You are on page 1of 21

CHAPTER 9

Models for Non-Ideal Systems

INTRODUCTION

In a non-ideal system, the distribution functions exhibit various


forms. Therefore, knowledge of these distributions for a given system
is important. By comparing the curves corresponding to ideal systems, it
is possible to determine the type of ideal behavior that a system most
closely resembles. If the differences are small, the system can be treated
as one of the ideal models discussed in Chapter 5. However, a large
difference may imply a complex system, in which case modification
will be required to reflect the real behavior of the system. The techniques
for measuring the RTD by means of a tracer offer a diagnostic approach
in analyzing the structure of flow in continuous systems.
The tank-in-series (TIS) and the dispersion plug flow (DPF) models
can be adopted as reactor models once their parameters (e.g., N, De, 1
and Npe ) are known. However, these are macromixing models, which
are unable to account for non-ideal mixing behavior at the microscopic
level. This chapter reviews two micromixing models for evaluating
the performance of a reactor-- the segregrated flow model and the
maximum mixedness m o d e l ~ a n d considers the effect of micromixing
on conversion.

BASICS OF NON-IDEAL FLOW

Three main factors affect either the interaction of fluids in a system


or the flow pattern. These are:

9 How early or late mixing of material occurs in the system.


9 The state of aggregation of the flowing material, its tendency to
clump, and for a group of molecules to move about together.
9 The residence time distribution of material that is flowing through
the system.

762
Models for Non-Ideal Systems 763

The fluid elements of a single flowing stream can mix with each
other either early or late in their flow through the system. This is
especially important with two reactants entering a vessel. Two extreme
states of aggregation of a process are referred to as macrofluid and
microfluid (Figure 9-1). In a macrofluid, molecules move together in
clumps. The clumps are distributed in residence times, but all the
molecules within a clump have the same age. In contrast, the clumps
are dispersed in a microfluid because all the molecules move inde-
pendently of one another regardless of age.
The RTD in a system is a measure of the degree to which fluid
elements mix. In an ideal plug flow reactor, there is no mixing, while
in a perfect mixer, the elements of different ages are uniformly mixed.
A real process fluid is neither a macrofluid nor a microfluid, but tends
toward one or the other of these extremes. Fluid mixing in a vessel,
as reviewed in Chapter 7, is a complex process and can be analyzed
on both macroscopic and microscopic scales. In a non-ideal system,
there are irregularities that account for the fluid mixing of different

0 ~ 0 _1# II e l e , , l l -0
O ltl , i mlo ~il tl

i ,..-.,.-i1
~176 " t.,lt I ,li,lll _i , ,
~- _~176 "t "'-1 " t ' I',' ~ l'.
~o o - o o l o ~ i,, ,, ' 1 , ' . . - , ,"
ooo-olo~ i o'.1",;''o'.
o io7_Oo1~ <->1 ' '

bOo o o )
tk.."
' ,,t' . . - , - . ,- 1' , , )
oD ~ "---Y_" ~

a b
Complete segregation Maximummixedness
(Molecules are kept grouped (Individual moleculesare
together in aggregates). free to move about and
intermix).

Figure 9-1. Two extreme states of fluid micromixing.


764 Modeling of Chemical Kinetics and Reactor Design

ages at the microscopic level. The RTD does not account for how fluid
elements at the microscopic scale interact with each other. There is
an obvious distinction between macromixing, or a set of mechanisms
associated with the RTD, and micromixing, which results from the
interactions between the fluid elements.
Macromixing defines the variation in the residence time experienced
by molecules flowing through a flow system, while micromixing
relates to the concentration history experienced by the molecules as
they traverse through the system. The conversion in a system can be
determined from the RTD with knowledge of the macromixing process.
Generally, the performance of non-ideal flow systems can be deter-
mined from known parameters, such as:

9 The kinetics of the system


9 The RTD of fluid in the system
9 The earliness or lateness of fluid mixing in the system
9 The nature of the fluid (i.e., whether a micro or macro fluid) in
the system

SEGREGRATED FLOW MODEL

In a mixing process where complete segregation occurs, the fraction


of fluid elements with residence times (t, t + dt) is E(t)dt, and the
average exit composition in the vessel is defined by:

mean concentration) _
fconcenationof /ffractiono,exit
[reactant remaining ][stream which is
of reactant in the / = C A = l~m~ (9-1)
all elements ]in an element of age[|of age between
exit stream J of exit stream| l/
~,between t and t + dt )~,t and t + dt

Consider a segregated flow system as illustrated in Figure 9-2, with


a known RTD, in which a reaction A ~ products. The disappearance
rate of A is

(--rA) = ( r A ) ( C A ) (9-2)

Since the fluid flows in clumps and is completely segregated by


age (i.e., each clump behaves like a batch reactor), it is possible to
determine the reactant concentration in every clump for a constant
density fluid as:
Models for Non-Ideal Systems 765

CAo CA
h~
Y

Figure 9-2. Reactor with segregated flow.

dC A
-k, dt )clump - ( r A ) (9-3)

Rearranging Equation 9-3 and integrating yields

CA,clump
~dC A
= t (9-4)
CAO rA

Therefore, CA,clump is a function of time for all nth order reactions.


Using Equation 9-3 for the reactant concentration remaining in an
element that has been in the reactor for a time t, E(t)dt is the frac-
tion of the exit stream with ages between t and t + dt. Equation 9-1
then becomes

oo

CA = f CA,clump(t)E(t)dt (9-5)
0

The reactant conversion X A is X A = 1 - CA/CAo . In terms of the


fractional conversion, Equation 9-5 becomes

1 - X A - ~--~A - i ( CA (9-6)
CAO CAO clumpE(t) dt

A zero order reaction in a batch reactor gives

CA ----1--k__t_t
CA~ CA~ (9-7)
766 Modeling of Chemical Kinetics and Reactor Design

Substituting Equation 9-7 into Equation 9-6 and integrating gives

CA = 1 - kt (l_e_CAo/kt) (9-8)
CAO CAO

For a first order reaction, we have

O)cump
( - e -kt (9-9)

and CA = ~ e-kt E(t)dt


CAO 0

CA
or = Z e-kt E(t)At (9-10)
CAO

The residence time density function E(t) is

E(t) - t1 e - t / ~ (8-25)

Substituting Equation 8-25 into Equation 9-10 gives

oo

CA _- I e-kt o - 1
e -t~ dt
CAO 0 t

=_ie- k+ t t dt (9-11)
t
0

Integrating Equation 9-11 by parts (i.e., ~wdv - wv - ~vdw) yields

CA 1
CAO - 1 + kt (9-12)

Equation 9-12 is identical to that derived in Chapter 5. There


appears to be two methods for determining the conversion in a
Models for Non-Ideal Systems 767

perfectly mixed reactor. In Chapter 5, complete mixing in the reactor


was assumed, and no RTD was explicitly used. Here, the first order
reaction gives identical results when the RTD is used as specified by
Equation 8-25.
For a second order reaction of a single reactant,

CA) _ 1
(9-13)
CAO clump 1 + ktCAo
The segregated flow model equation gives

1 -X A - CA _ 1 1 -t
CAO 1 + ktCAo " t e ~ dt (9-14)

where cz = 1/ktCAo and 0 = t/t. Equation 9-14 becomes

1 - XA - ( ~A ) - aeal
= e-(a+~
c~ +-------O
d(a + 0) - aea ei(a) (9-15)
CAo o

where Ei(x) = exponential integral ~7(e-Y/y)dy.


Equation 9-15 gives the conversion expression for the second order
reaction of a macrofluid in a mixed flow. An exponential integral,
ei(cz), which is a function of cz, and its value can be found from tables
of integrals. However, the conversion from Equation 9-15 is different
from that of a perfectly mixed reactor without reference to RTD. An
earlier analysis in Chapter 5 gives

CA _-1 + ~/1 + 4kCAot


CAO 2kCAo~ (5-245)

For an nth order reaction, the conversion in a batch reactor is:

1 -X A - [1 + (n - 1)C~,o1 kt] 1/(1-n) (9-16)

For an nth order reaction of a macrofluid, Equation 9-16 is sub-


stituted into Equation 9-6 to give
768 Modeling of Chemical Kinetics and Reactor Design

-t
oo e z,
1
1- X A = CA = [1 + ( n - 1)C~]~qCn-~ o t dt (9-17)
CAO
I0 -

For N-stirred tanks in series, the conversion is expressed by

CA __ 1 ! tN_ 1e
NN
dt
CA 0 ~N (N- 1)!

CA _ 1 _ 1
(8-160)

Complete segregation represents an extreme form of mixing on a


molecular scale. Molecules enter mixed and leave mixed, but no
mixing takes place within the system. Any RTD is possible in this
state. Complete segregated systems can be modeled as piston flow
reactors in parallel, as shown in Figure 9-3 or as single piston flow
reactor with side exits as in Figure 9-4. Other residence time distribu-
tions have some maximum possible level of micromixing, referred to

Ih,.I _llb.

h._! ~-

I
"1 t "

r'l . . . . . . I v

;J I-'
Figure 9-3. Models for segregated reactors" Piston flow elements in parallel.
Models for Non-Ideal Systems 769

.! Plug flow
"1
t
y

Figure 9-4. Single-piston flow element with side exits.

as maximum mixedness. Figure 9-5 shows an illustration of micro-


mixing. The abscissa (macromixing) measures the breadth of the
residence time distribution. It is zero for piston flow, fairly broad for
the exponential distribution of stirred tanks, and broader yet for
situations with stagnancy and bypassing. The ordinate (micromixing)
varies from none to complete and it measures the effects of micro-
mixing. This can be negligible for piston flow and have maximum
importance for stirred tank reactors. Soundly designed reactors will
fall in the normal region bounded by three apexes: piston flow, perfect
mixing, and complete segregation with an exponential distribution.

Perfectmixer

.m
Prohibited Region

~~ ~;~ \

~~/%"~Norm~regi~no\ ~ ~~
~ ~ ~~w~u~~~ ~` Bypassinregg~i .
Piston flow Segregatesti
drredtank
Macromixing
Figure 9-5. Schematic representation of mixing space. (Source: Nauman,
E. G., Chemical Reactor Design, John Wiley & Sons, 1987.)
770 Modeling of Chemical Kinetics and Reactor Design

A real system must lie somewhere along a vertical line in Figure


9-5. Performance is within the upper and lower points on this line;
namely maximum mixedness and complete segregation. Equation 9-10
gives the complete segregation limit. The complete segregation model
with side exits and the maximum mixedness model are discussed next.

COMPLETE SEGREGATION
M O D E L WITH SIDE E X I T S

Consider a plug flow reactor with side exits (Figure 9-6) through
which portions of the flow leave. Molecules are classified as they
traverse within the system according to their ages ~ (i.e., the life
expectancy of the fluid in the reactor at that point), implying that
mixing occurs as late as possible.
A material balance on an elemental volume AV of the reactor
between ages ~ and X + d~ follows.

Input of A by flow
in vessel at ~ = u[1 - f(L)]CA(~) (9-18)

Output of A by flow
in vessel at ~ + d L = u[1 - F(~, + dk,)]C A (X + d~) (9-19)

Output from the side exit = uCAE(~)d~ (9-20)

where the volume of fluid with a life expectancy between ~ and


+ d L is

..I uCAI~ ucAl~+~

I II I I1 uC~E(~)d~
Figure 9-6. A plug flow reactor with side exits.
Models for Non-Ideal Systems 771

AV-iu[1-F()~)ld)~- i u[1-F(~)]d)~-u[1-F()~)]A)~ (9-21)


0 ~ +d)~

The rate of loss of substance A


in this volume is - (-rA)AV

= (-rA)U[1 - f(~)lA~ (9-22)

A balance on component A between )~ and )~ + d)~ gives

rate of loss of]

I
Input of A /output of A by] /output of A by]
substance A ]
byflowin =[flow invessel [+]flowthrough [
by reaction /
vessel at )~ ~at )~+d)~ ) ~,sidesof vessel )
in AV )

which is"

u[1 - F(~)]CA(~ ) - u[1 - F(X + d~)]CA(~ + d~)

+ uCAE(~)d~ + (-rA)U[1 - F(~)]A~ (9-23)

Rearranging Equation 9-23 and dividing by u as A)~ ~ 0 gives"

+ CA()~)E()~) + (-r A1[1-F()~)] (9-24)


d~,

Equation 9-24 gives

de A [ \
= - (-rA) (9-25)
d)~

Each fluid element containing molecules of the same age during its
passage through the reactor can be considered as a small batch reactor.
The mean concentration overall fluid elements at the reactor outlet is
the expected concentration of all fluid elements over the RTD.
772 Modeling of Chemical Kinetics and Reactor Design

MAXIMUM MIXEDNESS MODEL (MMM)

Upon entering the reactor, the molecules of the feed immediately


mix and become associated with the molecules in whose company they
will eventually leave. Figure 9-7 shows a plug flow with multiple inlets
whose flow pattern is given by the same RTD. Zwietering [ 1] developed
the MMM as represented by a plug flow in which the feed enters
continuously and incrementally along the reactor length. Consider the
material balance of reactant A around the differential volume between
and ~ + d~.

(Input ofAby] (Input ofAby] (output of A] (rate of loss of]


/substance A /
/flow invessel/+[flowthroug h / - / b Y flowin / +
/bY reaction /
~at~,+d~, ) ~the side at ~, ) ~,vesselat~, ) ~,inAV )

Input of A by flow
in vessel at ~, + d~,- u[1 - F(~, + d~)]CA(~ + d~,) (9-26)
Input of A by flow
through the side at ~ - [uE(%)A%]CAo (9-27)

[uE(Z)AZ]CAO

CAO
CAO
~r ~r Ir ~r ~r ~r ~pr

UCA} L+dX uC ^t~


b..
jr-

.I I.
-]AV r
~,+d~,

Figure 9-7. A plug flow reactor with side entry.


Models for Non-Ideal Systems 773

Output of A by flow
in vessel at ~ , - u[1 - F(~,)]CA(~) (9-28)

The rate of loss of substance A


in this volume is = (-rA)AV

= (-rA)U[1 - F(L)IA)~ (9-29)

The steady-state material balance on A gives

u[1 - F(~, + d~)]CA(~, 4- d~)

+ [uE(~)A)~]CAo - u[1 -F()~)]CA0~ ) + u(--rA)[1 -- F(~)]A)~ (9-30)

Rearranging Equation 9-30 and dividing by u gives:

[1 - F(~ + d~)]CA(~ + d~,) + E(~)A~,CAo

- [1 - F(~)]CA(~, ) - (--rA)[(1 - F(E)]A~, - 0 (9-31)

Taking the limits as A~ ~ 0 gives"

d{[1- F(~)]CA (~)}


E(~)CAo q- - - 0 (9-32)
d~

or CAoE(~)+[1- F(~)] ~dCA


-CA(~')
E(~)-(-r A)[1- F(~,)]-0 (9-33)
d~

Rearranging Equation 9-33 yields"

dfA ( ~ ) (+C A o - C A ) E ( ~ )
(9-34)
d~ [1- F(~)]

Equation 9-34 does not use an initial value of C A as a boundary


condition. Instead, the usual boundary condition associated with Equa-
tion 9-34 is

Lim dCA - 0 (9-35)


z~0 d~,
774 Modeling of Chemical Kinetics and Reactor Design

The outlet concentration from a maximum mixedness reactor is


found by evaluating the solution to Equation 9-34 at ~, = 0; CAout =
CA(0). The analytical solution to Equation 9-34 is rather complex; for
reaction order n > 1, the (-r A) term is usually non-linear. Using
numerical methods, Equation 9-34 can be treated as an initial value
problem. Choose a value for CAout- CA(0) and integrate Equation 9-34.
If CA(~,) achieves a steady state value, the correct value for CA(0) was
guessed. Once Equation 9-34 has been solved subject to the appro-
priate boundary conditions, the conversion may be calculated from
CAout- CA(0).
EFFECT OF MICROMIXING ON CONVERSION

As stated earlier, the actual state of micromixing lies between two


extremes. However, further details are extremely difficult to use either
theoretically or experimentally. Kramers and Westerterp [2] gave
results for the extreme case of a perfectly macromixed reactor with
various reactions occurring. Figures 9-8 and 9-9 illustrate a comparison
between the segregated flow and complete mixing for n = 0 and n > 1,
respectively. It can be seen in Figure 9-9 that the differences between

PFR
I,O
/ (CSTR)mm . . . . . . .
A P /

.8 /
x .6
f.

8
"~ .4
8

.2

O
O . . . ,.b . . . 2'.0 3.0 4~0
Reaction number , K~./Co

Figure 9-8. Conversion for a zero order reaction in a CSTR and a PFT.
(Source: Wen, C. Y. and L. T. Fan, Models for Flow Systems and Chemical
Reactors, Marcel Dekker Inc., 1975.)
Models for Non-Ideal Systems 775

A----)-P

0.75 n=2

~"0.5
._o
L_

t-
O
U
S--complete segregation
M= maximum mixedness

0.1 0.4 I 4 I0 40 I02


Reaction number, KC~

Figure 9-9. Conversion for isothermal elementary reactions in a CSTR [2].

the levels of micromixing are small. For a first order reaction (or a
series of first order reactions), micromixing has no effect on the reactor
performance. Therefore, conversion is predicted exactly by the segre-
gated flow model (Equation 9-10). Micromixing decreases conversion
for reaction orders greater than 1 and increases for orders less than
1. The segregated flow model, consequently, predicts an upper bound
on conversion for n > 1 and a lower bound for n < 1.
Several types of models have been developed for intermediate levels
of micromixing with arbitrary macromixing RTD. Weinstein and Adler
[3], Villermaux and Zoulalain [4], and Ng and Rippin [5] have pro-
posed a model that divides the reactor into two environments: one in
a segregated state and the other in a maximum mixedness state. The
fraction in each state can be fitted to the actual reactor data, which
are then correlated. For bimolecular reaction, the maximum difference
between the conversion corresponding to complete segregation and that
corresponding to maximum mixedness can be as much as 50%. Tsai
et al. [6] have developed the two-environment model, which is an
empirical modification of the original two-environment model. It
considers material transfers from the entering environment to the
leaving environment at a rate that is proportional to the amount of
776 Modeling of Chemical Kinetics and Reactor Design

material remaining in the entering environment. They showed that the


reversed two-environment is more appropriate for the growth in a flow
reactor than other micromixing models. Chen and Fan [7] inferred that
a reversed two-environment model is more representative of polymer
reactors. In this case, the reactants initially enter the maximum mixed~
ness region, but become segregated as polymerization increases the
viscosity and decreases diffusivities.
Industrial reactors generally operate adiabatically. Cholette and
Blanchet [8] compared adiabatic plug flow reactor to the CSTRmm. For
exothermic reactions, they inferred that the performance of a CSTRmm
is better than that of a plug flow reactor at low values of conversion,
and vice-versa at high values of conversion. They further showed that
the design considerations for endothermic reactions are similar to those
for isothermal reactions.
For isothermal autocatalytic and adiabatic reactions, the effects of
macromixing and micromixing on the conversions are different from
those for isothermal elementary reactions. Figures 9-10 and 9-11
compare the performances of these reactions under the two extreme
states of micromixing. Worrel and Eagleton [9] performed mixing and
segregation studies in a CSTR. They deduced that the maximum
mixedness conversion is attained rapidly as impeller speed is increased.
However, the effect of micromixing on the conversion was negligible.

I.O

--' 0.8 (CSTR)mm


0 A* R"'R+ R

x 0.6

C
.g 0.4

C
0
~ o.2

O.Ol
0 2 4 6 8 I0
Mean holding time, T

Figure 9-10. Conversion for autocatalytic in a CSTR and a PFT. (Source:


Wen, C. Y. and L. T. Fan, Models for Flow Systems and Chemical Reactors,
Marcel Dekker Inc., 1975.)
Models for Non-Ideal Systems 777

1.0

(CSTR)mm
.8

.6

._o
$eg
=" .4
t-.
o
t~

.2
K = Koexp.(4x/l+O.l:x: )

0 .I .2 .3 .4 .5

Dimensionless residence time, Ko'f"


Figure 9-11. Conversion for a first order, adiabatic reaction in a CSTR and
a PFT. (Source: Wen, C. Y. and L. T. Fan, Models for Flow Systems and
Chemical Reactors, Marcel Dekker Inc., 1975.)

Their studies also showed the importance of the reacting system


kinetics for better design operation of CSTR reactors.
Generally, empirical models involving micromixing should be
viewed with caution because it is difficult to obtain valid correlations
for the parameters in the absence of sound experimental data. Also, a
model may not provide a unique measure of micromixing except for
the specific system being studied. Micromixing is a complex phe-
nomenon, which cannot be easily characterized by one or more adjust-
able parameters. A model with few parameters might provide an
excellent fit for one reaction and a poor fit for another reaction that
occurs in the same vessel. A comprehensive discussion and analyses
on micromixing are provided by Nauman and Buffham [10], and Wen
and Fan [11].
778 Modeling of Chemical Kinetics and Reactor Design

E x a m p l e 9-1
Consider the data of Hull and von Ronsenberg in Example 8-3 for
mixing in a fluidized bed. Suppose the solids in the fluidized bed were
not acting as a catalyst, but were actually reacting according to a first
order rate law (-r) - kC, k - 1.2 min -1. Compare the actual conversion
with that of an ideal plug flow.

Solution

For the plug flow system, the concentration of unreacted reactant


with respect to A is determined by CA/CAo -- e -kt where t is the
mean residence time; t - 2.5 min (see Example 8-3) and CA/CAo =
e -kt - e -(1"2)(25) = 0.05. The conversion, X A - 1 - CA/CAo - 0.95
or 95% conversion.
Using the m a c r o m i x i n g E q u a t i o n 9-10, an Excel s p r e a d s h e e t
(Example9-1.xls) was developed that gives CA/CAo - 0 . 1 1 . In terms
of conversion, X A = 0.89 or 89% conversion. Thus, the simula-
tion of the actual reactor suggests that it is less efficient than a plug
flow reactor. This may be due to several factors such as channeling
or bypassing of the fluid in the reactor. If the reactor had been
designed using the usual plug flow model, the actual performance in
the plant would be far less than expected. For a higher conversion in
the reactor, some modifications would be required to remedy the non-
ideal flow.

E x a m p l e 9-2

The flow through a reactor is 10 dm3/min. A pulse test gave the


following concentration measurements at the outlet.

t(min) c x 10 5 t(min) c x 10 5
0 0 15 238
0.4 329 20 136
1.0 622 25 77
2.0 812 30 44
3 831 35 25
4 785 40 14
5 720 45 8
6 650 50 5
8 523 60 1
10 418
Models for Non-Ideal Systems 779

1. What is the volume of the reactor? Calculate and plot E(0), F(0),
and I(0) versus 0.
2. If the reactor is modeled as a tank-in-series (TIS) system, how
many tanks are needed to represent this reactor?
3. If the reactor is modeled by a dispersion model, what is the
Peclet number (Npe)?
4. What is the reactor conversion for a first order reaction with
k = 0.1 min -1 for
a. The tank in series model
b. The segregation model
c. The dispersion model
d. Plug flow
e. A single CFSTR

Solution

An Excel spreadsheet (Example9-2.xls) was developed for the RTD


functions and the parameters. The calculated mean residence time is
9.878 min, and thus the volume of the reactor is

V-~ou

= 9.878 x 10

= 98.8 dm 3

Figure 9-12 shows plots of E(0), F(0), and I(0) versus 0.


Assuming that the reactor is modeled as TIS, the number N is
determined as

1
N~~

where ~ is the dimensionless variance. The value from the spread-


sheet is o~ - 0.765. Therefore, the number of tanks required to
represent this reactor is N - 1.31.
Assuming that the reactor is modeled by a dispersion model using
the closed boundary conditions, the Peclet number is determined by

(Y02 = 2 2 ( 1 - e-Npe ) (8-161)


Npe N2e
780 Modeling of Chemical Kinetics and Reactor Design

'2 x

=i E(Tt~TA)
0.8- -~. F(THETA)
- -'- I(THETA)
<~ 0.6-

0.4-

~ 0.2-

01'

d d d d d ,g ,--: e,i c,,i e6 e6 ,4 4 ~ ~6


DIMENSIONLESS TIME

Figure 9-12. Plots of E(e), F(0), and I(0) versus 0.

Expressing Equation 8-161 as a polynomial gives

2 + ~ 2 (l_e_Npe) (8-162)
f(Npe)- O~ Np e N2e

As an alternative to using the computer program (PROG16), Equa-


tion 8-162 can be expressed in a spreadsheet. Using the GOAL-SEEK
command from the Tools menu (Appendix B) from the Excel spread-
sheet menu to determine the Peclet number gives f(Npe) = 0. This
function is incorporated in the spreadsheet. The computed Peclet
number is Npe = 0.8638.
For a first order reaction with k = 0.1 min -1, the following con-
versions were determined:
For the TIS model,

1
XA - 1-
[1 N (8-160)

From the spreadsheet, X A "- 0.521 or 52.1%.


Models for Non-Ideal Systems 781

The segregation model gives"

CA ~" Z e-kt E(t)At (9-10)


CAO

XA - 0 . 4 7 or 47%

For the dispersion model, the conversion is determined by Equa-


tion 8-149 as

CA = 1 - X A =
4aexp[ e]
CAO (1 + a) 2 exp Npe - (1- a) 2 exp Npe

4kt ]0.5
where a = 1+
Npe

Using the value of Npe - 0.86, the calculated conversion from the
spreadsheet is X A - 0.53 or 53%.
The conversion for the plug flow is expressed as"

XA - 1_ e-k~

X A = 0.627 or 62.7%

For a single CSTR

k~
XA=I+ kt

(0.1)(9.878)
1 + (0.1)(9.878)

= 0.497 or 49.7%.

A summary of the conversion is given as"


782 Modeling of Chemical Kinetics and Reactor Design

Flow systems Conversion, X A (%)


Plug flow reactor 62.7
Dispersion 52.6
Tank-in-series 52.1
Segregated model 47.0
A single CSTR 49.7

REFERENCES

1. Zwietering, T. N., Chem. Eng. Sci., 11, 1-15, 1959.


2. Kramers, H. and Westerterp, K. R., Elements of Chemical Reactor
Design and Operation, Academic Press, New York, 1963.
3. Weinstein, H. and Adler, R. J., Chem. Eng. Sci., 22, 65, 1967.
4. Villermaux, J. and Zoulalain, A., Chem. Eng. Sci., 24, 1513, 1969.
5. Ng, D. Y. C. and Rippin, D. W. T., "The effect of incomplete
mixing on conversion in homogeneous reactions," Proc. 3rd
Europ. Symp. of Chem. Eng., Amsterdam, Netherlands, 1964.
6. Tsai., B. I., Fan, L. T., Erickson, L. E., and Chen, M. S. K., J.
Appl Chem. Biotechnol., 21, 307, 1971.
7. Chen, M. S. K. and Fan, L. T., Can. J. Chem. Eng., 49, 704-708,
1971.
8. Cholette, A. and Blanchet, J., Can. J. Chem. Eng., 39, 192, 1961.
9. Worrel, G. R. and Eagleton, L. C., Can. J. Chem. Eng., 39, 254, 1964.
10. Nauman, E. B. and Buffham, B. A., Mixing in Continuous Flow
System, John Wiley & Sons, 1983.
11. Wen, C. Y. and Fan, L. T., Models for Flow Systems and Chemical
Reactors, Marcel Dekker, Inc., 1975.

You might also like