5-103

5.4 DISTILLATION COLUMN (T-102)

5.4.1 INTRODUCTION

In this process, acetylated castor oil (ACO) which is also known as palmatic acid or
hexadecanoic acid (C
15
H
31
COOH), undergoes thermal cracking to produce 1-tetradecene
(C
14
H
28
), which is referred to as drying oil (DO); and acetic acid (CH
3
COOH). However, there
is an undesired reaction that produces 1-octacosene (C
28
H
56
), which is referred as gum, from
the drying oil dimers. The chemical reactions and reaction kinetics for the thermal cracking
process are as follows (Grummit & Fleming, 1945):

Chemical Reactions

DO acid acetic ACO
) ( ) ( ) (
28 14 3
1
31 15
l H C g COOH CH l COOH H C
k
+
(5.4.1)


gum DO
) ( ) ( 2
56 28 28 14
2
s H C l H C
k

(5.4.2)

There are two chemical reactions that occur in the thermal cracking of ACO. The first
(equation 5.4.1) is the main reaction where 1 mole of ACO (palmatic acid) reacts under high
temperature to produce 1 mole each of Acetic Acid & Drying Oil (1-tetradecene). The second
reaction (equation 5.4.2) is an undesired side reaction that produces 1 mole of Gum (1-
octacosene) from 2 moles of Drying oil (DO).








5-104

5.4.1.1 Process Description

Figure 5.4.1 shows the process flow diagram for this plant. This process begins with
the raw material, ACO, being fed from a holding tank where it is mixed with recycled ACO.
The mixed feed would then heated to reaction temperature. Since the reaction occurs at high
temperatures, there is no need for a catalyst. The heated feed then flows into a plug flow
reactor. The reactor is simply a vessel to promote radial mixing. Then, process fluid is feed to
the distillation column (T-101) to separate the gum from the product. Another stream from
first distillation column (T-101) is fed to a distillation column (T-102) where the ACO is
separated and recycled. The non-recycled stream again goes through another distillation
column (T-103) where the DO is purified from the acetic acid. The contents of Streams 18
(acetic acid) and 19 (DO) are cooled and sent to storage. Drying oil (DO) is the main product
for sale, while the acetic acid is sold as a by-product (Grummit & Fleming, 1945).
5-105

5.4.1.2 Process Flow Diagram


5-106

5.4.2 OBJECTIVE OF THE DESIGN

The objective of this chapter is to determine detail design parameter for distillation
column T-102. The purpose of this distillation column is to separate the recycled stream
(Stream 16), which consists of hexadecanoic acid (ACO) and tetradecene (DO), from stream
14. Hexadecanoic acid and tetradecene are then recycled back to the reactor to obtain
higher conversion in reactor 101 and 102. The non-recycled stream again goes through
another distillation column (T-103) where the DO is purified from the acetic acid.


5.4.3 CHEMICAL DESIGN PROCESS FOR DISTILLATION COLUMN (T-102)

Distillation is by far the most important separation process in the petroleum and
chemical industries. It is the separation of key components in a mixture by the difference in
their relative volatility or boiling points. In most cases, distillation is the most economical
separating method for liquid mixtures (Sinnott, 2005). In this process, the feed in stream 10
contains more than two components which are hexadecanoic acid, tetradecene, acetic acid
and octacosene. It is commonly referred to as multi-component distillation. Figure 5.4.2
below shows the schematic diagram of the distillation column unit (T-102).















5-107






























Figure 5.4.2: Distillation column (T-102)
Component
Flowrate
(kmol/hr)
Mole
Fraction
ACO 0.2653 0.00215
DO 61.4943 0.49889
AA 61.5026 0.49896
Gum 0.0001 0.00000
Component
Flowrate
(kmol/hr)
Mole
Fraction
ACO 191.4709 0.60744
DO 61.6862 0.19570
AA 61.5026 0.19512
Gum 0.5521 0.00175
Component
Flowrate
(kmol/hr)
Mole
Fraction
ACO 191.2056 0.99612
DO 0.19197 0.00100
AA 0.0000 0.00000
Gum 0.5520 0.00288


T-102
(S10)
315.21 kmol/hr
(S12)
123.26 kmol/hr

(S13)
191.95 kmol/hr
T = 125.5
o
C
P = 13 kPa
T = 82.8
o
C
P = 11 kPa
T =268.1
o
C
P = 12 kPa
Where
- ACO = hexadecanoic Acid
- DO = Tetradecene
- AA = Acetic Acid
- Gum = Octacosene
5-108

5.4.4 FLOWCHART FOR CHEMICAL DESIGN

NO
YES
5-109



Figure 5.4.3: Algorithm for distillation column design
5-110

5.4.5 THEORETICAL BACKGROUND

5.4.5.1 Activity coefficients (Non-ideal Content)

The activity coefficient is a unitless thermodynamic function. If
k
= 1 (an ideal
solution), the activity of component k is equal to its mole fraction and the behavior of k, from
the point of view of its chemical potential, is completely determined by its composition. If
k
>
1 (a nonideal solution where the component k is said to exhibit a positive deviation from
Raoult's law), then a
k
>x
k
and in the evaluation of its chemical potential, component k "acts as
if" the solution contains more of k than the mole fraction suggests. Similarly, if
k
< 1 (a
nonideal solution where the component k exhibits a negative deviation from Raoult's law) so
that a
k
<x
k
, the component "acts as if" there is less of it present than the composition
suggests. Therefore for non-ideal multicomponent system, the formula below is not valid
anymore.
(5.4.3)
Whereby, for non-ideal system, is needed. Thus formula becomes:
(5.4.4)

5.4.5.2 Relative Volatility

The relative volatility of two components can be expressed as the ratio of their k
values. From the K-values, the relative volatility for each component can be determined and
more accurate. (Sinnott, 2005):

j
i
j
K
K
i = o
(5.4.5)

Where, K
i
= light components
K
j
= heavy components




5-111

5.4.5.2.1 Key Component
Key component is component whose volatility that is vapor pressure characteristics relatives
to each other make them adjacent in the listing of the component in the feed. These key
components are separate into two as follows (Geankoplis,2003) :
1. Light key component is component whose percent recovery is greater in the
distillate than in the bottom. Acetic acid is chosen as light key component.
2. Heavy key component is the lightest component whose percent recovery is greater
in the bottom than in the distillate. Hexadecanoic acid is chosen as heavy key
component.

5.4.5.3 Bubble and Dew Point (Non-ideal system)
In order to estimate the number of stages, condenser and reboiler temperature, the
dew and bubble points must be calculated first. By definition, a saturated liquid is at its
bubble point (any rise in temperature will cause a bubble vapor to form), and a saturated
vapor is at its dew point (any drop in temperature will cause a drop of liquid to form) (Sinnott,
2005).

Bubble point:

= yi

= 0 . 1 kixi (5.4.6)
Dew point :

= xi

= 0 . 1 / ki yi (5.4.7)

5.4.5.4 Minimum theoretical Stages
The Fenske equation (Fenske,1932) is used to estimate the minimum stages required at total
reflux (Sinnott, 2005):

(5.4.8)
Where,
D = subscript for distillate
B = subscript for bottom
N
m
= minimum theoretical stages at total reflux
X
HK
= concentration of heavy key component
X
LK
= concentration of light key component

avg
= average relative volatility
5-112

5.4.5.5 Minimum Reflux Ratio
The number of stages required for separation will be dependent on the reflux ratio
used. As the reflux ratio is reduced a pinch point will occur at which separation can only be
achieved with an infinite number of stages. Colburn (1941) and Underwood (1948) have
derived equations for estimating the minimum reflux ratio for multicomponent distillations. As
the Underwood equation is more widely used , the equation can be stated in the form
(Sinnott, 2005):

u o
o
i
d i
i
, x
= R
m
+ 1 (5.4.9)
Where,
i
o

= the relative volatility of component i with respect to some reference
component, usually the heavy key
R
m
= the minimum reflux ratio
Xi,
d
= concentration of component i in the tops at minimum reflux

and u is the root of the equation:

q
x
i
f i i
1
,
u o
o
(5.4.10)
Where xi,
f
= the concentration of component i in the feed, and q depends on the condition of
the feed.


5.4.5.6 Column Efficiency
A quick estimate of the overall column efficiency can be obtained from correlation
given by OConnell (1946) in the form of an equation (Sinnott, 2005):

E
0
= 51 32.5 log (
a

a
) (5.4.11)
Where,

a
= molar average viscosity,mNs/m
2


a
= average volatility of the light key
The overall column efficiency is correlated with the product of the relative volatility of the light
key component and the molar average viscosity of the feed, estimated at the average column
temperature. The overall column efficiency is relatively high in the range 50-100% (Seader,
2004)
5-113

5.4.5.7 Number of stages using Lewis Matheson method
Estimation of the actual physical number of plates required in the distillation column is
usually done by dividing the actual theoretical stages by the overall plate efficiency (Sinnott,
2005):

(5.4.12)



Besides that, Lewis Matheson method is using in calculation of number of ideal stages. Feed
point also can be determined from this method:

1. For Top-Down:

(5.4.13)


2. For Bottom-up:

(5.4.14)












5-114

5.4.5.8 Approximate Column Sizing

A trial and error approach is necessary in plate design. Starting with a rough plate
layout, checking key performance and revising the design until a satisfactory design is
achieved (Sinnott, 2005). The design procedures are as follows:
1. Determination of the vapour and liquid rate, based on the reflux ratio and feed
condition.
2. Estimation of the system physical properties.
3. Selection on a trial plate spacing.
4. Based on the flooding condition, the column diameter is determined.
5. Try to make a plate layout with downcomer area, active area, hole diameter, hole
area, weir height, weir length, and plate thickness.
6. The weeping rate is checking.
7. The plate pressure drop is checking.
8. The down-comer backup is checking .
9. Determination of plate layout details
10. Confirmation on the percentage flooding based on the chosen column diameter.
11. Entrainment is checking.
12. Determination of the column wall thickness and column head selection.
13. The design is finalize with the drawing and data specification sheet.


5.4.5.9 Plate Spacing

The overall height of the column will depend on the plate spacing. Plate spacing from
0.15 m to 1m are normally used. In this distillation column, the plate spacing used is 0.50 m
as it is in the range of 0.15 m to 1 m (Sinnott, 2005).


5.4.5.10 Column Diameter

The column diameter is computed to avoid flooding where the liquid begins to fill the
column and leave with vapor because it cannot flow downward at the required rate (Seader,
2004).The principle factor on determining the column diameter is the vapor flow rate. The
column diameter can be calculated by calculating the top and the bottom net area at its
maximum volumetric flow rate.
5-115

The Liquid-vapor flow factor, F
LV
is given by:
F
Lv
=
5 . 0
|
|
.
|

\
|
|
.
|

\
|
L
V
Vw
Lw

(5.4.15)
Where:
L
w
= liquid mass flowrate, kg/s
V
w
= vapor mass flowrate, kg/s


5.4.5.10.1 Flooding velocity
The flooding condition fixes the upper limit of vapour velocity. A high vapour velocity
is needed for high plate efficiencies, and the velocity will normally be between 70 to 90 per
cent of that which would cause flooding. For design, a value of 80 to 85 per cent of the
flooding velocity should be used. The flooding velocity can be estimated from the correlation
given by Fair (1961,referred to Sinnott,2005):
(5.4.16)
Where,
U
f
= flooding vapour velocity, m/s, based on the net column cross-sectional
area,A
n

K
1
= a constant

5.4.5.11 Weir Liquid Crest
The height of the liquid crest over the weir can be estimated using the Francis weir
formula. For a segmental downcomer this can be written as (Sinnott, 2005):
h
ow
=
3 / 2
750
|
|
.
|

\
|
w
w
LxI
L

(5.4.17)

where l
w
= weir length, m
h
ow
= weir crest, mm liquid
L
w
= liquid flow-rate, kg/s


5-116

5.4.5.12 Weep Point
The vapor velocity at the weep point is the minimum value for stable operation.
Minimum vapor velocity through the holes based on the holes area,
h
(Sinnott, 2005)
:


h
=
5 . 0
2
) (
) 4 . 25 ( 9 . 0
v
dh
k

(5.4.18)

where

h
= minimum vapour velocity through the holes(based on the hole area), m/s,
d
h
= hole diameter, mm,
K
2
= a constant, dependent on the depth of clear liquid on the plate


5.4.5.13 Plate Pressure Drop
The pressure required is automatically developed by the reboiler, which generates
vapor at the pressure sufficient to overcome the pressure drop in the column and in the
condenser. The overall pressure drop is calculated to determine the pressure and
temperature in the reboiler and the pressure drop per plate must be checked to make sure
the plate will operate properly without flooding. The pressure drop across the plate can be
divided into two parts which are the friction loss in the holes and the pressure drop due to the
holdup of liquid on the plate. It is usually given as an equivalent head in millimeters or inches
of liquid (Sinnott, 2005):


l d t
h h h + =
(5.4.19)

Where
h
t
= total pressure drop per plate, mm of liquid
h
d
= friction loss for dry plate, mm of liquid
h
l
= equivalent head of liquid on plate, mm of liquid






5-117

5.4.5.13.1 Dry plate drop
Pressure drop through the holes can be predicted from a modification of equation for
flow through an orifice (Sinnott, 2005):

h
d
=
|
|
.
|

\
|
|
|
.
|

\
|
L
v
C
U
o
h

2
51
(5.4.20)


Where u
h
= vapor velocity throught holes, m/s

v
= vapor density

L
= liquid density
C
o
= orifice coefficient

The orifice coefficient is a function of the plate thickness, hole diameter and the hole of
perforated area ratio.


5.4.5.13.2 Residual head
Residual head is a function of liquid surface tension, froth density and froth height.
Taking residual head as a fixed value of 12.5 mm of water, the equation for estimating the
residual head proposed by Hunt et al. (1955, referred to Sinnott, 2005):

Residual head,h
r
=
|
|
.
|

\
|
L
3
10 5 . 12
(5.4.21)

Total Drop can be determined using this equation:
Pressure drop per plate,h
T
= h
d
+( h
w
+

h
ow
) + h
r
(5.4.22)

5.4.5.14 Downcomer Liquid Back Up
The liquid area and the plate spacing must be such that the level of the liquid and forth in the
down comer is well below the top of the outlet weir on the plate above. The backup of liquid
in the down comer is caused by the pressure drop over the plate and the resistance to flow in
the down comer itself.
5-118

Head loss by Cicalese et al (1947, referred to Sinnott, 2005):
h
dC
=
2
166
|
|
.
|

\
|
ap
w
LxA
L

(5.4.23)

where L
w
= liquid flow rate in downcomer, kg/s,
A
ap
= either the downcomer area A
d
or the clearance area under the downcomer
A
ap
; whichever is the smaller, m
2
.

The clearance area under the downcomer is given by:
A
ap
= h
ap
l
w
(5.4.24)

where h
ap
is height of the bottom edge of the apron above the plate. This height is normally
set at 5 to 10 mm below the outlet weir height (Sinnott, 2005):

h
ap
= h
w
(5 to 10 mm) (5.4.25)

Back up in the downcomer,h
b
= h
w
+

h
ow
+ h
T
+ h
dc
(5.4.26)



5.4.5.15 Downcomer residence time
Sufficient residence time must be allowed in the down comer for the entrained vapor
to disengage from the liquid stream; this is to prevent heavily aerated liquid being carried
under the downcomer. At time of at least 3 second is recommended. The down comer
residence time, t
r
is given by (Sinnott, 2005):

(5.4.27)
Where,
t
r
= residence time, s,
h
bc
= clear liquid back-up, m.



5-119

5.4.5.16 Check Entrainment by Fair
Entrainment can be estimated from the correlation given by Fair (1961), Figure 5.4.13, which
gives the fractional entrainment (kg/kg gross liquid flow) as a function of the liquid-vapour
factor F
LV
, with the percentage approach to flooding as a parameter. The percentage flooding
is given by (Sinnott, 2005):
Flooding percentage =
F
V
u
U % 100
(5.4.28)

5.4.5.17 Perforated Area

The area available for perforation will be reduced by the obstruction caused by
structural members (the support rings and beams), and by the use of calming zones.
Calming zones are unperforated strips of plate at the inlet and outlet sides of the plate. The
width of the support ring for sectional plates will normally be 50 to 75mm: the support ring
should not extend into down comer area. A strip of imperforated plate will left round the edge
of cartridge-type trays to stiffen the plate. A cartridge type construction is used; allowing 50
mm unperforated strips round. Formula used is list below (Sinnott, 2005):

Mean length, unperforated edge strips = (D
c
0.05) x
t
x (5.4.29)
Mean length of calming zone, approx = weir length + width of unperforated strip (5.4.30)
Total area available for perforation, A
p
= Active area (Area of unperforated edge + Area of
calming zones) (5.4.31)











5-120

5.4.6 CALCULATION FOR CHEMICAL DESIGN

5.4.6.1 Step 1: Obtained composition and flow rate

Composition and flow rate at feed,F; distillate,D and bottom,W are obtained from figure 5.4.2.


5.4.6.2 Step 2: Obtained K-values

From table 5.4.1, the average activity coeffients is calculated based on K-values. Therefore,
based on the value, it is show that the system is non-ideal system.

Table 5.4.1 : Average activity coefficients
Number of tray ACO DO Acetic Gum
Tray1 10.4332 22.0111 5.2563 46.5074
Tray2 10.4668 13.9794 4.9210 38.3746
Tray3 10.3886 12.0192 4.7993 35.8138
Tray4 10.2905 18.9198 4.7438 35.0642
Tray5 10.1914 22.7024 4.6991 34.6213
Tray6 10.1110 23.6861 4.6128 26.7363
Tray7 10.0327 42.6745 4.4954 24.3179
Tray8 9.9548 9.9136 4.2639 18.9130
Tray9 9.8422 9.5379 4.0870 14.1124
Tray10 9.7408 9.4201 4.0128 12.7745
AVERAGE 10.1452 18.4864 4.5891 28.7236
STDEV 0.2396 9.6974 0.3654 10.5931


5.4.6.3 Step 3: Determination of relative volatilities

Table 5.4.2 shows the average relative volatility for each component. The average relative
volatility is based on light key. Acetic acid is chosen as light key. So the average relative
volatility,
avg
is 497.26. It shows that the component is volatile as it has greater volatility. The
greater relative volatility, the mixture is easier to separate.





5-121

Table 5.4.2: average relative volatility
No.of trays (ACO) (DO) (Acetic) (Gum)
Tray1 1 85.6000 1053.3788 0.6599
Tray2 1 47.4561 728.3269 0.6112
Tray3 1 39.3427 650.7129 0.5953
Tray4 1 62.0185 638.1988 0.5921
Tray5 1 75.0035 635.8290 0.5911
Tray6 1 66.2164 438.9700 0.5189
Tray7 1 111.5893 372.6323 0.4967
Tray8 1 20.6083 229.8526 0.4347
Tray9 1 14.2100 124.8799 0.3613
Tray10 1 12.3553 99.7943 0.3373
AVERAGE 1 53.4400 497.2576 0.5199
STD 0 31.0754 285.0185 0.1050


5.4.6.4 Step 4: Determination of dew and bubble point

The system is a non-ideal system and the bubble point and dew point is calculated
based on the relationship between K-values at each tray and temperature that are plotted
into graph and equation forms. Figure 5.4.4 to 5.4.7 shows variation of K-values with
temperature for all components.


Figure 5.4.4: graph of K value for hexadecanoic acid

5-122


Figure 5.4.5: graph of K value for Tetradecene



Figure 5.4.6: graph of K value for Acetic acid


5-123


Figure 5.4.7: graph of K value for Octacosene


5.4.6.4.1 Dew Temperature (Top column)

Operating Temperature = 355.93 K
Operating Pressure = 11 kPa

Based on equation 5.4.7 for dew point, value close to 1.0, therefore dew temperature
355.93 K is accepted.

Table 5.4.3: Method to find dew temperature at the top of the column
Component yi = XD K y=KXD Xi = y/K
ACO 0.00215 -0.66217 -0.00143 0.00215
DO 0.49889 -5.66899 -2.82820 0.49889
Acetic Acid 0.49896 -4.53711 -2.26382 0.49896
TOTAL 1.00000 1.00000



5-124

5.4.6.4.2 Bubble Point (Bottom Column)
Operating Temperature = 570.43 K
Operating Pressure = 12 kPa

Based on equation 5.4.6 for bubble point,
i i
x k value should close to 1.0. This is trial and
error method as
i i
x k value is not 1 at temperature 541.21 K. Therefore, using goal and
seek function, the bubble point is at 570.43 K where
i i
x k value is closed to 1.


Table 5.4.4: Method to find bubble temperature at the bottom of the column

Component Xi=XB K yi = KXi
ACO 0.9961 0.9895 0.9857
DO 0.0010 13.3364 0.0133
Gum 0.0029 0.3342 0.0010
TOTAL 1.0000 1.0000


5.4.6.4.3 Bubble Point for Feed
Operating Temperature = 450.70 K
Operating Pressure = 13.00 kPa

Based on equation 5.4.6 for bubble point,
i i
x k value should close to 1.0. This is trial and
error method as
i i
x k value is not 1 at temperature 398.62 K. Therefore, using goal and
seek function, the bubble point at feed is 450.70 K.

Table 5.4.5: Method to find bubble temperature of the feed
Component Xi=XF K yi = KXi
ACO 0.60744 0.06758 0.04105
DO 0.19570 2.72792 0.53385
Acetic Acid 0.19512 2.17852 0.42506
Gum 0.00175 0.02292 0.00004
TOTAL 1.00000 1.00000

5-125

5.4.6.5 Step 5: Calculate minimum theoretical stages using Fenske equation
Minimum number of stages is calculated using Equation 5.4.8:

N
m
=
avg
B D
x
) 25 . 497 log(
) 11 ^ 10 08 . 5 / 9961 . 0 ( ) 00215 . 0 / 49896 . 0 log(

N
m
= 5 stages

5.4.6.6 Step 6: Estimation of Minimum Reflux Ratio
By using Underwood method (equation 5.4.10) to estimate minimum reflux, q = 1.

q
x
i
f i i
1
,
u o
o

u o
o
i
f i i
x ,
=1 1 = 0

Table 5.4.6: minimum reflux ratio base on q value


Component Xif i i Xif 2.51
ACO 0.60744 1.0000 0.6074 -0.4013
DO 0.19570 53.4400 10.4581 0.2054
Acetic Acid 0.19512 497.2576 97.0224 0.1961
Gum 0.00175 0.5199 0.0009 -0.0005
Total 0.0003

This is trial and error method, so, the value of u = 2.51 which is close to 0.

Therefore for the equation 5.9:

u o
o
i
, ixi
d
= R
m
+ 1, R
m
+1 can be determine using = 2.51.
Table 5.4.7 shows the summary calculation to determine R
m
+1 and also based on mole
fraction and relative volatilities at top tray.


5-126

Table 5.4.7: Summary calculation for the equation to determine R
m
+1
Component Xid i i Xid i- i Xid / i-
ACO 0.00215 1.000 0.00215 -1.51362 -0.0014
DO 0.49889 53.440 26.66069 50.92640 0.5235
Acetic Acid 0.49896 497.258 248.11020 494.74394 0.5015
Total 1.0236

R
m
+ 1 = 1.0236
Therefore, the minimum reflux ratio, R
m
= 0.0236
1 R
R
m
m
+
= 0.0230
For specimen calculation, for R=2,
R+1 = 3
R/(R+1)= 0.67


Figure 5.4.8: Erbar-Maddox correlation (Sinnott, 2005)

5-127

Figure 5.4.8 is a Erbar-Maddox correlation. From Erbar- Maddox correlation, value N
m
/N can
be determined. Therefore,

N
N
m
= 0.20
N
m
/N = 0.20


N =23stages

However, iteration needs to be performed to obtain optimum reflux ratio. For other reflux
ratio, the summary is in table 5.4.8:

Table 5.4.8: Reflux ratio
R 2 3 4 5 6
Nm/N 0.2 0.25 0.28 0.35 0.42
N 23 18.776 16.764 14.412 11.176
Deviation 4.6940 2.0117 3.3529 2.2353

Table 5.4.8 shows iteration between reflux ratio and stages in order to determine optimum
reflux ratio. Value for Nm/N is from figure 5.4.8. The optimum reflux ratio is near this value
which is 5. Therefore, the required number of stages is 14 stages.


5.4.6.7 Step 7: Determination of Column Efficiency using OConnells correlation
At average temperature,
T avg = (5.4.32)
=
= 448.57 K
Then, log is obtained from equation 5.4.33,
log = Ax (5.4.33)


5-128

Table 5.4.9: Viscosity data
Component Vis A Vis B log Viscosity mol frac
ACO 0.428231
-
0.448606 0.955536 9.026849 0.607436
DO 0.017500
-
0.424580 0.041256 1.099654 0.195698
Acetic Acid -0.011000
-
0.450630 -0.024435 0.945290 0.195115
Gum -0.039509
-
0.636560 -0.062154 0.866654 0.001752

Using equation 5.4.33, viscosity at average temperature for each component can be
determined as summarize in table 5.4.9 above. Therefore,
Molar average viscosity of feed ,
avg
= 0.607436 (9.026849) + 0.195698 (1.099654) +
0.195115(0.945290) + 0.001752 (0.866654)
= 5.8844 mNs/m
2

Therefore, from equation 5.4.11, column efficiency,
E
0
= 51 32.5 log (
a

a
)
From previous,
a
= 497.26
Hence the column efficiency, E
o
= 51 32.5 log (5.8844 x 497.26)
= 61.65 %

5.4.6.8 Step 8: Determination of number of stages using Lewis Matheson method

Constant molar overflow is assumed and the material balance and equilibrium
relationship equations are solved stage by stage starting at the top and bottom of the
column. The procedure is to start the calculation at the top and bottom of the column and
proceed toward the feed point. The initial estimates of the component distributions in the
products are then revised and the calculations repeated until the iteration stop when ratio of
Light key/ Heavy key at top down and bottom up is close to ratio of light key / heavy key at
feed. Table 5.4.10 to 5.4.12 shows summarization of top down, bottom up and results
obtained after iteration performed. Summarize table for stage by stage is shown in Appendix
A4.

5-129

Top down Equation:

Table 5.4.10: Summary of top-down calculation
L
o
616.31 kmol/ hr
D 123.26 kmol/ hr
R 5
V
1
739.57 kmol/ hr
L
o
/V
1
0.83

From equation 5.4.13:

y
2
= 0.83 (x
1
-x
o
) + y
1

Bottom up Equation:
Table 5.4.11: Summary of bottom-up calculation
V = V
o
739.57 kmol/ hr
F 315.21 kmol/ hr
L = L
o
+F 931.52 kmol/ hr
B 191.95 kmol/ hr
B / L 0.206
V / L 0.794

From equation 5.4.14:

X
B1
= 0.794 y
B
+ 0.206 x
B





5-130

Table 5.4.12: Results after Iteration
Component Composition (Top-down) Composition (Feed) Composition (Bottom-up)
ACO 0.9778 0.9398 7.389E-06
DO 0.0000 0.0049 1.0000
Acetic acid 0.0735 0.0264 7.578E-09
Gum 0.0102 0.0289 1.754E-09

Summary:
1. At Top-down: 3 stages
2. At Bottom-up: 11 stages
3. Feed Tray is at Tray 4
4. Number of ideal stages = Top-down + Bottom-up
= 14 stages

5. Actual stages = Number of ideal stages (5.4.34)

0
E
= (14-1) / 0.62
= 21 stages
Therefore, Number of actual stages is 21 stages

6. Feed tray location:
a) From top = (3-0.5) / (0.62)
= 4 trays

b) From bottom= (11-0.5) / (0.62)
=17 trays
Therefore, feed tray is on the 4
th
Tray.
5-131

5.4.6.9 Step 9: Calculation for density and relative molar mass (RMM)

Table 5.4.13: Properties of components at feed, distillate and bottom
Component MW feed,Xf distillate,XD Bottom, XB Liquid density (kg/m
3
)
ACO 256.43 0.60744 0.00215 0.99612 881.58
DO 196.38 0.19570 0.49889 0.00100 774.10
Acetic Acid 60.05 0.19512 0.49896 5.0774E-11 1051.50
Gum 392.73 0.00175 7.7961E-07 0.00288 804.59

Table 5.4.13 shows properties of component at feed, distillate and bottom. The properties
are then used to calculate relative molar mass and density at top and bottom product. The
formulas used are listed as below (Sinnott, 2005):

1. Relative molar mass = MW(X) (5.4.35)
2. Mean liquid density,
L
=
i i
x

(5.4.36)
3. Mean Vapor density,
V
=
STP
OP
OP
STP
STP
P
P
x
T
T
x
V
MW

(5.4.37)

Calculation for average relative molar mass, RMM :
Using equation 5.4.35,

RMM of Feed = 0.60744 (256.428) + 0.19570 (196.378) + 0.19512 (60.052)
+ 0.00175 (392.730)
= 206.5991 kg / kmol

RMM of Top Product = 2.1522x10
-3
(256.428) + 0.4988900 (196.378) +
0.4989571 (60.052)
= 128.49 kg / kmol

RMM of Bottom Product = 0.99612 (256.428) + 0.00100 (196.378) +
2.88x10
-3
(392.730)
= 257 kg / kmol



5-132

Calculation for density :
1. Top product
Using equation 5.4.36:
Liquid density,
L
=
i i D
x
,

= 2.1522x10
-3
(881.5780) + 0.49889 (774.0980) + 0.4989571
(1051.50)


= 912.74 kg / m
3

Using equation 5.4.37:
Vapor density,
V
=
STP
OP
OP
STP
STP
P
P
x
T
T
x
V
MW

= (128.49 / 22.4) kg/m
3
(273 / 355.93) K (0.11 / 1.0) bar
= 0.4840 kg / m
3

2. Bottom product
Using equation 5.4.36:
Liquid density,
L
=
i i B
x
,

= 0.99612 (881.5780) + 0.00100 (774.098) + 2.88x10
-3

(804.593)
= 881 kg / m
3

Using equation 5.4.37:
Vapor density,
V
=
STP
OP
OP
STP
STP
P
P
x
T
T
x
V
MW

= (257/ 22.4) kg/m
3
(273 / 541.21) K (0.12 / 1.0) bar
= 2.7269 kg / m
3






5-133

5.4.6.10 Step 10: Flow rate Calculation
The liquid and vapor mass flow rate above and below feed point is calculated and then used
to obtain Liquid-vapor flow factor, F
LV
.

Above the feed point:


1. liquid mass flow rate,L
m
= D(R+1)
= [(315.21 kmol/hr x 206.60 kg/kmol)/3600 s] x 3
= 13.20 kg/s

2. Vapor mass flow rate,V
m
= L
m
+ D
= 13.20 + 4.3993
= 17.60 kg/s

Below the feed point:

1. Liquid mass flow rate,L
n
= F + L
m

= 18.09 + 13.20
= 31.29 kg/ s
2. Vapor mass flow rate, V
m
= L
n
W
= 31.29 13.69
= 17.60 kg/ s
5-134

Table 5.4.14 below shows summary of physical properties at distillate and bottom
product.

Table 5.4.14: Physical properties at distillate and bottom product

Distillate product Bottom product
Temperature (K) 355.93 541.21
L, kg/m
3
912.74 881
V, kg/m
3
0.4840 2.7269
Liquid flowrate (kg/s) 13.20 31.29
Vapor flow rate (kg/s) 17.60 17.60


5.4.6.11 Step 11: Calculation of Column Diameter
This step is to determine the diameter of the column. Procedure in Column diameter
calculation is:
1. Determination of Liquid-vapor flow factor,F
LV
to get K
1
value
2. Based on K
1
value, flooding velocity can be obtained
3. Finding maximum volumetric flow rate, Q. From this, net area required, A
n
are
estimated
4. Column cross area sectional,
As
are determined
5. Then, column diameter can be obtained

The Liquid-vapor flow factor, F
LV
at top product can be obtained from equation 5.4.15:
F
Lv
=
5 . 0
|
|
.
|

\
|
|
.
|

\
|
L
V
Vw
Lw

= 0.0172










5-135

The Liquid-vapor flow factor, F
LV
at bottom product can be obtained from equation 5.4.15:
F
Lv
=
5 . 0
|
|
.
|

\
|
|
.
|

\
|
L
V
Vw
Lw

= 0.10

Take plate spacing as 0. 50 m, from figure 5.4.9 below (Sinnott, 2005), value for k
1
can be
obtained.

Figure 5.4.9: Flooding velocity, sieve plates
Hence, from figure 5.4.9 value of K
1
for top and bottom are:
Top k
1
= 0.09
Bottom k
1
= 0.07




5-136

Correction for surface tension:

k
1
= (o / 0.02)
0.2
(5.4.38)
which liquid surface tension is 0.02 N/m (Sinnott, 2005).

Value of surface tension,o at top is 0.03117 N/m and 0.0268 N/m at bottom. Surface tension
is obtained as calculation in Appendix A4. Hence:

1. = 0.098
2. = 0.075

Using equation 5.4.16, Flooding velocity,U
f
:

Therefore, for top and bottom pruduct, flooding velocity are:

1. = 4.2692 m/s

2. = 1.3514 m/s

Design for 85% flooding at maximum flow rate (Sinnott, 2005):
Therefore,
v
for top column,
= 3.6288 m/s

And for bottom column,
= 1.1487 m/s



5-137

Therefore, Maximum volumetric flow rate at top (Using equation 5.4.39):
Q =
V
V
m

(5.4.39)
Q =
4840 . 0
60 . 17
= 36.36 m
3
/s

And maximum volumetric flow rate at bottom (Using equation 5.4.40):
Q =
V
V
n

(5.4.40)
Q =
7269 . 2
60 . 17
= 6.45m
3
/s


In order to calculate the column diameter, an estimate of the net area is required:
Net area required, A
n
= Q / (5.4.41)
Using equation 5.4.41, for the top column:
A
n
=
6288 . 3
36 . 36
= 10.02 m
2

and for the bottom column:
A
n
=
1487 . 1
45 . 6
= 5.62 m
2




5-138

Column cross area sectional,
As
Take down comer area as 12% of total net area, A
n
(Sinnott, 2005).
The column cross area sectional,
As
for top column,
A
s
=
88 . 0
n
A
=
88 . 0
02 . 10

= 11.39 m
2
The column cross area sectional,
As
for bottom column,
A
s
=
88 . 0
n
A
=
88 . 0
62 . 5

= 6.38 m
2

Column diameter,D
c
Column diameter for the top column:

D
c
=
5 . 0
4
|
.
|

\
|
t
s
xA
=
5 . 0
39 . 11 4
|
.
|

\
|
t
x
= 3.807 m

Column diameter for the bottom column:

D
c
=
5 . 0
4
|
.
|

\
|
t
s
xA
=
5 . 0
38 . 6 4
|
.
|

\
|
t
x
= 2.8511 m

5.4.6.12 Step 12: Provisional Plate Design
For plate design, down comer area, active area, holes area, hole size and weir height are
calculated first. The equation involve for calculation are shown below (Sinnott, 2005):
(5.4.42)
Downcomer area: A
d
= 12% x A
C
(5.4.43)
Active area: A
a
= A
c
2 A
d
(5.4.44)
Hole area, take design 10% of active area: 10% x A
a
(5.4.45)
Weir length: l
W
= D
C
x (L
W
/ D
C
) (5.4.46)
5-139

5.4.6.12.1 Column Area, A
c
Column area is calculated using equation 5.4.42,
Column Area, A
c
=
4
2
D t

= (2.8511)
2
/ 4
= 6.38 m
2

5.4.6.12.2 Down-comer area, A
d
Take downcomer area as 12% of total. From equation 5.4.43
A
d
= 12% x A
C



= 0.12 x 6.38
=0.766 m
2

5.4.6.12.3 Net area,A
n

An = A
c
A
d

= 6.38 - 0.766
= 5.614 m
2

5.4.6.12.4 Active Area, A
a

Active area is calculated by using equation 5.4.44
A
a
= A
c
-2A
d

= 6.38 - 2(0.766)
= 4.851 m
2





5-140

5.4.6.12.5 Hole area, A
h

Using equation 5.4.45, take 10% of active area, A
a
(Sinnott, 2005):
A
h
= 10% x A
a

=0.1 x 4.851
= 0.4851 m
2

Figure 5.4.10: Relation between downcomer area and weir length

(A
d
/ A
c
) x 100% = (0.766/6.38) = 12 %
From the figure 5.4.10, I
w
/ D
c
= 0.77
From equation 5.4.46, weir length,
I
w
= 0.77 x 2.8511
= 2.195 m


5-141

Table 5.4.15 shows the summary for Provisional Plate Design.
Table 5.4.15: Summary of Provisional Plate Design
Column diameter,D
c
2.8511 m
Column area,A
c
6.38 m
2

Downcomer area,A
d
0.766 m
2

Net area,A
n
5.62 m
2

Active area ,A
a
4.85 m
2

hole area, A
h
0.485 m
2

Weir length,I
w
2.195 m

5.4.6.13 Step 13: Check Weeping
Take this values as suggested in Chemical Engineering Design volume 6,
Weir height, H
w
= 50mm
Hole diameter,d
h
= 5mm
Plate thickness = 5mm

Maximum liquid rate, L
w
= L
n
= 31.29 kg/s

Minimum liquid rate at 70% turn down (Sinnott, 2005),
L
w
= 0.7 x 31.29
= 21.90 kg/s







5-142

5.4.6.14 Step 14: Weir Liquid Crest
From equation 5.4.17, Maximum Weir Liquid Crest:
H
ow
=
3 / 2
750
|
|
.
|

\
|
w
w
LxI
L

=
3 / 2
195 . 2 881
29 . 31
750
|
.
|

\
|
X
= 47.31 mm liquid

and Minimum weir liquid crest:
H
ow
=
3 / 2
750
|
|
.
|

\
|
w
w
LxI
L

=
3 / 2
195 . 2 881
90 . 21
750
|
.
|

\
|
X
= 37.25 mm liquid

At minimum rate,
Clear liquid depth, H
w
+

H
ow
= 50 + 37.25 = 87.25 mm
Therefore, from figure 5.4.11, K
2
= 30.8
















Figure 5.4.11: Weep point Correlation (Eduljee, 1959)

5-143

5.4.6.15 Step 15: Weep Point
The vapor velocity at the weep point is calculated Using equation 5.4.18,
Minimum vapor velocity through the holes based on the holes area,U
h
:

U
h
=
5 . 0
2
) (
) 4 . 25 ( 9 . 0
v
dh
k

=
5 . 0
) 7269 . 2 (
) 0 . 5 4 . 25 ( 9 . 0
8 . 30


= 7.53 m/s


5.4.6.16 Step 16: Plate Pressure Drop














Figure 5.4.12: Relation between per cent perforated area with orifice (Sinnott, 2005)


5-144

For plate thickness/hole diameter = 5/5 =1
Percent perforated area = (A
h
/ A
a
) x 100 = (0.485/ 4.85) x 100 = 10
Therefore, from the figure 5.4.12, C
o
= 0.84
Maximum vapor velocity through holes, U
h
max =
h
A
Q
(5.4.47)
Therefore, U
h max
= 6.45 / 0.485
= 13.30 m/s

Dry plate drop, h
d

From equation 5.4.20:
h
d
=
|
|
.
|

\
|
|
|
.
|

\
|
L
v
C
U
o
h

2
51
=
|
.
|

\
|
|
.
|

\
|
881
7266 . 2
84 . 0
30 . 13
51
2

= 39.57 mm liquid

Residual Head, hr
From equation 5.4.21:
Residual head,h
r
=
|
|
.
|

\
|
L
3
10 5 . 12

=
|
|
.
|

\
|
881
10 5 . 12
3

=14.20 mm liquid

From equation 5.4.22:

Pressure drop per plate,h
T
= h
d
+( h
w
+

h
ow
) + h
r

= 39.57 + 87.25 + 14.20
= 141 mm liquid


5-145

5.4.6.17 Step 17: Downcomer Liquid Back Up
The height of the bottom edge of the apron above the plate,H
ap
From equation 5.4.25:
The height of the bottom edge of the apron above the plate,H
ap
= H
w
- 10
= 50 10
= 40 mm
= 0.04 m

Area under apron, A
ap
From equation 5.4.24:
Area under apron, A
ap
= h
ap
x I
w

= 0.04 x 2.195
= 0.0878 m
2

Since it is less than
d
A

which is 0.766m
2
, thus it can be used to calculate the head loss in
down comer

Head loss in down comer,h
dc

From equation 5.4.23:
Head loss by Cicalese et al (1947),
h
dC
=
2
166
|
|
.
|

\
|
ap
w
LxA
L

=
2
0878 . 0 881
29 . 31
166
|
.
|

\
|
x

= 27.14 mm
Using equation 5.4.26,
Back up in the downcomer,h
b
= h
w
+

h
ow
+ h
T
+ h
dc

= 87.25 + 141 + 27.14
= 255.39 mm liquid
= 0.255 m
5-146

Rule: hb < (plate spacing + weir height)
0.26 < 0.28
So tray spacing is acceptable

Downcomer Residence Time, t
r
From equation 5.4.27, Downcomer Residence Time, t
r
:


= 5.51 s

Therefore, t
r
is greater than 3.0 s which is recommended. So t
r
is satisfactory

5.4.6.18 Step 18:Check Entrainment by Fair
U
v
=
n
A
Q
=
62 . 5
453 . 6
= 1.149 m/s


From equation 5.4.28,
Flooding percentage =
F
V
u
U % 100
=
351 . 1
% 100 149 . 1
= 85 %













5-147

From Figure 5.4.13, Value = 0.035. Value is below 0.1 so the column diameter proposed
earlier is acceptable (Sinnott, 2005).

Figure 5.4.13: Entrainment correlation for sieve plates (Fair, 1961)










5-148

5.4.6.19 Step 19: Perforated Area














Figure 5.4.14: Relation between L
h
/D
c
vs L
w
/D
c


From figure 5.4.14,
At I
w
/D
c
= 0.77
u = 99
0

Angle subtended at plate = 180 99 = 81
o

Mean length, unperforated edge strips, L
s

Allowing 50 mm perforated strip round plate edge with 50 mm wide calming zone. From
equation (5.4.29):
Mean length,unperforated edge strips ,L
s
= (2.851 0.05) x t x
180
81

= 3.960 m

5-149

Area of unperforated edge strips, A
es

= (50 x 10
-3
) m x L
s
= (50 x 10
-3
) m x 3.960
= 0.198 m
2

Mean length of calming zones
Using equation 5.4.30:
Mean length of calming zones = 2.195 + 0.05
= 2.245 m

Area of calming zone, A
u

Area of calming zone, A
u
= 2 x 0.05 x (mean length of calming zones) (5.4.48)
= 2 x 0.05 x 2.245
= 0.225 m
2

Total area available for perforation, A
p

Using equation 5.4.31:
Total area available for perforation, A
p
= Active area (Area of unperforated edge + Area of
calming zones)
= 4.852 - 0.198 - 0.225
= 4.429 m
2




5-150

The ratio of A
h
/ A
p
= 0.485/ 4.429 = 0.11
From Figure 5.4.15,
I
p
/d
h
= 2.7
Acceptable because range between 2.5 4.0 (Sinnott, 2005)











Figure 4.12: Relation between Ah/ApvsIp/dh

Figure 5.4.15: Relation between A
h
/A
p
vs I
p
/d
h



5.4.6.20 Step 20: Number of Holes
Area of one hole,A
1
=
4
2
H
D t
=
4
) 005 . 0 (
2
t
= 1.9635 x 10
5
m
2
Number of holes =
1
A
A
h
=
5
10 9635 . 1
485 . 0

= 24,700 holes




5-151

5.4.6.21 Step 21: Column Height
Total Column Height = tray spacing x no. of stages
=0.5 x 21
=11 m
Actual height:
= 11 m + 2 m
= 13 m





















5-152

Table 5.4.16: Chemical Design Sheet for Distillation Column
Chemical Design Specification Sheet

Item No : T-102
Function: To obtaine higher conversion in reactor 101 and 102
Operating Condition

Operating Temperature (
o
C)
Feed
Distillate
Bottom


125.47
82.78
268.06

Operating Pressure (kPa)
Feed
Distillate
Bottom


13
11
12
Stream Flow rate (kg/s)
Feed Flow rate 18.09 Distillate flow rate 4.40
Bottom Flow rate 13.69
Column specifications
Design type
Material of construction
Tray
Number of trays
Reflux Ratio
Vertical
Stainless Steel
Sieve trays
21
5
Column Diameter (m)
Column Height (m)
Column cross sectional
area (m
2
)
tray spacing (m)
2.851
13
6.384

0.5
Sieve plate Design
Plate thickness (mm)
Downcomer area (m
2
)
Net Area (m
2
)
Active Area (m
2
)
5
0.77
5.62
4.85
Hole diameter (mm)
Hole area (m
2
)

5
0.49
Weep Point
Weir Design

Minimun vapor velocity (m/s)
Maximum liquid flowrate (kg/s)
Turn Down

7.53
31.29
70%

Weir length (m)
Weir Height (mm)
Weir liquid crest:
Maximum (mm liquid)
Minimum (mm liquid)

2.20
50

47.31
37.25


5-153

Hole Design Plate pressure Drop
Area of one hole (m
2
)
Number of holes
1.963 x 10
-5
24,711
Total Pressure drop (mm
liquid)
141
Plate layout Downcomer Liquid backup
Angle,
Mean length unperforated edge
strips (m)
Area of unperforated edge
strips (m
2
)
Mean length of calming
zone (m)
Area calming zone (m
2
)
Area of perforation (m
2
)
81
o
3.96

0.198

2.25

0.22
4.429
Height of liquid backup (m)
Area under apron (m
2
)
Residence time in
downcomer, t
r
(s)
0.04
0.0878
5.51


References:

Sinnott, R.K., 2005. Chemical Engineering Design. 4
th
edition: Vol 6. Coulson &
Richardsons.

Perry, R.H. and Green, D.W. (1997). Perrys Chemical Engineering Handbook. 7th ed.
USA: McGraw-Hill, Inc

Seader J.D & Henley E. (1998), Separation Process Principles, John Wiley & Sons, New
Jersey.