Академический Документы
Профессиональный Документы
Культура Документы
153
We now focus our attention on accounting for the pressure drop in the rate law. For an ideal gas, the concentration of reacting species i is
gas-phase reactions pressure drop may be very important
(3-46)
18)
We now must determine the ratio as a function of volume V or the catato account for pressure drop. We then can combine the lyst tration, rate law, and design equation. However, whenever accounting for the effects of pressure drop, the differential form the mole balance (design equation) must be used. for example, the second-order reaction
When
one
is being carried out in a packed-bed reactor, the differential form of the mole balance equation in terms of catalyst weight is
-The
(2-17)
law is
Chap. 4
(4-20)
Note from Equation (4-20) that the larger the pressure drop the smaller P) from frictional losses, the the reaction rate! with the mole balance (2-17) and assuming Combining Equation gives isothermal operation (T =
Dividing by
yields
P)
(4-21)
We now need to relate the pressure drop to the catalyst weight in order to determine the conversion as a function of catalyst weight.
4.4.2 Flow Through a Packed Bed
The majority of gas-phase reactions are catalyzed by passing the reactant used most to calculate through a packed bed of catalyst particles. The pressure drop in a packed porous bed is the Ergun
Er gun equation
Phenomena (NewYork:
4.4
155
where
P = pressure,
= porosity =
.
lb,
factor)
= 1.0)
4.17 X
= viscosity of gas passing through the bed, = length down the packed bed of pipe, ft = superficial velocity = volumetric flow
h cross-sectional
area of pipe,
p = gas density, = = superficial mass velocity,
s) or
h)
In calculating the pressure drop using the Ergun equation, the only parameter that varies with pressure on the right-hand side of Equation (4-22) is the gas density, p. We are now going to calculate the pressure drop through the bed. Because the reactor is operated at steady state, the mass flow rate at any point down the reactor, is equal to the entering mass flow rate, equation of continuity),
=m
=
Recalling Equation
we have
Equations
56
Chap. 4
Simplifying yields
(4-24)
I
where
(4-25)
For tubular packed-bed reactors we are more interested in catalyst weight rather than the distance down the reactor. The catalyst weight up to a distance of down the reactor is
W
weight of catalyst
(1
volume of solids density of
(4-26)
bulk density of the where A, is the cross-sectional area. (mass of catalyst per volume of reactor bed), is just the product of the solid density, , the fraction of solids, - :
[Equation
we can change
(4-27)
Further simplification yields
(4-28)
where
(4-29)
157
Equation will be the one we use when multiple reactions are occurring or when there is pressure drop in a membrane reactor. However, for single reactions in packed-bed reactors it is more convenient to express the Ergun equation in terms of the conversion X . Recalling Equation (3-42) for
+
and development leading to Equation
F,,
(3-42)
-=
where, as before,
- _- - _ ( 1 + EX)
2 T o P I P,
(4-30)
We note that when is negative the pressure drop will be less higher pressure) than that for E = When E is positive, the pressure drop AP will greater than when E = 0. For isothermal operation, Equation (4-30) is only a function of conversion pressure:
Recalling Equation
Two coupled
equations to be solved numerically
(4-21)
we see that we have two coupled first-order differential equations, (4-31) and that be solved simultaneously. A variety of software packages and numerical integration schemes are for this purpose.
Analytical Solution. If E 0, or if we can neglect ( E X ) with respect i o 1 1 can obtain an analytical solution to Equation (4-30) for isothermal operation (Le., T isothermal operation with E = 0, Equabecomes tion
158
Isothermal w i t h
Chap. 4
--- -a
Taking inside the derivative, we have
Integrating with P =
at W = 0 yields
2
(4-33)
where again
Equation (4-33) can be used to substitute for the pressure in the rate law, in which case the mole balance can be written solely as a function of conversion and catalyst weight. The resulting equation can readily be solved either analytically or numerically. If we wish to express the pressure in terms of reactor length we can use Equation (4-26)to substitute for Win Equation (4-33). Then (4-34)
Calculate the pressure drop in a 60 ft length of 1 schedule 40 pipe packed in diameter when 104.4 with catalyst pellets of gas is passing the bed, The temperature is constant along the length of pipe at 260C. The void fraction is 45% and the properties of the gas are similar of at this temperature. The entering pressure is 10 atm.
4.4
159
Evaluating
the pressure drop parameters
For
0.01414 7383.3
104.4 0.01414
For air at and atm,
0.0673 = 0.413
From the problem statement,
= 0.0208 ft
4.17 X
lb
7383.3
1 - 0.45)
h)
0.0208 ft
ft3
0.01244 164.1
160
2
Chap. 4
0.0775
atm
60 ft
(E4-5.5)
P=
AP
=
= 2.65 atm
P=
= 7.35 atm
(E4-5.6)
to relate conversion and catalyst weight. Recall our mole balance, rate law, and
(2-17)
(4-19)
E
0:
(4-35)
P -
Using Equation (4-33) to substitute for PIP, in terms of the catalyst weight, we obtain
Only for
Combining:
- =(1 -
[(I
Separating variables:
161
(4-36)
second-order
1- {I -
We now proceed (Example 4-6) to combine pressure drop with reaction in a packed bed for the case where we will that E X 1 in the Ergun equation in the rate law in order to obtain an analytical solution. Example 4-7 but removes this assumption and solves Equations (4-21 ) and (4-3 1) numerically.
Example Calculating X in a Reactor with Pressure Drop
The economics
7 billion pounds of ethylene oxide were produced in the United States in 1997.The 1997 selling price was $0.58 a pound, amounting to a commercial value of $4.0 billion. Over 60% of the ethylene oxide produced is used to make polyesethylene glycol. The major end uses of ethylene oxide are antifreeze ter surfactants and solvents (5%). We want to calculate the catalyst weight necessary to achieve 60% conversion when ethylene oxide is to be made by the vapor-phase catalytic oxidation of ethylene with air.
+
A +fB
Ethylene and oxygenare fed in stoichiometricproportions to a packed-bed reacat a prestor operated isothermally at 260C. Ethylene is fed at a rate of 0.30 sure of 10 atm. It is proposed to use 10 banks of 1 -in.-diameter schedule 40 packed with catalyst with 100 tubes per bank. Consequently,the molar flow rate to The properties of the reacting fluid are be each tube is to be 3 considered identical to those of air at this temperature and pressure. The density of the -in.-catalystparticles is 120 and the bed void fraction is 0.45.The rate law is
=
lb
cat. h
with3
k = 0.0141
Chem., 45, 234 (1953).
162
Solution
Chap. 4
1.
mole balance:
(E4-6.1)
2. Rate law:
=
113 213
(E4-6.2)
1/3
The algorithm
+
(E4-6.4)
(E4-6.8)
= where k 5. Developing the design equation. For a packed-bed reactor, the relationship 1 is between and W when
is valid
E
=0
PO
(4-33)
or
(E4-6.9)
Combining Equations
and (E4-6.
we have
4.4
163
Integrating gives us
Solving for
we obtain
(E4-6.11)
1.08 lb
= 0.54 X = 2.03
1.5 X
1.5 X
lb
0.79 mol
inerts
= 5.64 X
Summing:
3.65
1)
atm
k' =
= 0.0141
3 atm
1 - [1
-0.15)
a
1 (1 In order to calculate
Evaluating the drop parameters
(E4-6.12)
we need the superficial mass velocity, G. The mass flow rates of each entering species are:
164
Chap.
30.24
mol
- 17.28
56.84
This is essentially the same superficial mass velocity, temperature, and pressure as in Example 4-5. Consequently, we can use the value of calculated in Example 4-5. atm 0.0775 (0.01414
lb mol cat 45.4 lb of catalyst per tube or 45,400 lb of catalyst total This catalyst weight corresponds to a pressure drop of approximately 5 atm. If we had neglected pressure drop, the result would have been 1
k
0.0266
=
1 1 - 0.6
of catalyst per tube (neglectingpressure drop)
,
35.3
and we would have had insufficient catalyst to achieve the desired Substitutingthis catalyst weight 35,300 total) into Equation (E4-6.10) gives a conversion of only 53%.
4.4
Example 4-
Rework Example for the case where volume change is nor Ergun equation and the two coupled differential equations conversion and are solved
I
Solution Rather than rederive everything starting the etry, and pressure drop equations, we will use th 4-6 Combining (E4 and
Next. we
.3)
For the reaction conditions in Example 4-6, we the 0, and y 1.0 and the parameter values W 0, cat, -0.15, k' 0.0266 1.08 E cat, and large number of ordinary differential equation solver software are extremely user friendly have become available. We shall ODE solvers) use POLYMATH4 to solve the examples in the text. However, the CD-ROM as all the MATLAB and contains an example that uses ASPEN, as MATH solution programs to the example With POLYMATH simply enters Equations (E4-7.3) and (E4-7.4) and the corresponding parameter value into the computer (Table with the they (rather, boundary) conditions are solved and displayed as shown Figure E4-7.1, the Ergun equation in Example We note that neglecting E X -0.09) to obtain an solution resulted in less than a 10% error. Developed by Professor M. Cutlip of the University of Connecticut. and Professor M. Corporation, Shacham of Ben Gurion University. Available from the Box 7939, Austin, TX 78713.
166
TABLE
Reactor Design
Chap. 4
Equations
15 0266
a,
w f = 60
Scale:
0.000
ooo
U
However, larger errors will result if large values of E X are neglected! By taking into E = -0.15) in the pressure account the change in the volumetric flow rate drop term, we see that 44.0 lb of catalyst is required per tube as opposed to 45.4lb when E was neglected in the analytical solution, Equation (E4-7.4). Why was less The that catalyst required when was not neglected in Equation the numerical solution accounts for the fact that the pressure drop will be less because E is negative.
4.4
167
It is also interesting to learn what happens to the volumetric flow rate along the length of the reactor. Recalling Equation
=
----
(3-44)
let f be the ratio of the volumetric flow rate, to the entering volumetric flow rate, at any point down the reactor. For isothermal operation Equation (3-44) becomes
E4-7.2 shows X, y y = and f down the length of the reactor. We see that both the conversion and the volumetric flow increase along the length of the reactor while the pressure decreases. For gas-phase reactions with orders greater than zero, this decrease in pressure will cause the reaction rate to be less than in the case of no pressure drop.
4.000
3.200
2.400
1.600
0.800
0.000
W
Figure of added
Output in graphical
from POLYMATH
conversion
We note from Figure that the catalyst weight necessary to raise the conversion the last 1% from 65% to 66% (3.5 lb) is 8.5 times more than that (0.41 lb) required to raise the conversion 1% at the reactor's entrance. Also, during the last 5% increase in conversion, the pressure decreases from 3.8 atm to 2.3 atm.
168
Chap. 4
When small catalyst pellets are required, the pressure drop can be significant. In 4-6 we saw that significant design flaws can result if pressure drop is or if steps are not taken to minimize pressure drop. One type of reactor that minimizes pressure drop and is also inexpensive to build is the spherical reactor, shown in Figure 4-8. In this reactor, called an ultraformer, dehydrogenation reactions such as paraffin are carried out. aromatic t
Figure 4-8
Products.) This reactor
petroleum naphtha.
one in a series of by K. R
Another advantage of spherical reactors that they are the most economical shape for high pressures. As a first approximation we will assume that the fluid moves down through the reactor in plug Consequently, because
4.4
169
of the increase in cross-sectional area, A,, as the fluid enters the sphere, the will decrease. From the Ergun equation superficial velocity, G = [Equation
we that by decreasing G, the pressure drop he reduced significantly, resulting in higher conversions. Because the cross-sectional area of the reactor is small near the inlet and outlet, the presence of catalyst there would cause substantial pressure drop; thereby reducing the efficiency of the spherical reactor. To solve this problem, (Figscreens to hold the catalyst are placed near the reactor entrance and ures 4-9 and 4-10). Were is the location of the screen from the center of the
Feed
I
Products
Figure 4-9 of a
axis
the inside
variables used with a spherical reactor. The initial and final integration values are slhown
as
and
reactor. We can use elementary geometry and integral calculus to derive the following expressions for cross-sectional area and catalyst weight as a function of the variables defined in Figure 4-10: A,
Spherical reactor catalyst weight
By using these formulas and the standard pressure drop algorithm, one can a variety of spherical reactor prablems. Note that Equations and
170
Chap. 4
(4-39) make use of L and not L'. Thus, one does not need to adjust these formulas to treat spherical reactors that have different amounts of empty space at L L ' ) . Only the upper limit of integration needs the entrance and exit to be changed, = L + .
Example 4-8 Dehydrogenation Reactions in a Spherical Reactor
Reforming reactors are used to increase the octane number of petroleum. In a reforming process 20,000 barrels of petroleum are to be processed per day. The corand 440 molls, respectively. In the responding mass and molar feed rates are 44 reformer, dehydrogenationreactions such as paraffin olefin occur. The reaction is first-orderin paraffin.Assumethat pure paraffin enters the reacand a corresponding concentration of 0.32 tor at a pressure of 2000 Comparethe and conversion when this reaction is carried out in a tubular packed bed 2.4 in diameter and 25 m in length with that of a sphericalpacked bed 6 m in diameter. The catalyst weight is the same in each reactor, 173,870 kg.
k'
Additional information:
= 0.032 = 0.02 = 0.4
.
= L' = 27
= 2.6
Solution
We begin by performing mole balance over the cylindrical core of thickness Az shown in Figure
4.4
Pressure Drop in
Reactors
171
Following
the
I . Mole balance:
algorithm
In - out generation = 0
Dividing by and taking the limit as
0 yields
(E4-8.2)
(E4-8.3)
x (1 + 1 where
P
(E4-8.4)
that ( y with a subscript) represents the mole fraction and y alone represents the pressure ratio, The variation in the dimensionless pressure, y, is given by incorporating the variable y in Equation (4-24):
The equations in boxes are the key equations used in the ODE solver
The units of
(E4-8.7)
program
(E4-8.8)
For a
reactor
(E4-8.10)
172
Chap.
(E4-8.11)
X
0.02
+ (25,630
0.01
I
The last term in, brackets converts Recalling other params) to L = 27 dm, R = 30 dm, and eters, m 44 = 2.6 Table E4-8.1 shows the POLYMATH input used to solve the above equations. The MATLAB program is given as a living example problem on the CD-ROM.
TABLE Equations
POLYMATH PROGRAM
Initial Values
0
.02
rhoca
= 0,
54
For the spherical reactor, the conversion and the pressure at the exit are
A comparison between reactors
0.81
= 1980
If similar calculations are performed for the tubular packed-bed reactor (PBR), one finds that for the same catalyst weight the conversion and pressure at the exit are
= 0.71
308
Figure E4-8.2 shows how conversion, and pressure, vary with catalyst weight in each reactor. Here and represent the tubular reactor and
4.4
Pressure Drop
Reactors
173
IKEY:
0.800
y2
0.40 0
n. 200
oao
0.
1.200
1 . 6 0 0
Figure
and represent the reactor In addition to the higher spherical reactor has the economic benefit of reducing the at the exit sion cost because of higher
Because pressure drop in the spherical reactor is very small, one could increase the reactant flow rate significantly and still maintain adequate pressure at the exit. In fact, Amoco uses a reactor with similar specifications to process 60,000 barrels of petroleum naphtha per day.
(4-40)
where D
pipe diameter, cm
= average velocity of gas,
G=
The friction factor is a function of the Reynolds number and pipe roughness. The mass velocity G is constant along the length of the pipe. Replacing with and combining with Equation for the case of constant T and Equation (4-40) becomes