Introduction to Gassolid Fluidized Bed Reactors Professor M

Introduction to Gas-solid Fluidized Bed Reactors Professor M. H. Al-Dahhan CHEMICAL REACTION ENGINEERING LABORATORY

Outline/Contents ü ü ü ü Introduction. Fluidization Flow Regimes. Overall Gas (Voidage) and solids Hold-up. Radial and Axial Solids Hold-Up Profiles. Radial and Axial voidage distribution. Gas and Solid Mixing. Scale-Up. Reactor Modeling. CHEMICAL REACTION ENGINEERING LABORATORY

INTRODUCTION CHEMICAL REACTION ENGINEERING LABORATORY

Fluidized Bed Reactor Components Inlet to cyclone The material fluidized is a solid (catalyst). The fluidizing medium is either a gas or a liquid. Gas distributor CHEMICAL REACTION ENGINEERING LABORATORY

Advantages v It has the ability to process large volumes of fluid. v Excellent gas-solid contacting. v Heat and mass transfer rates between gas and particles are high when compared with other modes of contacting. v No hot spot even with highly exothermal reaction. v Ease of solids handling. Disadvantages v Broad residence time distribution of the gas due to dispersion and bypass in the form of bubbles. v Broad residence time distribution of solids due to intense solids mixing. v Erosion of internals. v Attrition of catalyst particles. v Difficult Scale-up due to complex hydrodynamics. CHEMICAL REACTION ENGINEERING LABORATORY

Industrial Applications of Fluidized Bed Reactor ü ü ü ü Acrylonitrile by the Sohio Process. Yang 2003 Fischer-Tropsch Synthesis. Phthalic anhydride synthesis. Methanol to gasoline and olefin processes. Cracking of Hydrocarbons (Fluid Catalytic Cracking, etc). Coal combustion. Coal gasification Cement clinker production. Titanium dioxide production. Calcination of AL(OH)3. Granulation drying of yeast. Heat exchange Absorption Nuclear energy (Uranium processing, nuclear fuel fabrication, reprocessing of fuel and waste disposal). CHEMICAL REACTION ENGINEERING LABORATORY

Fluidization Flow Regimes CHEMICAL REACTION ENGINEERING LABORATORY

Geldart's Classic Classification of Powders ü Group A (Aeratable) : - (e. g. , Ammoxidation of propylene) small mean particle size and/or low particle density (<~1. 4 g/cm 3), gas bubbles appear at minimum bubbling velocity (Umb). Kunii and Levenspiel (1991) ü Group B (Sand-Like) : - (e. g. , Starch) particle size 40 μm to 500 μm and density 1. 4 to 4 g/cm 3, gas bubbles appear at the minimum fluidization velocity (Umb). ü Group C (Cohesive) : - very fine particle, particle size < 30 μm, difficult to fluidize because inter-particle forces are relatively large, compared to those resulting from the action of gas. ü Group D (Spoutable) : - (e. g. , Roasting coffee beans) large particle, stable spouted beds can be easily formed in this group of powders. Diagram of the Geldart classification of particles, Geldart (1973 ).

Flow Regimes in Fluidized Beds J. Ruud van Ommen, 2003 CHEMICAL REACTION ENGINEERING LABORATORY

Minimum Fluidization Velocity This equation can be used to calculate the minimum fluidization velocity U if the void fraction emf at incipient fluidization is known. Experimentally, the most common method of measurement requires that pressure drop across the bed be recorded as the superficial velocity is increased stepwise through Umf and beyond, Umf is then taken at the intersection of the straight lines corresponding to the fixed bed and fluidized bed portions of the graph obtained when is plotted against U on log-log coordinates. Kunii and Levenspiel (1991)

Bubbling Fluidization ü This type of fluidization has been called ‘aggregative fluidization’, and under these conditions, the bed appears to be divided into two phases, the bubble phase and the emulsion phase. ü The bubbles appear to be very similar to gas bubbles formed in a liquid and they behave in a similar manner. The bubbles coalesce as they rise through the bed. CHEMICAL REACTION ENGINEERING LABORATORY

Turbulent Fluidization Turbulent regime has the following features: - ü High solid hold-ups (typically 25 -35 % by volume). ü Limited axial mixing of gas. ü Suitable for exothermic and fast reactions. ü Good gas-solid contact and hence, favors reactant conversion. ü high gas flow-rates operation and good for isothermal operation. ü Favorable bed to surface heat transfer. Canada et al. 1978 CHEMICAL REACTION ENGINEERING LABORATORY

Some commercial processes in turbulent fluidization Process Particle classification Typical gas velocity (m/s) FCC regenerators Group A 0. 5 -1. 5 Acrylonitrile Group A ~0. 5 Maleic anhydride Group A ~0. 5 Phthalic anhydride Group A ~0. 5 Ethylene dichloride Group A ~0. 5 Roasting of zinc sulfide Group A ~1. 5 Bi et al. 2000 CHEMICAL REACTION ENGINEERING LABORATORY

Fast Fluidized Bed ü The fast fluidization occurs as a result of continuing increasing in operating velocity beyond that required at turbulent fluidization, a critical velocity, commonly called the transport velocity (Utr), will be reached where a significant particle entrainment occurs. ü The CFB has significant industrial applications because of its efficiency, operational flexibility, and overall profitability (Berruti et al. , 1995). CHEMICAL REACTION ENGINEERING LABORATORY

Transition between Fluidization Regimes. ü Grace (1986 a) summarized the effects of particles properties and operating conditions on fluidization behavior and prepared a flow regime diagram. The flow regime diagram was further modified by Kunii and Levenspiel (1997). ü For given particles and operating velocity, the gas-solid contact pattern can be determined using this diagram. Likewise, for a given flow regime, this diagram could provide available combinations of particle properties and gas velocity. Yang 2003

Fluidization diagram Solid hold-up Yerushalmi and Cankurt, 1970

Methods for Regime Transition Identification Several measurement methods have been utilized to determine the transition from bubbling or slugging to turbulent fluidization which can be classified into three groups: - ü ü Visual Observation, . Pressure Drop-versus Velocity diagram. local and overall bed expansion. Based on signals from pressure transducers, capacitance probes, optical fiber probes, X-ray facilities. Bi et al. 2000 CHEMICAL REACTION ENGINEERING LABORATORY

Generalized effect of operating and design parameters on flow regime transition Parameter Effect on flow regime transition Pressure In general, pressure accelerates the flow regime transition, thereby decrease transition velocity (Lanneau , 1960, Cai et al. 1989, Yates 1996). Temperature Transition velocity increases as the temperature is increased, (Peeler et al. , 1999, Cai et al. , 1989 and Foka et al. , 1996). Static Bed Height The transition velocity was almost independent of the static bed height, which varied from 0. 4 to 1. 0 m (Grace and Sun 1990). Similar results were reported by Cai (1989) and Satija and Fan (1985) with (Hmf/Dt) > 2. On the other hand, for a shallow fluidized bed of (Hmf/Dt) < 2 with Group B and D particles, Canada et al. (1978) and Dunham et al. (1993) found that Uc increased with static bed height. This could be related to the undeveloped bubble flow in shallow beds before transition to turbulent fluidization can occur (Bi et al. 2000). Particle Size and Density Uc increases with increasing mean particle size and density (Cai et al. 1989, Bi et al. 2000). Column Diameter Transition velocity decreases with increasing column diameter for small column, becoming insensitive to column diameter for Dt > 0. 2 m, (Cai, 1989). Similar trends were observed by (Zhao and Yang, 1991) with internals. Internals Transition to turbulent fluidization tends to occur at lower gas velocities in the presence of internals which usually restrict bubble growth and promote bubble breakup.

Effect of column diameter Cai (1989) ü Uc decreases with increasing column diameter for small columns (less than 2 m), becoming insensitive to column diameter for Dt > 0. 2 m. ü Similar trends were observed by Zhao and Yang (1991) in columns with internals. CHEMICAL REACTION ENGINEERING LABORATORY

Some Selected References ü Cai et al. , 1989, “Effect of operating temperature and pressure on the transition from bubbling to turbulent fluidization”, AICHE Symposium series, 85, 37 -43. ü Chehbouni et al. , (1994), “Characterization of the flow transition between bubbling and turbulent fluidization”, Ind. Eng. Chem. Res. , 33, 1889 -1896. ü Bi et al. , (2000), “A state-of-art review of gas-solid turbulent fluidization”, Chemical engineering science, 55, 4789 -4825. ü Andreux et al. (2005), “New description of fluidization regimes”, AICHE Journal, 51, No. 4, 1125 -1130. CHEMICAL REACTION ENGINEERING LABORATORY

OVERALL GAS (VOIDAGE) AND SOLID HOLDUP CHEMICAL REACTION ENGINEERING LABORATORY

Overall gas holdup It is an important hydrodynamic parameter which is defined as the fraction of reactor dynamic volume occupied by the gas. Typical relationship between overall gas (voidage) holdup and superficial gas velocity in where is shown in following schematic Avidan and Yerushalmi, 1970

Effect of operating and design parameters on gas holdup or bed density Inertial bed height It is independent on initial bed height (Hilal et al. , 2002). Particle size The dimensionless density ( / mf) decreases as the particle size is reduced. The bed expansion is larger for a wide than a narrow distribution of particles. (Grace and Sun, 1991). Particle density / mf decreases as the particle density decreases. Column diameter The bed expansion increases with increasing bed diameter. Temperature The voidage increases with increasing temperature. CHEMICAL REACTION ENGINEERING LABORATORY

Hilal et al. 2002 Effect of column diameter ü The bed expansion increases with increasing bed diameter (Volk et al. 1962, Xavier et al. , 1978). ü The bed expansion decreases with increasing beds, a condition he attributed to the development of bubble channeling in the larger beds (De-Groot 1967). ü The bed density is greatest for the smaller diameter bed at the same excess velocity (Hilal et al. , 2002). Matsen 1996

Effect of pressure ü Higher operating pressures reduced the bed expansion (H/Hmf) (Miller et al. , 1981). ü The increase of bed expansion with pressure (Chiba et al. , 1986, and Chitester et al. , 1984). ü The physical properties of the fluidizing gas, density and viscosity did not have any significant effect on bed expansion (Denloye, 1982), and Knowlton, 1977). ü Bed expansion increased significantly with pressure but this influence, very strong at low pressures, seemed to reach a maximum at approximately 800 k. Pa and decreased thereafter up to 1200 k. Pa (Llop et al. , 1995; 2000, and Olowson and Almstedt, 1990). Some conflict !!!!!

Some Selected References ü Avida and Yerushalmi (1982), “Bed expansion in high velocity fluidization”, Powder technology, 32, 223 -232. ü Meller et al. , (1984), “The effect of particle density on the hold-up in a fast fluid bed”, AICHE Symposium series, No. 234, 80, 52 -59. ü Lee and Kim (1990), “Bed expansion characteristics and transition velocity in turbulent fluidized beds”, 62, 207 -215. ü Hilal et al. , (2002), “Solid hold-up in gas fluidized beds”, Chemical engineering and processing, 41, 373 -379. CHEMICAL REACTION ENGINEERING LABORATORY

Radial and Axial Solids Hold-Up Profiles CHEMICAL REACTION ENGINEERING LABORATORY

Radial Profile ü Although, overall gas holdup has been traditionally used for characterization of hydrodynamics of fluidized bed columns, it is a single lumped parameter. Hence, for detailed characterization, one need to look at the way solid is distributed across the reactor. Mabrouk et al. 2005 U=0. 53 m/s, sand particle (250 microns, 2. 5 g/cm^3) Bubbling regime, Fiber optical needle probe ü The local solid holdup was greater near the wall compared to that near the centerline and that the radial particle velocity was nearly parabolic (Van Zoonen, 1962; Mabrouk et al. 2005). CHEMICAL REACTION ENGINEERING LABORATORY

Axial Profile The axial solid hold-up obtained by fiber optical needle probe and CARPT shows a quasi linear profile (Mabrouk et al. 2005). Mabrouk et al. 2005 CHEMICAL REACTION ENGINEERING LABORATORY

Measurement techniques of Radial and Axial Solids Hold-Up Profile CARPT Mabrouk et al. 2005 CHEMICAL REACTION ENGINEERING LABORATORY

Some Selected References ü Bi et al. , (2000), “A state-of-art review of gas-solid turbulent fluidization”, Chemical engineering science, 55, 4789 -4825. ü Mabrouk et al. , “Scale effects on fluidized bed hydrodynamics” Inter. J. of Chemical Reactor Eng, 2005. ü Schweitzer et al. , (2001), “Local gas hold-up measurement in fluidized bed and slurry bubble column. CHEMICAL REACTION ENGINEERING LABORATORY

Gas and Solid Mixing CHEMICAL REACTION ENGINEERING LABORATORY

(a) Axial Solid Mixing Lee and Kim 1990 Du et al. 2002 CHEMICAL REACTION ENGINEERING LABORATORY

(b) Radial Solid Mixing Du et al. 2002 CHEMICAL REACTION ENGINEERING LABORATORY

Gas Mixing (a) Axial Gas Mixing Foka et al. 1996 CHEMICAL REACTION ENGINEERING LABORATORY

Selected gas mixing studies Investigators Model Tracer injection dp (µm) D(m) U (m/s) Uc (m/s) Dzg (m 2/s) Lee and Kim (1989 b) (Air-CO 2) Diffusion process with axial and radial dispersion coefficients Steady state 362 0. 1 0. 88 1. 00 1. 08 1. 20 0. 85 0. 22 0. 235 0. 230 0. 245 0. 215 Li and Weinstein (1989) (Air-He) One dimensional dispersion Steady state 59 0. 152 0. 1 0. 5 1. 3 0. 43 0. 1 0. 55 0. 60 Li and Wu (1991) (Air-H 2) 1 D pseudohomogeneous diffusion Non-ideal pulse 58 0. 09 1. 0 0. 44 0. 45 0. 51 0. 56 Foka et al. (1994) (Air-Ar) One dimensional dispersion Pulse 75 0. 1 0. 417 0. 516 0. 614 0. 691 0. 792 0. 892 0. 977 1. 051 1. 142 0. 47 0. 080 0. 102 0. 11 0. 195 0. 130 0. 167 0. 097 0. 060 0. 075 Foka et al. (1996) (Air-Ar) Two-phase model of van Deemter (1980) Pulse (less than 0. 5 s) 75 0. 1 0. 21 0. 4 0. 5 0. 6 0. 7 0. 8 0. 94 0. 55 0. 09 0. 16 0. 19 0. 175 0. 14 0. 13 0. 14 Zhang et al. (1996) (Air-O 2) Pseudo-homogeneous model with axial and radial dispersion Steady state 77. 6 0. 19 0. 392 0. 588 0. 784 1. 078 0. 5 0. 374 0. 514 0. 619 0. 783 Wei et al. (1993) (Air-flue gas) One dimensional dispersion Steady state 58 5. 76 1. 26 1. 41 0. 41 3. 05 3. 4 Warsito et al. (2002) (helium and phosphor) 2 -D Dispersion model Unsteady state 60 0. 203 0. 21 -1. 5 0. 5 Plotted in Fig.

(b) Radial Gas Mixing For turbulent fluidized beds, almost all gas mixing studies have been concentrated on the axial mixing, very limited information is available regarding the radial gas mixing (Du et al. 2002). Lee and Kim 1989 Du et al. 2002

Solids flow pattern and mixing Radioactive particle tracking technique for solids mixing investigations Mostoufi and Chaouki, 2001 152 mm ID, 1500 mm in height Experimental setup and the used detectors configuration

Radioactive particle tracking selected results Mostoufi and Chaouki, 2001

Solids diffusivities Mostoufi and Chaouki, 2001

Velocity field, velocity gradient and axial solid diffusivity Mostoufi and Chaouki, 2001

Some Selected References ü Lee and Kim (1989), “Gas mixing in slugging and turbulent fluidized beds”, Chem. Eng. Comm. , 86, 91 -111. ü Foka et al. , (1996), “Gas phase hydrodynamics of a gas-solid turbulent fluidized bed reactor”, Chemical engineering science, No. 5, 51, 713 -723. ü Du, B. , Fan, L. -S. , Wei, Fan, Warsito, W. , “Gas and solids mixing in a turbulent fluidized bed”, AICh. E Journal, 48, No. 9, 1896 -1909. CHEMICAL REACTION ENGINEERING LABORATORY

Fluidized Bed Scale-up CHEMICAL REACTION ENGINEERING LABORATORY

Scale-up criteria Glicksman et al, 1993, 1998 Sanderson and Rhodes, 2005 Horio et al. , 1986 van den Bleek and Schouten, 1996 CHEMICAL REACTION ENGINEERING LABORATORY

Sanderson and Rhodes, 2005 Properties of the Silica Sand Bed Materials Used in the Similarity Experiments Vertical distance from top surface of distributor plate to each pressure tapping point. The tapping point heights correspond to the same dimensionless probe height (h/Hs) at each scale.

Scale-up criteria evaluation in small scale fluidized beds Results for the average absolute deviation of dimensionless pressure for correct and misscaled beds. Materials A and B in the 146 - and 300 -mm beds, respectively, are correctly scaled. Materials A* and B* in the 146 - and 300 -mm beds, respectively, are also correctly scaled, but different from the A–B pair. Comparison of the dimensionless average cycle frequency for the pressure fluctuation data for all preliminary experiments. Sanderson and Rhodes, 2005 CHEMICAL REACTION ENGINEERING LABORATORY

Scale-up criteria evaluation in large scale fluidized beds Ranges of Superficial and Dimensionless Superficial Gas Velocities and Particle Reynolds Number for the Hydrodynamic Similarity Experiments* Comparison of the normalized ensembleaveraged amplitude spectra for the dimensionless pressure fluctuations from the 146 -mm bed with material A and the 300 -mm bed with mismatched bed material B* at low gas velocity. Sanderson and Rhodes, 2005 CHEMICAL REACTION ENGINEERING LABORATORY

Sanderson and Rhodes, 2005 Comparison of the dimensionless average absolute deviation of pressure measured from pressure probes located at h/Hs=0. 77 and r/R=0 in all five fluidized beds for a range of dimensionless gas velocities. Comparison of the dimensionless average cycle frequency of pressure measured from pressure probes located at h/Hs=0. 46 and r/R = 0 in all five fluidized beds for a range of dimensionless gas velocities. All beds, with the exception of the 600 -mm bed with material D, have been scaled using the simplified scaling criteria. CHEMICAL REACTION ENGINEERING LABORATORY

Sanderson and Rhodes, 2005 Agreement map showing qualitatively how well the pressure fluctuations from the various probe locations and superficial gas velocities from 1. 25 to 3. 5 Umf match for the scaled fluidized beds. Black dots indicate the location of the probe tips in the actual measurement runs; the results have been extended across the bed width assuming the behavior to be axisymmetric (excellent agreement trends are indistinguishable; good agreement trends are similar with some scatter; poor agreement trends are only marginally better than for the misscaled scenario). CHEMICAL REACTION ENGINEERING LABORATORY

Sanderson and Rhodes, 2005 Comparison of the normalized probability distributions for the correctly scaled beds (300 mm, material B; 1560 mm, material D) with the mismatched bed (600 mm, material D) at low gas velocity for the probe located at r/R=0 and h/H=0. 2. Comparison of the normalized probability distributions for the correctly scaled beds (146 mm, material A; 300 mm, material B; 1560 mm, material D) at high gas velocity for the probe located at r/R=0 and h/H=0. 77. CHEMICAL REACTION ENGINEERING LABORATORY

Additional evaluation for scale-up criteria, Glicksman et al. , 1993 CHEMICAL REACTION ENGINEERING LABORATORY

Low velocity High velocity Solid fraction profiles, glass particles Solid fraction profiles, plastic particles

Selected References 1. Sanderson, John, and Rhodes, Martin, Bubbling Fluidized Bed Scaling Laws: Evaluation at Large Scales, AICh. E Journal, 2005; 51 (10): 2686 -2694. 2. Glicksman LR, Hyre M, Woloshun K. Simplified scaling relationships for fluidized beds. Powder Technol. 1993; 77: 177 -199. 3. Horio M, Nonaka A, Sawa Y, Muchi I. A new similarity rule for fluidized bed scale-up. AICh. E J. 1986; 32: 1466 -1482. 4. Glicksman LR. Scaling relationships for fluidized beds. Chem Eng Sci. 1988; 43: 1419 -1421. 5. van den Bleek CM, Schouten JC. Deterministic chaos: A new tool in fluidized bed design and operation. Chem Eng J. 1993; 53: 75 -87. 6. Schouten JC, van der Stappen MLM, van den Bleek CM. Scale-up of chaotic fluidized bed hydrodynamics. Chem Eng Sci. 1996; 51: 1991 - 2000. 7. Glicksman LR, Hyre MR, Farrell PA. Dynamic similarity in fluidization. Int J Multiphase Flow Suppl. 1994; 20: 331 -386. 8. Glicksman LR. Fluidized bed scale-up. In: Yang W-C, ed. Fluidization Solids Handling and Processing—Industrial Applications. Park Ridge, NJ: Noyes; 1999. CHEMICAL REACTION ENGINEERING LABORATORY

Reactor Modeling CHEMICAL REACTION ENGINEERING LABORATORY

Review of Fluidized bed reactor modeling (Mahecha and Grace et al. 2006). Predicting the behavior of a gas-solid fluidized-bed reactor requires information on the stoichiometry, thermodynamics, heat and mass transfer, reaction rates and flow pattern of the different phases in the reactor (Kunii, Levenspiel, 1990). Many reactor models have been proposed for fluidized bed reactors. ü In addition to those reviewed by Yates (1983), Crace (1986) and Ho (2003), more recent ones include (Thompson, Bi et al. 1999), (Abba, Grace et al. 2003) and (Chen, Yang et al. 2004). ü Each of these incorporate a different set of assumptions leading to a different set of mathematical expression to simulate the reactor. ü Most models are developed for a specific process, or else so simplified that they cannot adequately describe all important features of reactors and processes of real practical interest. Moreover, the available models are overwhelmingly restricted to steady state operation. ü While progress has been made in adding some of the complexities encountered in practice, e. g. allowance for gradual transitions between flow regimes (Thompson, Bi et al. , 1999; Abba, Grace et al. , 2003), volume change due to reaction (Abba, Grace et al. , 2002), membranes to selectively introduce or remove one species (Chen, Prasad et al. , 2003), and use of a sorbent to selectively capture one product component (Prasad, Elnashaie, 2004). ü Until 2005 there are no models general enough to incorporate all of these features. Recent work has been done to handle and include all these features (Mahecha and Grace et al. 2006), while also facilitating the analysis of dynamic behavior.

FUNDAMENTAL DIFFERENTIAL DYNAMIC MODEL FOR CATALYTIC SYSTEMS ü “The model is initially developed in rectangular coordinates for simplicity, but can be transformed to any other coordinate system (e. g. cylindrical curvilinear) using elementary vector calculus theory of vector operators (Mahecha and Grace et al. 2006). ü This model includes most existing fluid bed reactor models as special cases, allowing clear connections to be established among the models and showing the significance and implications of each simplifying assumption. This will lead to a more systematic approach to fluidized-bed reactor modeling, facilitating what has been called the “optimum degree of sophistication” (Aris, 1961). ü Once the more general model has been developed and debugged, we will be in a position to apply it to important and potentially viable industrial processes such as partial oxidation reactions and hydrogen production processes (Mahecha and Grace et al. 2006). ” CHEMICAL REACTION ENGINEERING LABORATORY

Generalization of Models (Mahecha and Grace et al. 2006). The set of generalizations for the model is as follows: 1) “The dynamic equations take into consideration in a rigorous manner the heat and mass capacities of the gases and solids in each pseudo-phase (Elnashaie, Elshishini, 1993). 2) The model equations can be written in any coordinate system. 3) The development is for a system of “NC” components and “NR” reactions, depending on the feedstock/reactions. 4) The model is not restricted to a single flow regime. Its hydrodynamic parameters can be calculated as proposed by (Abba, Grace et al. , 2003) for several adjacent flow regimes. 5) Both mass and heat dispersion are included along all coordinate axes (Bird, Stewart et al. , 2002). 6) The model deals with anisotropic mass diffusion and heat conduction. 7) The model takes into consideration three-dimensional convective velocities (Bird, Stewart et al. , 2002). 8) The convective velocities can be calculated using any function (e. g. accounting for changes in the number of moles and gas volumetric flow (Abba, Grace et al. , 2002)). Changes with time, temperature, pressure and chemical reaction are also covered. ” CHEMICAL REACTION ENGINEERING LABORATORY

Generalization of Models (cont. ) (Mahecha and Grace et al. 2006). 9) “The model accounts for catalyst chemisorption (Elnashaie, Elshishini, 1993) and solid capture of any species. 10) Hydrodynamic parameters are obtained from appropriate correlations and equations relevant to the different flow regimes (Grace, Abba et al. , 1999). 11) The model accounts for deactivation of catalyst (Chen, Yan et al. , 2004). 12) The model considers the use of membranes to remove certain products (i. e. to break thermodynamic barrier) or to supply certain reactants (i. e. to improve the system selectivity to a desired product). Membrane deactivation fuctions can also be included (Raich & Foley, 1995). 13) The catalyst effectiveness factor may differ from “ 1” (Elnashaie, Elshishini, 1993). 14) In the energy balance, different expressions for calculating the internal energy (Smith, Van Ness et al. , 1996) can be used including, where appropriate, sensible and latent heats (in case of change of phase). 15) The reactor cross-sectional area can vary along the height of the reactor. The model does not need to be modified when using different geometries. ” CHEMICAL REACTION ENGINEERING LABORATORY

Pseudo-phase approach Control volumes for the conservation balances include both gas and solid phases, without ignoring the effect of the solids on the system dynamics (Gas carried inside the solids and the heat and mass capacitances of the solids are included in the mole and energy balances). Solid sorbent (seq) Terms are included for any non-catalytic solid phase, which sorbs/captures any of the species in the reactor (i. e. for carbon dioxide capture to enhance steam reforming and separate CO 2 for subsequent sequestration). CHEMICAL REACTION ENGINEERING LABORATORY

Mole and Energy Fundamental balances Mahecha and Grace et al. 2006). CHEMICAL REACTION ENGINEERING LABORATORY

Mole Balance The molar rate balance over a differential element for phase (p) is given by: The number of mole balance equations is NC. N(P) where NC is the number of chemical species and N(P) is the number of pseudo-phases. The generalized mole balance of each compound in phase (p) is as follows: - CHEMICAL REACTION ENGINEERING LABORATORY

Energy Balance The differential energy balance for phase (p) is given by: Energy dissipation due to viscous effects is neglected. The number of energy balance equations is N(P) where N(P) is the number of pseudo-phases. The generalized energy balance for phase (p) is as follows: - CHEMICAL REACTION ENGINEERING LABORATORY

Pressure Balance A simplified differential pressure balance in the z direction for phase (p) is given by: The density of phase (p) can be calculated using the void fraction as: CHEMICAL REACTION ENGINEERING LABORATORY

Boundary and Initial Conditions ü ü ü The differential control volume of pseudo-phase (p) has no external exchange with the surroundings. The interaction of the pseudo-phase with its surroundings should thus be included in the boundary conditions. The boundary conditions should be specified according to the geometric arrangement of the system, and may vary from case to case. The boundary conditions (i. e. for the simplest single-phase case) may assume axial symmetry, zero flux at the walls and Danckwerts criteria when the diffusion in the fore and aft sections is negligible (Danckwerts, 1953). A base set of boundary conditions is displayed in Table 1. Other details of the model can be found elsewhere (Mahecha-Botero, Grace et al. , 2005). (Mahecha and Grace et al. 2006).

CASE STUDY: APPLICATION OF MODEL TO AN OXYCHLORINATION FLUIDIZED-BED REACTOR (Mahecha and Grace et al. 2006) “Here, as an example of application of the comprehensive model, it simulates an industrial scale fluidized bed reactor which is carried out with special emphasis on the oxychlorination process as a means of producing ethylene dichloride (EDC) from ethylene (ETY). While this represents a simplified special case of the full model, it demonstrates many of the features of the model, while also facilitating verification of the numerical code (written in Matlab 7), since this case has already been solved previously (Abba et al. , 2002) using g-PROMS. The ethylene oxychlorination process involves complex reactions with non -linear temperature dependence (Abba, Grace et al. , 2002). Despite the great industrial impact of oxychlorination reactions, few studies are available in the literature (Carrubba, Spencer, 1970) and detailed studies (e. g. (Ellis, Abba et al. , 2000) are proprietary. ” CHEMICAL REACTION ENGINEERING LABORATORY

CASE STUDY: APPLICATION OF MODEL TO AN OXYCHLORINATION FLUIDIZED-BED REACTOR (Cont’d) The reaction network was simplified as suggested by (Abba, Grace et al. , 2002). We assume that the main product is EDC. Byproducts include a few percent of carbon oxides (COx) and less than one percent chlorinated hydrocarbons (IMP) that exclude EDC. Reactor parameters CHEMICAL REACTION ENGINEERING LABORATORY

Results Predicted steady-state ETY molar flows in the high - and low-density pseudo-phases vs height in the reactor. Predicted steady-state oxygen molar flows in the high- and low-density pseudo-phases vs height. (Mahecha and Grace et al. 2006). Predicted steady-state HCl molar flows in the high- and low-density pseudo-phases vs height. Predicted steady-state EDC molar flows in the high- and low-density pseudo-phases vs height. Mahecha and Grace et al. 2006).

Results (Cont’d) (Mahecha and Grace et al. 2006). Predicted steady-state H 2 O molar flows in the high- and low-density pseudo-phases vs height. Predicted steady-state impurity molar flows in the high- and lowdensity pseudo-phases vs height. Predicted steady-state COx molar flows in the high- and low-density pseudo-phases vs height. Pressure vs reactor height. Predicted axial profile of steadystate overall ETY conversion.

Remarks (Mahecha and Grace et al. 2006). • The generalized dynamic model provides a new approach for simulating complex fluidized–bed catalytic systems. • The model is able to describe fluidized bed reactor systems relying on fewer assumptions than other models in the literature. When different combinations of assumptions are incorporated in the model, it simplifies to a number of fluid bed reactor models previously presented in the literature. CHEMICAL REACTION ENGINEERING LABORATORY