the extraction phase separator,which carries the liquid and gas phases toward the correspondingflow meters:the air is then released into the atmosphere,while the water is pumped to the water reservoir,which also acts as the main separator.The other two-phase streams that134 A.Marchitto et al./International Journal of Multiphase Flow34(2008)128–144depart from the test section outlet ports are sent,viaflexible hoses,to the main separator.The loop is closed at the mixing junction,where the air coming from the air supply line and the water pumped by the main pump are mixed to generate the two-phaseflow.Characterization of the operating conditions and investigation of phase distribution inside the pair of chan-nels are based on the measurement of the gas and liquidflow rates.The main and extracted gas streams are metered by two thermal devices,able to directly measure the massflow rate with an overall accuracy of±2% of the reading.The main and extracted liquid streams are metered by two magnetic devices with an overall accuracy of±0.8%of the reading.In order to deduce the inlet gas superficial velocity,local pressure and tem-perature must be known.The gas temperature is measured by a K-type thermocouple connected to a rack dis-play,which converts the signal into a4–20mA current for the acquisition system.The overall accuracy of temperature measurement is±1°C.Additional measurements are those water temperature(by a K-type ther-mocouple)and the gas pressure after the pressure regulator,upstream of the mixer,by an absolute pressure transducer.The test section(Fig.2)consists of a distributor,an interchangeable orifice plate and a system of n=16 vertical channels.The upper outlet ports of the channels are connected in pairs in order to allow the stream from each pair to be separately collected(channel pairs from j=1to j=8).The distributor has a circular cross-section of D=26mm i.d and was machined and polished from a rectangular block of acrylic resin. The transparent block facilitates visualisation,while theflat external surface minimises the distortion due to refraction.Horizontally oriented,the distributor is equipped with an interchangeable plate with16orifices supplying an equal number of vertical channels,which are connected to the header with a pitch of18mm (Fig.2).The channel dimensions(length,depth,width)are500,15,18mm,respectively.Theflange connecting the header to the upstream supply pipe can befitted with an orifice plate(inlet nozzle),in order to investigate the remixing effects of such singularity on theflow distribution inside the header and the channels.Four pres-sure taps(connected to a differential pressure transducer)allow the pressure gradient at the header inlet,and inside it,to be recorded.Four impedance probes are simultaneously used to obtain the instantaneous cross-sectional void fraction upstream the distributor and inside it.The void fraction sensor adopted in this investigation consists of ringelectrode pairs placed on the internal wall of the cylindrical test duct,flush to the pipe surface;one sensor(S1) is placed upstream of the distributor,10.6diameters upstream of the inlet port,and the other three sensors (S2,S3and S4)are placed inside the distributor,equally spaced72mm apart.The description of the procedure and the uncertainty analysis on void fraction measurement have been reported elsewhere(Fossa,2001;Fossa and Guglielmini,2002),the uncertainty being found to be about 4%.From the analysis of the probe signal,the time-average cross-sectional void fraction a and the void frac-tion probability density function(PDF)can be inferred;this latter is a keyfigure inflow pattern identification (Zuber and Jones,1975).3.2.Geometrical factors and operating conditionsThe experiments were carried out for differentflow conditions and for different geometrical configurations. The operating conditions,evaluated at the distributor inlet,cover the V sg=1.50–16.50m/s and V sl=0.20–1.20m/s,gas and liquid superficial velocity ranges,respectively.Intermittentflows(plug,slug)and annular flow were visually observed.Analysis of the PDFs of the void fraction signals(as obtained from the probe upstream of the inlet port section)and the inspection of the Taitel and Dukler(1976)map confirmed the exis-tence of intermittent and annularflow regimes(Fig.3).The pressure at the test section inlet was varied in the range from1.4to2.2bar to control the superficial velocities,also taking into account the pressure drops through the orifice plate and the nozzle,when present.At the connection of the header to the parallel channels,four different orifice plates with the diameter d of2, 3,4and6mm,were employed.The resulting four different header-channel area ratios A R¼1ðD=dÞ2were 10.56,4.69,2.64and1.17,respectively.The straight pipe and the distributor were connected either by orifice nozzles(20,16and12mm i.d.)or without any area restriction(r=(d/D)2=1).3.3.Data processingThe gas and liquidflow rates measured are hereafter presented in a non-dimensional way.The non-dimen-sional gasflow ratioð_mÃg;j Þor liquidflow ratioð_mÃl;jÞinside the pair of channels No.j is the ratio of the mea-sured gas or liquidflow rateð_m k;jÞin the pair of channels under consideration(No.j)over the mean gas(or liquid)flow rate calculated for uniform distribution:_mÃk;j ¼_m k;jPi¼1_m k;i=Nðk¼g;lÞwhere N is the overall number of channel pairs.A.Marchitto et al./International Journal of Multiphase Flow34(2008)128–144135Further physical quantities were measured:the instantaneous value of the cross-sectional void fraction at the inlet port and inside the header,the absolute pressure and the temperature at the inlet port.From the statistical analysis of the void fraction time series,information was inferred on the actual two-phaseflow pattern at the inlet and inside the header.The absolute pressure and the temperature at the inlet port were used as references to obtain the local gas and liquid superficial velocities.The phase distributions inside the channels were also analysed in terms of the standard deviation of the k-phaseflow ratio,here defined as follows:STD k¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiX Nj¼1ð_mÃk;jÀ1Þ2=N v uu tFor every geometrical configuration and operating condition STD k represents a synthetic,single-value,index of liquid or gasflow rate mal-distribution.When local gas or liquidflow rate is close to the mean calculated value,STD k is small and tends to increase as theflow distribution worsens.A further index was then introduced to analyse and quantifyflow rate mal-distribution:NSTD k indicates the‘‘Normalized Standard Deviation’’for the k-phase,which is the ratio between the actual STD k and the maximum value of standard deviation for a certain number,N,of vertical channels.NSTD k¼STD kSTD k;max¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiP Nj¼1ð_mÃk;jÀ1Þ2ðNÀ1ÞNsBy utilizing STD or NSTD for both phases it is possible to compare the uniformity of gas and liquidflow rate for different conditions,thus determining the relative effectiveness of the various configurations investigated. Several other indexes can be used to synthetically rank the behaviour of header-channels systems with regard offlow rate distribution.Here,STD and NSTD were chosen because of their simplicity and because STD was previously utilized by other authors(Vist and Pettersen,2004;Hrnjak,2004;Fei et al.,2002;Cho et al.,2003; Fei and Hrnjak,2004)for similar analysis.4.Results and discussion4.1.Effects of the operating conditions on two-phaseflow distributionsThe distribution of the massflow rate of each phase among the channels,for a given geometrical configu-ration,strongly depended on the superficial gas and liquid velocities at the header inlet.Typical distribution profiles at low values of the liquid superficial velocity(V sl=0.45m/s)are presented in Fig.4a and b,in terms of gas and liquidflow ratios,respectively,with the gas superficial velocity as a param-eter.The orifice diameter d in these representations is4mm(A R=2.64)and in these cases no inlet nozzle is present(r=1).At the lowest gas superficial velocity(V sg=1.50m/s),the gas phase was mainly diverted into thefirst channel pairs,promoting a rush of water into thefirst three pairs of channels,while the liquidflow ratio decreased below the value of1in the remaining channels.The liquidflow ratio was characterised by a maximum,which tended to move toward the downstream channels as the gasflow rate increased.At higher values of the gas superficial velocity,the gasflow rate tended to become more uniform,while the waterflow rate tended to feed only the last pairs of channels.All these behaviours are well described in Fig.4c,in which the liquidflow ratios are reported as a function of gas superficial velocity of the mixture for each pair of chan-nels.The situations corresponding to V sl=0.45m/s,V sg=1.50and5.25m/s are also reproduced in the images in Fig.5,in which the differentflow patterns and the phase distributions in each channel are clearly visible.At the liquid superficial velocity of V sl=1.20m/s,the gas and liquid phases were again unevenly distrib-uted.In this case,however,the gas phase was preferentially distributed into thefirst channels,while the liquid phase was generally distributed into the last channels,even at low gas superficial velocities.On increasing the gasflow rate,the gasflow ratio profile improved,thus reducing the mal-distribution.By contrast,the liquid 136 A.Marchitto et al./International Journal of Multiphase Flow34(2008)128–144A.Marchitto et al./International Journal of Multiphase Flow34(2008)128–144137Fig.5.Gas and liquidflow ratio inside the channel pairs for low liquid and gas superficial velocities(V sl=0.45m/s,V sg=1.50m/s(a) and5.25m/s(b)).Orifice plate diameter d=4mm.phase tended to feed only the last channels.All these behaviours are represented in Fig.6a and b and also in the images in Fig.7.The gas and liquidflow ratio profiles shown in Fig.6a and b have been described by many authors as typ-ical of horizontal headers connected to vertical upward channels.However,when the liquid and the gas super-ficial velocities are quite low,the liquidflow ratio exhibits a maximum in the channels located in the upstream part of the header(Figs.4and5).This last behaviour was also observed by Horiki and Osakabe(1999)in their study on horizontal protruding-type headers in which the waterflow was contaminated with bubbles.The effects of the operating conditions on gas and liquidflow ratio profiles can be represented synthetically in terms of STD for both phases.As shown in Fig.8,on increasing the gasflow rate,the gas STD g decreased, for any liquid superficial velocity examined,thus reducing the mal-distribution of the gas phase.By contrast, the liquid phase distribution tended to worsen when the gas superficial velocity increased,and the waterFig.7.Gas and liquidflow ratio inside the channel pairs for high gas and liquid superficial velocities(V sl=1.20m/s,V sg=1.50(a)and 5.25m/s(b)).Orifice plate diameter d=4mm.tended to feed only the last channels.As a result of these effects,the quality mixture STD decreases as both liquid and gas superficial velocities increase.Finally,the effects of the operating conditions described above were generally observed for all the other orifice plate configurations.However,as the ratio A R was reduced,the occurrence of the maximum of the liquidflow ratio in the upstream or intermediate channels was observed for a wider range of liquid and gas flow rates.For A R=10.56(d=2mm)the maximum was observed in the range V sl60.25m/s, V sg61.35m/s;for A R=1.17(d=6mm)however,it was observed in the wider range V sl60.45m/s, V sg612.75m/s.4.2.Effects of the diameter of the orifices connecting the distributor to the channelsReducing the orifice diameter generally had a typical dual effect:the gas mal-distributions diminished,while the liquid mal-distributions tended to increase.Concerning the liquidflow ratio,at low gas superficial veloc-ities and small orifice diameters,the entrainment of the water into thefirst channels tended to disappear;at higher gas superficial velocities,the shift of the liquid toward the last channels became more evident.As an example,the diagrams in Fig.9show a comparison between the results obtained with orifices of2,3,4 and6mm,for the same values of gas and liquid superficial velocity(V sl=0.45m/s,V sg=1.50and5.25m/s).The effects of orifice diameter d and header-channel distribution area ratio A R on the gas and liquid distri-butions are also well represented by the standard deviation STD g and STD l.As shown in Fig.10,while enlarg-ing the orifice diameter generally improved the liquid uniformity inside the channels,the liquid standard deviation STD l always increased as the gas superficial velocity rose,as already seen in Fig.8.On the other hand,the gas standard deviation STD g showed an opposite behaviour:its value increased as the orifice diam-eter increased and as the gasflow rate decreased.Similar STD profiles were also observed at higher liquid velocities(V sl=0.80and1.20m/s),but the effect of the orifice diameter was less marked.4.3.Effects of the presence of an orifice nozzle upstream of the distributorThe presence of a nozzle at the inlet of the distributor significantly modified phase distribution into the par-allel channels.The main effect of such a restriction was to produce a jet inside the distributor that modified the flow patterns.This effect was also noticeable with regard to the liquid distribution in single-phaseflow:a strong reduction of the inletflow area(d=12m,r=(d/D)2=0.21)forced the water to the rear part of the header.A CFD analysis of the single-phase distributor,together with the connected upstream duct,was carried out by means of the Fluent6.1Ócode usingfirst-order approximation for derivatives and Standard k–e model for turbu-lence.CFD numerical results were obtained for different orifice nozzle diameters.The comparison between CDF results and the corresponding experimental data is reported in Fig.11with reference to the liquidflow,。
GRE测试颠峰精选单词词汇之二步曲
GRE测试颠峰精选单词词汇之二步曲感谢大家对小女子上一篇注水文的厚爱,现推出后半部分的分类词汇人及其职业alderman amanuensis bursar caddie concierge constable eunuch docent cosmetologist geisha joiner majorette matron provost stevedore ventriloquist teamster 工具设备abutment adze alembic altimeter ammeter anemometer atomizer cask derailleur derrick diesel gantry gaff gimlet hygrometer machete martingalemattock mimeograph punk quirt semaphore cudgel /shillelagh sickle sonar spade thimble trestle trivet washer家居生活bassinet brazier cuspidor dentifrice dormer marquee valance服装beret barrette brogan buskin toga slicker mufti食品ale bouillon compote cruet delicatessen grit jalopy scull prow keel ckshaw sedan gondola身体chignon entrails excreta follicle gulletgut/viscera sinus shank labial医疗auscultation chiropody chiropractic douche etiology euthanasia physic swab poultice健康abscess ague chilblain contusion excrescence lumbago malaria malocclusion migraine presbyopia rheum pustule suppurate rheumatism宗教friar doxology evangel cassock litany manse muezzin officiant vestry synod sexton postulant 神鬼传说poltergeist纺织batiste bobbin chenille damdsk fancywork ginfham purl rayon ticking warp woof农牧bran grist guano hacienda malt paddy工业erratum pyrotechnics军事enfilade娱乐alto jig cabaret derby oboe mazurka lute polka prom roulette体育goalie prizefight自然地理azimuth bight butte everglade fen magma stalactite solsticestalagmite stratosphere turf材料燃料物质anthracite brimstone emery pasteboardpewter plywood pug puttystucco talc wainscot建筑pavilion belfry campanile学术数学algorithm Braille ergonomics lyceum polytechnic prosody动物armadillo beagle bison cranium dromedary gazelle gibbon jackal lynxmarmot orangutan porpoise opossum protozoan rhinoceros runt snout vicunavole yak artiodactyls/cloven hoof ewe 鸟martin thrush水族sepia虫chrysalis cicada louse (praying) mantis植物bulrush larch rosin poplar花与颜色azalea begonia camellia epiphyllum fringed iris fuchsia garnetlilac magnolia peony tawny蔬菜chive fennel parsley radish水果jujube。
- 1、下载文档前请自行甄别文档内容的完整性,平台不提供额外的编辑、内容补充、找答案等附加服务。
- 2、"仅部分预览"的文档,不可在线预览部分如存在完整性等问题,可反馈申请退款(可完整预览的文档不适用该条件!)。
- 3、如文档侵犯您的权益,请联系客服反馈,我们会尽快为您处理(人工客服工作时间:9:00-18:30)。
Flow patterns past two circular cylinders in proximityKyongjun Lee,Kyung-Soo Yang *Department of Mechanical Engineering,Inha University,253Yonghyun-Dong,Nam-Gu,Incheon 402-751,Republic of Koreaa r t i c l e i n f o Article history:Received 18February 2008Received in revised form 6June 2008Accepted 24July 2008Available online 5August 2008Keywords:Circular cylinderImmersed boundary method Flow pattern Vortex shedding Wake interactiona b s t r a c tFlow patterns past two nearby circular cylinders of equal diameter immersed in the cross-flow at low Reynolds numbers (Re 6160),were numerically studied using an immersed boundary method.We con-sidered all possible arrangements of the two cylinders in terms of the distance between the two cylinders and the inclination angle of the line connecting the cylinder centers with respect to the direction of the main flow.Ten distinct flow patterns were identified in total based on vorticity contours and streamlines,which are Steady,Near-Steady,Base-Bleed,Biased-Base-Bleed,Shear-Layer-Reattachment,Induced-Sep-aration,Vortex-Impingement,Flip-Flopping,Modulated Periodic,and Synchronized-Vortex-Shedding.Collecting all the numerical results obtained,we propose a general flow-pattern diagram for each Re ,and a contour diagram on vortex-shedding frequency for each cylinder at Re =100.The perfect symmetry implied in the geometrical configuration allows one to use these diagrams to identify flow pattern and vortex-shedding frequencies in the presence of two circular cylinders of equal diameter arbitrarily posi-tioned in physical space with respect to the main-flow direction.Ó2008Elsevier Ltd.All rights reserved.1.IntroductionCross-flow past a group of cylinders is often found in practical engineering applications.For instance,air flow past a bundle of pipes in a chemical plant,coolant flow past tubes in a heat exchan-ger,sea-water flow past columns of a marine structure,and wind past an array of chimneys of a power plant,to name a few.Flow characteristics past each cylinder are affected by its neighbors via wake interaction,resulting in alteration of the overall flow pattern.Therefore,flow pattern past multiple cylinders heavily depends on their relative positions with respect to the main-flow direction.Consequently,vortex-shedding frequency of the individual cylin-der is accordingly determined,and serves as an important factor in generation of flow noise [1].Being motivated by this,many researchers have been involved in studying wake interaction be-tween two circular cylinders of equal diameter immersed in a cross-freestream as a basic wake-interaction model.Spatial arrangement of two cylinders can be classified into three catego-ries,namely,aligned with the direction of the main flow (in tan-dem),placed side-by-side,and placed in a staggered arrangement.When two circular cylinders of equal diameter are placed side-by-side,flow pattern varies with Reynolds number (Re )and the distance between the cylinders.Kang [2]identified six distinct flow patterns depending on Re and the distance between the cylinders.It was also found that when the surface-to-surface distance is long-er than 5D (here,D is the cylinder diameter),wake interaction van-ishes and flow past each cylinder behaves just like the flow past the single cylinder placed in the freestream [2].Williamson [3]classified the types of vortex shedding by using flow visualization.When Re =100and the surface-to-surface distance is longer than D ,the flow pattern past the two cylinders is periodic,and vortex shedding from each cylinder is either in phase or out of phase by 180°with the other one.However,when the surface-to-surface distance is shorter than D ,the flow pattern past the two cylinders is non-periodic and completely irregular [2,3].When the two cylinders are placed in tandem,there exists the critical distance between the two cylinders,below which vortex shedding of the upstream cylinder (hereafter,called ‘‘main cylin-der”,MC)does not occur;the value of the critical distance varies depending on Re [4–7].Mizushima and Suehiro [4]revealed that the flow pattern with L /D =2.0is different from that with L /D =4.0at Re =100mainly due to suppression of vortex shedding from MC.Here,L denotes the streamwise distance between the two cylinder centers (Fig.1(a)).They also predicted existence of bistable solutions in the range of 3.16L /D 63.5for Re =100.Sub-sequently,Tasaka et al.[5]performed experimental investigation to verify the prediction of Mizushima and Suehiro [4],and reported 4.66L/D 65.0as the range of L /D for existence of bistable solu-tions,being shifted from the numerical prediction of Mizushima and Suehiro [4].Tasaka et al.[5]attributed the discrepancy to the finite length of the cylinders in their experiments.Carmo and Meneghini [6]investigated the three-dimensional structures be-hind two circular cylinders when they are in tandem for the range of 160<Re <320.Sharman et al.[7]performed numerical predic-*Corresponding author.Tel.:+82328607322;fax:+puters &Fluids 38(2009)778–788Contents lists available at ScienceDirectComputers &Fluidsj o u r n a l h o m e p a g e :w w w.e l s e v i e r.c o m /l o c a t e /c om p flu idrevealing a sudden change in keyflow parameters between L/D= 3.75and4.0.Sumner et al.[8,9]experimentally studiedflow-induced forces andflow patterns for high-Re(8506Re61900)flows past two cir-cular cylinders in staggered arrangements;they identified nine dis-tinctiveflow patterns.Akbari and Price[10]identifiedfive distinct flow patterns at Re=800.Gu and Sun[11]and Sun et al.[12]also reportedflow patterns in turbulentflow regime.In spite of the numerous studies carried out so far,the case where the two cylinders are placed in a staggered position has not been systematically studied,especially in the laminarflow re-gime.In this investigation,flow patterns past two nearby circular cylinders of equal diameter immersed in the cross-flow at Re6160,based on the freestream velocity(U)and D,were numer-ically studied as a basic model for laminar wake interaction.An im-mersed boundary method[13]was employed for effective treatment of the cylinders on a Cartesian grid system.We consider all possible arrangements of the two circular cylinders in terms of the distance between them and the inclination angle of the line connecting their centers with respect to the mainflow direction. The current authors recently reported their numerical results on theflow-induced forces in the sameflow setting[14].As a fol-low-up study,the present paper classifies theflow patterns ob-served,and describes distinctive characteristics of eachflow pattern.Collecting all the results obtained,we propose aflow-pat-tern diagram(‘‘map”)for prediction of theflow pattern associated with a particular arrangement of the two cylinders as well as a Strouhal-number diagram for vortex shedding of each cylinder to provide an overall picture on the wake interaction.The perfect symmetry implied in the geometrical configuration allows one to use these diagrams to identifyflow pattern and vortex-shedding frequencies in the presence of two circular cylinders of equal diam-eter arbitrarily positioned in physical space with respect to the mainflow direction.2.Formulation and numerical methodologyThe current investigation requires a parametric study where numerous numerical simulations must be performed with various values of L and T.Here,T represents the vertical distance between the two cylinder centers(Fig.1(a)).This kind of parametric study demands considerable amount of computing resources;the com-puting efforts can be significantly reduced by employing an im-mersed boundary method which facilitates implementing the solid surfaces of the arbitrarily-positioned cylinders on a Cartesian o u jo x jÀq¼0ð1Þo u iþo u i u jj¼Ào piþ1o2u ij jþf i i¼1;2ð2Þwhere u i(or u,v),p,q and f i represent velocity component in x i(or x,y)direction,pressure,mass source/sink,and momentum forcing, respectively.All the physical variables except p are non-dimension-alized by U and D;pressure is non-dimensionalized by far-field pressure(P1)and the dynamic pressure.The governing equations were discretized using afinite-volume method in a nonuniform staggered Cartesian grid system.Spatial discretization is second-or-der accurate.A hybrid scheme is used for time advancement;non-linear terms are explicitly advanced by a third-order Runge–Kutta scheme,and the other terms are implicitly advanced by the Crank–Nicolson method.A fractional step method[15]was em-ployed to decouple the continuity and momentum equations.The Poisson equation resulted from the second stage of the fractional step method was solved by a multigrid method.For detailed description of the numerical method used in the current investiga-tion,see Yang and Ferziger[16].3.Boundary conditions and numerical parametersThe main cylinder isfixed at the origin of the coordinate system, and the downstream cylinder(hereafter,called‘‘surrounding cylin-der”,SC)is placed at various locations relative to MC,which are represented by dots in Fig.1(b).The total number of cases com-puted is208for Re=100,and58for Re=40,50,and160.In the case where the center distance(R,Fig.1(a))is minimum, L/D=0.25,T/D=1.0,and the inclination angle(/,Fig.1(a))is 75.96°.The entire computational domain was defined as À35D6x635D,andÀ50D6y650D.For each cylinder,32Â32 uniform grid cells in x-and y-directions,respectively,were allo-cated,and uniform grid cells of the same cell size as in the cylinder region were employed between the cylinders.In the other region of the domain,nonuniform grid cells were used.The numerical resolution was determined by a grid-refinement study to ensure grid-independency.In the case of L/D=1.5,T/D=0.5(Fig.2),for in-stance,the total number of grid cells was368Â208,while 480Â352grid cells were allocated in the most demanding case (L/D=5.0,T/D=5.0).No-slip condition was imposed on the cylinder surfaces;a Dirichlet boundary condition(u=U,v=0)was used on theK.Lee,K.-S.Yang/Computers&Fluids38(2009)778–7887794.Verification of the codeFor verifying our numerical methodology and resolution,the flow past a single circular cylinder at Re =100was computed and our results were compared with other authors’numerical [2,18]and experimental [19]results.The size of computational domain and the boundary conditions employed were as described in the previous section.The same number of uniform grid cells (32Â32)as in the two-cylinder cases was allocated for the cylin-der,and the total numerical resolution was 320Â192.Table 1pre-sents mean drag coefficient ðC d Þ,maximum magnitude of lift-coefficient fluctuation ðC 0l ;max Þ,and Strouhal number (St )of the vor-tex shedding in the literature,along with the present results.The present results are in excellent agreement with those of other authors,verifying that the numerical scheme and resolution used in the current simulation are adequate and reliable.5.Results and discussionIn the present investigation,we classified types of flow pattern based on streamlines and contours of spanwise vorticity (x z ).Spanwise vorticity is defined as 1o v Ào u,and contour lines cover À4.06x z 64.0with increment of 0.1in the following contour plots.Solid and dashed lines represent positive and negative vor-ticity,respectively.Four different values of Re (40,50,100,and 160)were considered.For Re =100and 160,the general trend can be summarized as follows.When the two cylinders are ‘‘very”close to each other,the vortex shedding occurs as if the two cylinders were just one object.When they stand ‘‘reasonably”apart,wake interaction takes place,resulting in various types of flow pattern depending on how the two cylinders are positioned.When they are ‘‘far”away from each other,wake interaction is weak and flow characteristics around each cylinder resemble those of a single cylinder at the same Re .In the current investigation,10distinct flow patterns were identified in total for Re 6160.To name the flow patterns,weadopted the terms in the previous studies [2,9],if applicable.Otherwise,we named the new flow patterns with our own terms.In the following section,detailed description of flow patterns is gi-ven for Re =100.The effect of Re on wake interaction is addressed in Section 5.2.5.1.Classification of flow pattern for Re =1005.1.1.Base bleed (BB)BB pattern is noticed when (L /D ,T/D )=(0.0,1.25),(0.25,1.0),(0.25,1.25),(0.5,1.0),(0.75,0.75).Even though the distance be-tween the two cylinders is quite short,the flow through the ‘‘gap”is significant owing to the large /.Fig.3presents snapshots of pressure contours together with streamlines through the gap at the four selected instants with equal time interval during one per-iod of vortex shedding (T p ).Fig.4shows the corresponding time histories of drag coefficient (C d )for MC and SC,respectively.In Fig.3(a),a low pressure region is seen behind MC,attracting the gap flow into this region and also causing a high drag force on MC (see A in Fig.4).The opposite is the case in Fig.3(c);pressure is low behind SC.As a result,the gap flow is diverted towards the low pressure region (Fig.3(c)),and SC experiences a high drag force (see B in Fig.4).In Fig.5,contours of x z are presented at the same instants as in Fig.3.The short distance between the cylinders pre-vents ‘‘individual”shedding;the vortices are shed in a fashion sim-ilar to that behind a single object.5.1.2.Biased base bleed (BBB)BBB pattern is found when (L /D ,T /D )=(1.0,0.25),(1.0,0.5).The gap between the cylinders is very narrow like the BB pattern,but /is small unlike the BB pattern.Fig.6shows snapshots of instanta-neous pressure contours and streamlines through the gap at the four selected instants with equal time interval during T p .The flow through the gap is always heading downwards due to the small /(Fig.6),while in the BB pattern,the direction of the gap flow is alternating up and down (Fig.3).In Fig.7,contours of x z are shown at the same instants as in Fig.6.Like in the BB pattern,vortices are shed as if the two cylinders were just one object.5.1.3.Shear layer reattachment (SLR)SLR pattern is observed approximately when 1.256L /D 64.0,T/D 60.75;in this range,/is less than 22°.Instantaneous stream-lines of a typical case of SLR pattern (L /D =3.0,T/D =0.5)are pre-sented in Fig.8.The shear layer formed above MC reattaches on SC.A lower portion of the shear layer is deflected by SC,and to-gether with the shear layer formed below MC,surrounds the recir-culation region behind MC.In Fig.9where instantaneous contours of x z are taken at four instants with equal time interval during T p ,it is seen that SC significantly interfere with the upper shear layer from MC.It is also noticed that vortex shedding from MC has been completely suppressed by SC,and vortex shedding from SC is somewhat delayed.This is one of the important features of SLR pat-tern.Sharman et al.[7]reported that in tandem arrangements of the two cylinders,vortex shedding from MC resumes at a certain ‘‘critical”distance between them which is in the range of 3.756L /D 64.0depending on Re .5.1.4.Induced separation (IS)IS pattern is noticed approximately when 1.06L /D 63.5,0.756T /D 61.75.In this region,/is relatively large compared to that of SLR pattern.Fig.10presents instantaneous contours of x z taken at the four instants with equal time interval during T p .TheTable 1Validation of numerical method:flow past a single cylinder at Re =100ReC d C 0l ;max St 780K.Lee,K.-S.Yang /Computers &Fluids 38(2009)778–788one vortex (C in Fig.10).However,negative vortices are mainly shed from SC (Fig.10).Fig.11shows a snapshot of streamlines for L /D =1.5,T /D =1.0.The gap flow actually induces separation on SC,enhancing vortex shedding from SC.Consequently,vortex shedding from SC is much more active in IS pattern than in SLR pattern.5.1.5.Vortex impingement (VI)VI pattern appears in the range of approximately L /D P 4.0,T /D 61.75,where the two cylinders are far apart (R /D P 4.0)and /is less than 24°.When the two cylinders are placed in tandem withK.Lee,K.-S.Yang /Computers &Fluids 38(2009)778–788781a short distance between them(L/D63.75for Re=100),SC sup-presses the vortex shedding of MC[7].With a larger distance,how-ever,SC fails to suppress the vortex shedding of MC,and vortices are shed from both cylinders.Vortices of positive and negative signs, shed from MC in an alternating manner,impinge upon SC.However, when the two cylinders are slightly‘‘off-tandem”(say,T/D P0.25),shedding with equal time interval when L/D=4.0,T/D=1.0.The negative vortex which developed in the upper shear layer of MC im-pinges upon SC(A in Fig.12(a)),and then split into two parts(A0and A00in Fig.12(b)).One of them(A00)is combined with the negative vor-tex shed from SC,and the other remains as a separate secondary vor-tex(B in Fig.12(c),and C in Fig.12(d)).782K.Lee,K.-S.Yang/Computers&Fluids38(2009)778–788where the distance between the two cylinders(1.346R/D63.20) is somewhat larger than those in BB or BBB patterns,and/P38°. In this pattern,both cylinders shed vortices,yielding two pairs of counter-rotating vortices.Unlike BB or BBB pattern,however,FF pattern is irregular in time,being consistent with Meneghini et al.[20]and Kang[2].Kang reported that in side-by-side arrange-ments of two cylinders,flow pattern is irregular when1.46T/D62.2.Fig.13shows temporal variations of drag and lift coefficients(C d,C l),respectively,for L/D=0.0,T/D=2.0;irregular force coeffi-cients reflect irregularflow behavior in FF pattern.5.1.7.Modulated periodic(MP)MP pattern occurs in the range of approximately0.56L/ D63.0,T/D P2.5where the two cylinders are far apart (2.696R/D65.59),and approximately/P45°.Combination of the large distance and the relatively high inclination angle yields weak interaction between the two wakes,resulting in slight mod-ulation in vortex shedding from each cylinder.It is noticed that the vortex-shedding frequency of MC is slightly higher than that of SC.K.Lee,K.-S.Yang/Computers&Fluids38(2009)778–788783Fig.14presents temporal variation of C l for L /D =1.5,T /D =3.0,revealing modulation of amplitude and frequency of the force coef-ficient.Instantaneous contours of x z at two distinct instants are illustrated in Fig.15.It is seen that two positive vortices shed from the cylinders are side-by-side in Fig.15(a).However,the opposite is the case at a later time (Fig.15(b)),indicating the difference be-tween the two frequencies of vortex shedding.MP pattern is sim-ilar to ‘VPE’(vortex pairing and enveloping)or ‘VPSE’(vortex pairing,splitting and enveloping)patterns in Sumner et al.[9]for 8506Re 61900.5.1.8.Synchronized vortex shedding (SVS)SVS pattern is noticed in the arrangements where the two cylinders is far apart (R /D P 2.5)and /P 20°.Due to the long dis-tance and high inclination angle,the interaction of the wakes is negligible,and vortices are shed just like the single cylinder in the freestream.Fig.16shows contours of x z at four instants of equal time interval during one period of vortex shedding for L /D =0.0,T /D =3.0.The cylinders shed vortices independently;interaction of wakes such as vortex merging is hardly seen.In this particular case,the vortices from the two cylinders are out of phase by 180°owing to the perfect geometrical symmetry with respect tothe center line between the two cylinders.However,in a staggered arrangement of the cylinders,the phase difference changes,and the wake interaction is still negligible in SVS pattern.5.2.Flow patterns for low-Re flowsThe following two flow patterns are uniquely observed in low-Re flows (Re =40and 50).784K.Lee,K.-S.Yang /Computers &Fluids 38(2009)778–7885.2.1.Steady(S)In this pattern,flow is steady in the entire domain.5.2.2.Near steady(NS)NS pattern is characterized by‘‘delayed”unsteadiness.In Fig.17,streamlines are presented at four instants of equal time interval during one period of vortex shedding for L/D=0.0, T/D=1.5at Re=50.In NS pattern,flow is steady around the two cylinders,but becomes unstable in the far downstream wake.It should be noted that NS pattern is also found at Re=40which is slightly below the critical Reynolds number(Re c%43)for the sin-gle-cylinder case.5.3.Flow-pattern diagramCollecting all the results obtained,we propose theflow-pattern diagrams on L–T-plane in Fig.18to provide an overall picture on theflow patterns past two nearby circular cylinders of equal diam-eter.All the patterns discussed earlier are indicated together with their‘‘schematic”boundaries for convenience.The cases actually computed are denoted by the dots.One can identify the effect of Re by comparing the diagrams.At a subcritical Reynolds number(Re=40,Fig.18(a)),theflow would be entirely steady,if there were only single cylinder present.at Re=50which is above Re c(Fig.18(b)).FF and VI patterns do not occur in the low-Reflows.For the high Reynolds numbers(Re=100 and160),S and NS patterns are not found(Fig.18(c)and(d)).As Re increases from Re=100to Re=160,FF and VI patterns occur in more arrangements as expected.5.4.Frequency of vortex sheddingFig.19presents contour diagrams of Strouhal number(St)of vortex shedding for MC(Fig.19(a))and SC(Fig.19(b)),respec-tively,for all possible arrangements of the two cylinders at Re=100.The Strouhal number was computed based on C l using fast Fourier transform(FFT),and normalized with that of the single cylinder(St s=0.165)immersed in the freestream at Re=100.When multiple dominant frequencies were identified (for example,in FF pattern),the most dominant one was reported in Fig.19.For the main cylinder(Fig.19(a)),the maximum St(=0.204)is identified for L/D=1.0,T/D=2.0.In the range of approximately T/D P2.0,the Strouhal number of MC is larger than St s,implying that SC induces more frequent vortex shedding from MC in this region.The opposite is the case when T/D62.0because SC suppresses vortex shedding from MC.It should be noted that in the range of approximately0.06L/D62.5,1.56T/D62.0,St ofK.Lee,K.-S.Yang/Computers&Fluids38(2009)778–788785786K.Lee,K.-S.Yang/Computers&Fluids38(2009)778–788other that the‘‘effective”characteristic length scale is larger than D.As a result,the‘‘effective”Reynolds number increases and St drops.Although the distance between the cylinders is identical (R/D=1.12)in both patterns,the inclination angle in BB pattern is larger than that in BBB pattern,yielding a bigger‘‘effective”char-acteristic length,thus smaller St.In the cases of SLR,IS,and VI patterns where/645°,the Strou-hal numbers of MC and SC are identical,but smaller than St s,indi-cating strong wake interaction due to small/.Nevertheless,the Strouhal number in VI pattern is larger than those in SLR and IS patterns,and more closer to St s,which implies that wake interac-tion is rather weak in VI pattern because the two cylinders are far apart in VI pattern compared to the other two patterns.In the cases of FF and MP patterns where/is large,the Strouhal number of MC is larger than that of SC as well as St s.One can con-jecture that in these patterns,vortex shedding from MC is more en-hanced by the presence of SC via wake interaction.However,as the two cylinders become far apart and/also increases,such as in SVS pattern,the Strouhal numbers of the two cylinders become identi-cal again,and approach St s.5.5.Extension of usage of the diagrams to other arrangementsEven though the diagrams(Figs.18and19)were constructed only for one quarter of all possible arrangements of MC and SC with MCfixed at the origin,one can still use them for the other arrangements using the symmetry of the geometrical configura-tion.The constant freestream velocity allows the symmetry of SC with respect to the x-axis(Fig.20(a)).For example,when SC is placed in the fourth quadrant,the diagrams are still effective as they are,as long as the vertical distance between SC and the x-axis is used for T.Furthermore,the symmetry with respect to the origin is valid as the two cylinders are identical(Fig.20(b)).Thus,as long as the upstream cylinder is designated as MC,and the downstream cylinder as SC according to our‘‘definitions”of MC and SC,and to-gether with the symmetry with respect to the x-axis if necessary, the diagrams can be used in any possible arrangement of the two cylinders.6.ConclusionIn this investigation,we considered cross-flow past two arbi-trarily-positioned nearby circular cylinders of equal diameter for Re6160,and classified itsflow pattern into10distinct ones based on streamlines and contours of spanwise vorticity.They are steady (S),near steady(NS),base bleed(BB),biased base bleed(BBB), shear layer reattachment(SLR),induced separation(IS),vortex impingement(VI),flipflopping(FF),modulated periodic(MP), and synchronized vortex shedding(SVS).Characteristics of each pattern were identified and described in detail.For high-Reflows(Re=100and160),when the two cylinders are very close to each other(such as in BB or BBB patterns),vortices are shed as if they were just one object;the Strouhal number of the vortex shedding is less than that of the single cylinder immersed in the freestream.Wake interaction becomes complicated depending on the distance between the two cylinders and the inclination an-gle of the line connecting the cylinder centers with respect to the mainflow.When the surrounding cylinder is located in the near-wake region of the main cylinder,the frequency of the vortex shed-ding is identical for both cylinders,and is less than that of the sin-gle cylinder.In FF and MP patterns where the inclination angle is large,the vortex-shedding frequency of the main cylinder is larger than that of the surrounding cylinder as well as that of the single cylinder,implying that the vortex shedding of the main cylinder is enhanced by the presence of the surrounding cylinder in the re-gion of high inclination angle.However,both frequencies approach that of the single cylinder as the two cylinders get far apart.It also turned out that presence of a neighboring cylinder can destabilize a subcriticalflow,and stabilize a slightly supercritical flow in the particular arrangements of the two cylinders.Collecting all the computed results,we propose theflow-pattern diagrams and the contour diagrams of Strouhal number of vortex shedding in thefirst quadrant of physical space with the main cylin-derfixed at the origin.Due to the symmetry implied in the geomet-K.Lee,K.-S.Yang/Computers&Fluids38(2009)778–788787sponding vortex-shedding frequencies sheds light on developing a wake-interaction model in engineeringflows.AcknowledgementThis work was supported by Inha University Research Grant. References[1]Kim CM,Conlisk AT.Flow induced vibration and noise by a pair of tandemcylinder due to buffeting.J Fluids Struct1990;4:471–93.[2]Kang S.Characteristics offlow over two circular cylinders in a side-by-sidearrangement at low Reynolds numbers.Phys Fluids2003;15:2486–98.[3]Williamson CHK.Evolution of a single wake behind a pair of bluff bodies.JFluid Mech1985;159:1–18.[4]Mizushima J,Suehiro N.Instability and transition offlow past two tandemcircular cylinders.Phys Fluids2005;17:104107-1–104107-11.[5]Tasaka Y,Kon S,Schouveiler L,Le Gal P.Hysteretic mode exchange in the wakeof two circular cylinders in tandem.Phys Fluids2006;18:084104-1–084104-6.[6]Carmo BS,Meneghini JR.Numerical investigation of theflow around twocircular cylinders in tandem.J Fluids Struct2006;22:979–88.[7]Sharman B,Lien FS,Davidson L,Norberg C.Numerical predictions of lowReynolds numberflows over two tandem circular cylinders.Int J Numer Meth Fluids2005;47:423–47.[8]Sumner D,Richards MD,Akosile OO.Two staggered circular cylinders of equaldiameter in cross-flow.J Fluids Struct2005;20:255–76.[9]Sumner D,Price SJ,Païdoussis MP.Flow-pattern identification for twostaggered circular cylinders in cross-flow.J Fluid Mech2000;411:263–303.[10]Akabari MH,Price SJ.Numerical investigation offlow patterns for staggeredcylinder pairs in cross-flow.J Fluids Struct2005;20:533–54.[11]Gu ZF,Sun TF.On interference between two circular cylinders in staggeredarrangement at high subcritical Reynolds numbers.J Wind Eng Indust Aerodyn 1999;80:287–309.[12]Sun TF,Gu ZF,He DX,Zhang LL.Fluctuating pressure on two circular cylindersat high Reynolds numbers.J Wind Eng Indust Aerodyn1992;41:577–88. [13]Kim J,Kim D,Choi H.An immersed boundaryfinite-volume method forsimulations offlow in complex geometries.J Comput Phys 2001;171:132–50.[14]Lee K,Yang KS,Yoon DH.Flow-induced forces on two circular cylinders input Fluids2009;38:111–120.[15]Kim J,Moin P.Application of a fractional-step method to incompressibleNavier–Stokes equations.J Comput Phys1985;59:308–23.[16]Yang KS,Ferziger rge-eddy simulation of turbulent obstacleflow using adynamic subgrid-scale model.AIAA J1993;31(8):1406–13.[17]Kim DH,Yang KS,Senda rge eddy simulation of turbulentflow past asquare cylinder confined in a put Fluids2004;33:81–96.[18]Park J,Kwon K,Choi H.Numerical simulations offlow past a circular cylinderat Reynolds numbers up to160.KSME Int J1998;12:1200–5.[19]Williamson CHK.Oblique and parallel modes of vortex shedding in thewake of a circular cylinder at low Reynolds numbers.J Fluid Mech 1989;206:579–627.[20]Meneghini JR,Saltara F,Siqueira CLR,Ferrari Jr JA.Numerical simulation offlow interference between two circular cylinders in tandem and side-by-side arrangements.J Fluids Struct2001;15:327–50.788K.Lee,K.-S.Yang/Computers&Fluids38(2009)778–788。