I NTERNATIONAL J OURNAL OF C HEMICAL R EACTOR E NGINEERING

Volume 5 2007 Article A17

A Comprehensive Approach to Reaction Engineering

Andres Mahecha-Botero Said S.E.H. Elnashaie John R. Grace C. Jim Lim

University of British Columbia, andresm@chml.ubc.ca University of British Columbia, jgrace@chml.ubc.ca Pennsylvania State University, nashaie@chml.ubc.ca University of British Columbia, cjlim@chml.ubc.ca ISSN 1542-6580 Copyright c 2007 The Berkeley Electronic Press. All rights reserved.

A Comprehensive Approach to Reaction Engineering

Andres Mahecha-Botero, John R. Grace, Said S.E.H. Elnashaie, and C. Jim Lim

Abstract
A generalized modeling approach is used to develop a systematic algorithm for formulating and solving chemical/biochemical reaction engineering problems. This systematic approach is general enough that it can treat different systems with varying degrees of complexity utilizing the same methodology. The procedure can be used in both introductory and advanced chemical/biochemical reaction engineering courses. This will provide the students with a powerful toolkit to tackle a wide range of academic and industrial engineering problems as well as a solid starting point for developing research projects in this eld. This may also allow the students to have a better understanding of the multiple phenomena encountered in chemical/biochemical engineering systems and encourage them to prepare models at an optimum level of sophistication for design, optimization, and exploration of novel ideas. KEYWORDS: reaction engineering, reactor models, mechanistic modeling, education, chemical reaction engineering, uidized-bed reactors

Mahecha-Botero et al.: A Comprehensive Approach to Reaction Engineering

1. INTRODUCTION
In the teaching of science and engineering, it is very important to provide a solid background to students. This allows a better understanding of the underlying physical/chemical phenomena and encourages students to advance their knowledge in this field of study. It has been difficult to formulate a robust modeling approach in chemical engineering education. Most learning has been oriented to particular cases only. Although individual case analysis is essential to the development of a chemical engineer, this could limit the attainment of an overall integrated picture of physical and chemical phenomena. A systematic approach is introduced here which we believe can facilitate student understanding of multiple phenomena encountered in chemical engineering systems. The final model presented in this study includes standard forms found in the literature as special cases, allowing for clear connections to be established among the models and showing the significance and implications of each simplifying assumption. This approach for solving chemical reaction engineering problems should encourage students to build more sophisticated models and simulate more complex systems.

2. SYSTEM THEORY AND ITS APPLICATION TO REACTOR ENGINEERING

Mechanistic modeling is an iterative process of representing a system found in nature by an abstract mathematical description based on physical and chemical principles in order to make predictions and gain insights about the systems underlying phenomena. This process attempts to match observations with a set of equations describing and explaining what is observed and/or measured in nature, and predicting the behaviour of a system. Such models are currently utilized in virtually all fields of knowledge, including psychology, medicine, politics, economics, as well as in all branches of science and engineering.

2.1. System Characterization

The first step in mechanistic modeling is to define the system and its boundaries. This step should include a description of how the system interacts with its surroundings. Thus in reaction engineering systems one needs to emphasize the exchange of species and energy to include this information in mole and energy balances.

2.2. Identification of State Variables

The modeler should define what kind of information he/she wants to obtain from the model. The state variables should be chosen to describe the key features of the system and all the relevant information required to define the system. The most important state variables are the concentrations or molar flows of each species, system temperature, and pressure.

2.3. Identification of Independent Variables

This step defines the functional relation of the state variables to the system geometry and their time dependence. This is a very important step which sets the tone for the complexity/simplicity of the system. These independent variables should be chosen carefully depending on the system because they determine the model robustness and usefulness. If time dependence is included, the model is said to be dynamic; otherwise it is assumed to be steadystate. If variation within the system geometry is considered, the model is said to be distributed; otherwise it is lumped. The most widely used dependent variables are time and a single distance coordinate, e.g. distance, volume passed, or weight of catalyst passed in travelling along a tubular reactor.

2.4. Model Development

The modeler should define the relation among the system variables and the system parameters based on a set of governing equations. In general, all systems involve multiple phenomena of different complexity and nature. One of the most important duties of the modeler is to understand, organize and couple these phenomena to represent the

Published by The Berkeley Electronic Press, 2007

International Journal of Chemical Reactor Engineering

Vol. 5 [2007], Article A17

system. A good model must represent key elements of the most important phenomena and the interactions of the state variables in order to reproduce synergistic effects. One may begin with a very simple representation, adding additional elements until an adequate representation of the particular case at hand has been formulated. We prefer to develop a very general model in order to account for the most important characteristics of an overall system. This model can then be simplified by judicious introduction of different assumptions, depending on the particular case to be studied. In reaction engineering the overall process usually involves a combination of fundamental/conservation equations, differential balances and empirical relations. The goal is to reach the optimum degree of sophistication (Aris, 1961), models which satisfy Occams razor principle, i.e. with the elements that are needed, but without extra embellishments.

2.5. Parameter Values (design, operational and physico-chemical)

A set of parameter values is required to solve the model. These parameters can be taken from the literature or from separate but linked experimental efforts designed to reproduce or simulate the conditions of the system under study. Obtaining accurate reaction kinetics and thermo-chemical data is a key requirement for reliable predictive chemical reaction engineering models.

2.6. Simplifying Assumptions

If the same predictions can be obtained by different models, one should prefer the simplest one. For many systems, many different assumptions can be introduced without significantly affecting the accuracy of the results. For example, if the overall rate of reaction is dominated by chemical kinetics, then many different mass transfer approaches could be used with negligible overall impact on the ultimate predictions. Each assumption should be justified either by physical reasoning, experimental findings, mathematical derivations or by the experience of the modeler. Robustness and accuracy should be balanced by the modeler.

2.7. Simulations and Numerical Analysis

After providing a complete set of equations describing the interactions of all system phenomena in terms of the main system variables, a solution is required to obtain the desired results. Solution methods vary in complexity. In the simplest cases this solution may be obtained analytically, but in most applications of practical interest, especially in reaction engineering, numerical simulation is required. Step 2.3 of the above sequence determines the type of numerical technique and software required to solve the model. If the number of independent variables is zero, the resulting equations are transcendental. When only one independent variable is considered, the system can be described by ordinary differential equations (ODEs). If more than one independent variable is included, partial differential equations (PDEs) must be solved. The number of state variables defines the number of equations to be solved. The number of linearly independent equations should equal the number of state variables.

2.8. Model Validation

Every model should be tested against experimental data in order to check its accuracy. In some cases, experimental data must be utilized to establish values of parameters by fitting the model to experimental results. The ultimate test for a model is to be able to predict experimental results without any adjustable parameters. Benchmarks using wellknown systems, limiting case analysis and experimental tests for individual elements of the model should be performed whenever possible. Statistical analysis of the results is always desirable. Any discrepancy should be analyzed and fed back into the previous steps in order to improve the model. Final results can be documented and published for future improvements when exposed to the scrutiny of the scientific community. Figure 1 depicts the iterative process for model development. In this paper, the first 6 steps for the generalized mathematical modeling of reaction engineering systems are treated in a systematic way in order to facilitate learning and implementation.

http://www.bepress.com/ijcre/vol5/A17

Mahecha-Botero et al.: A Comprehensive Approach to Reaction Engineering

System Characterization (Based in onObservation) Observation)

No

Good Agreement

Yes

Identification of State Variables Identification of Independent Variables Parameter Values

(design, operational and physico-chemical)

Experiments

Compare Predictions with Experimental Results Coding Simulations

Model Development

Simplifications

Verification of Results

Figure 1. Comprehensive modeling iterative process.

3. GENERALIZED GOVERNING EQUATIONS FOR REACTION ENGINEERING

The backbone for any reactor engineering model is developed in this section. This backbone may be enriched with many additional relations in order to describe the reacting system. In order to write a single set of equations for many different cases, multidimensional control indices are used in the mole and energy balances. These control indices act as switches to include or discard terms in the balance equations. These indices may then have a value of 0 or 1, depending on the degree of complexity of the model equations. In Section 3.1 a set of balances is developed for a lumped system. These balances are extended to onedimensional distributed systems in Section 3.2. The systems of equations derived in this section are intended for multi-phase systems. Equations for a general pseudo-phase () are derived below allowing for exchange of mass and heat with other pseudo-phases nearby. The pseudo-phase may contain matter in any physical state such as solid, liquid and gas, as well as a combination of them. The balance equations described below should be applied to all pseudo-phases involved in the system of study.

3.1. Lumped Systems (Continuous Stirred Tank Reactors, Batch and Semi-Batch Reactors)
3.1.1 Mole Balance The mole balances for each species in pseudo-phase () (Figure 2) are given by: [Input Output]() + [Chemical reaction generation/consumption]() + [Exchange with other pseudo-phases] () = [Molar accumulation rate] () Initially the mole balance is developed in terms of concentrations as follows:

N NO NR NI + v . C v . C V . . . a . r V . + f ( ) k if ( )k ( ) l i( ) l ( ) ij j ( ) j ( ) j ( ) ( ) aI ( n ) .kci . Ci( n ) Ci( ) l =1 j =1 n =1 k =1 n d = ( ) .V( ) .Ci( ) dt

(1)

i =1,2,NC

Published by The Berkeley Electronic Press, 2007

International Journal of Chemical Reactor Engineering

Vol. 5 [2007], Article A17

where for pseudo-phase (): v( ) is the volumetric flow rate, pseudo-phase (), V( ) is the total volume, reaction j, j
( )

Ci( ) is the concentration of species i in the fluid part of

is the stoichiometric coefficient of species i in

( ) is the void fraction, ij

( )

is the overall effectiveness factor, a j

is the catalyst activity, rj is the rate of reaction j, aI ( n ) is ( )

the interphase transfer area per unit volume between phases () and (n),

k c is the interphase mass transfer coefficient,

and Ci( n ) is the concentration of species i in an adjacent pseudo-phase (n) in contact with pseudo-phase (). In addition, NC is the number of chemical species, whereas N is the number of phases.

(C

i( n )

, T( n ) , T( ) , T(Cool ) , T( n )

(v (v

f ( ) k

, C if ( ) , Ei( ) , T f ( )
k k

) )

External exchange

( k =1)

(v (v

( ) l

, C i( ) , E i( ) , T ( ) , C i( ) , E i( ) , T ( )

) )

( l =1)

f ( ) k

, C if ( ) , Ei( ) , T f ( )
k k

( k = 2)

( ) l

(l = 2)

(v

f ( ) k

, C if ( ) , Ei( ) , T f ( )
k k

V( ) , ( )

(v

( ) l

, C i( ) , E i( ) , T ( )

(k = N I )

(l = N O )

Figure 2. Mole and energy fundamental balances for lumped systems. The following expressions are obtained after expanding the time derivatives and introducing the control indices m for future simplifications:

1. v f ( ) .Cif 2 . v( ) l .Ci + 3 .V( ) .ij . j .a j .rj

k =1
k ( )k

NI

NO

NR

l =1

( )l

j =1

( )

( )

( )

+ 4 .V( ) . aI ( n ) .kci . Ci( n ) Ci( ) = 5 . ( ) .V( ) .

n =1 n

d d Ci( ) + 6 .Ci( ) . ( ( ) .V( ) ) dt dt

( )

(2)

i =1,2,NC In order to account for changes in the volume and/or number of moles, the mole balances can be re-written in terms of molar flow-rates. This form of the equations simplifies the simulations when the overall volume and/or number of moles vary significantly over the course of time. It would be advisable to express the concentrations of species as: Ci( ) =

Fi( ) v ( )

to obtain:

http://www.bepress.com/ijcre/vol5/A17

Mahecha-Botero et al.: A Comprehensive Approach to Reaction Engineering

1. Fif 2 . Fi + 3 .V( ) .ij . j .a j .rj + 4 .V( ) . aI

k =1
( )k

NI

NO

NR

l =1

( )l

j =1

( )

( )

( )

n=1 n

( n )

Fi Fi .kci . ( n ) ( ) v ( n ) v( )

d Fi = 5 . ( ) .V( ) ( ) dt v( )
where

F + 6 . i( ) v ( )

d . ( ( ) .V( ) ) dt

(3)

i =1,2,NC

Fi( ) represents the molar flow rate of species i.

3.1.2 Energy Balance The energy balance based on the First Law of thermodynamics for pseudo-phase () (Figure 2) gives [Rate of flow of heat to the system ]() - [Rate of work done by the system] () + [Energy input by mass flow Energy output by mass flow] () = [Heat accumulation rate] () Initially the mole balance is developed in terms of concentrations and internal energy:

N V . a U .(T T ) + U .V .a .(T T )+ .V .a . . T 4 T 4 + U .V .a .(T T ) (n) ( ) ( ) SR ( ) I ( n ) I SR ( ) SR S M S Cool ( ) Cool Cool ( ) n =1 n

N NO NC (4) . I NC Ws + v f ( ) .Cif( ) .Eif( ) + v f Solids ( ) . Solids .E f Solids ( ) v f ( )l .Ci( ) .Ei( ) + vSolids ( )l . Solids .ESolids ( ) k k k k k k =1 i =1 l =1 i =1

d NC = ( ) .V( ) .Ci( ) .Ei( ) + (1 ( ) ).V( ) . Solids .ESolids ( ) dt i =1

where U I , U SR and U Cool are the interphase, surroundings and cooling heat transfer coefficients respectively. T() represents the temperature,

Ei( ) is the internal energy of species i, ESolids ( ) is the internal energy of the solids in

pseudo-phase () and vSolids ( ) is the volumetric flow rate of solids. The fundamental thermodynamic relation, Ei( ) = H i( ) P.Vi( ) , is introduced. Assuming that the residual value of the internal energy (i.e.

d NC Ci .P.Vi( ) ) is negligible and expanding the time derivative we obtain: dt i =1 ( )

N V . a U .(T T ) + U .V .a .(T T )+ .V .a . . T 4 T 4 + U .V .a .(T T ) (n) ( ) ( ) SR ( ) I ( n ) I SR ( ) SR S M S Cool ( ) Cool Cool ( ) n =1 n

N NO NC . I NC Ws + v f ( ) .Cif( )k .H if( ) + v f Solids ( ) . Solids .H f Solids ( ) v f ( )l .Ci( ) .H i( ) + vSolids ( )l . Solids .ESolids ( ) k k k k k =1 i =1 l =1 i =1 NC d d NC = ( ) .V( ) .Ci( ) . H i( ) + H i( ) . ( ) .V( ) .Ci( ) dt dt i =1 i =1 d d + (1 ( ) ).V( ) . Solids . (H Solids ( ) ) + H Solids ( ) . ((1 ( ) ).V( ) . Solids ) dt dt

( )

(5)

Published by The Berkeley Electronic Press, 2007

International Journal of Chemical Reactor Engineering

Vol. 5 [2007], Article A17

where

H i( ) is the enthalpy of species i and H Solids ( ) is the enthalpy of solids in pseudo-phase ().
Next we substitute

d ( ) .V( ) .Ci( ) from Equation (1) and add it to Equation (5). Using the definition of dt

heat of reaction ( H RX j ) and rearranging we obtain:

V( ) . a I( n )U I .(T( n ) T( ) ) + U SR .V( ) .aSR .(T TS )+ M .V( ) .aSR . . T TS + U Cool .V( ) .aCool .(TCool T( ) )
4 4 n =1 n NR NI NC Ws v f ( ) .Cif( ) . H i ( ) H if( ) + V( ) . H RX j . j ( ) .a j ( ) .rj ( ) k j =1 k =1 i =1 NC d d d = ( ) .V( ) .Ci( ) . H i( ) + (1 ( ) ).V( ) . Solids . (H Solids ( ) ) + H Solids ( ) . ((1 ( ) ).V( ) . Solids ) dt dt dt i =1

( )

(6)

N NC + H i( ) V( ) . aI ( n ) .kci . Ci( n ) Ci( ) i =1 n =1 n

) + (v
NI

k =1

NO . . H vSolids ( )l . Solids .H Solids ( ) f Solids ( ) Solids f Solids ( ) k k l =1

) (

The following expression is obtained after introducing the definition of enthalpy and control indices facilitate future simplifications:

m to

1.V( ) . aI
n =1 n

( n )

U I .(T( n ) T( ) ) + 2 .U SR .V( ) .aSR .(T TS ) + 3 . M .V( ) .aSR . . T TS

4

NI NC + 4 .U Cool .V( ) .aCool .(TCool T( ) ) 5 .Ws 6 . v f ( ) .Cif( ) .Cpi( ) . T( ) T f( ) k k k =1 i =1 NR NC d + 7 .V( ) . H RX j . j .a j .rj = 8 . ( ) .V( ) .Ci( ) .Cpi( ) . (T( ) ) ( ) ( ) ( ) dt i =1 j =1

(7)

d d + 9 . (1 ( ) ).V( ) . Solids .CpSolids( ) (T( ) ) + CpSolids( ) .T( ) . ((1 ( ) ).V( ) . Solids ) dt dt NO NI 10 v f Solids ( )k . Solids .CpSolids( ) .T f( )k vSolids ( )l . Solids .CpSolids( ) .T( ) l =1 k =1 N NC + 11 Cpi( ) .T( ) .V( ) . aI( n ) .kci . Ci( n ) Ci( ) n =1 i =1 n

) (

where Cpi( ) is the heat capacity of species i in pseudo-phase (). The energy balance is re-written in terms of molar flow-rates in order to account for changes in volume or molar flow rate/number:

http://www.bepress.com/ijcre/vol5/A17

Mahecha-Botero et al.: A Comprehensive Approach to Reaction Engineering

1.V( ) . aI
n=1 n

( n )

U I .(T( n ) T( ) ) + 2 .U SR .V( ) .aSR .(T TS ) + 3 . M .V( ) .aSR . . T TS

4 NI

NC + 4 .U Cool .V( ) .aCool .(TCool T( ) ) 5 .Ws 6 . Fif( ) .Cpi( ) . T( ) T f( ) k k k =1 i =1 NR NC Fi d + 7 .V( ) . H RX j . j .a j .rj = 8 . ( ) .V( ) . ( ) .Cpi( ) T ( ) ( ) ( ) v dt ( ) j =1 i =1 ( ) d d + 9 . (1 ( ) ).V( ) . Solids .CpSolids( ) (T( ) ) + CpSolids( ) .T( ) . ((1 ( ) ).V( ) . Solids ) dt dt

( )

(8)

NO NI 10 v Cp T . . . f Solids ( )k Solids Solids( ) f( )k vSolids ( )l . Solids .CpSolids( ) .T( ) l =1 k =1

) (

N Fi Fi + 11 Cpi( ) .T( ) .V( ) . aI( n ) .kci . ( n ) ( ) v i =1 n =1 ( n ) v( ) n

NC

3.1.3 Initial Conditions Except when there is steady-state, the initial conditions to solve this set of ODEs must be specified, i.e. we need to specify the concentrations and temperature inside the reactor at t=0. 3.1.4 Simplifications The general model for lumped systems consists of two main balances (Equations (2) and (7)). These general mole and energy balances can be simplified in many special cases by means of control indexes. Some common simplification strategies are listed below. In real problems, a combination of these simplifications is commonly applicable: a) Continuous Stirred Tank Reactor (CSTR) In this case the change in total volume is negligible: Set b) Fluid pseudo-phase (No solids) In this case the void fraction ( ) = 1 . Also set

6 = 0 .

9 = 10 = 0 6 = 0 .

c) Batch reactor (Closed system) Set 1 = 2 = 6 = 0 . If the change in total reactor volume is negligible: Set d) Semi-Batch reactor (No output) Set

2 = 0

and

d ( ( ) .V( ) ) = v f ( ) . dt

e) Steady state operation Set 5 = 6 = 8 = 0 . f) Isothermal system Set T( ) = T f ( ) . There is no need for an energy balance.

Published by The Berkeley Electronic Press, 2007

International Journal of Chemical Reactor Engineering

Vol. 5 [2007], Article A17

g) Adiabatic operation Set 2 = 3 = 4 = 0 . h) Isolated system This means that the system is adiabatic and closed. i) Single phase Set N = 1 or 4 = between phases). j) Single reaction Set N R = 1 . k) Single input/output Set N I = 1 or/and N O = 1 . l) Appreciable change in number of moles Equations (3) and (8) should be utilized in order to work with molar flows as state variables.

1 = 11 = 0

(For multi-phase systems it is necessary to account for mass and heat transfer

3.2. Distributed Systems (Plug Flow Reactors, Packed Beds, Fluidized-Beds, etc)
In order to account for axial variation, we implement the balances developed in the previous section with the addition of axial dispersion into a differential control volume of length z : 3.2.1 Mole Balance The mole balances for each species in pseudo-phase () (Figure 3): [Convective input Convective output]()+ [Diffusive input Diffusive output]() + [Chemical reaction generation/consumption]() + [Exchange with other pseudo-phases] () = [Molar accumulation rate] () Initially the mole balance is developed in terms of concentrations:

[(v

( ) .Ci( )

) (v
( z)

( ) .Ci( )

) (

( z + z )

]+

NR ( ) ( ) + . A . N N A . z . ( ) ij . j ( ) .a j ( ) .rj ( ) i ( ) ( z ) i ( ) ( z + z ) ( ) ( ) j =1

N + A( ) .z. aI ( n ) .kci . Ci( n ) Ci( ) n =1 n

= . A .z. d C ( ) ( ) i dt ( )

( )

(9)

i =1,2,NC Replacing the diffusive mass flux by its corresponding Fickian form

C N i ( ) = Di . i( ) ( ) z

, dividing by

z and taking the limit as z 0 (note that V = A( ) .z ) the following expression is obtained:

http://www.bepress.com/ijcre/vol5/A17

Mahecha-Botero et al.: A Comprehensive Approach to Reaction Engineering

(C

i( n )

, T( n ) , T( ) , T(Cool ) , T( n )

External exchange

(v

( )

, C i( ) , N i( ) , q( ) z , Ei( ) , P( )
z

(z)

(v
V( ) , ( ) A

( )

, Ci( ) , Ni( ) , q( ) z , Ei( ) , P( )

z

( z + z )

Figure 3. Mole and energy balances for distributed systems

Ci( ) 1 . v( ) .Ci( ) + ( ) . D . i ( ) z z A( ) z

N NR + ij . j .a j .rj + aI .kc . Ci Ci i ( n p ) (n) ( ) j =1 ( ) ( ) ( ) n =1 n

)
(10) i =1,2,NC

( ) .Ci( ) t

)
m

were Di ( ) is the diffusion coefficient of species i in pseudo-phase (). The following expressions are obtained after expanding the time derivatives and introducing control indices to allow for future simplifications:

Ci( ) 1 . v( ) .Ci( ) + 2 . ( ) . D . 1 . i ( ) A( ) z z z

NR + 3 .ij . j .a j .rj ( ) ( ) ( ) j =1

+ 4 . aI ( n ) .kci . Ci( n ) Ci( )

n =1 n

= 5 . ( ) . Ci( ) t

( )

(11)

i =1,2,NC Alternatively, the mole balances can be expressed in terms of molar flow-rates in order to account for changes in volume/number of moles:

Published by The Berkeley Electronic Press, 2007

10

International Journal of Chemical Reactor Engineering

Vol. 5 [2007], Article A17

1 Fi . Fi( ) + 2 . Di( ) . ( ) 1 . A( ) z z z v( )

( )

Fi Fi + 4 . aI ( n ) .kci . ( n ) ( ) v n =1 ( n ) v( ) n
N

NR + 3 .ij . j .a j .rj ( ) ( ) ( ) j =1 F = 5 . ( ) . i( ) t v( )

(12)

i =1,2,NC 3.2.2 Energy Balance The energy balance based on the first law of thermodynamics for pseudo-phase () (Figure 3) can be written: [Rate of flow of heat to the system ]() - [Rate of work done by the system] () + [Energy input by mass flow Energy output by mass flow] () +[Chemical reaction heat generation/consumption] ()+[Diffusive heat input Diffusive heat output]()= [Heat accumulation rate] () Initially the mole balance is developed in terms of concentrations and internal energy:

N A .z. a U .(T T ) + U . A .z.a .(T T )+ . A .z.a . . T 4 T 4 + U . A .z.a .(T T ) (n) ( ) ( ) ( ) ( ) ( ) I ( n ) I SR SR S M SR S Cool Cool Cool ( ) n =1 n . . NC NC A( ) .z. w s + v( ) .Ci( ) .Ei( ) + vSolids ( ) . Solids .ESolids ( ) v( ) .Ci( ) .Ei( ) + vSolids ( ) . Solids .ESolids ( ) ( z + z ) ( z ) i =1 i =1

NR + A( ) .z H RX j . j ( ) .a j ( ) .r j ( ) + A( ) . (q( ) )( z ) (q( ) )( z + z ) j =1

NC d = ( ) . A( ) .z. Ci( ) .Ei( ) + (1 ( ) ). A( ) .z. Solids .E Solids ( ) dt i =1

(13)

Replacing the diffusive heat flux by its corresponding Fourier form and taking the limit as z 0 (note that

T q ( ) = K . ( ) ( ) z

, dividing by z

V = A( ) .z ), we obtain:

a
n =1 n

I ( n )

U I .(T( n ) T( ) ) + U SR .aSR .(T TS )+ M .aSR . . T TS + U Cool .aCool .(TCool T( ) ) w s

4 4 .

T 1 NC NR . v( ) .Ci( ) .Ei( ) + vSolids ( ) . Solids .ESolids ( ) + H RX j . j ( ) .a j ( ) .rj ( ) + K ( ) . ( ) A( ) z i =1 z z j =1 NC = ( ) .Ci( ) .Ei( ) + (1 ( ) ). Solids .ESolids ( ) t i =1

(14)

where

K ( ) is the thermal conductivity of pseudo-phase ().

http://www.bepress.com/ijcre/vol5/A17

Mahecha-Botero et al.: A Comprehensive Approach to Reaction Engineering

11

After plugging in the fundamental thermodynamic relation: Ei( ) = H i( ) P.Vi( ) , assuming that the

d NC residual value of the internal energy (i.e. Ci .P.Vi( ) ) is negligible and expanding the time derivative, we dt i =1 ( )
may write:

a
n=1 n

I ( n )

U I .(T( n ) T( ) ) + U SR .aSR .(T TS )+ M .aSR . . T TS + U Cool .aCool .(TCool T( ) ) w s

4 4 .

T( ) NR 1 NC . v( ) .Ci( ) .H i( ) + vSolids ( ) . Solids .H Solids ( ) + H RX j . j ( ) .a j ( ) .rj ( ) + K . ( ) A( ) z i =1 z z j =1 NC NC = ( ) . Ci( ) . H i( ) + ( ) . H i( ) . Ci( ) + ((1 ( ) ). Solids .H Solids ( ) ) t t t i =1 i =1

( )

( )

(15)

The following expression is obtained after introducing the definition of enthalpy and the control indices m to facilitate future simplifications:

1. a I
n =1 n .

( n )

U I .(T( n ) T( ) ) + 2 .U SR .aSR .(T TS ) + 3 . M .aSR . . T TS + 4 .U Cool .aCool .(TCool T( ) )

4 4 NR T( ) 1 NC . v( ) .Ci( ) .H i( ) + 7 . H RX j . j ( ) .a j ( ) .r j ( ) + 8 . K . ( ) z z A( ) z i =1 j =1

5 .w s 6 .
NC

= 9 . ( ) . Ci( ) .
i =1

NC 1 . (vSolids ( ) . Solids .H Solids ( ) ) H i( ) + 10 . ( ) . H i( ) . Ci( ) + 11 t t A( ) z i =1

( )

( )

(16)

+ 12

((1 ( ) ). Solids .H Solids ( ) ) t

The energy balance can again be expressed in terms of molar flow-rates, in order to facilitate accounting for changes in volume/number of moles:

1. aI
n =1 n .

( n )

U I .(T( n ) T( ) ) + 2 .U SR .aSR .(T TS ) + 3 . M .aSR . . T TS + 4 .U Cool .aCool .(TCool T( ) )

4 4 NR T 1 NC . Fi( ) .Cpi( ) T( ) + 7 . H RX j . j ( ) .a j ( ) .r j ( ) + 8 . K ( ) . ( ) z z A( ) z i =1 j =1

5 . ws 6 .

(17)

NC F NC Fi i = 9 . ( ) . ( ) . Cpi( ) .T( ) + 10 . ( ) . Cpi( ) .T( ) . ( ) t i =1 v( ) t i =1 v( ) (1 ( ) ). Solids .CpSolids( ) .T( ) + 12 t

+ 11 1 . vSolids ( ) . Solids .CpSolids .T( ) ( ) A( ) z

3.2.3 Initial and Boundary Conditions The initial conditions for solving this set of partial differential equations include specification of the concentrations and temperature inside the reactor at t=0. The boundary conditions for distributed systems depend on the degree of complexity in the balance equations. The boundary conditions may assume axial symmetry, zero flux at the walls

Published by The Berkeley Electronic Press, 2007

12

International Journal of Chemical Reactor Engineering

Vol. 5 [2007], Article A17

and/or Danckwerts criteria when diffusion in the fore and aft sections is negligible (Danckwerts, 1953). Table 1 gives typical specifications of these boundary and initial conditions. Table 1. Boundary and Initial Conditions

Di ( ) z .
At z = 0

Ci( ) z

= U gas( ) . Ci( ) ]0 Ci( ) ] 0 +

)
)

Kz.

T( ) z

= U gas( ) . A. ( ) . C p . T( ) ]0 T( ) ] 0 +
P( ) = (P( ) )( z =0 ) =0 =0 T( ) z T( ) r T( ) =0 =0

At z = L At r = 0 At

Ci( ) z Ci( ) r Ci( )

r=

D 2

At t = 0 3.2.4 Simplifications

=0 r C i ( ) = C i( )

( )

t =0

=0 r T( ) = (T )t = 0

The general model for distributed systems consists of two main balances (Equation (11) and (16)). These general mole and energy balances can be simplified to special cases using the control indexes. Some common simplifications are listed below. In real problems, a combination of these simplifications is commonly appropriate. a) Plug Flow Reactor (PFR) If axial dispersion is negligible, set

2 = 0 .
11 = 12 = 0 .

b) Fluid pseudo-phase (No solids) In this case the void fraction ( ) = 1 . Also set c) Packed bed reactor (PBR) If axial dispersion is negligible, set accomplished by using
'

2 = 0 . Everything should be expressed per unit mass of catalyst. This can be

rj , which can be calculated by: rj = cat .(1 ( ) ).rj' . Special attention should be paid to

the reactor pressure drop as discussed in section 3.3, below.. d) Fluidized bed reactor (FBR) This type of reactor is very complex to model, but a simple modeling approach can be based on the above development. At least two phases should be considered, i.e. N = 2 . Two sets of balances should be carried out for the high and low-density phases. Many hydrodynamic considerations must be included. For catalytic fluidized bed reactions rj =

cat .(1 ( ) ).rj' .

e) Steady state operation Set 5 = 9 = 10 = 0 . The governing equations then become ordinary differential equations. f) Isothermal system

http://www.bepress.com/ijcre/vol5/A17

Mahecha-Botero et al.: A Comprehensive Approach to Reaction Engineering

13

Set T( ) = T f ( ) . There is no longer any need for an energy balance. g) Adiabatic operation Set 2 = 3 = 4 = 0 . h) No shaft work Set 5 = 0 . i) Single phase Set N = 1 or 4 = j) Single reaction Set N R = 1 . k) Appreciable change in number of moles Equations (12) and (17) should be used in order to work with molar flows, rather than concentrations as state variables.

1 = 0 . For multi-phase systems should account for mass and heat transfer between phases.

3.3. Additional Relations

Predicting the behaviour of reaction engineering systems requires information or assumptions with respect to: reaction kinetics and stoichiometry, thermodynamics, flow patterns, heat and mass transfer, reactor size and geometry, etc. Some useful relations, some of which may be included in the above balances are: a) Pressure in lumped systems: For a lumped system the pressure is assumed to be constant. b) Pressure balance in distributed systems: For distributed systems there are many different ways to calculate the pressure drop. For homogeneous pipe flow, empirical relations representing the Moody friction factor diagram are most useful. Ergun-type equations should be used for packed beds. For fluidized beds, the pressure drop is assumed to be equal to the weight (less buoyancy) per unit bed cross sectional area. c) Concentration of species i in pseudo-phase ():

Fi C i( ) = ( ) v ( )

d) Volumetric flow-rate for gases:

FT v( ) = v f ( ) . ( ) FT ( ) f

T( ) . T( ) f

P( ) f . P( )

Z ( ) . Z ( ) f

e) Volumetric flow-rate for liquids: v( ) = v f ( ) is often assumed to be constant. f) Total molar flow-rate:

FT( ) = Fi( )
i =1

NC

Published by The Berkeley Electronic Press, 2007

14

International Journal of Chemical Reactor Engineering

Vol. 5 [2007], Article A17

g) Heat of reaction as a function of temperature for reaction j:

H RX j = H

0 RX j

(TR ) + Cp j .dT
TR

h) Variation of heat capacity for reaction j:

Cp j = ij .Cpi( )
i =1

NC

i) Reaction rates:

r j ( ) = Function Ci( )
Ea R .T( )

( )
.

j) Reaction rates for elementary reaction j:

rj

( )

= k 0 .e

Re ac tan ts

ij
i( )

. (Arrhenius relation, power law kinetics).

k) Empirical catalyst deactivation function such as:

a j( ) = e

( .C )
c c( )

. (This is but one example among many).

3.4. Summary of Mechanistic Modeling Algorithm for Reaction Engineering Systems

The following algorithm should be used in the mechanistic model of real reaction engineering systems: 1) Identify the system to model. Define whether it is lumped (Section 3.1) or distributed (Section 3.2). 2) For the lumped systems apply Equations (2) and (7). For the distributed systems apply equations (11) and (16). 3) Introduce the desired simplifications from 3.1.4 or 3.2.4 using multidimensional control indexes. 4) Introduce additional relations, such as those from section 3.3 as needed.

4. CASE STUDY: USE OF THE APPROACH TO SIMULATE A FLUIDIZED-BED REACTOR FOR PRODUCTION OF PHTHALIC ANHYDRIDE
In this section, the above developed strategy is applied for the simulation of an industrial scale process. A two-phase distributed fluidized-bed reactor model is used for the production of phthalic anhydride (Overall equipment height = 13.7 m, D = 2.1 m T = 360oC, P = 2.5 atm) with naphthalene as feedstock (Johnsson et al., 1987).

4.1 Reactor Model

The modeling algorithm described in section 3.4 is applied in combination with the analysis from section 2. The system has been characterized and studied following the five steps suggested in section 2. Afterwards, the model equations from section 3.2 were used coupled with many additional relations appropriate for fluidized-bed reactors (Mahecha-Botero et al., 2005a; 2005b, 2006, 2007). To apply the model from section 3.2, the following considerations are introduced:

http://www.bepress.com/ijcre/vol5/A17

Mahecha-Botero et al.: A Comprehensive Approach to Reaction Engineering

15

Steady state operation. This assumption allows for transformation of model equations into a set of ODEs. In Equation 11, set 5 = 0 . Axial dispersion is neglected. In Equation 11, set 2 = 0 . Isothermal operation. This is a common assumption for fluidized-bed reactors given their small temperature gradients in practice. The reactor is divided in two pseudo-phases. The high density (also called emulsion) and low density (also referred to as bubble, dilute, void) pseudo-phases occupy a fraction of the total volume ( H ) and

( L ) respectively.
Fluidized-bed reactors use fine powders as catalysts allowing for effectiveness factors close to 100%. In addition, catalyst deactivation is not expected to be an issue for this reactor system. Therefore we set

j ( H ) .a j ( H ) = j ( L ) .a j ( L ) = 1 .
The resulting mole balance equations for the two pseudophases are: High density pseudo-phase (emulsion):
NR 1 d . v( H ) .Ci( H ) + ( H ) .ij .rj + ( H ) .aI ( H L ) .kci . Ci( L ) Ci( H ) = 0 (H ) A dz j =1

(18) i =1,4

Low density pseudo-phase (bubble):

NR 1 d . v( L ) .Ci( L ) + ( L ) .ij .rj + ( L ) .aI ( LH ) .kci . Ci( H ) Ci( L ) = 0 (L) A dz j =1

(19) i =1,4

where A is the reactor cross sectional area. The bed volume fractions add up to unity:

( H ) + ( L) = 1

(20)

Each pseudo-phase contains gas and solids. The total molar flow-rate of each chemical species in the reactor is divided between the two pseudo-phases, i.e.

Fi = v( H ) .Ci( H ) + v( L ) .Ci( L )
4.1.1 Reaction Kinetics

(21)

A reaction kinetic model for the oxidation of naphthalene was proposed by (DeMaria et al., 1961) and refined by (Johnsson et al., 1987).
r3 r1 NA NQ PA r2 r4 NA PA MA, CO, CO2

where MA, NQ, NA and PA denote Maleic anhydride, Naphthoquinone, Naphthalene and Phthalic anhydride, respectively. The corresponding rate expressions are:

r2 ( ) = cat .(1 ( ) ).k 2 C NA( ) C O2 ( )

r1( ) = cat .(1 ( ) ).k1C NA( ) CO2 ( )

Published by The Berkeley Electronic Press, 2007

16

International Journal of Chemical Reactor Engineering

Vol. 5 [2007], Article A17

0.8 r4 ( ) = cat .(1 ( ) ).k 4 C PA ( ) C O

r3 ( ) = cat .(1 ( ) ).k 3 C NQ( )

2 ( )

4.1.2 Probabilistic Averaging of Hydrodynamic Parameters A general probabilistic approach for covering the three most common fluidization flow regimes (bubbling, turbulent and fast fluidization) is used to calculate bed hydrodynamic parameters (Abba et al., 2002). The probabilities of being above or below the regime transition boundaries are computed by imposing appropriate probability density functions, as well as regime boundary correlations and uncertainties associated with them. These probabilities are used as proxies for the probabilities of the applicability of the regime-specific models in the different flow regimes. For details of this approach see (Thompson et al., 1999; Abba, 2001; Abba et al., 2002). The fluidized-bed system may operate in one of the three regimes. In the transition zones the system may present characteristics found in more than one flow regime following the summation rule:

Pturb + Pfast + Pbubb = 1,

(22)

The probabilities can be calculated depending on the fluidizing gas velocity, as described by Abba (2002). These probabilities are used as weighting factors to obtain point estimates of the hydrodynamic parameters. The model parameters (coefficients in the mole and energy balance equations for each separate fluidization regime), are then weighted according to:

= P, r r
r =1

(23)

where r is the value of for regime r and Pr is the probability of being in regime r, with r = 1 for bubbling fluidization, 2 for turbulent fluidization and 3 for fast fluidization. Table 2 contains some regime specific correlations for hydrodynamic parameters. Table 2. Some hydrodynamic parameters Bubbling regime Turbulent regime Fast fluidization regime
1

A.(1 mf ) bubb= 1 U U mf 1 + 0.711 g .db

(Clift and Grace, 1985)

1/ 2

U +1 turb= U +2
(King, 1989)

G fast= 1 + so slip pU

(Patience et al., 1992)

(a .k )
I

ci bubb

U Di . mf .ubr = mf + 2. .d 3 b
(Sit and Grace, 1981)

6 . db

(a

.k ci

turb

= 1.631 .Sc i0.37 .U

(a

.k ci

fast

(Foka et al., 1996)

4.Di . ( H ) .U 2 = . L r . t C

1/ 2

(Pugsley et al., 1992)

( L ) bubb = ( L ) turb = ( mf ) /( L 0 mf )
(Abba et al., 2002; Abba et al., 2003)

( L ) fast =

rC2

Dt 2

(Abba et al., 2002; Abba et al., 2003)

http://www.bepress.com/ijcre/vol5/A17

Mahecha-Botero et al.: A Comprehensive Approach to Reaction Engineering

17

4.2. Solution of Model Equations

These model equations can be solved using many computational packages. For their flexibility and user friendliness, Matlab and Mathematica provide appropriate teaching tools. Mathematica (Wolfram, 2003) is a fully integrated environment for technical computing, combining interactive calculation (both numeric and symbolic), visualization tools, and a complete programming environment, with many built-in subroutines to solve sets of equations, such as those encountered in reaction engineering. In our case study, the program solves a system of 9 ODEs and generates a table containing the values of the state variables as z is varied from 0 to the expanded bed height.

4.3. Results and Discussion

The molar flows, concentration and pressure profiles were obtained using Mathematica 5.0. Figure 4 shows the rapid disappearance of naphthalene in both pseudo-phases along the reactor. As naphthalene is consumed, naphthoquinone is generated as an intermediate component. The desired partial oxidation product phthalic anhydride is generated very quickly. Further oxidation consumes all the remaining reactant. The intermediate component, naphthoquinone, reaches a maximum concentration and decreases as the reaction proceeds. Since the product is the phthalic anhydride, it is desirable to consume all intermediate components. The phthalic anhydride reaches a maximum at approximately 1.4 moles per second. Later, some maleic anhydride is produced, decreasing the yield of phthalic anhydride in the reactor. It is important to note that the flow rates in the two pseudo-phases have different values as expected. The low-density pseudo-phase contains most of the gas and accounts for most of the molar flows in the reactor. From Figure 4 and 5, it is observed that the profiles in the high-density pseudo-phase develop faster (i.e. achieve their maximum values at a shorter distance) than their low-density counterparts. This behaviour is due to its higher solid content and therefore higher reaction rate. Most of the reaction products are produced in the emulsion pseudo-phase. The generated product has to diffuse to the low-density pseudo-phase due to a concentration gradient. These products diffuse to the bubbles via inter-phase mass transfer. This is why the concentration profiles in the two phases differ in the lowest section of the bed. For example, the concentration of pthalic anhydride in the denser pseudo-phase is greater than in the bubbles. As this product diffuses, the concentration in the bubble-phase gradually increases, following the trend of the high-density pseudo-phase until the diffusion driving force dies out. In the case of the intermediate component naphthoquinone, something similar occurs for the initial steps of the reaction phenomena. When r1( H ) r3( H ) is greatest, the concentration of naphthoquinone increases quickly. It is observed

that the maximum concentration value for low-density in Figure 5 occurs after the maximum for high-density. This is expected since most of the reaction occurs in the emulsion phase. Then naphthoquinone is consumed at a rate r3( H ) r1( H ) , decreasing this intermediate concentration faster in the emulsion pseudo-phase. The interphase

diffusion turns out to be fast, making the concentrations uniform in both pseudo-phases beyond the lowest section of the bed Figure 6 shows that the pressure decreases rapidly and linearly with height. The overall pressure drop is modest showing one of the advantages of fluidized bed reactors. Figure 6 also shows rapid conversion of naphthalene in the lower part of the reactor. The maximum conversion approaches 100% in agreement with the values reported in the literature for this industrial reactor of 99.9%. The predicted phthalic anhydride yield is approximately 75%, in reasonable agreement with measured values of around 80%.

Published by The Berkeley Electronic Press, 2007

18

International Journal of Chemical Reactor Engineering

Vol. 5 [2007], Article A17

(a)

1.4 Naphthalene (L) 1.2

Phthalic anhydride (L)

Molar Flows (mol/s)

1 0.8 0.6 0.4 0.2 0 0 0.5 1 1.5 2 Naphthalene (H)

Phthalic anhydride (H)

Reactor Height (m)

(b)

0.25

0.2

Molar Flows (mol/s)

Naphthoquinone (L)

0.15

0.1 Maleic anhydride (L) 0.05 Naphthoquinone (H) Maleic anhydride (H) 0 0 0.5 1 1.5 2

Reactor Height (m)

Figure 4. Molar flows in the low and high density phases vs reactor height for phthalic anhydride process in fluidized bed reactor. For reactor size and operating conditions, see (Johnsson et al., 1987): (a) Molar flows of naphthalene and phthalic anhydride. (b) Molar flows of maleic anhydride and naphthoquinone. From the shape of these curves one might suggest that a shorter reactor may give the same naphthalene conversions reducing the formation of total oxidation products thus increasing the phthalic anhydride yield. Heat transfer requirements should be considered in order to determine the optimum/safe reactor height.

http://www.bepress.com/ijcre/vol5/A17

Mahecha-Botero et al.: A Comprehensive Approach to Reaction Engineering

19

(a)

1 0.9 Phthalic anhydride (H)

Concentrations (mol/m3)

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 Naphthalene (H) 0.5 Naphthalene (L) Phthalic anhydride (L)

1.5

Reactor Height (m)

(b)
Concentrations (mol/m3)

0.2 0.18 0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 0 0.5 1 1.5 2 Naphthoquinone (H) Maleic anhydride (L) Maleic anhydride (H) Naphthoquinone (L)

Reactor Height (m)

Figure 5 Species concentrations in the low and high density phases vs reactor height for phthalic anhydride process in fluidized bed reactor. For reactor size and operating conditions, see (Johnsson et al., 1987): (a) Concentrations of naphthalene and phthalic anhydride. (b) Concentrations of maleic anhydride and naphthoquinone.

Published by The Berkeley Electronic Press, 2007

20

International Journal of Chemical Reactor Engineering

Vol. 5 [2007], Article A17

300 1.00 Conversion 280 Yield 260

Conversion and Yield (-)

0.80

0.60

0.40 Pressure

240

0.20

220

200 0 1 2 3 4 5 6 7 Figure 6. Conversion, yield and pressure vs reactor height. For reactor size and operating conditions, see (Johnsson et al., 1987)

0.00

5. CONCLUSIONS
A general modeling procedure was developed for reaction engineering systems. This procedure may be employed in teaching introductory and advanced reaction engineering. It provides students with a powerful toolkit to tackle a wide range of academic and industrial engineering problems, as well as a starting point for research projects in this field. It should also provide better understanding of the multiple phenomena encountered in chemical/biochemical engineering systems and encourage a higher level of sophistication. The approach is illustrated with a two pseudo-phase distributed fluidized-bed reactor model applied to an industrial-scale phthalic anhydride process in order to show the capabilities of the approach. This simple model represents the main aspects of this complex system. Simulation results give reasonable agreement with the limited available industrial data.

NOTATION
A

A( )
aCool

Reactor cross sectional area, (m2) Area of pseudo-phase (), (m2) Cooling heat transfer surface area per unit reactor volume, (m-1) Interphase transfer area per unit volume between phases () and (n), (m-1) Catalyst activity (associated with catalyst deactivation), (dimensionless) External heat transfer surface area per unit reactor volume, (m-1) Concentration of species i in pseudo-phase (), (mol/m3) Heat capacity of species i in pseudo-phase (), (J.mol-1.K-1)

a I ( n )

a j ( )
a SR

C i( )
Cpi( )

http://www.bepress.com/ijcre/vol5/A17

Pressure (kPa)

Mahecha-Botero et al.: A Comprehensive Approach to Reaction Engineering

21

db
Di ( )

Bubble diameter (m) Diffusion coefficient of species i in pseudo-phase (), (m2/s) Internal energy of species i in pseudo-phase (), (J/mol) Molar flow rate of species i in pseudo-phase (), (mol/s) Solids flux (kg.m-2.s-1) Enthalpy of species i in pseudo-phase (), (J/mol) Heat of reaction j, (J/mol) Heat conduction coefficient of pseudo-phase (), (J.m-1.K-1) Interphase mass transfer coefficient (m/s) Vessel height (m) Number of species in pseudo-phase (), (dimensionless) Number of inputs to pseudo-phase (), (dimensionless) Number of outputs from pseudo-phase (), (dimensionless) Diffusive molar flux in pseudo-phase (), (mol.m-2.s-1) Total number of reactions, (dimensionless) Total number of pseudo-phases, (dimensionless) Diffusive heat flux in pseudo-phase (), (W/m2) Pressure in pseudo-phase (), (Pa) Probability of being in regime r (dimensionless) Rate of production by chemical reaction j in pseudo-phase (), (mol.kg-1.s-1) Rate of production by chemical reaction j in pseudo-phase (), (mol.(m of pseudo-phase ()) -3.s-1) Time, (s) Cooling temperature, (K) Phase () temperature, (K) External surface temperature, (K) Ambient air temperature, (K) Superficial gas velocity (m/s) Bubble rise velocity (m/s) Heat transfer coefficient for cooling system, (W.m-2 K-1) Interphase heat transfer coefficient, (W.m-2 K-1) Minimum fluidization velocity (m/s) Heat transfer coefficient to the surroundings, (W.m-2 K-1) Volumetric flow rate of pseudo-phase (), (m3/s)

E i( ) Fi( )
Gso

H i( )
H RX j

K ( ) kc i
Lt Nc
NI No N i( )
NR

N q ( )
P( )

Pr r j ( )
r j ( )
'

t
TCool

T( )
TS
T

U ubr

U Cool
UI U mf

U SR v ( )

Published by The Berkeley Electronic Press, 2007

22

International Journal of Chemical Reactor Engineering

Vol. 5 [2007], Article A17

vSolids ( )
V( )
V( )
.

Volumetric flow rate of solids in pseudo-phase (), (m3/s) Total volume of pseudo-phase (), (m3) Differential volume of pseudo-phase (), (m3) Rate of work done by system on the surroundings, (W) Height coordinate, (m) Compressibility factor in pseudo-phase (), (dimensionless) Heat balance control indexes (dimensionless, may be 0 or 1) Mole balance control indexes (dimensionless, may be 0 or 1) Voidage in pseudo-phase () , (dimensionless) Voidage of minimum fluidization , (dimensionless) Surface emissivity, (dimensionless) Averaged value dependent of the operating conditions and physical properties Density (kg/m3) Boltzman constant, (W/m2.K4) Stoichiometric coefficient of species i in reaction j, (dimensionless) Volume fraction of pseudo-phase () Slip factor (dimensionless) Pseudo-phase () Overall effectiveness factor, (dimensionless) Bulk Bubbling flow regime Catalyst Feed Fast fluidization flow regime High-density pseudo-phase Compound i Reaction j Low-density pseudo-phase Sub-index for multi-dimensional control indices in model equations Gas mixture Phase in contact with pseudo-phase () Regime r Turbulent flow regime Vertical direction Maleic anhydride Naphthalene Naphthoquinone

Ws
z

Z ( )
Greek Symbols

m m ( ) mf
M

ij ( )
slip
()

j( )
Subscripts b bubb cat f fast H i j L m mix (n) r turb

Abbreviations MA NA NQ

http://www.bepress.com/ijcre/vol5/A17

Mahecha-Botero et al.: A Comprehensive Approach to Reaction Engineering

23

PA XNA YPA

Phthalic anhydride Conversion of naphthalene Yield of phthalic anhydride from naphthalene

REFERENCES
Abba, I.A., "A generalized fluidized bed reactor model across the flow regimes", Ph.D. Dissertation. University of British Columbia, Vancouver, Canada, (2001). Abba, I.A., Grace, J.R., Bi, H.T., "Variable-gas-density fluidized bed reactor model for catalytic processes", Chem. Eng. Sci., 57, 4797-4807 (2002). Abba, I.A., Grace, J.R., Bi, H.T., Thompson, M.L., "Spanning the flow regimes: A generic fluidized bed reactor model", AIChE Journal, 49, 1838-1848 (2003). Aris, R., "The optimal design of chemical reactors; a study in dynamic programming", New York, Academic Press (1961). Clift, R., Grace, J.R., "Continuous bubbling and slugging." Chapter 3 in Fluidization. (J.F. Davidson, R. Clift and D. Harrison, Eds.) London, Academic Press (1985). Danckwerts, P.V., "Continuous flow systems: Distribution of residence times", Chem. Eng. Sci., 2, 1-13 (1953). DeMaria, F., Longfield, J.E., Butler, G., "Catalytic reactor design", Ind. Eng. Chem., 53, 259 (1961). Foka, J., Chaouki, J., Guy, D.K., "Gas phase hydrodynamics of gas-solid turbulent fluidized-bed reactors", Chem. Eng. Sci., 55, 713-723 (1996). Johnsson, J.E., Grace, J.R., Graham, J.J., "Fluidized bed reactor model verification on a reactor of industrial scale", AIChE Journal, 33, 619-627 (1987). King, D.F., "Estimation of dense bed voidage in fast and slow fluidized beds of FCC catalyst", Fluidization VI. (J.R. Grace, L.W. Schemilt and M.A. Bergougnou, Eds.) New York, Engineering Foundation (1989). Mahecha-Botero, A., Grace, J.R., Elnashaie, S.S.E.H., Lim, C.J., "Comprehensive modelling of gas fluidized-bed reactors allowing for transients, multiple flow regimes and selective removal of species", International Journal of Chemical Reactor Engineering, 4, A11 (2005a). Mahecha-Botero, A., Grace, J.R., Elnashaie, S.S.E.H., Lim, C.J., "FEMLAB simulations using a comprehensive model for gas fluidized-bed reactors", COMSOL Multiphysics (FEMLAB) Conference proceedings., Boston, USA (2005b). Mahecha-Botero, A., Grace, J.R., Elnashaie, S.S.E.H., Lim, C.J., "A generalized dynamic model for fluidized-bed reactors and its application to the production of pure hydrogen", Asian Pacific Confederation of Chemical Engineering Congress, Kuala Lumpur, Malaysia (2006). Mahecha-Botero, A., Grace, J.R., Elnashaie, S.S.E.H., Lim, C.J., "Time scale analysis of a fluidized-bed reactor based on a generalized dynamic model", Fluidization XII: The 12th International Conference on Fluidization. New Horizons in Fluidization Engineering, Harrison Hot Springs, BC, Canada. (2007). Patience, G.S., Chaouki, J., Berruti, F., Wong, R., "Scaling considerations for circulating fluidized bed risers", Powder Technology, 72, 31-37 (1992).

Published by The Berkeley Electronic Press, 2007

24

International Journal of Chemical Reactor Engineering

Vol. 5 [2007], Article A17

Pugsley, T.S., Patience, G.S., Berruti, F., Chaouki, J., "Modeling the catalytic oxidation of n-butane to maleic anhydride using a circulating fluidized bed reactor", Ind. Eng. Chem. Res., 31, 2652-2660 (1992). Sit, S.P., Grace, J.R., "Effect of bubble interaction on interphase mass transfer in gas fluidized beds", Chem. Eng. Sci., 36, 327-335 (1981). Thompson, M.L., Bi, H.T., Grace, J.R., "A generalized bubbling/turbulent fluidized bed reactor model", Chem. Eng. Sci., 54, 2175-2185 (1999). Wolfram, S., "The Mathematica Book", Cambridge, Wolfram Media/C.U.P. (2003).

http://www.bepress.com/ijcre/vol5/A17