Professional Documents
Culture Documents
Short-Cut: Absorber Culculatr'ons
Short-Cut: Absorber Culculatr'ons
OWENS
ROBERT N. M A D D O X
Short-cut
Absorber
Culculatr'ons
A bsorption is essentially a vapor pressure phenomenon.
The driving force is the difference between the
partial pressure of a constituent in the gas and the vapor
pressure due to the portion absorbed in the liquid ."
These were the words of Alois Kremser (9) that started
theoretical analysis of the absorption process in 1930.
Since then, many useful and helpful additions have been
made to the solution of absorption problems.
Absorption is a mass transfer operation, principally
occurring in one direction. I n the petroleum industry,
absorption describes a multicomponent separation with
countercurrent flow to increase concentration driving
forces. A gas stream rich in heavy components is fed
into the bottom of the absorber to rise through contacting
devices. I t is the in-gas, wet gas, or rich gas. A
sponge or absorbing oil is fed in the top. I t is the lean
oil. The two streams pass countercurrently and the gas
stream leaving the absorber is denoted the discharge gas,
lean gas, or dry gas; the oil leaving, the rich oil.
The design or evaluation of an absorber represents
time and capital investment. Thus, speed and accuracy
are plied against one another. The accuracy of solu-
tion, in turn, is a function of the accuracy of the data
supplied and the method of calculation.
Design calculations began with little data and easy to
use methods. Time brought improved equilibrium
and enthalpy data and led to more exact calculation
methods. The advent of the computer made possible
rigorous calculation methods; however, not everyone
has access to a computer and a program for such calcula-
tions. Need still exists for a simple hand method that
incorporates absorption calculations, heat balance
calculations, and a minimum number of assumptions
for a broad range of absorber operating conditions.
Such a method would find application in preliminary
designs, economic feasibility studies, initial profiles for
rigorous calculation, or even final designs in the absence
of exact methods. Figure 7. A simple absorber
14 INDUSTRIAL A N D E N G I N E E R I N G CHEMISTRY
A simple hand method that incorporates absorption calculations,
heat balance calculations, and a minimum number of assumptions
for a broad range of absorber operating conditions
-ZA+1
= fraction not recovered (22) stream. Holland (6) adapted this method to a com-
puter solution for absorbers using the absorption factor
and approach.
An iterative tray-by-tray method was described by
T A
$A = 1 - ~
ZA +1 = lean oil fraction lost (23) Sujata (77) and programmed for computer application
by Spear (76). The Sujata method presents no new
They may be evaluated either rigorously by individual equations, but rather applies the absorption factor tech-
tray absorption factors or approximately by effective nique combined with a simultaneous solution for the
absorption factors based on the two-tray model as
defined by Equation 19.
Presenting the first short-cut method to incorporate
heat balance equations, Hull and Raymond (8) pointed AUTHORS William R. Owens is a graduate assistant and
out that the key to over-all absorber heat balance was Robert N . Maddox is Professor and Head of the School of
establishing either the discharge gas or the rich oil Chemical Engineering, College of Engineering, Oklahoma State
temperature. I n light hydrocarbon fractionators, the University, Stillwater, Okla. The authors thank the Depart-
problem is solved by dew- and bubble-point calculations. ment of Health, Education, and Weyare for the NDEA
I n absorbers this procedure is complicated by the Traineeship which supported Mr. Owens during this work,
presence of components with widely varying boiling and the Oklahoma State Thiversity Computing Center for com-
points. puting time used.
first trial, the temperature of tray two is assumed equal Equation 28 represents a heat balance around the middle
to TDQ;for subsequent trials, Tz is calculated from n - 2 trays of the absorber. After TAvEhas been esti-
20 I N D U S T R I A L A N D E N G I N E E R I N G CHEMISTRY
mated for the tower, Equation 29 analysis of the temperature profile from approximately
50 tray-by-tray solutions of multicomponent absorption
problems.
(TBTM -
n
TTOP) TTOP
2
TBTM
) (29)
The average L/V ratio is determined from Equation 30
22 I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y
TABLE V I . EQUILIBRIUM CONSTANT VALUES FOR T H E EQUATION
In K = A + BIT + C / T a+ D/Ta
Component A B C
-
D
200 psia
Methane 0.62936 41.67177 -222.35380 302.69330
Ethane 1 ,89355 27.73705 -254.60420 408.75500
Propane 1.41560 31 ,66986 -302.31800 442,47130
i-Butane 1.06618 37.21840 -363.29280 531 .47560
n-Butane 1.70875 28.09457 -337.54400 504.89700
i-Pentane 1.55110 26.85196 -345.38220 474.47930
n-Pentane 1.45431 28.93027 -365.60330 500.95320
200 mw oil 6.27465 - 65.07498 -174.09260 245 ,73240
Nitrogen - 4,32824 -15.77304 84.92576 -151.42850
500 psia
Methane 0.86989 26,59536 -149.53520 180.98960
Ethane 2.78127 1.13197 -122.21500 211.40340
Propane 2.67452 0.07913 -143.93070 205.87790
&Butane 2.10888 6.29061 -189.73140 241.37350
nButane 2.84696 - 3.14933 -166.11700 225.79970
i-Pentane 2.28437 3.43893 -217.79740 278.51110
n-Pentane 2.78643 - 3.21470 - 199.56500 257.36800
200 mw oil 6.81646 -34.68934 -508.67470 701.43180
Nitrogen -5.18179 - 15.63904 78.55727 -130.90850
Decane 12.66661 -164.81940 368.43090 -402.67520
250 mw oil -5.45264 - 12.38642 66.67615 -118.86150
1000 psia
Methane -0.60247 42.93807 -248.92420 375.15760
Ethane 1.98596 2.60610 -122.12060 235.75920
Propane 1.74780 1.27734 -124.64890 182.44810
&Butane 1.21320 7.51930 -169.46260 222.95100
n-Butane 1.71057 1.19572 -155.51380 212.94260
i-Pentane 1.11933 9.60301 -219.32000 296.40310
n-Pentane 1.78504 - 0.54964 -177.99260 227.70500
200 mw oil 19.15638 -276.73060 829.38210 -798.32900
Nitrogen - 6.65472 - 4.10338 22.03521 -39.18790
evaluated using the effective absorption factor proposed in Tables IV, V, a n d VIII. I n all calculations, the same
by Edmister, was compared with the results of tray-by- equilibrium and enthalpy data were used for all methods.
tray calculations for 44 sample problems. T h e tray- Equation coefficients for these properties are shown in
by-tray calculation procedure was that of Sujata (77) Tables V I and VII. Data for the curve fits were
as programmed and tested by Spear (76). taken from the 1957 edition of the NGSMA Engineering
Table I shows the complete list of problems run to Data Book (72).
evaluate the proposed procedure. T h e number of Results from the Kremser-Brown calculation proce-
theoretical plates varied from two to 12, with eight dures are the poorest of the short-cut procedures in-
being used most frequently because most industrial vestigated. This is to be expected because of the assump-
absorbers are designed with six to 10 theoretical trays. tions necessary in making the calculations. Dry-gas and
Tower pressures ranged from 200 to 1000 psia, with 500 rich-oil temperatures were provided from tray-by-tray
psia being used most often. Temperatures ranged from results on that problem. T h e average absorption
-20” to 120°F. Key component recovery ranged temperature was assumed to be the average of the four
from 20 to 90%. I n most cases, absorption oil was terminal stream temperatures. The L / V ratio was
stripped, but both nonstripped and presaturated oils assumed to be the ratio of the lean oil to the rich gas.
were tested. Gas compositions were varied, with The relatively large errors in component recovery and
arbitrarily assigned designators of lean, average, and lean oil rate indicate that the Kremser-Brown technique
rich. The range of lean oil and rich gas compositions should find major application in providing “first-guess”
used is shown in Table 11. information for one of the other calculating techniques.
Results of these calculations are summarized in Table Equations 8, 9, and 11 evaluated using the two-tray
111. Comparative results for key variables are shown effective factors presented by Edmister give, over-all,
200 psia
Methane 1293.00 636.43 24,000
Ethane 2808.80 1008.30 39.328
Propane 4344.10 1277.50 60.454
i-Butane 6590.80 1346.40 92.632
n-Butane 6590.80 1346.40 92.632
i-Pentane 8353.80 1541.10 116.350
n-Pentane 8353.80 1541.10 116.350
200 m w oil 28028.00 1311.70 561 ,760
500 psia
Methane 612.61 777.21 15.010
Ethane 594,36 1450.00 15.633
Propane 4928.90 573.48 136.510
i-Butane 7759.90 132.25 218.040
+Butane 7759.90 132.25 218.040
i-Pentane 9254.40 212.11 251.650
n-Pentane 9254,40 212.11 251.650
Decane 12873.00 1413.60 390.560
200 mw oil 16600.00 1985.20 549.190
250 mw oil 16600.00 1985.20 549.190
1000 psia
Methane 518.18 716.74 21.001
Ethane 528.86 1317.40 28.629
Propane 5615.70 170.98 166.830
i-Propane 8108.50 -190.91 234.360
+Butane 8108.50 -190.91 234.360
i-Pentane 10630.00 -525.51 297.240
n-Pentane 10630.00 -525.51 297.240
200 mw oil 37289.00 - 5889.70 11160.600
24 I N D U S T R I A L A N D E N G I N E E R I N G CHEMISTRY
duction of the terminal fractional absorption constant
TABLE V I I I . PREDICTED LEAN O I L RATE w. Average results of the proposed method predicted all
Krernser- temperatures within 1.8”F. of Sujata calculation results
Run Suiata Probosed Edmister Brown and all component and oil rates within 1.0%. The
1 26.568 26.771 27.574 35,360 method is recommended for use in the space between
2 14.705 14.907 15.361 20.878 rough guesses and rigorous calculation for any simple
3 7.006 6.988 6.934 11.174 absorber. The time requirement, either by hand or
4 27.614 27.674 28.880 37.773 computer, will be only slightly greater than for other
5 32.018 31 ,972 32.392 40,227 methods with heat balance included.
6 8.606 8.643 8,868 12.717
7 8.036 7.860 7.607 12.863 Nom en clat ure
8 6.146 6.274 6.605 9.918 A = absorption factor defined as A = L / K V
9 0.317 0.265 . . .a 2.997 A, = effective absorption factor defined by Equation 17
A’ = absorption factor defined by Equation 18
10 38.679 38.737 39.707 50.661 c = number of components
11 11.137 11.056 11.075 15,879 C, = heat capacity of specified stream
12 26.710 26.914 27.610 35.328 F = component fraction absorbed
~ heat
A H A= ~ of absorption
13 6.996 6.985 6.944 11.170 K = component equilibrium constant defined by K = y / x
14 25.673 25.886 26.716 36.336 L = total liquid rate leaving a tray, moles
15 7.797 7.666 7.283 12.225 I = component liquid rate leaving a tray, moles
n = number of trays in the column
16 27.875 28.046 28.877 25.054 S = total wet gas shrinkage, moles
17 6.982 7.042 7.341 10.592 T = temperature
18 60.370 60.690 61.094 68.990 V = total vapor rate leaving a tray, moles
V = component vapor rate leaving a tray, moles
19 15.765 16.010 16.050 18,030 X = liquid mole fraction
20 16.783 16.873 18.101 28.156 Y = vapor mole fraction
21 4.533 4.389 4.436 11.312
22 27.259 27,487 28.200 35.650 Greek Symbols
23 7.195 7,192 6.987 11.396 KA = AiAzAs. .A, .
24 26.573 26.728 27,578 35.303 LIA = AiAzAa.. . A , + AzA3.. .A, f ... + A,
+A = fraction of any wet-gas component not recovered
25 6.785 6.768 6.958 11.074 $4 = fraction of any lean-oil component leaving in the dry gas
26 40.222 40,204 40.008 54.947. w = fraction of total gas absorption that occurs on the ter-
27 IO. 374 10.417 10,498 17.223 minal trays
28 26.021 26.353 27.343 34.756
Subscripts
29 6.766 6.916 6.942 10.944
30 20.042 20.280 20.938 27,490 AVE = column average
BTM = bottom tray
31 9.710 9.759 9.938 14.531 DG = discharge gas stream
32 8.311 8.330 8.384 12.771 i = tray reference
33 5.796 5.740 5.566 9.731 IG = in-gas stream
LO = lean-oil stream
34 4.670 4.597 4.274 8.430 n = total number of trays (last tray)
35 32.474 32.714 32.777 43.581 0 = reference for stream entering tray I
36 8.352 8.453 8.368 13.753 OIL = average of lean-oil and rich-oil streams
! RO = rich oil stream
37 29.543 29.781 30,070 39.521 TOP = top tray
38 7.660 7.640 7.580 12.469 1 = first tray
39 28.138 28.304 28.806 37.454 2 = second tray
40 7.350 7.302 7.223 11.601
BI BL IOG RAPHY
41 8.374 8.269 7.712 12.601
(1) Brown, G. G., and Souders, M., Jr., IND.ENC.CHEM.,24,519 (1932).
42 9.345 9.350 9.462 13.104 (2) Edmister, W. C., i b d , 35, 837 (1943).
43 30.815 31.200 30.863 35.123 (3) Edmister, W. C., ThePetroleum Engineer, 18, No. 13 (September 1947).
(4) Edmister, W. C., A.I.Ch.E. J., 3, No. 2, 165 (June 1957).
44 8.218 8.398 7,848 11.222 ( 5 ) Erbar, J. H., Oklahoma State University, Stillwater, Okla., private communica-
tion.
a Did not converge. (6) Holland, C. D., “Multicomponent Distillation,” Prentice-Hall, Inc., 1963.
(7) Horton, G., and Franklin, W. B., IND.ENO.CHEM.,92,1384 (1940).
(8) Hull, R. J., and Raymond, K., Oil B Gus J. (Nov. 9, 16,.23, and 30, Dec. 7, 14,
and 28, 1953).
(9) Kremser, Alois, Not. Petrol. News, 22, 48, (May 21, 1930).
(10) Landes, S. H., and Bell, F. W., Petrol. Refiner, 39, No. 6 , 201 (1960).
(11) Lewis, W. K., and Matheson, G. L., IND. ENC.CHEM.,24,494 (May 1932).
(12) Natural Gasoline Supply Men’s Association, “Engineering Data Book,” 7th
Conclusions ed., Tulsa, Okla., 1957.
The proposed method for making short-cut absorption (13) Nielsen, K.L., “Methods in Numerical Analysis,” Macmillan Co., 1964.
(14) Ravicz, A. E., “Non-Ideal Stage Multicomponent Absorber Calculations by
calculations provides design information closer to Automatic Digital Computer,” Ph.D. Thesis, University of Michigan, Ann Arbor,
Mich. (1959).
rigorous calculation results than any other short-cut (15) Smith, B. D., “Design of Equilibrium Stage Process,” McGraw-Hill, 79 pp,
1963.
method tested. This method incorporates heat balances (16) Spear, R. R., “An Evaluation of the Sujata Absorption Calculation Method,”
and absorption factors with no empirical graphs or Master’s Thesis, Oklahoma State University, Stillwater, Okla., 1966.
(17) Sujata, A. D., Petrol. Refiner, 40 (12), 137 (December 1962).
charts. The only nonexplicit assumption is the intro- (18) Thiele, E. W., and Geddes, R. L., IND. ENO.CHEM.,25,289 (March 1933).
To simplify understanding, an illustrative example will be Solutions can be reached without estimates but will require
presented. Explanatory notes and comments will be inter- more iterations. Perhaps the best way is to perform a
spersed with the calculations. simple Kremser-Brown type absorption calculation assum-
The calculation procedure can be divided into five dis- ing an average tower temperature of 90°F. For purposes
tinct areas: of this solution assume a total absorption of 9.6 moles/100
moles rich gas.
Preliminary calculations The second step is to assume a value for the percentage of
Absorption calculations the total absorption that takes place on the two terminal
Terminal tray calculations trays. This value can range from about 75y0to 90% with
Heat balance calculations 80% being a good over-all average value. Now we must
Adjusting section estimate the ( L / V ) ratio for the tower.
In order to carry out these short-cut calculations, the same
information must be known as is generally required for ( L / V ) = (lean oil +
shrinkage)/rich gas
other short-cut methods of calculation. The single excep- 20f96
( L / V ) = A = 0.296
tion to this is terminal-stream temperatures. Neither the 100
dry-gas nor the rich-oil temperature is required. Both are
independently calculated during the course of the calcula- I n the preliminary calculations, this will be assumed con-
tion. stant across the tower.
B. We now must calculate the enthalpy and heat
Problem Statement capacity of the feed streams to the tower. These are
An absorber operates at 300 psia and contains the equiva- calculated as the mole fraction average of the individual
lent of eight theoretical trays. Lean oil (assumed to have component values.
characteristics of octane) enters at G O O F and a fixed rate of
20 moles per 100 moles of rich gas. The rich gas enters at Rich gas enthalpy = Z(mo1es of component X component
70°F. The composition of the rich gas is shown in column enthalpy)
1 of the tabular calculations attached (Table IX). Calcu- = Z: [(I) x ( 3 ) ] (referring to columns in
late the product stream compositions and temperatures. table of calculations)
= 637,200 Btu
Solution
I n like manner:
Preliminary calculations. A. The first step in the solu-
tion is the assumption of the total moles of material ab- Rich-gas heat capacity = Z[(1) x (4)] = 1041 Btu/OF
sorbed in the tower. A good initial estimate is important Lean-oil enthalpy = 2 [ ( 2 ) X (5)] = 422,700 Btu
since the total shrinkage is the major iteration variable. Lean-oil heat capacity = Z[(2) x (6)] = 2082 Btu/OF
Terminal (Top) Tray Evaluation I. Calculate a new estimate of the top tray temperature,
The initial temperatures should be improved for the next
TDG
from Equation 27 :
iteration. The top tray temperature is refined by combin-
ing pseudo-dew-point and heat-balance calculations.
This procedure is iterative, but experience and insight
TDG= TLO + AHTop - (CPNEXT)
CPLO
- Tz)
(TDG
When the two values are the same, the top tray temperature = 60 + 13.08
has been found. = 73.08'F
= 76.99'F
+ 900 + 2082 1 Total dry gas = Z(23) = 90.364 moles
Total rich oil = Z(24) = 29.636 moles
Adjusting Section
Shrinkage = 100.0 - 90.364 = 9.636 moles
For the foregoing calculation, the assumed value of
shrinkage was 9.6 moles. The value calculated was 9.637 The estimated shrinkage (9.637 moles) and the calculated
moles. A new value should be assumed and the calcula- value (9.636 moles) are almost equal, indicating a solution
tions repeated to bring the assumption and final result into to the problem. The solution to this problem was aided
better agreement. Direct iteration will suffice. by fortuitous initial shrinkages and temperature assump-
A. Determine the L / V values a t the top tray, bottom tions. This sample problem presents only the calculational
tray, and the average L / V for the column. format of the proposed short-cut absorber calculation
Top tray method.
liquid leaving tray 1 Tray-by-tray results of the Sujata procedure programmed
(L/V)TOP = dry gas by Spear (76) predict the rich-oil temperature to be 76.3'F
and the dry-gas temperature to be 74.2'F. The dry-gas
26.410 composition was 77.562 moles methane, 8.231 ethane,
= ~ = 0.2923
90.363 4.545 propane, and 0.156 octane.