AbstractFlow of a gas–solid two-phase mixture through a packed bed is relevant to a number of industrial processes such as heat recovery andfiltration of dusty flue gases, iron making in shaft reactors, gas purification, and sorption enhanced reaction processes. In spite of theindustrial relevance, little work has been reported in the literature. The limited amount of research work has mainly addressed the mac-roscopic hydrodynamics in terms of pressure drop and solids hold-ups at the ambient temperature. Very little is done, until fairly recent,on solids motion at the single particle level, hydrodynamics at elevated temperatures and heat transfer. This paper reviews the recentdevelopment in the field including both the hydrodynamics and heat transfer of gas–solid two-phase mixtures flowing through packedbeds, which is believed to represent the state-of-the-art in the field. The review is not aimed to be exhaustive but rather focused on ourown work carried out over the past few years in the Institute of Particle Science & Engineering at the University of Leeds. And some ofour results are compared with that of other groups.  2008 National Natural Science Foundation of China and Chinese Academy of Sciences. 41848

Published by Elsevier Limited and Science inChina Press. All rights reserved.Keywords: Gas–solid two-phase flow; Hydrodynamics; Heat transfer; PEPT; Euler–Euler multi-fluid model 1. IntroductionFlow of a gas–solid two-phase mixture through apacked bed is relevant to a number of industrial processessuch as heat recovery and filtration of dusty flue gases,ironmaking in shaft reactors, gas purification, and sorptionenhanced reaction processes. In spite of the industrial rele-vance, only a limited amount of work has been reported inthe literature in the past. These studies can be broadlypided into two categories, co-current flows in which gasflow and particles move on average in the same directionand counter-current flows in which particles move on aver-age in an opposite direction to the gas flow. The co-currentflows are more preferred in many cases due to better con-trollability of particles. The work on the co-current flows of gas–solid two-phase mixtures through packed beds hasaddressed some aspects of the flow including axial velocitydistribution of suspended particles [1–4], solids hold-ups inthe bed and pressure drop [4–15], transient accumulation ofsuspended particles in the initial stage [16], non-uniformityof solids hold-up in the entrance region of the bed [10,11],flow behaviour with a lateral inlet [17], and solids behav-iour in a dilute gas–solid two-phase mixture [2,4].This paper reviews recent development in this field. Thefocus of the review is on the recent work carried out in theInstitute of Particle Science & Engineering at the Univer-sity of Leeds though comparisons will be made with theresults published by other groups. The reason for focusingon our own work is that our approach is much more sys-tematic including experimental work, theoretical analysesand numerical modelling, and covering flow hydrodynam-ics at both room temperature and elevated temperatures,and solids motion at both the single particle and bulkscales.  2. Solids circulation technology for non-periodic sorptionenhanced reaction processesFig. 1 schematically illustrates a solids’ circulating tech-nology recently proposed for non-periodic adsorptionenhanced chemical reaction processes. The system is forgaseous reactants and products, and consists of a reactorand a desorber. The reactor has a structured packing ofcatalyst which also serves as an adsorber. Relatively fineadsorbent particles are pneumatically conveyed throughthe packed structure by gaseous reactants/products forcontinuous (in-situ) removal of a product. Adsorbentregeneration is carried out outside the reactor (ex-situ),thus decoupling the reaction and regeneration phases andenabling a steady-state (non-periodic) operation if twodesorbers are used. The decoupling of the adsorption anddesorption steps is important because many high capacityadsorbents have a slow desorption kinetics [18]. The novelprocess can be viewed as the adsorptive-reactor equivalentof the fluid catalytic cracking (FCC) process in whichadsorbent is the transported medium. Like the FCC pro-cess, the benefits of this process are expected to be substan-tial, with excellent control of adsorbent residence time, theenhanced heat and mass transfer, the continuous supply offeed to a single unit, and an integrated energy supplysystem.The technological challenges for the proposed systeminclude (a) control of gas–solid mixture flowing throughthe structured packing and solids circulation, (b) possibleattrition/erosion of adsorbent and catalyst particles, (c) heattransfer to and within the reactor and (d) heat supply to thepacked bed reactor. The systematic work carried out atLeeds aimed at addressing the above challenges. For doingso, two systems based on the concept shown in Fig. 1 areconstructed (Fig. 2). One of the systems is for cold stateexperimental work studying the hydrodynamic aspectsand the solids motion in the system, and the other one forhot state experiments investigating both the hydrodynamicsand heat transfer and sorption enhanced chemical reactionsusing the steam-methane reforming (SMR) for hydrogenproduction as the model reaction. The details of the experi-mental systems are briefly described in Section 3.Fig. 1. Non-periodic adsorption enhanced reaction processes: the concept(Interface – refers to fresh-spent adsorbent interface, product D isadsorbed onto adsorbent). 3. Experimental systems and techniques3.1. Experimental system for cold state workThe experimental systems and techniques for the coldstate work have been well described [2,13,14,19]. The sys-tem consisted of an acrylic Perspex glass packed column,two cyclones in series for particle separation, a particleinjection unit for introducing suspended particles into thepacked bed, a hopper for collecting particles from thecyclones and for dispensing the particles to the particleinjection unit, and various flow measurement and controldevices (Fig. 2a). The column for the cold state experimentswas either 50 mm or 100 mm in inner diameter and1000 mm long packed with either 5 mm or 10 mm glassballs. The test section was 600 mm long and located inthe middle part of the column. Various devices were usedto measure the solids and gas flow rates, pressure profilein the axial direction, and both the static and dynamichold-ups. The flow for both the single gas phase and gas–solid two-phase mixtures has been shown to be fully devel-oped in the test section [13]. Compressed air was used asthe conveying gas and glass ballotini with average diame-ters of 55 and 112.5 lm and a material density of2500 kg/m3were used as the suspended particles. Theexperiments were carried out under approximately ambientconditions (19  C, 1.10 bar).Solids motion was studied under the ambient condi-tions by using the Birmingham non-intrusive positronemission particle tracking (PEPT) techniques. The detailsof PEPT technique can be found in Refs. [20,21]. In brief,the PEPT technique makes use of a single radioactive tra-cer that carries positrons. Positrons annihilate with localelectrons. This results in emission of back-to-back511 keV c-rays. Detection of the pairs of c-rays enablesthe tracer location to be found as a function of time bytriangulation. The detectors of the PEPT facility at Bir-mingham cover a field of view of approximately40 cm   50 cm, and have a maximum separation of 80 cm. As the PEPT technique tracks single particlemotion, it can be used to map solids velocity field in adevice that lies within the field of detection(40 cm   50 cm   80 cm) and operated in batch under asteady state or a very slow unsteady state condition.The particle flow system used in this work was operatedcontinuously, and had a dimension in the flow directionexceeding the field of detection. As a consequence, only 500 mm of the test section of the packed column waslocated between the most sensitive part of the two detec-tors. Due to solids circulation, the tracer particle could gothrough the test section many times. This allowed collec-tion of sufficient data for analysing solids motion bothmicroscopically and macroscopically. Resin beads with215 lm diameter were used as tracers. They were acti-vated by an ion exchange process with radioactive waterproduced in a cyclotron (16O+ 3He = 18F + neutron).18F decays by positron emission with a half-life of 110 min. The produced tracers had an activity of 300–600 lCi, thus ensured the collection of high qualitydata over the duration of each experimental run typicallylasting for  1.5 h.3.2. Experimental system and techniques for hot state workThe experimental system used for the hot-state work isshown schematically in Fig. 2(b). The details can be foundin Refs. [22–24]. It is similar to the cold state experimentalsystem except that (i) there were various temperature mea-surement and control devices; (ii) there were only pressuretransducers for measuring the inlet and outlet pressures ofthe column; (iii) the size of the packed column wasdifferent.The packed column and all the pipeline of the hotstate experimental system were made of stainless steel.The packed column had an internal diameter of 41 mm,an external diameter of 48 mm and a length of1100 mm. It was heated by a three-zone ceramic heaterwith each zone controlled independently (Watlow, UK).Glass balls with 5 mm diameter were randomly packedinto the column. Two thermocouple assemblies (TCDirect, UK), each consisting of 12 Type J thermocouples,were used to measure the temperature fields in the inte-rior of the packed bed. A 1 mm stainless steel rod withtwo radial supporting arms (1 mm) was inserted in thecentral part of the column to position the thermocouples.The thermocouples were carefully wired along the rodwith thermocouple tips protruding into the bed side toensure that they were bathed in the surrounding flow.Particular attention was also paid to connecting the ther-mocouple wires to the data-logger through the supportingrod to minimise disturbance to the flow and temperaturefields. Axial temperature profile was measured in the col-umn centre by seven thermocouples located at seven axialpositions of 0, 188, 379, 579, 764, 964 and 1100 mm fromthe inlet. Radial temperature profiles in two axial posi-tions of 579 and 764 mm from the inlet were obtainedby five thermocouples in each of the axial position, wherethe thermocouples were located and supported by the twotiny supporting arms. The external surface temperature ofthe packed bed was measured by seven thermocoupleslocated in different axial and tangential positions to mon-itor the wall temperature uniformity. A thermocouple wasalso mounted onto the surface of a packed particlelocated approximately half way between the centre andthe column wall, where another one was positionednearby to measure the fluid temperature passing acrossthe particle so that the temperature difference betweenthe fluid and packed particles could be investigated. Alltemperature signals were collected by a data acquisitionsystem (NI PCI-6052E) inside a PC. A SCXI-1102 32-channel thermocouple amplifier was used to achieve ahigh accuracy of temperature measurements. A Labviewsoftware was used for system configuration and datalogging. 4. Experimental results4.1. Hydrodynamics in the cold stateInitial experiments were carried out using gas phase onlyto check the performance of the experimental system, andto determine the voidage of the packed beds. Fig. 3(a)and (b) show the pressure drop results for both the 50and 100 mm columns, respectively. The one-dimensionalErgun equation [25] is used for determining the voidage.Apparently, this is based on an assumption of constant voi-dage throughout the packed space without the wall effect.The voidage values obtained are therefore the apparentones, which contain the effect of the column wall (seeRef. [14] for more details).Having built up the confidence in the system, gas–solidtwo-phase mixtures were then used in the experiments.Fig. 4 shows the pressure drop (DP) of gas–solid two-phasemixtures flowing through the 50 mm packed column. Anapproximately linear relation is seen between the pressuredrop (DP/L) and solids flux (Gs) given the particle sizeand superficial gas velocity (Ug) under the conditions ofthese runs. As will be discussed later on the effect of tem-perature (hot state experiments), such a linearity only holdsfor a narrow range of the solids loading. Efforts were alsomade to unify the experimental data under different condi-tions. Fig. 5 is the pressure drop given in the format ofEuler number (Eu) as a function of solid-to-gas mass fluxratio (Gs/Gg), where Eu ¼ DP=ðqgU2g=2Þ with qg being thegas density. It can be seen that the data points fordp = 5 mm packed particles collapse into a single curvethough data point scattering is apparent. Effort was thenmade to develop a relationship based on the Ergun equa-tion (see Section 4.3 for more details).Gas–solid two-phase mixtures flowing through packedbeds are complicated. Two cases could occur to the fineparticles moving through the interstices of packed particles,either suspended in the gas flow or trapped between thepacked particles. At the steady state, the amount of fineparticles trapped in the interstices of packed particles isconstant (on the macroscopic scale). The volume ratio ofthese trapped particles to the total volume of the particlesin the gas stream is defined as the static solids hold-up,whereas the volume fraction of the suspended particles is termed dynamic hold-up. The methods for measuring thetwo hold-ups have been detailed elsewhere [13]. From themicroscopic view, fine particles are likely to be trappedfor a certain period time before re-enter the gas streamand the particles may be trapped several times during mov-ing through the packed structures. At a given time, theamount of particles re-enter the gas stream is equal to theamount of particles get trapped, hence the constant staticsolids hold-up is a dynamic concept depending on the timescales of the particle trapping and re-entering the gasstream, or in other words, the time scale of solids exchangebetween the static and dynamic hold-ups. Under the exper-imental conditions of our work, the static hold-up is verysmall in comparison with the dynamic hold-up. This isdemonstrated in Fig. 6 for a superficial gas velocity of1.83 m/s [15]. More experimental results can be found inWang et al. [13] and Ding et al. [14].4.2. Solids motion in the cold stateSolids motion at the single particle level was measuredby using the PEPT technique as outlined in Section 3.1. The results of PEPT experiments are particle position asa function of tracking time, which can be processed to givethe 3-D velocity of the tracer particles in the Lagrangianframework. As the tracer particle passes through thepacked bed many times over the tracking period ( 2 h),we assume that the tracer can visit all positions of theintensities in the packed bed. One can obtain solids velocityin the Eulerian framework by using local average thusenables reconstruction of the velocity vector map of thetracer. Fig. 7 shows an example of the velocity vectormap for 112.5 lm particles flowing through the 50 mm col-umn packed with 10 mm particles. It can be seen that par-ticles move in a tortuous way in most part of the packedbed except for a region close to the wall where the voidagehas the highest value due to the wall effect. Fig. 8 shows theradial distribution of axial solids velocity. One can see thatthe axial velocity changes with the radial position in a peri-odic manner, which is believed to be associated with voi-dage distribution in the packed bed (see Section 5 formore discussion).The PEPT results also allow calculation of the occu-pancy distribution of the tracer particle, where the occu-pancy is defined as the total time the tracer spends in a voxel pided by the total tracking time. Note that theoccupancy is different from particle concentration butshould be an indication of particle concentration in awell-mixed system without dead zone. Fig. 9 shows anexample of the radial distribution of the occupancy wherethe radial position, r, has been normalized by the packedparticle diameter. It is clear that the suspended particlesspend most of their time in the region close to the walldue to high voidage.4.3. Hydrodynamics in the hot stateExperiments in the hot state were performed at 100  C.Fig. 10(a) shows the pressure drop at 100  C as a functionof solid mass flux, Gs, for 112.5 lm glass beads under var-ious superficial gas velocities. One can see that, given thesuperficial gas velocity, pressure drop at 100  C increases,in a nonlinear manner, with increasing solids loading, par-ticularly at low solids loadings. At high solids loadings, therelationship between the pressure drop and solids loadingis approximately linear, in agreement with the cold stateexperimental results as discussed in Section 流经填料床的气固两相流的流体力学和传热英文文献和中文翻译:http://www.751com.cn/fanyi/lunwen_42059.html