CFD Model of Dense Gas-solid Systems in Jetting Fl uidized Beds

CHEM. RES. CHINESE U.2002. 18(2), 117- 120CFD Model of Dense Gas-solid Systems in Jetting Fluidized Beds *Kai Zhang°"，Jiyu Zhang and Biiang Zhang06 AInstitute of Coal Chermistry, Chinese Academy of Sciences, P.O. Box 165, Taiyuan 030001, P. R. ChinaJohn YatesDepartment of Chemical Engieering , University College London, Lomdon, WC1E 7JE, UKReceived Nov.6, 2001A CFD code has been developed based on the conservation principles describing gas and solid flow in flu-idized beds. This code is employed to simulate not only the spatiotemporal gas- and solid -phase velocities andvoidage profiles in a two-dimensional bed but also fluid dynamics in the jet region. The computational resultsshow that gas flow direction is upward in the entire bed accompanied with random local eirculations, whilst solidflow direction is upward at the center and downward near the wall. The radical reason of strong back -mixing ofsolid particles and good transfer behavior between two phases is that the jet entrains solid particles. Numericalcalculation indicates that gas velocity, solid velocity and pressure profile have a significant change when thevoidage is 0. 8. The simulated timne-averaged voidage profiles agree with the experimental results and simulateddata reported by Gidaspow and Ettechadich (1983). Therefore, CFD model can be regarded as a useful tool tostudy the jet characteristics in dense gas-solid fluidized beds.Keywords Jetting fluidized bed. CFD model, Gas- solid dynamics, Modified SIMPIEArticle ID 1005- 9040(2002)-02-117-04Introductionrole in commercial reactor, such as in an ash agglom-As modification of gas-solid fluidized beds, jet-erating fluidized-bed coal gasifier. In this study weting fluidized beds are widely used in a variety oftry to simulate the hydrodynamic characteristics inphysical and chemical processes because of their goodthe dense gas- solid jetting fluidized bed. A modifiedmixing characteristics, high heat and mass transferSIMPLE algorithm is used to solve the governingrates, and fast chemical reaction. However, the lackequations of the gas-solid fluidized bed. CASICC, aof complete understanding of the fluid dynamicsCFD software, is developed based on the algorithm.makes it difficult to scale up or improve on a jettingThis code produces lots of information concerning gasfluidized bed by judicious selection and control of the and solid dynamics in the bed, especially in the jet re-operation parameters. Computational fluid dynamicsgion.(CFD) is an emerging technique for predicting the Theoretical Model and Numerical Methodbehavior in this system as it can offer an attractive al-Hydrodynamic models of fluidization adopt theternative approach to obtain useful information aboutfirst principles of conservation of mass, momentumhydrodynamic characteristics, especially in those casesand energy. The mass conversion and Navier -Stokeswhere the use of experimental methods is severely re-equations, describing gas and solid flow in the coldstricted by technical constrains. In the last threemodel of fluidized beds, are given in the vector form:decades, many gas- solid CFD models have been putcontinuity equationsforth1-n，which originate from the governinga(e,a)KA2+ V●(epAu)=0,i= (g,s)equations proposed by Anderson and Jackson[8] oIshi[C]. These models have been successfully utilizedmomentum equationsto analyze the hydrodynamics in overall bed although(54)+ v .pl)=-eVP- β(i_ u)+they often differ in terms of the momentum equationsand the closure relations. However , lttle attention isV .i+ epg- G(eg)Ve, i= (g,s), k= (s,g)paid to the jet region, which plays a very importantwhere e represents the volume fraction (εq 十e,=1).中国煤化工* Supported by National Natural Science Foundation of Chine(NdYHCN M H GLondon,# #To whom correspondence should be addressed. Present address: Leparument o1 uneuncas biuguering, University Col-lege London, 1.ondon, WC1E 7JE, UK. E-mail: k. zhang@ucl. ac. uk118CHEM. RES. CHINESE U. .Vol. 18G(Eg) Ve, is ncgligible in the gas momentum equa-vergence in the iterative solution of strongly nonlincarion. In order to solve the above equations, the basic densc multiphase flow equations. CASICC code pro-variables including voidage Eg. pressure P, and phasevides a series of significant computational results ofvelocity vector u should be specified. All other vari- the porosity, the pressure, the gas and solid phascables in the equations need to be derived from thesevelocity fields in either two-dimensional Cartesian orbasic variables(the details see Ref. [3]) according to axisymmetric cylindrical coordinate under either athe closure principles of governing cquations.steady or unsteady state condition. Numerous dataIt is important to select an appropriate numerical are also stored into the text files for the post process-scheme for multidimensional multi-phase fluid analy- ing, such as contour，vector plots, and animations.ses, as it affects not only the solution accuracy but al-Computational Sectionso the efficient use of the computer resourccs. SIM-In order to verify the reliability of CASICC soft-PLE method is regarded as one of the most effcctiveware, the simulation was carried out in a 0. 3937 malgorithms to deal with single fluid flow[I0]. Thiswide and 0. 5844 m high jetting fluidized bed reportedmethod is modified to explore the dense heteroge-by Gidaspow and Ettehadichl6]. Resin and sandsneous two-phase fluidization condition in the study.were used as solid materials, whose physical proper-Both gas phase and solid phasc are treated as inter-ties are listed in Table 1. In a typical computationalpenetrating continua. Scalar quantities (pressure andproccss, solid was loaded into the bed to a specialvoidage) are computed at each cell center, whereasstatic bed height. Air was then introduced into thevelocity components on a staggered grid cainciding bed. The gas inlet velocity was kept at the minimumwith the cell boundaries. Gas continuity equation isfluidization velocity exccpt for a jet nozzle in the cen-employed as a pressure-correction formula and solidter of the bed, i.e.，a jet gas velocity is specified atcontinuity equation is employed to calculate eachthe jet nozzle. When axisymmetrical concept wasphase volume fraction. This new method permits the adopted， the actual computational region wascalculation t0 proceed smoothly to unity and avoids di-0. 19685 mX0.5844 m.Table 1 The physical properties of solid particles'Solid materialParicle site/m Particle density/(kg●m-)Min, fuidization fractioa(->Min. luidization velocity/(m.gr )Resin744X10-614740.4240. 220Sand(No. 1)503X 10~826100, 4020. 282Sand(No.2)630x 10-5660).4400.4161770X10-825500.4220.940， The boundary conditions(B. C.) are; B.C.1 at y=0, a constant gas mas fIux through the distributor plate and the jet nouzle. B.C.2aly= 0.5844 m. P= atmospheric. B.C.3 at r-0, U,s=Up.s=0, due 1o the symmetry about the jet center. B. C.4 at x=0.19685 m, ug=0 andslip boundary for solid phasetI1],Results and DiscussionHowever, the local circulations exist randomly in theTwo- phase calculation produces enormous infor-global circulating flow in both gas and solid velocitymation about the transient and time-averaged gas-fields. It has been found from the transient gas veloc-and solid-velocity vectors, pressure contours anity fields that the shearing layer appears at a certainvoidage distributions. The other data, such as jet re-location depending on the high velocity gas jet. Thegion bydrodynamics, are obtained from the post-pro-instability of shearing layer gradually increases aftercessing program.start-up，which results in jet eruption and bubbleThe gas ising from an orifie might be in theformation. Then the bubble enters the upper diluteform of a bubble , a pulsating jet, or a permanent jet,phase region from the emulsion phase. Moreover ，depending on the operating condition[12]. Whether itthere is gas momentum exchange between jet regionis a bubble, a pulsating jet or a permanent jet, whatand its surrounding emulsion phase. From the tran-is important is that extensive mixing and contactingsient solid velocity fields, it is found that the solidbetween solid and gas immediately appears in the jet-word nrimarilv hy jet action in theting region. This is called the jet in this paper.centr中国煤化工the outer region1 Transient Velocity Fieldsnear|YHCNMHGjetorthebubbleOn the whole, the gas flow direction is upwardinduccs the transport of sold particles in the voidagein the entire bed, whilst the solid flow direction iswake, which contributes to strong back mixing ofupward at the center and downward near the wall.the bed particles, and good mass and heat transfersNo. 2ZHANG Kai et al.119between gas and solid phases. These simulated reequal to the voidage at minimum fluidization state.sults agree with the experimental data of Yang etThe dilute region at the top of the bed represents theal. [3].splash zone where bubbles burst on the dense phase2 Transient and Time-averaged Voidage Profilesinterfacc.Fig. 1 is a typical example of the computational0.28厂0. 28results, which employs No. 1 sand as solid medium.0o.24|(A)0. 24,(B)This Figure shows the visual representation of the0. 20一E 0.20simulated gas and solid distribution in the bed. The1E0.16calculated dstributions of solid volume fraction areconverted into small dots. These dots are distributed”0. 080.08[randomly throughout each computational cell in such0.0a manner that the dot density in the cell [i,j] corre-0.180.120.08 0.04 0. 000.180.120.080.040.00sponds to the computed transient distribution of theWidsh/mWidth/msolid volume fraction, e[i, j]，in the same cell0.280.30[ T[i,j]. This Figure can also reveal jet evolution.0.240.201-0.10a1=0.15 81=0.255S 0.161。100. 080.04-0.20.00.2-0.20.00.2-0.80.0 D.0.000.040.080.120.160 0.000.050.100.15 0.200.5e- 0.60。1-0.50st=C.70。Widuh/m0.4Fig.2 Comparison of experimental time- averagedporosity in the bed,10.(A) Circular ofice(exp. ); (B) rectangular orilice(exp. ); (C) simulation(gidaspow et al. ); (D) Sim-.rularion(CSAICC).wanb/m2-3. widnh/m”“Width/m'3 Jet DevelopmentFig. 1 Representation of the gas solidWhen jet velocity is 13. 57 m/s, the process ofdistribution in the bed.formation and development of jet in the two-dimen-Jet gas velocity 13. 57 m/s.sional fluidized bed is shown in Fig. 3. The jet ap-Fig. 2 shows the comparison of experimental andpears at t=0.05 s, and begins to grow at a latersimulated time- averaged voidage profiles in the bed.time. At t=0.20 s, the first jet has detached fromFig 2(A) and Fig. 2(B) are experimental results bythe top of nozzle to form a bubble, then, a new cycleusing circular orifice and rectangular orifice, respec-starts. Yang et al. [I4] and Luo[is]， using pitot-tubetively, when jet velocity is 5. 773 m/s. Fig. 2(C) istechnique，measured and analyzed the gas velocitysimulated data when jet velocity is 5. 78 m/s56].field in the semi-cylindrical fluidized bed and the two-Fig. 2(D) is the simulation result using CASICC codedimensional fluidized bed，respectively. Their resultswhen jet velocity is 5.78 m/s. It is clear that theindicated that gas-phase velocities changed consider-computational result stated above appears in agree-ably at the jet boundary. In CASICC code, it isment with the experimental results by Gidaspow and2/s 0.20Ettehadieh[6]. The isoporosity contours 6 = 0. 45,t/seq=0.55, es=0. 65 and so on are drawn in the bed.It is noticed that from the inlet of the nozzle to a cer-tain height the voidage is very high(0. 85). Around0.1540.10this dilute region, there exist eliptic isoporosity con-中国煤化工tours. The eliptic contours extend when the jet gas0.0 0.1velocity increases. A slumped region of high solidMHCNMHGconcentration(e= 0.45) is found near the side wallsFig.3 The growth and propagation ofthe jet(Sand No. 1).in the lower part of the bed, which is approximately20CHEM. RES. CHINESE U.Vol.18found that gas velocity ，solid velocity and pressurcparticles. The simulated time-averaged voidage is inprofile change significantly when the voidage is 0. 8,fair agreement with the simulated and experimentalwhich appears to confirm the arbitrary definition byresults reported in the literature. It is reasonable toGidaspow and Ettehdiehto].choose 0.8 as a void boundary because basic flow4 Gas Velocity Profile in the Jet Regionfield variables, including gas velocity, solid velocityFig. 4 shows gas velocity vector profile in the jetand pressure profile, have distinct changes. The jetregion at different time with a typical jet velocity ofregion can be divided into two zones: the lower zone13.57 m/s. It is obvious that the jet region enlargesand the upper zone. The boundary is stable in theas time passes. Here the symbol ‘+’represents thelower zone, whilst it becomes unstable with the in-velocity size and direction. A jet region can be divid-crease of height in the upper zone. A bubble finallyed into two zones: the lower zone and the upperdetaches from the top of the jet.zone. At the lower zone, the jet boundary is conicalMairlomenclatureand quite stable. Particles are entrained into the jetG(t) particle-particle interaction elastic modulus/Paand propelled upward by the high velocity gasstream. In this zone, the axial gas momentum f[luxtime/sdrops rapidly with height. The radial gas momentumvelocity vector/(m "s~ 1)flux also falls very steeply with radial distance from3 gas-solid friction coefficient/(kg. m-1●s")the jet center. The upper zone is the bubble formingvolume fractionregion, the boundary becomes unstable, the axial gasshear viscosity/Pa ●smomentum flux profiles fall gradually and the parti-cles with relatively high momentum flux are thrownp_ density/(kg/m )Operatorback to the emulsion phase，which eventually resultsVgradientin the bubble detaching from the top of the jet.又●divergence一Veclor scale，10 m/sReferencess 0.10叶1] Gidaspow D.. Appl. Mech. Rev.. 1986, 39(1), 12] Gidaspow D.. Muliphase Florw und Fluidication, AcademicPress, 19940. 05-[ 3] Zhang K.. Ph. D. Diseration, Institute of Coal Chenistry.CAS(in Chinee), 19960. 00L[ 4] KuipersJ. A. M. van Swaij w. P. M.，Advunce in Chem.0.025 0.05 0.10 0.15 0.17 0.19Eng., 1998, 24, 227Fig.4 Cas velocity vector profiles in the[5] Hijertager B. H.,ttal.，Ed. : Arastoopour H. AIChE Sym-jet region(Sand No. 1).poiumn Series, 199, 95(321), 1-6[6] Gidaspow D.. Ettehadieh B.. I & EC Fundam., 1983. 22,Conclusion93The CASICC software developed is used to ana-[7] van WachemB. C. M. ral, AIChE J., 2001，47(5)，lyze the flow dynamics in the dense fluidized bed ,1035which indicates that CFD modeling is an effective[8] Anderson T. B, Jackson R.. Ind. Eng. Chem. Fundam.，1967，6, 527method to solve the strongly nonlinear multiphase[9] [shi M. . Thermo-fuid Dynamic Theory of Trwo- Phase Floru,flow equations in the dense fluidized bed. None of theDirection des Eirudes et Rechtrces d' Elecricie de France. Eyparameter values are fitted in this model; all of themrolles, Paris, 1975are obtained directly from available theory or well-es-[10] Patankar s. V.. Numericadl Heat Transfer und FIuid Flow.Hemnisphere Publishing Corporation, 1980tablished empiricism.1] Ding J.. Gidaspow D. AIChE J.. 1990. 36(4). 523The simulated result shows that gas flow direc-[12] Msimil L. , Eds: DavidsonJ. F.. Cift R.. Harrison D..tion is upward in the entire bed, whilst solid flow di-Fluidiration, 2 Editiom. Acadenic Pres, London, 1985. 133rection is upward at the center and downward near[13]C. B，AIChEJ,中国煤化工the wall. Gas momentum exchanges between jet re-[14]d Flow Scaleup Faciligion and its surrounding emulsion phase. TransportMHC N M H Gire Teti DOE/Mc/of solid particles in the bubble wake is induced by jet19122-T1.1983or bubble to form a very strong back -mixing of solid[15] LuoG. H. Ph. D. Dissertation, Institute of Cou! Chemistry,