ET FI

mu É IEL ETILI .
Distillation is the process of separating components or substances from

ET a liquid mixture through the use of selective boiling and condensation.

Distillation may result in essentially complete separation (almost pure
components), or it may be a partial separation that increases the
concentration of the selected components in the mixture. In any case,
the process exploits the differences in the volatility of the components
of the mixture. In industrial chemistry , distillation is a unit operation
of virtually universal importance, but it is a process of physical
separation; not a chemical reaction.

Graphic method.
In graphical methods, the analysis of a distillation tower with a binary mixture is only
possible, and they are only useful for a very preliminary design. The most commonly used
graphical methods are McCabe-Thiele and Ponchon-Savarit. The first uses a mole fraction
diagram in liquid and vapor for the analysis, while the second uses an enthalpy versus
mole fraction diagram at constant pressure.

Numerical methods. (rigorous methods)

Rigorous methods involve mathematical property prediction models in their calculations
to obtain tower parameters. These types of methods are quite complex, such that they are
already programmed in process simulators. These methods are increasingly used given
their accuracy and also because, although the disadvantages are their laboriousness and
long calculation time, computer programs now leave these aspects aside. Some examples
of these methods are: bubble point, simultaneous correction and inside-out.
Rigorous methods for distillation tower design are very complex, since they integrate
rigorous property prediction models within the method. They are necessary for a final
design of the equipment since with them more precise data about the process is obtained.
These methods are composed of material, energy balances and equilibrium relationships
for each stage of the tower, and their complexity lies in the fact that these equations are
not linear and are strongly related to each other. In any case, these models are already
programmed in process simulators, so their use is relatively simple.

Fundamentals of the numerical and graphic

method
Numerical methods : there are a wide variety of interactive methods for solving nonlinear
algebraic equations. In general, they make use of the separation of equations together with
the trial equation and/or linearization using the Newton Raphson technique, described in
detail by Myers and Seider. A first attempt at resolution was to make step-by-step
calculations and the Lewis-Matheson and Thiele-Geddes equation, based on the trial and
error method for solving simple separators with one feed and two products. The Lewis-
Matheson method was widely used for manual calculations, but also showed numerical
instability when used in digital computers.
Graphical methods : Graphical methods are very widely used to visualize the
relationships between a set of variables and, therefore, generally used in chemical
engineering. It is useful in the design of staged contact systems because the calculation
procedures involve the simultaneous resolution of equilibrium relationships and material
and enthalpy balances.

Numerical methods
Short Fenske-Underwood-Gulliland (FUG) design method

Short methods are used to make an approximate design of a distillation tower.

These methods make such assumptions that significantly simplify the calculation of
tower parameters. However, this makes them less accurate, although their results are
quite acceptable, and their application is convenient if you do not have a rigorous
method programmed or if you do not have the time to program it.

The Fenske-Underwood-Gilliland (FUG) method is based mainly on four

equations, which predict the tower parameters by giving as data the distribution of key
components in the bottom and the
 tower dome, the reflux ratio in the tower, the feed plate and the pressure profile
in the tower.

The equations that make up the FUG method are the following:

• Fenske equation: calculates the minimum number of equilibrium stages needed in

the tower to achieve the desired separation. This number of stages corresponds
to a total reflux in the tower. To use this equation, the key components light
(represented by i) and heavy (represented by j) and their distribution in the first
(plate 1) and in the last plate (plate N+1) of the tower (counted from below)
must be indicated. upward), as well as the relative volatility between both
compounds, which is assumed constant throughout the tower. The Fenske
equation is as follows:

(xi , N + 1M x j1 )

log I x i ,1_l x
j,N+1I

Il A )\
N min log
a
i,j
The average relative volatility in the column between the light and heavy key components
(□ i,j) is calculated as the geometric mean of the relative volatilities between compounds i
and j in the first and last plate of the column, that is:

a i,j = [( a i,j ) N ( a i,j ) 1 ] 1/2 (Equation 2)

By solving equation 1, the distribution of non-key components ( i ) in the

distillate ( d i ) and in the bottoms ( b i ) can be obtained, as shown in equations
1a and 1b.

bi= fi
dn
min
HK
i,
b H.K.
HK

d HK
at N

b i,
H.K.

1 + d HK at N min

bHK i,
H.K.

• Underwood equation: with this equation (Equation 3) the minimum reflux in

the tower is determined, which corresponds to infinite equilibrium stages,
and even though this minimum reflux is an impossible condition to operate,
it serves as a reference to know from which reflux the tower can be
operated. To obtain it, the following equations must be solved
simultaneously:

1 - q =, V a i , HK x F , , i => f
q2
ai
, H.K. 0
 V min
->- i . HK x D , i

d
to
i , HK -0

• Empirical Gilliland correlation: it is used to calculate the tower stages given a

real reflux, which is a multiple of the minimum reflux obtained from the
Underwood equation. This correlation is expressed as:

N - N min = 1 - exp
1+ 54.4 X v X - 1
1
N + 111 + 117.2

L< ) )

Where:

R - R min
X=

This correlation is valid as long as the number of components is between 2

and 11, the pressure is maximum 40.82 atm, the thermal feed condition ( q )
has a value between 0.28 and 1.42, and the relative volatility is between 1.11
and 4.05. .

• Kirkbride equation: from this equation (Equation 7) the optimal feeding stage
is obtained assuming ideal distribution of the components throughout the
tower.

0.206

F , HK M x
N B,
rectification
N II x
L.K.

exhaustion _\ F , LK/ D , H.K.J ^ /

However, if the depletion zone has more stages than the rectification zone, this equation
loses accuracy, since the boilup ratio is not considered important in the tower design.
Rigorous distillation column design methods

Rigorous methods for the design of distillation towers are very complex, since they integrate
rigorous property prediction models within the method. They are necessary for a final design of the
 equipment since with them more precise data about the process is obtained. These methods are
composed of material, energy balances and equilibrium relationships for each stage of the tower, and
their complexity lies in the fact that these equations are not linear and are strongly related to each other.
In any case, these models are already programmed in process simulators, so their use is relatively
simple.

Graphic methods
ponchon method
Generally, graphical methods have been used to solve problems in binary systems (Ponchon and Savarit,
1921; McCabe-Thièle, 1925), although graphical procedures have also been developed for ternary systems,
such as that of Torres-Marchal (1981), and even for multicomponent systems (Hengstebeck, 1961).
While in the McCabe-Thièle graphical method certain simplifications are made (molar latent heats of
vaporization independent of the composition and negligible heats of mixing), the Ponchon and Savarit
graphical method rigorously relates all the variables that intervene in the calculation of a rectification
column, taking into account the variations in enthalpies depending on the composition, both of the liquid and
the vapor. For the application of the Ponchon and Savarit method to binary systems, the enthalpy-
composition diagram of the mixtures is needed, in a range of temperatures that covers the mixtures formed
by both phases, liquid and vapor, at the distillation pressure (which as first approximation, it will be assumed
constant throughout the entire column).
Such a diagram is depicted in Figure 4-1. The upper curve corresponds to the variation of the enthalpy of the
saturated vapor versus the composition, and separates in the diagram the areas corresponding to vapor (V)
and liquid plus vapor (L+V). For its part, the lower curve represents the variation in the enthalpy of the
saturated liquid as a function of the composition and delimits the regions of liquid plus vapor (L+V) and
liquid (L). The diagram in Figure 4-1 has also represented an equilibrium line or distribution line, which
joins points a and b, which respectively represent two phases in equilibrium (liquid and vapor). Thus, point c
of the L+V zone splits into two phases whose compositions and enthalpies will be given by the coordinates
of points a and b.
 Curve
of]

Crowd
ed

Mole Fractions

fífitUibriti enthalpy-composition I started a si.ffettia binOtia.

steam
crowded
Consider a general rectification column, such as that shown in Figure 4-2. This column consists of a total
condenser and a series of zones, within each of which the net flows of matter and enthalpy are constant, that
is, they are adiabatic zones without lateral currents of feed or extraction of matter or heat. Above each of
these zones, for example k, a series of feeds is introduced represented by A ssk ( ) = - å 0 1 of composition
zA(i,s) and enthalpy HA(s). The possibility of carrying out a series of product extractions is also considered,
represented by P ssk ( ) = - å 0 1 , of composition and enthalpy zP (i,s) and HP (s) respectively, heat inputs
(interboilers) represented by Q s A sk ( ) = - å 0 1 and heat removals (intercondensers) represented by Q s E
sk ( ) = - å 0 1 . In all cases, i refers to the component, k a specific sector or zone and s refers to the 4 zones
preceding k (so it can take values from 0 to k-1). Zone k has been explicitly represented in Figure 4-2.

MCCABE – THIELE METHOD.

Background: McCabe and Thiele have developed a graphical mathematical method to determine the number
of theoretical plates necessary for the separation of a binary mixture of A and B. This method employs
material balance with respect to certain parts of the column, producing operating lines and the yx
equilibrium curve for the system. The main assumption is that there must be an equimolar spill through the
column, between the feed inlet to the upper plate and the feed inlet to the lower plate. Liquid and vapor
streams enter a plate, establish their equilibrium, and leave it.
We perform a total material balance on plate “n” which results,
Vn+1 + Ln-1 = Vn + Ln
In this plate “n” a balance of component with respect to A (more volatile) gives,
Vn+1.Yn+1+Ln-1.Xn.Yn-1 = Vn .Yn + Ln .Xn
This figure shows us a continuous distillation column with feed F at an intermediate point, a distillate
product D that comes out from the top and a liquid product W that comes out from the bottom; The column
operates in a steady state.
D, Xp
We carry out a total material balance in the column,

And a global balance for component A,

F.xF = D.xD + W.xW
W,X
w

Thus, the slope is R/(R+1) or L n /V n+1 . xd

.
When x = 0, y
(0) R
□1
□

When x = y, x = x D (Starting point of the LOE).

We determine the theoretical stages starting at x D and staggering the first plate up to x 1 (See Figure 3).
So y 2 is the composition of the vapor passing through the liquid x 1 . We proceed similarly with the rest
of the theoretical plates which are stepped down the column in the enrichment section until reaching the
feed plate.

Exhaustion section.
It corresponds to the section of the column below the feed inlet and includes the feeding plate.
 Xw XN x
m

Figura 4. Exhaustion section. Figura 5. Construction of the

LOA.
To find its line of operation, we perform a total material balance and component A (Figure 4). A total
material balance results,
L m □ V m □1□ W
□ V m □1□ L m □W
And a material balance of component A,
V m □1. and m □1□ L m . x m or W . xW
When we solve for y m+1 we obtain the line of operation of the depletion section (LOA),
L W.
and m +
m xW

Since we assume equimolar flow, L m = L1 N= and V m+1 = V N .

The slope of the LOA is L m /V m+1 .
When x = 0: y (0)
□ □ W . xW
Vm
□1
When y = x: x = x W .
We determine the theoretical stages starting at x W and staggering the first plate up to x N (See Figure 5).
Then, and W is the composition of the vapor passing through the liquid x N. We proceed similarly with the
rest of the theoretical plates which are stepped up the column in the enrichment section until reaching the
feed plate.

Feeding conditions.
The conditions of the feed stream F entering the column determine the relationship between the vapor Vm in
the depletion section and the vapor Vn in the enrichment section, in the same way between the liquids Lm
and Ln.
For calculations, the feed conditions are presented with the quantity “q”, which is defined as the heat
necessary to vaporize 1 mole fed at the input conditions, divided by the molar latent heat of vaporization of
the feed.
In terms of enthalpy we can write the equation of q, q□ H V - hF HV - hL

Where, H V : feed enthalpy at dew point

h F : feed enthalpy at inlet conditions h L : feed enthalpy at bubble point

When Food is

q>1 Subcooled liquid (h F < h L )

q=1 Saturated liquid (h F = h L )
0<q<1 Mixture (h L < h F < H V )
q=0 Saturated vapor (h F = H V )
q<0 Superheated steam (h F > H V )
Table 1. Feeding conditions.

We also consider “q” as the ratio of the moles of saturated liquid in the feed plate to each mole fed to the
column.
In figure 6 we see how if the feed has part of steam, it will be added to V m and gives V n ; If the feed has
some liquid, it will be added to L n to give L m .
 So,
Lm = Ln + qF (YO)

Vn = Vm + (1-q)F (II)

Figure 6. Feeding area. We obtain the intersection point between the depletion and enrichment
trading lines on a y – x graph as follows:
a) We write the equations of the LOE and LOA omitting the subscripts of the plates,
LOE: (III)
Vn . and
□Ln.x□
D. xd

LOA: Vm . and □ L m . x □ W (IV)

. xW

Where “x” and “y” gives us the point of intersection of the two operation lines.
b) We subtract the equations (IV) – (III),
( V m □ Vn ) y □ ( L m □ L n ) x □ ( W + D . xD
. xW )

c) We substitute equations (I), (II) in equation (V) as well as the material balance for component A in
the column and rearrange,
From equation (I): (Lm – Ln) = qF
From equation (II): (Vm – Vn) = (q-1)F
From the balance of A in the column: Fx F = (Wx W + Dx D )
Substituting and solving for “y” gives us the line of operation of the feed,
( q 1) F. and q . F. x F. xF

qx
y= x- F

x
_(q-1) ( q - 1)
Slope: q/(q-1)
When y = x: x = x F
In Figure 7 we see the various slopes of this operating line depending on the feeding condition. (For your
understanding, see also Table 1).
 Figure 7. Construction of the power line q.

Calculation of reflux ratio .

- Total reflux.
The number of ideal plates needed depends on the lines of operation. To fix these lines, the reflux
relationship R = Lo/D, of the upper part of the column, is established.
One of the limiting values of this reflux ratio R is that of total reflux or R = ∞.
Since R depends on Lo, then Lo is very large which means that the slope of the LOE R/(R+1) is one (1) and
that the operating lines of both sections coincide with the 45º diagonal (line y = x).

Figure 8. Determination of the minimum number of stages

ideals (Nm).

We obtain the number of theoretical stages by staggering the plates from x D to x W (Figure 8). This
 This full reflux condition can be interpreted as a requirement for infinite condenser, reboiler, and column
diameter sizes for a given feed rate.

- Minimal reflux.
It is the reflux ratio (Rm) that will require an infinite number of plates for the desired separation of distillate
and residue. This corresponds to a minimum vapor flow in the column and therefore to minimum condenser
and reboiler sizes.

(to) (b)
Figure 9. Determination of minimum reflux.
Decreasing R the slope of the LOE decreases and the intersection of the LOE with the LOA approaches the
equilibrium curve. When this intersection touches the equilibrium line, a compressed point is produced
where the number of steps required becomes infinite (Figure 9(a)).
ET.......................................................................................................................................................................1
mu É IEL ETILI ................................................................................................................................................1
FI ET................................................................................................................................................................1
Graphic method.......................................................................................................................................1
Numerical methods. (rigorous methods).................................................................................................1
Numerical methods.........................................................................................................................................2
Short Fenske-Underwood-Gulliland (FUG) design method...................................................................2
Il A )\...........................................................................................................................................................3
d
to...................................................................................................................................................5
i , HK - 0..............................................................................................................................................5
1..................................................................................................................................................................5
L< ) )......................................................................................................................................................5
Graphic methods.............................................................................................................................................6
y= x - F..........................................................................................................................................11
_(q-1) x ( q - 1)...................................................................................................................................11
In some cases when the equilibrium line has an inflection (See Figure 9(b)), the operating line at minimum
reflux will be tangent to the equilibrium line, that is, the compressed point will be at the inflection point.

- Reflux operation.
The operating reflux ratio that we should actually apply is located between the total reflux limit (minimum
number of plates) and the minimum reflux limit (infinite number of plates). Selecting the appropriate value
of R requires a complete economic balance of the column's fixed costs and operating costs.
 It has been shown in many cases that to achieve this, the operating reflux ratio must be between 1.2
and 1.5, the minimum reflux value.

Advantages and disadvantages of graphical

methods
and numerical in distillation.
Method Advantage Westadvantage
Graphic Well, although we know
with this type of method we
can make calculations and It has certain problems, the
carry out designs using main one being the
programs for the simulation inaccuracy in the method as
of an industrial process, just well as the time it takes to
as they are useful for the complete it. If the 4
preliminary design of a components of the waterfall
distillation column. are not known, it is not
possible to graph the line of
operation. When the molar
heats of vaporization of all
mixtures are not constant and
if the other values are taken
into consideration in
comparison with the first, for
each mole of liquid that
evaporates, one mole of
vapor is not condensed,
therefore

which L and V do not remain

constant in the cascade.
 numeric They require many laborious
They are exact and are
calculations.
automated in programs so
their implementation is
quick.

In conclusion, we can say that the different methods are very extensive in terms of information
and likewise within those methods there are other methods which can help us have a good distillation
and likewise each method is intended for a certain type of distillation. Some are very long, others
shorter, the purpose of these is to make our work easier for us as chemical engineers and also for the
industry to save money and have good quality.

Bibliography:
file:///C:/Users/jesus/Downloads/destilacion_unlocked.pdf
https://rua.ua.es/dspace/bitstream/10045/10023/6/Reyes-Labarta-Juan-Antonio_5.pdf
https://reaccionesunefm.files.wordpress.com/2009/10/destilacic3b3n-fraccionada.pdf