01/12/2009

1
Multicomponent distillation Multicomponent distillation
F. Grisafi
01/12/2009
2
Introduction Introduction
The problem of determining the stage and reflux
i t f lti t di till ti i h requirements for multicomponent distillations is much more
complex than for binary mixtures.
With a multicomponent mixture, fixing one component
d l d h h
p g p
composition does not uniquely determine the other
component compositions and the stage temperature.
Also when the feed contains more than two components it is Also when the feed conta ns more than two components t s
not possible to specify the complete composition of the top
and bottom products independently.
The separation between the top and bottom products is The separation between the top and bottom products is
usually specified by setting limits on two "key components",
between which it is desired to make the separation.
01/12/2009
3
Calculation procedure Calculation procedure
The normal procedure for a typical problem is to solve the MESH
(Material balance, Equilibrium, Summation and Heat) balance
equations stage-by-stage, from the top and bottom of the column
toward the feed point.
For such a calculation to be exact, the compositions obtained from For such a calculation to be exact, the compositions obtained from
both the bottom-up and top-down calculations must mesh at the
feed point and mesh the feed composition.
The calculated compositions will depend on the compositions The calculated compositions will depend on the compositions
assumed for the top and bottom products at the commencement of
the calculations.
Th h i i ibl h h k h h Though it is possible to match the key components, the other
components will not match unless the designer was particularly
fortunate in choosing the trial top and bottom compositions.
01/12/2009
4
Calculation procedure Calculation procedure
For a completely rigorous solution the compositions must be p y p
adjusted and the calculations repeated until a satisfactory match
at the feed point is obtained by iterative trial-and-error
calculations.
Clearly, the greater the number of components, the more difficult
the problem.
For other than ideal mixtures the calculations will be further For other than ideal mixtures, the calculations will be further
complicated by the fact that the component volatilities will be
functions of the unknown stage compositions.
f h f d b If more than a few stages are required, stage-by-stage
calculations are complex and tedious.
01/12/2009
5
Short-cut" methods Short cut methods
Before the advent of the modern digital computer, various "short- p
cut" methods were developed to simplify the task of designing
multicomponent columns.
Though computer programs will normally be available for the
rigorous solution of the MESH equations, short-cut methods are
still useful in the preliminary design work, and as an aid in defining
problems for computer solution.
Intelligent use of the short-cut methods can reduce the computer
time and costs.
01/12/2009
6
Short-cut" methods Short cut methods
The short-cut methods available can be divided into two classes:
1. Simplifications of the rigorous stage-by-stage procedures to
enable the calculations to be done using hand calculators, or
graphically. g p y
Typical examples of this approach are the methods given by
Hengstebeck (1961), and the Smith-Brinkley method (1960); which
are described in Section 11.7 (C&R Vol. VI).
2. Empirical methods, which are based on the performance of
operating columns, or the results of rigorous designs.
Typical examples of these methods are Gilliland's correlation, yp p ,
which is given in (C&R Vol. II, Chapter 11) and the Erbar-Maddox
correlation given in Section 11.7.3 (C&R Vol. VI).
01/12/2009
7
Key components Key components
The designer must select the two "key components between which it is
desired to make the separation.
The light key will be the component that it is desired to keep out of the The light key will be the component that it is desired to keep out of the
bottom product, and the heavy key the component to be kept out of the
top product.
Specifications will be set on the maximum concentrations of the keys in
the top and bottom products the top and bottom products.
The keys are known as "adjacent keys" if they are "adjacent" in a listing of
the components in order of volatility, and "split keys" if some other
component lies between them in the order; they will usually be adjacent. p y y j
If any uncertainty exists in identifying keys components (e.g. isomers),
trial calculations should be made using different components as the keys to
determine the pair that requires the largest number of stages for
separation (the worst case). p ( w ).
The "non-key" components that appear in both top and bottom products
are known as "distributed" components; and those that are not present, to
any significant extent, in one or other product, are known as "non-
distributed" components distributed components.
01/12/2009
8
Number and sequencing of columns Number and sequencing of columns
In multicomponent distillations it is not possible to obtain more than one
pure component one sharp separation in a single column pure component, one sharp separation, in a single column.
If a multicomponent feed is to be split into two or more virtually pure
products, several columns will be needed.
Impure products can be taken off as side streams; and the removal of a Impure products can be taken off as side streams; and the removal of a
side stream from a stage where a minor component is concentrated will
reduce the concentration of that component in the main product.
For separation of N components, with one essentially pure component taken
overhead or from the bottom of each column (N 1) columns will be overhead, or from the bottom of each column, (N 1) columns will be
needed to obtain complete separation of all components.
For example, to separate a mixture of benzene, toluene and xylene two
columns are needed (3-1), Benzene is taken overhead from the first column
and the bottom product, essentially free of benzene, is fed to the second
column.
This column separates the toluene and xylene.
01/12/2009
9
Costs considerations
The order in which the components are separated will determine p p
the capital and operating costs.
Where there are several components the number of possible
sequences can be very large; for example with five components the sequences can be very large; for example, with five components the
number is 14, whereas with ten components it is near 5000.
When designing systems that require the separation of several
components efficient procedures are needed to determine the components, efficient procedures are needed to determine the
optimum sequence of separation.
Separation schemes
f for a 4 components
mixture
01/12/2009
10
Heuristic rules for optimum sequencing Heuristic rules for optimum sequencing
1. Remove the components one at a time.
2. Remove any components that are present in large
excess early in the sequence.
3. With difficult separations, involving close boiling
components, postpone the most difficult separation to
late in the sequence.
Difficult separations will require many stages, so to
reduce cost, the column diameter should be made a
small as possible. As the column diameter is dependent small as poss ble. As the column d ameter s dependent
on flow-rate, the further down the sequence the smaller
will be the amount of material that the column has to
handle.
01/12/2009
11
Tall and vacuum columns Tall and vacuum columns
Where a large number of stages is required, it may be g f g q , y
necessary to split a column into two or more separate
columns to reduce the height of the column, even
though the required separation could theoretically though the required separation could, theoretically,
have been obtained in a single column.
This may also be done in vacuum distillations, to reduce
the column pressure drop and limit the bottom
temperatures temperatures.
01/12/2009
12
Short-cuts methods Short cuts methods
01/12/2009
13
Short-cut methods for stage and reflux
i t requirements
Some of the more useful short-cut procedures which p
can be used to estimate stage and reflux requirements
without the aid of computers are given in this section.
M t f th h t t th d d l d f th Most of the short-cut methods were developed for the
design of separation columns for hydrocarbon systems
in the petroleum and petrochemical systems industries,
d ti t b i d h l i th t and caution must be exercised when applying them to
other systems (as it is assumed almost ideal behavior of
mixtures).
They usually depend on the assumption of constant
relative volatility, and should not be used for severely
non ideal systems non-ideal systems.
01/12/2009
14
Pseudo-binary systems Pseudo binary systems
If the presence of the other components does not significantly
affect the volatility of the key components the keys can be affect the volatility of the key components, the keys can be
treated as a pseudo-binary pair.
The number of stages can then be calculated using a McCabe-
Thiele diagram or the other methods developed for binary Thiele diagram, or the other methods developed for binary
systems.
This simplification can often be made when the amount of the non-
key components is small or where the components form near-ideal key components is small, or where the components form near-ideal
mixtures.
Where the concentration of the non-keys is small, say less than
10% they can be lumped in with the key components 10%, they can be lumped in with the key components.
For higher concentrations the method proposed by Hengstebeck
(1946) can be used to reduce the system to an equivalent binary
system system.
01/12/2009
15
Hengstebeck's method Hengstebeck s method
For any component i the Lewis-Sorel material balance
equations and equilibrium relationship can be written in q q p
terms of the individual component molar flow rates; in
place of the component composition:
Stripping section Rectifying section Stripping section Rectifying section
01/12/2009
16
Hengstebeck's method Hengstebeck s method
To reduce a multicomponent system to an equivalent p y q
binary it is necessary to estimate the flow-rate of the
key components throughout the column.
Hengstebeck considers that in a typical distillation the
flow-rates of each of the light non-key components
approaches a constant limiting rate in the rectifying approaches a constant, limiting, rate in the rectifying
section; and the flows of each of the heavy non-key
components approach limiting flow-rates in the stripping
ti section.
01/12/2009
17
Hengstebeck's method Hengstebeck s method
Putting the flow-rates of the non-keys equal to these
li iti t i h ti bl th bi d limiting rates in each section enables the combined
flows of the key components to be estimated.
Stripping section Rectifying section
h i i
lighter
Stripping section Rectifying section
heavier species
lighter
species
01/12/2009
18
Hengstebeck's method Hengstebeck s method
The method used to estimate the limiting flow-rates is
th t d b J (1939) Th ti that proposed by Jenny (1939). The equations are:
Stripping section Rectifying section
d
i
and b
i
= corresponding top and bottom flow rate of component i.
01/12/2009
19
Hengstebeck's method Hengstebeck s method
Estimates of the flows of the combined keys enable operating lines to be
drawn for the equivalent binary system drawn for the equivalent binary system.
The equilibrium line is drawn by assuming a constant relative volatility for
the light key:
where y and x refer to the vapor and liquid concentrations of the light key.
Hengstebeck shows how the method can be extended to deal with
it ti h th l ti l tilit t b t k t t d situations where the relative volatility cannot be taken as constant, and
how to allow for variations in the liquid and vapor molar flow rates.
He also gives a more rigorous graphical procedure based on the Lewis- g g g p p
Matheson method (see Section 11.8).
01/12/2009
20
Bubble and Dew Point calculations Bubble and Dew Point calculations
K
i
=F(T) [De Priester charts]
Bubble point temperature (x known): Bubble point temperature (x
i
known):
Dew point temperature (y known):
1 1
( ) 1 ( ) ( ) 1
c c
i i i i
i i
K T x f T K T x
= =
= =

Dew point temperature (y
i
known):
1 1
1 ( ) 1
( ) ( )
c c
i i
i i
i i
y y
f T
K T K T
= =
= =

Solution: find the temperature T by trial-and-error procedure as f(T)=0
01/12/2009
21
Example 11.5 Example 11.5
Estimate the number of ideal stages needed in the butane-pentane
splitter defined by the compositions given in the table below.
The column will operate at a pressure of 8 3 bar with a reflux ratio The column will operate at a pressure of 8.3 bar, with a reflux ratio
R=2.5. The feed is at its boiling point (q=1).
Assumed composition
(for DP & BP calculations)
x
d
x
b
C
3
0.111 0.000
iC 0 333 0 000
(for DP & BP calculations)
iC
4
0.333 0.000
nC
4
0.533 0.018
iC
5
0.022 0.345
nC
5
0.000 0.636
Note: a similar problem has been solved by Lyster et al (1959) using a
nC
5
0.000 0.636
Note: a similar problem has been solved by Lyster et al. (1959) using a
rigorous computer method and it was found that 10 stages were needed.
01/12/2009
22
Example 11.5 - solution
The top and bottom temperatures (dew points and bubble points)
were calculated by the methods illustrated in Example 11.9.
Relative volatilities are given by equation 8.30:
Equilibrium constants were taken from the De Priester charts.
Relative volatilities estimated:
Li ht k
Light non-key comp.
Light non-key comp.
Heavy non-key comp.
01/12/2009
23
Example 11.5 - solution Example 11.5 solution
Stripping section Rectifying section
Calculations of
non-key flows:
Stripping
section
Rectifying
section section
01/12/2009
24
Example 11.5 - solution Example 11.5 solution
S i i i R if i i
Flow of combined keys:
Stripping section Rectifying section
L=RD
V=(R+1)D
V=V-(1-q)F
L=L+qF L =L+qF
R= reflux ratio; q= feed thermal index (1 for boiling liquid; 0 for saturated vapour)
01/12/2009
25
Example 11.5 - solution Example 11.5 solution
Equilibrium curve Equilibrium curve
Equilibrium points
01/12/2009
26
Example 11.5 - solution
The McCabe-Thiele diagram is shown in Figure: 12 stages required; g g g q
feed on seventh from base.
01/12/2009
27
Empirical correlations methods
The two most frequently used empirical methods for estimating the q y p g
stage requirements for multicomponent distillations are the
correlations published by Gilliland (1940) and by Erbar and Maddox
(1961).
These relate the number of ideal stages required for a given
separation, at a given reflux ratio, to the number at total reflux
(minimum possible) and the minimum reflux ratio (infinite number ( p ) (
of stages).
Gilliland's correlation is given in C-R Vol. 2, Chapter 11.
The Erbar-Maddox correlation is given in this section, as it is now
generally considered to give more reliable predictions.
Their correlation is shown in Figure 11.11; which gives the ratio of he r correlat on s shown n F gure . ; wh ch g ves the rat o of
number of stages required to the number at total reflux, as a
function of the reflux ratio, with the minimum reflux ratio as a
parameter. p
To use Figure 11.11, estimates of the number of stages at total
reflux and the minimum reflux ratio are needed.
01/12/2009
28
Number of stages calculation
Fig.11.11
Erbar-Maddox Erbar Maddox
correlation
(1961)
R/(R+1)=0.6
Example:
R
m
=1.33
R=1.5
R
m
/(R
m
+1)=0.57
R/(R+1)=0.6
N /N 0 34
N
m
/N=0.34
N
m
/N=0.34
01/12/2009
29
Number of stages calculation Number of stages calculation
Gilliland correlation (1940), C&R II Vol.
Log-log plot
Linear plot
( | | + | | + + 1 4 54 1 n n R R
(

|
.
|

\
|
+
+
|
.
|

\
|
+ +
+ +
=
+

5 . 0
1
2 . 117 11
4 . 54 1
exp 1
1 n
n n
m
1 +

= +
R
R R
m
where:
01/12/2009
30
Minimum number of stages (Fenske Equation) M n mum num r of stag s (F ns Equat on)
The Fenske equation (Fenske, 1932) can be used to estimate the
minimum stages required at total reflux.
The derivation of this equation for a binary system is given in C R The derivation of this equation for a binary system is given in C-R
Vol. 2, Chapter 11. The equation applies equally to multicomponent
systems and can be written as:
01/12/2009
31
Minimum number of stages (Fenske Equation) M n mum num r of stag s (F ns Equat on)
Normally the separation required will be specified in terms of the
key components, and equation 11.57 can be rearranged to give an
estimate of the number of stages.
where o
LK
is the average relative volatility of the light key with
respect to the heavy key, and x
LK
and x
HK
are the light and heavy key
concentrations.
Th l ti l tilit i t k th t i f th l The relative volatility is taken as the geometric mean of the values
at the column top and bottom temperatures.
To calculate these temperatures initial estimates of the
compositions must be made so the calculation of the minimum compositions must be made, so the calculation of the minimum
number of stages by the Fenske equation is a trial-and- error
procedure.
01/12/2009
32
Large relative volatilities and feed location Larg r at o at t s an f ocat on
If there is a wide difference between the relative volatilities at
the top and bottom of the column the use of the average value in
the Fenske equation will underestimate the number of stages the Fenske equation will underestimate the number of stages.
In these circumstances, a better estimate can be made by
calculating the number of stages in the rectifying and stripping
ti t l sections separately.
The feed concentration is taken as the base concentration for the
rectifying section and as the top concentration for the stripping rectifying section and as the top concentration for the stripping
section, and estimating the average relative volatilities separately
for each section.
Thi d ill l i ti t f th f d i t l ti This procedure will also give an estimate of the feed point location
cannot be taken as constant.
01/12/2009
33
Components split estimate Components split estimate
If the number of stages is known, equation 11.57 can be used
to estimate the split of components between the top and
bottom of the column at total reflux bottom of the column at total reflux.
It can be written in a more convenient form for calculating
the split of components:
Fenske Equation
where d
i
and b
i
are the flow-rates of the component i in the tops
and bottoms, d
r
and b
r
are the flow-rates of the reference (light
k ) t i th t d b tt key) component in the tops and bottoms.
Note: from the column material balance:
d + b = f d
i
+ b
i
= f
i
where f
i
is the flow rate of component i in the feed.
01/12/2009
34
Minimum reflux ratio Underwood equation
Colburn (1941) and Underwood (1948) have derived equations for Colburn (1941) and Underwood (1948) have derived equations for
estimating the minimum reflux ratio for multicomponent
distillations. These equations are discussed in C-R Vol. 2, Chapt. 11.
As the Underwood equation is more widely used it is presented in
thi ti Th ti b t t d i th f this section. The equation can be stated in the form:
1
LK
u o < <
The value of u must lie between the values of the relative volatility of the light and The value of u must lie between the values of the relative volatility of the light and
heavy keys, and is found by trial and error.
01/12/2009
35
Minimum reflux ratio Minimum reflux ratio
In the derivation of equations 11.60 and 11.61 the
l ti l tiliti t k t t relative volatilities are taken as constant.
The geometric average of values estimated at the top
and bottom temperatures should be used This requires and bottom temperatures should be used. This requires
an estimate of the top and bottom compositions.
Though the compositions should strictly be those at
minim m fl x th l s d t min d t t t l fl x minimum reflux, the values determined at total reflux,
from the Fenske equation, can be used.
A better estimate can be obtained by replacing the A better est mate can be obta ned by replac ng the
number of stages at total reflux in equation 11.59 by an
estimate of the actual number; a value equal to N
m
/0.6
is often used.
01/12/2009
36
Feed-point location
A limitation of the Erbar Maddox and similar empirical methods is A limitation of the Erbar-Maddox, and similar empirical methods, is
that they do not give the feed-point location.
An estimate can be made by using the Fenske equation to calculate
the number of stages in the rectifying and stripping sections
t l b t thi i ti t f th f d i t separately, but this requires an estimate of the feed-point
temperature.
An alternative approach is to use the empirical equation given by
Kirkbride (1944): ( )
01/12/2009
37
Distribution of non-key components
(graphical method) (graphical method)
The graphical procedure proposed by Hengstebeck (1946), which is
based on the Fenske equation, is a convenient method for based on the Fenske equation, is a convenient method for
estimating the distribution of components between the top and
bottom products.
Hengstebeck and Geddes (1958) have shown that the Fenske
equation can be written in the form: equation can be written in the form:
Specifying the split of the key components determines the
constants A and C in the equation.
h d b f h h b d l The distribution of the other components can be readily
determined by plotting the distribution of the keys against their
relative volatility on log-log paper, and drawing a straight line
through these two points through these two points.
01/12/2009
38
Example 11.6: Geddes-Hengstebeck method
Use the Geddes-Hengstebeck method to check the component
distributions for the separation specified in Example 11 5 distributions for the separation specified in Example 11.5
01/12/2009
39
Example 11.6: solution Example 11.6 solution
The average volatilities will be taken as those estimated in Example
11 5 Normally the volatilities are estimated at the feed bubble 11.5. Normally, the volatilities are estimated at the feed bubble
point, which gives a rough indication of the average column
temperatures. The dew point of the tops and bubble point of the
bottoms can be calculated once the component distributions have bottoms can be calculated once the component d str but ons have
been estimated, and the calculations repeated with a new estimate
of the average relative volatilities, as necessary.
The distribution of the non-keys are read from Figure.
As these values are close to those assumed for the calculation of As these values are close to those assumed for the calculation of
the dew points and bubble points, there is no need to repeat with
new estimates of the relative volatilities.
01/12/2009
40
Component distribution
o
Light key d
i
/b
i
=24, o
LK
= 2
Heavy key d
i
/b
i
=0.053, o
HK
= 1
iC
4
The distribution of the non-keys are read from
LK
C
3
y
Figure at the appropriate relative volatility and
the component flows are calculated from the
following equations:
d
i
/
b
i
d
i
/
b
i
Overall column material balance d
i
+b
i
=f
i
d d
From which
HK
nC
5
LK
o
01/12/2009
41
Example 11.7 - Erbar-Maddox method
For the separation specified in Example 11.5, evaluate the effect
of changes in reflux ratio on the number of stages required.
Solution
The relative volatilities estimated in Example 11 5 and the The relative volatilities estimated in Example 11.5, and the
component distributions calculated in Example 11.6 will be used for
this example.
01/12/2009
42
Example 11.7 - solution Example 11.7 solution
Minimum number of stages; Fenske equation, equation 11.58:
Minimum reflux ratio; Underwood equations 11.60 and 11.61.
As the feed is at its boiling point q = 1
Trial and error
calculation
01/12/2009
43
Example 11.7
solution solution
Underwood equation (11.60)
u= 1.35

Case of R=2.0:
From E-M diagram (Fig. 11.11):

01/12/2009
44
Example 11.7 - solution Example 11.7 solution
01/12/2009
45
Example 11.8 Feed point estimation p p
Estimate the position of the feed point for the separation
considered in Example 11.7, for a reflux ratio of 3.
Solution
Use the Kirkbride
ti (11 62) equation (11.62):
Product distributions
taken from Example 11 6: taken from Example 11.6:

01/12/2009
46
Summary of empiric methods Summary of empiric methods
Data: feed composition and thermal condition, operating pressure.
1 Choose LK and HK and fix their distribution on top and bottom 1. Choose LK and HK and fix their distribution on top and bottom
product
2. Estimate overall top and bottom flow rate compositions (assume
li ht k t i t d h k t i light non-key components in top and heavy non-key component in
bottom as first attempt)
3. Estimate dew and bubble points on top, bottom and feed
(K
i
taken from De Priest chart).
01/12/2009
47
Summary of empiric method - cnt
4. Estimate relative volatility with respect to the HK
of all comps. at the T
dew
(top) and T
boiling
(bottom) :
and calculate average volatilities o= (o *o )
0 5
and calculate average volatilities o
i
= (o
i,top
*o
i,bot
)
0.5
5. Recalculate overall top and bottom flow rate compositions on
th b i f th H t b k d G dd ti the basis of the Hengstebeck and Geddes equation:
by plotting on log-log diagrams knew points relevant to LK & HK
(assume first trial relative volatilities)
Please note that A=log(d
HK
/b
HK
) and C=[log(d
LK
/b
LK
) A]/logo
LK
Check first attempt composition and go back to step 3 to update
T
d
and T
b ili
and recalculate relative volatilities if necessary T
dew
and T
boiling
and recalculate relative volatilities if necessary.
01/12/2009
48
Summary of empiric method - cnt
6. Estimate the minimum number of stages for the separation N
m
F k Fenske eq.:
7. Estimate the minimum reflux ratio R
m
U d d Underwood eq.:
where u is the root of eq :
q = (L L )/F
where u is the root of eq.:
Please note that 1 < u < o
LK
q = (L
R
-L
S
)/F
01/12/2009
49
Summary of empiric method - cnt
8. Choose an operative reflux ratio R and calculate the number of
(ideal) stages N with the
a) Erbar-Maddox diagram
or
b) with Gilliland Correlation:
(

|
.
|

\
|
+
+
|
.
|

\
|
+ +
+ +
=
+

5 . 0
1
2 . 117 11
4 . 54 1
exp 1
1 N
N N
m
1 +

= +
R
R R
m
9. Calculate the feed stage position
Ki kb id Kirkbride eq.:
01/12/2009
50
De Priester charts - values for hydrocarbons
High temperature Low temperature
01/12/2009
51
Exact stage by stage
calculation methods
01/12/2009
52
Example 11.9 Stage by stage calculation
This example illustrates the complexity and trial and error nature
of stage-by-stage calculation.
The same problem specification has been used in earlier examples
to illustrate the shortcut design methods to illustrate the shortcut design methods.
A butane-pentane splitter is to operate at 8.3 bar with the
following feed composition:
LK
HK
Specification: not more than 1 mol of the light key in the bottom product
d h 1 l f h h k i h d fl i and not more than 1 mol of the heavy key in the top product, reflux ratio
of 2.5. Make a stage-by-stage calculation to determine the product
composition and number of stages required.
01/12/2009
53
Example 11.9 solution
Only sufficient trial calculations will be made to illustrate the
method used.
B i 100 l f d Basis 100 mol feed.
Estimation of dew and bubble points:
The K values, taken from the De
Priester charts are plotted in Priester charts, are plotted in
Figure for easy interpolation.
01/12/2009
54
Example 11.9 solution Example 11.9 solution
To estimate the dew and bubble points, assume that nothing
heavier than the heavy key appears in the tops and nothing lighter heavier than the heavy key appears in the tops, and nothing lighter
than the light key in the bottoms.
Bubble-point calculation bottoms
Bottom & top composition
(first attempt assumption)
Bubble-point calculation, bottoms
D p int l l ti n t p B bbl p int l l ti n f d (li id) Dew-point calculation, tops Bubble-point calculation, feed (liquid)
01/12/2009
55
Example 11.9 solution
Top down calculations, assume total condensation with no sub-cooling
y
1
= x
d
x
0
It is necessary to estimate the composition of the "non-keys" so
that they can be included in the stage calculations.
First trial top composition:
In each stage calculation it will necessary to estimate the stage
temperatures to determine the K values and liquid and vapor temperatures to determine the K values and liquid and vapor
enthalpies. The temperature range from top to bottom of the column
will be approximately 120 60 = 60C. An approximate calculation
(Example 11 7) has shown that around 14 ideal stages will be needed; (Example 11.7) has shown that around 14 ideal stages will be needed;
so the temperature change from stage to stage can be expected to
be around 4 to 5C.
01/12/2009
56
Example 11.9 Stage 1 Example 11.9 Stage 1
Estimation of stage temperature and outlet liquid composition (x
1i
),
Imp s ilib i m b t n V nd L : Impose equilibrium between V
1
and L
1
:
01/12/2009
57
Example 11.9 Stage 1 Example 11.9 Stage 1
Summary of stage equations:
Enthalpy [kJ/kmol] diagram of liquid
and vapor hydrocarbons
01/12/2009
58
Example 11.9 Stage 1 Example 11.9 Stage 1
Before a heat balance can be made to estimate L
1
and V
2
an estimate
of y
1
and T
1
is needed. V
2
is dependent on the liquid and vapor flows,
so as a first trial assume that these are constant and equal to L
0
and
V
1
; then, from equations (i) and (ii):
Approximate mass
balance over the stage
01/12/2009
59
Example 11.9 Stage 1 Example 11.9 Stage 1
Enthalpy data from figure:
01/12/2009
60
Example 11.9 Stage 1 Example 11.9 Stage 1
01/12/2009
61
Example 11.9 Stage 1
Could revise calculated values for y but L /V is close enough to Could revise calculated values for y
2
but L
1
/V
2
is close enough to
assumed value of 0.71, so there would be no significant difference
from first estimate.
01/12/2009
62
Example 11.9 Stage 2
Estimation of stage temperature and
outlet liquid composition (x ) outlet liquid composition (x
2
).
01/12/2009
63
Example 11.9 Stage 2
01/12/2009
64
Example 11.9 Stages 3-4
As the calculated liquid and vapor flows are not changing much from stage
to stage the calculation will be continued with the value of L/V taken as
constant at 0.7.
3 Stage 3
Stage 4
01/12/2009
65
Example 11.9 Stages 5-6
Stage 5
Stage 6
01/12/2009
66
Example 11.9 Stage 7
Stage 7 composition
Feed composition
01/12/2009
67
Example 11.9 Bottomup calc.
To illustrate the procedure the calculation will be shown for the
reboiler and bottom stage, assuming constant molar overflow.
With th f d t it b ili i t d t t l fl th With the feed at its boiling point and constant molar overflow the
base flows can be calculated as follows:
It will be necessary to estimate the It will be necessary to estimate the
concentration of the non-key components in
the bottom product; as a first trial take:
01/12/2009
68
Example 11.9 reboiler
Material balance
01/12/2009
69
Example 11.9 Stage 1 (from bottom)
The calculation is continued stage-by-stage up the column to the
feed point (stage 7 from the top) feed point (stage 7 from the top).
If the vapor composition at the feed point does not mesh with the
top-down calculation, the assumed concentration of the non-keys in p , m f y
the bottom product is adjusted and the calculations repeated.
01/12/2009
70
Rigorous computer methods Rigorous computer methods
The application of digital computers has made the rigorous solution
of the MESH equations a practical proposition and computer of the MESH equations a practical proposition, and computer
methods for the design of multicomponent separation columns will
be available in most design organizations.
A considerable amount of work has been done over the past twenty A considerable amount of work has been done over the past twenty
or so years to develop efficient and reliable computer-aided design
procedures for distillation and other staged processes.
Several different approaches have been taken to develop programs
th t ffi i t i th f t ti d it bl f th that are efficient in the use of computer time, and suitable for the
full range of multicomponent separation processes that are used in
the process industries.
A design group will use those methods that are best suited to the A design group will use those methods that are best suited to the
processes that it normally handles.
In this section a brief outline will be given of the methods that
have been developed. p
01/12/2009
71
Rigorous solution procedure: basic steps Rigorous solution procedure basic steps
The basic steps in any rigorous solution procedure will be:
1. Specification of the problem; complete specification is essential
for computer methods.
2. Selection of values for the iteration variables; for example,
estimated stage temperatures, and liquid and vapour flows (the
column temperature and flow profiles).
3. A calculation procedure for the solution of the stage equations.
4. A procedure for the selection of new values for the iteration
variables for each set of trial calculations variables for each set of trial calculations.
5. A procedure to test for convergence; to check if a satisfactory
solution has been achieved.
01/12/2009
72
Available computer methods Available computer methods
It is convenient to consider the methods available
under the following four headings:
1 L i M th th d 1. Lewis-Matheson method.
2 Thiele Geddes method 2. Thiele-Geddes method.
3. Relaxation methods. 3. Relaxat on methods.
4. Linear algebra methods.
01/12/2009
73
Rating and design methods Rating and design methods
With the exception of the Lewis-Matheson method, all the
methods listed above require the specification of the number of methods listed above require the specification of the number of
stages below and above the feed point.
They are therefore not directly applicable to design: where the
designer wants to determine the number of stages required for a
ifi d ti specified separation.
They are strictly what are referred to as "rating methods"; used
to determine the performance of existing, or specified, columns.
Given the number of stages they can be used to determine product
compositions.
Iterative procedures are necessary to apply rating methods to the
design of new columns design of new columns.
An initial estimate of the number of stages can be made using
short-cut methods and the programs used to calculate the product
compositions; repeating the calculations with revised estimates till compositions; repeating the calculations with revised estimates till
a satisfactory design is obtained.
01/12/2009
74
Lewis-Matheson method Lewis Matheson method
The method proposed by Lewis and Matheson m p p y
(1932) is essentially the application of the
Lewis-Sorel method (Section 11.5.1) to the
solution of multicomponent problems solution of multicomponent problems.
Constant molar overflow is assumed and the Constant molar overflow is assumed and the
material balance and equilibrium relationship
equations are solved stage by stage starting at g y g g
the top or bottom of the column (refer to the
Example 11.9 C&R Vol. VI).
01/12/2009
75
Lewis-Matheson method Lewis Matheson method
To define a problem for the Lewis-Matheson method the
f ll i i bl t b ifi d d t i d f following variables must be specified, or determined from
other specified variables:
1. Feed composition, flow rate and condition.
2. Distribution of the key components.
3. One product flow.
4. Reflux ratio.
5. Column pressure.
6. Assumed values for the distribution of the non-key
components components.
01/12/2009
76
Lewis-Matheson method Lewis Matheson method
The usual procedure is to start the calculation at the top and
b tt f th l d d t d th f d i t 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
h l l d f h d b
p p
the compositions calculated from the top and bottom starts
mesh, and match the feed at the feed point.
Efficient procedures for adjusting the compositions to Eff c ent procedures for adjust ng the compos t ons to
achieve a satisfactory mesh at the feed point are given by
Hengstebeck (1961).
Good descriptions of the Lewis-Matheson method with Good descriptions of the Lewis Matheson method, with
examples of manual calculations, are also given in the books
by Oliver (1966) and Smith (1963); a simple example is given
in C&R Vol. 2, Chapter 11.
01/12/2009
77
Lewis-Matheson method Lewis Matheson method
In some computer applications of the method, where the
assumption of constant molar overflow is not made, it is convenient
to start the calculations by assuming flow and temperature to start the calculations by assuming flow and temperature
profiles.
The stage component compositions can then be readily determined
and used to revise the profiles for the next iteration. p
In general, the Lewis-Matheson method has not been found to be
an efficient procedure for computer solutions, other than for
relatively straightforward problems.
It is not suitable for problems involving multiple feeds, and side-
streams, or where more than one column is needed.
The method is suitable for interactive programs run on
bl l l t d P l C t programmable calculators and Personal Computers.
As the calculations are carried out one stage at a time, only a
relatively small computer memory is needed.
01/12/2009
78
Lewis-Matheson method: details
(C&R Vol 2) (C&R Vol. 2)
In a binary system, the relationship between the composition of the vapour
y
A
and of the liquid x
A
in equilibrium may also be expressed in a way, which is
particularly useful in distillation calculations. part cularly useful n d st llat on calculat ons.
If the ratio of the partial pressure to the mole fraction in the liquid is
defined as the volatility, then:
volatility of A =P
A
/x
A
and volatility of B = P
B
/x
B
volatility of A =P
A
/x
A
and volatility of B = P
B
/x
B
The ratio of these two volatilities is known as the relative volatility o given
by (ideal systems):
o
B
= P x
B
/P
B
x = P /P
B
o
AB
= P
A
x
B
/P
B
x
A
= P
A
/P
B
Being P the total pressure, substituting Py
A
for P
A
, and Py
B
for P
B
:
o
AB
= Py
A
x
B
/Py
B
x
A
= y
A
x
B
/y
B
x
A
y
A
/y
B
= o
AB
x
A
/x
B
The extension of this concept to multicomponent (ideal) systems with
components A B C D leads to: components A, B, C, D, leads to:
y
A
/y
B
= o
AB
x
A
/x
B
; y
C
/y
B
= o
CB
x
C
/x
B
; y
D
/y
B
= o
DB
x
D
/x
B
;
01/12/2009
79
Lewis-Matheson method: details Lewis Matheson method details
If a mixture of components A, B, C, D, and so on has mole
fractions x
A
, x
B
, x
C
, x
D
, and so on in the liquid and y
A
, y
B
, y
C
, y
D
,
and so on in the vapour, then:
1 ... = + + + +
D C B A
y y y y (1)
B i iB B i
y x y y o =
B B D B C B B B A
y y y y y y y y y 1 ... = + + + +
1 + + + +
(2)
(3)
(4)
as
b t
B B
N
i
i iB
y x x =

=1
o
B B D DB B C CB B B BB B A AB
y x x x x x x x x 1 ... = + + + + o o o o (4)
(5)
subst.:

=
=
N
i
i iB B B
x x y
1
o
N N N
(6)
expliciting:
; ; ;
1 1 1

= = =
= = =
i
i iD D D
i
i iC C C
i
i iA A A
x x y x x y x x y o o o (7)
similarly:
01/12/2009
80
Lewis-Matheson method: example Lewis Matheson method example
A mixture of ortho, meta, and para-mononitrotoluenes containing
60 % mol ortho-mononitrotoluene 60 % mol ortho mononitrotoluene
4% mol meta-mononitrotoluene
36% mol para-mononitrotoluene 6 mo para monon troto u n
Is to be continuously distilled to give a top product of x
do
=98% mol
ortho, and the bottom is to contain x
wo
=12.5% mol ortho.
The mixture is to be distilled at a bottom temperature of T
B
=410K
requiring a pressure in the boiler of about P=6.0 kN/m
2
.
If a reflux ratio of R=5 is used, how many ideal plates will be
i d d h ill b h i i i f h
y p
required and what will be the approximate compositions of the
product streams?
The volatility of ortho relative to the para isomer may be taken as
1 70 d f th t 1 16 th t t o
op
=1.70 and of the meta as o
om
=1.16 over the temperature range
of 380415 K.
01/12/2009
81
Column scheme Column scheme
V
Specs:
x
fo
, = 0.4
x
fm
, =0.04
0 36
D, x
do
, x
dm
, x
dp
V
m
x
fp
=0.36
x
do
=0.98
x
wo
=0.125
F
,
do
,
dm
,
dp
L
m
P = 6 kN/m
2
T
B
= 410 K
R =L
m
/D = 5
x
fo
, x
fm
, x
fp
W x
wo
x
wm
x
wp
V
n L
n
W, x
wo
, x
wm
, x
wp
01/12/2009
82
Lewis-Matheson method: example
As a first estimate, it is supposed that the distillate contains 0.6%
Solution
meta and 1.4% para.
A material balance then gives the composition of the bottoms.
For 100 kmol of feed with D and W kmol of product and bottoms,
respectively and x
do
and x
wo
the mole fraction of the ortho in the
distillate and bottoms, then an overall material balance gives: g
100 = D + W
An ortho balance gives:
60 = Dx + Wx 60 = Dx
do
+ Wx
wo
and:
60 = (100 W)0.98 + 0.125W
f h h from which:
D = 55.56 kmol and W = 44.44. kmol
01/12/2009
83
Lewis-Matheson method: example
The compositions and amounts of the streams are then be obtained by
the overall mass balance as follows:
Equations of operating lines
The liquid and vapour streams in the column are obtained (assuming
that McCabe and Thiele conditions hold) as follows: that McCabe and Thiele conditions hold) as follows
Above the feed-point:
Liquid downflow L
n
= 5D = 277.8 kmol
Vapour up V
n
= 6D = 333.4 kmol
01/12/2009
84
Lewis-Matheson method: example
Equations of operating lines
Below the feed point assuming the feed is liquid at its boiling point then: Below the feed-point, assuming the feed is liquid at its boiling point then:
Liquid downflow L
m
= L
n
+ F = (277.8 + 100) = 377.8 kmol
Vapour up V
m
= L
m
W = (377.8 44.44) = 333.4 kmol
The equations for the operating lines may then be written as: q p g y
below the feed plate:
h
(8)
ortho:
meta:
(i)
meta:
para:
(i)
01/12/2009
85
Lewis-Matheson method: example
Equations of operating lines
Above the feed plate: Above the feed plate:
(9)
ortho:
meta:
para:
(ii)
01/12/2009
86
Lewis-Matheson method: example
Composition of liquid on first plate
The temperature of distillation is fixed by safety considerations at 410 K
and from a knowledge of the vapour pressures of the three components and, from a knowledge of the vapour pressures of the three components,
the pressure in the still is found to be about 6 kN/m
2
.
The composition of the vapour in the still is found from the relation The composition of the vapour in the still is found from the relation
y
so
= o
o
x
so
/Eo
s
x
s
(eqns. 6-7)
The liquid composition on the first plate is then found from equation (i).
As example and for ortho:
0.191 = (1.133x
1
0.0166) x
1
= 0.183
The values of the other compositions are found in this way (see following p y g
tables).
01/12/2009
87
Lewis-Matheson method: example
01/12/2009
88
Lewis-Matheson method: example
01/12/2009
89
Lewis-Matheson method: example Lewis Matheson method example
The liquid on plate 7 has a composition with the q p p
ratio of the concentrations of ortho and para about
that in the feed, and the feed will therefore be
introduced on this plate introduced on this plate.
Above this plate the same method is used but the
operating equations are equation (ii).
The vapour from the 16
th
plate has the required
concentration of the ortho isomer and the values concentration of the ortho isomer, and the values
the meta and para are sufficiently near to take this
as showing that 16 ideal plates will be required.
01/12/2009
90
Thiele-Geddes method Thiele Geddes method
Like the Lewis-Matheson method, the original method of Thiele
and Geddes (1933) was developed for manual calculation. It has
subsequently been adapted by many workers for computer subsequently been adapted by many workers for computer
applications.
The variables specified in the basic method, or that must be
derived from other specified variables are: derived from other specified variables, are:
1. Reflux temperature.
2 Reflux flow rate 2. Reflux flow rate.
3. Distillate rate.
4 Feed flows and condition 4. Feed flows and condition.
5. Column pressure.
6. Number of equilibrium stages above and below the feed point. 6. Num r of qu r um stag s a o an ow th f po nt.
01/12/2009
91
Thiele-Geddes procedure Thiele Geddes procedure
The method starts with an assumption of the column
temperature and flow profiles. p p
The stage equations are then solved to determine the
stage component compositions and the results used to
revise the temperature profiles for subsequent trial revise the temperature profiles for subsequent trial
calculations.
Efficient convergence procedures have been developed
f h Thi l G dd h d
g p p
for the Thiele-Geddes method.
The so-called "theta method", described by Lyster etal.
(1959) and Holland (1963) is recommended (1959) and Holland (1963), is recommended.
The Thiele-Geddes method can be used for the solution
of complex distillation problems, and for other multi-
t ti component separation processes.
01/12/2009
92
Relaxation methods Relaxation methods
With the exception of this method, all the methods described solve
the stage equations for the steady state design conditions the stage equations for the steady-state design conditions.
In an operating column other conditions will exist at start-up, and
the column will approach the "design" steady-state conditions after
a period of time. a period of time.
The stage material balance equations can be written in a finite
difference form, and procedures for the solution of these
equations will model the unsteady-state behaviour of the column.
Relaxation methods are not competitive with the "steady-state"
methods in the use of computer time, because of slow convergence.
However, because they model the actual operation of the column,
convergence should be achieved for all practical problems convergence should be achieved for all practical problems.
The method has the potential of development for the study of the
transient behaviour of column designs, and for the analysis and
design of batch distillation columns. g
01/12/2009
93
Linear algebra methods Linear algebra methods
The Lewis-Matheson and Thiele-Geddes methods use a stage-by-
stage procedure to solve the equations relating the component stage procedure to solve the equations relating the component
compositions to the column temperature and flow profiles.
However, the development of high-speed digital computers with
large memories makes possible the simultaneous solution of the
complete set of MESH equations that describe the stage
compositions throughout the column.
If the equilibrium relationships and flow-rates are known (or
assumed) the set of material balance equations for each component
is linear in the component compositions is linear in the component compositions.
Whit the aim of a numerical method these equations are solved
simultaneously and the results used to provide improved estimates
f h d fl fil of the temperature and flow profiles.
01/12/2009
94
Linear algebra methods Linear algebra methods
The set of equations can be expressed in matrix form and solved p
using the standard inversion routines available in modern computer
systems. Convergence can usually be achieved after a few
iterations.
It is possible to include and couple to the distillation program, some
thermodynamic method for estimation of the liquid-vapour
relationships (activity coefficients) as the UNIFAC method (see relationships (activity coefficients) as the UNIFAC method (see
Chapter 8, Section 16.3).
This makes the program particularly useful for the design of
l f h l d f h columns for new processes, where experimental data for the
equilibrium relationships are unlikely to be available.