cognizancejournal.

com
Adnan Hadžihasanović et al, Cognizance Journal of Multidisciplinary Studies, Vol.4, Issue.12, December 2024, pg. 381-394
(An Open Accessible, Multidisciplinary, Fully Refereed and Peer Reviewed Journal)
ISSN: 0976-7797
Impact Factor: 4.843
Index Copernicus Value (ICV) = 77.57

Optimization of the Process of Batch

Distillation of Binary Mixtures
Adnan Hadžihasanović1; Muhamed Bijedić2
¹Plamingo doo, Bosnia and Herzegovina
²Department of Chemical Engineering, Tehnološki fakultet u Tuzli, Bosnia and Herzegovina
1
[email protected]; 2 [email protected]

DOI: 10.47760/cognizance.2024.v04i12.036

Abstract: The method of dynamic optimization, known as maximum principle of Pontryagin, is presented. It
is applied to the process of batch rectification of binary mixtures, with aim of solving the problem of
minimum time (for given distillate quantity of specified composition) or maximum distillate quantity of
specified composition (for given batch time). The optimum profiles of reflux ratio are found for seven
different separation tasks, based on semi-rigorous model without holdups. The results obtained for distillate
quantity are compared to respective results obtained with constant reflux ratio and constant distillate
composition working policies, and also with the results of other authors. The solution algorithms for all the
three working policies are given. Computer implementation of the algorithms is performed in programming
language FORTRAN.

Keywords: Batch distillation, Mathematical modeling, Dynamic optimization, Maximum principle of Pontryagin.

I. INTRODUCTION
The optimization of batch distillation processes, particularly for binary mixtures, is a critical area of study in
chemical engineering, aiming to enhance efficiency and output quality. Batch distillation, unlike continuous
distillation, operates over a series of discrete steps, making it ideal for the production of high-purity specialty
chemicals in smaller quantities. The process involves heating a mixture until one component vaporizes more
than the other, then condensing the vapor to separate it from the mixture. The challenge lies in determining the
optimal conditions—such as the reflux ratio, heating rates, and the number of stages—for the separation to be
most effective. Advanced mathematical models and simulations have been developed to predict the behavior of
these systems, incorporating factors like relative volatility and mass balance. These models are essential for
industries that require rapid and robust methods for separating volatile organic compounds, as environmental
regulations become increasingly stringent. Researchers have employed various methodologies, including the
Fenske-Underwood-Gilliland (FUG) equations and sensitivity analysis, to design and optimize batch distillation
columns. Experimental validation of these models ensures their reliability and applicability in real-world
scenarios. The ultimate goal of this optimization is to achieve the desired degree of separation while minimizing
energy consumption and maximizing production efficiency. As computational power increases, so does the
ability to solve complex optimization problems, leading to more sophisticated and efficient batch distillation
operations. The continuous evolution of optimization techniques in batch distillation reflects the dynamic nature
of the chemical industry and its adaptation to market demands and environmental considerations.
The semi-rigorous model of multicomponent batch distillation [1] is based on the following assumptions:
relative volatilities are constant, molar holdups of the plates and of the condenser are neglected, vapor flowrate
is constant, heat balance is neglected. The model is composed of the total material balance and the component
material balance:

(1)

(2)

©2024, Cognizance Journal, ZAIN Publications, Fridhemsgatan 62, 112 46, Stockholm, Sweden, All Rights Reserved 381
 cognizancejournal.com
Adnan Hadžihasanović et al, Cognizance Journal of Multidisciplinary Studies, Vol.4, Issue.12, December 2024, pg. 381-394
(An Open Accessible, Multidisciplinary, Fully Refereed and Peer Reviewed Journal)
ISSN: 0976-7797
Impact Factor: 4.843
Index Copernicus Value (ICV) = 77.57

where: MN+1 is the holdup of the reboiler, t is the time, V is the vapor flowrate, R is the reflux ratio, xi,N+1 is the
fraction of the i-th component in the liquid phase in the reboiler, xi,0 is the fraction of the i-th component in the
liquid phase in the condenser.
The current composition of the liquid in the condenser is obtained by solving the following system of nonlinear
equations:

(3)
∑

where: yi,n is the fraction of the i-th component in the vapor phase on the n-th plate, Ki,n is the equilibrium ratio
of evaporation of the i-th component on the n-th plate, xi,n is the fraction of the i-th component in the liquid
phase on the n-th plate, xi,n-1 is the fraction of the i-th component in the liquid phase on the plate above the n-th
plate.
This model can be solved relatively easily in the case of a binary mixture. The equations for the calculation
plate-by-plate down the column are [2]:

( ) (4)

(5)

where: xn is the fraction of the more volatile component in the liquid phase on the n-th plate, yn is the fraction of
the more volatile component in the vapor phase on the n-th plate, α is the relative volatility of the more volatile
component, yn+1 is the fraction of the more volatile component in the vapor phase on the plate below the n-th
plate, y1 is the fraction of the more volatile component in the vapor phase on the first plate.
In this case, the fact that the plate-by-plate calculation must arrive at a given value of the composition in the
reboiler, xN+1, means that y1 must be assumed and then solved iteratively.
The following operating modes of batch distillation are known: (1) mode with a constant reflux ratio, (2)
mode with a constant distillate composition, and (3) mode with an optimal reflux ratio profile. The mode with a
constant reflux ratio is most often applied in practice, because its control is the simplest. However, at a constant
reflux ratio, the mole fraction of the more volatile component in the distillate decreases with time. Therefore, it
is necessary to find such a reflux ratio at which the average composition of the collected distillate will be equal
to the desired value. The mode with a constant distillate composition is less often applied in practice. This mode
of operation requires an expensive composition sensor and its control is more complicated. In order to maintain
the composition of the distillate at the desired value, it is necessary to continuously increase the reflux ratio,
because the liquid in the reboiler is depleted of more volatile component as the process progresses. The mode
with an optimal profile of the reflux ratio, in addition to achieving the desired composition of the distillate, also
aims to obtain its maximum amount in the given time of the process [3], [4], [5], [6] or the minimum duration of
the process for a given amount of distillate [2], [7], [8], [9], [10], [11]. This reflux ratio profile will generally lie
between the profiles of the previous two operating modes. Also, by combining the minimum time and maximum
distillate amount problems, a reflux ratio profile that maximizes profit can be found [12].
Converse and Gross solved the maximum distillate amount problem using dynamic programming [13], the
calculus of variations [14], and the maximum principle [15]. Murty et al. solved this problem using dynamic
programming [13], the maximum principle [15], the quasilinearization technique [16], the method of conjugate
gradients [17] and the method of sequential conjugate gradients [18]. Divekar et al. solved the problem using the
maximum principle [15], and Logsdon and Biegler using nonlinear programming [19], [20], [21], [22]. The
minimum time problem was solved by Coward, Robinson, Mayur and Jackson, and Hansen and Jorgensen using
the maximum principle [15].
We consider in this paper the problem of maximum distillate quantity of specified composition for given
batch time. The optimum profiles of reflux ratio are found for seven different separation tasks, based on semi-

©2024, Cognizance Journal, ZAIN Publications, Fridhemsgatan 62, 112 46, Stockholm, Sweden, All Rights Reserved 382
 cognizancejournal.com
Adnan Hadžihasanović et al, Cognizance Journal of Multidisciplinary Studies, Vol.4, Issue.12, December 2024, pg. 381-394
(An Open Accessible, Multidisciplinary, Fully Refereed and Peer Reviewed Journal)
ISSN: 0976-7797
Impact Factor: 4.843
Index Copernicus Value (ICV) = 77.57

rigorous model without holdups. The results obtained for distillate quantity are compared to respective results
obtained with constant reflux ratio and constant distillate composition working policies, and also with the results
of other authors [4]. We show that with optimal reflux ration mode we can achieve maximum distillate quantity.

II. MATERIALS AND METHODS

A. Maximum principle of Pontryagin
For a given system of differential equations
( ) (6)
with initial conditions
( ) (7)
and final conditions
( ) (8)
it is necessary to find a control scheme u = u(t) that transfers the system from conditions xi0 to conditions xi1 in
the shortest time. In order to find this solution, according to Pontryagin's maximum principle, it is necessary to
introduce additional variables p1, p2, ... , pn, where each is defined

( )
∑ (9)

It is also necessary to define the Hamiltonian function

( ) ∑ ( ) (10)

from which it follows:

(11)

(12)

According to Pontryagin's maximum principle, the Hamiltonian function reaches its maximum value at the
point u = u(t), where u is the value of the control parameter along the optimal path. In addition, the Hamiltonian
function is positive and constant in the time range t0 ≤ t ≤ t1. Applying these equations to batch distillation
allows solving the problem of optimal time or optimal amount of distillate (minimum time or maximum amount
of distillate). For binary batch distillation without holdup, the basic system of differential equations, which
corresponds to equation (6), is:

(13)

(14)

These equations make it possible to define the Hamiltonian function according to equation (10), as [7]
(15)
where two additional variables p1 and p2 can be defined according to equations (9) or (12) as [7]

(16)

©2024, Cognizance Journal, ZAIN Publications, Fridhemsgatan 62, 112 46, Stockholm, Sweden, All Rights Reserved 383
 cognizancejournal.com
Adnan Hadžihasanović et al, Cognizance Journal of Multidisciplinary Studies, Vol.4, Issue.12, December 2024, pg. 381-394
(An Open Accessible, Multidisciplinary, Fully Refereed and Peer Reviewed Journal)
ISSN: 0976-7797
Impact Factor: 4.843
Index Copernicus Value (ICV) = 77.57

(17)

provided V is constant throughout the process.

The optimal time problem requires that at the end of the distillation, the required amount of distillate M0 of
the required purity xi,0 is obtained in the minimum time. This is equivalent to fixing MN+1 and xi,N+1 in minimum
time, so the problem can be seen as finding the reflux ratio profile that transfers the system from state (M0N+1,
x0i,N+1) to state (M1N+1, x1i,N+1) in minimum time. In order to find the optimal reflux ratio profile, p1, p2, MN+1 and
xi,N+1 must be found as functions of time, so that [2]:
(a) the differential equations (13), (14), (16) and (17) are satisfied, along with the initial and final conditions
at MN+1 and xi,N+1;
(b) the value of the reflux ratio at each instant of time is such that it maximizes the Hamiltonian function at that
instant of time, whereby in this maximization p1, p2, MN+1 and xi,N+1 are considered constants;
(c) the maximum value of the Hamiltonian function in (b) is positive and constant throughout the time.
The procedure begins by assuming the values of the additional variables p1 and p2 at an initial time t = 0.
Equations (13), (14), (16) and (17) are then integrated with respect to time, ensuring that above condition (b) is
satisfied at each step. Since Hamilton's function is homogeneous in terms of p1 and p2, it is necessary to assume
their ratio only at t = 0, where it can be taken that p1(0) = 1. Since dMN+1/dt is always negative, xi,N+1 must
decrease to x1i,N+1 at some point in time. At that time MN+1 will generally not equal M1N+1, so the initial
assumption of the ratio of additional variables must be adjusted to achieve this equality. The minimum time for
batch distillation is the time corresponding to that value of p2(0) which simultaneously gives MN+1 = M1N+1 and
xi,N+1 = x1i,N+1 during integration. It is not necessary to pay special attention to condition (c), although checking
the constancy of H(p1, p2, MN+1, xi,N+1) is useful for the accuracy of the solution, so p2(0) should be assumed with
this condition in mind.
The procedure described above simultaneously solves the problem of the maximum amount of distillate,
because the given amount of distillate is the maximum for the calculated minimum time. A reflux ratio profile
that minimizes process time for a given amount of distillate simultaneously maximizes the amount of distillate
for that same process time.

B. Algorithms of solution for binary mixture

1) Operating mode with a constant reflux ratio:
( ) ( )

step 1

( ) ( )
step 2

( )
step 3

( )
( )
( )
( )

©2024, Cognizance Journal, ZAIN Publications, Fridhemsgatan 62, 112 46, Stockholm, Sweden, All Rights Reserved 384
 cognizancejournal.com
Adnan Hadžihasanović et al, Cognizance Journal of Multidisciplinary Studies, Vol.4, Issue.12, December 2024, pg. 381-394
(An Open Accessible, Multidisciplinary, Fully Refereed and Peer Reviewed Journal)
ISSN: 0976-7797
Impact Factor: 4.843
Index Copernicus Value (ICV) = 77.57

( ) ( ) ( )

( ) ( )
| | ( )
( ) ( )

( )

( ) ( )

( )

( ) ( ) ( ) ( )
( )
̅ ( ) ( )

( )
| ̅ | (if not go to step 1)

2) Operating mode with constant distillate composition:

( ) ( )

( )
step 1

( )

( ) ( )
step 2

( )

( )
( )
( )
( )
( )
( ) ( ) ( )
( ) ( )

( ) ( )
| | ( )

©2024, Cognizance Journal, ZAIN Publications, Fridhemsgatan 62, 112 46, Stockholm, Sweden, All Rights Reserved 385
 cognizancejournal.com
Adnan Hadžihasanović et al, Cognizance Journal of Multidisciplinary Studies, Vol.4, Issue.12, December 2024, pg. 381-394
(An Open Accessible, Multidisciplinary, Fully Refereed and Peer Reviewed Journal)
ISSN: 0976-7797
Impact Factor: 4.843
Index Copernicus Value (ICV) = 77.57

( ) ( )

( )

( ) ( )

( )

3) Operating mode with optimal reflux ratio profile

( ) ( ) ( )

step 1
( )

step 2

( )
step 3

( )
step 4

( )
( )
( )
( )
( )
( ) ( ) ( )
( ) ( )

( ) ( )
| | ( )

( ) ( )
( ) ( ) ( )
( ) ( ) ( )

( )

( ) ( ) ( )

( )

©2024, Cognizance Journal, ZAIN Publications, Fridhemsgatan 62, 112 46, Stockholm, Sweden, All Rights Reserved 386
 cognizancejournal.com
Adnan Hadžihasanović et al, Cognizance Journal of Multidisciplinary Studies, Vol.4, Issue.12, December 2024, pg. 381-394
(An Open Accessible, Multidisciplinary, Fully Refereed and Peer Reviewed Journal)
ISSN: 0976-7797
Impact Factor: 4.843
Index Copernicus Value (ICV) = 77.57

( ) ( )

( )
step 5

( )
( )
( )
( )
( )
( ) ( ) ( )
( ) ( )

( ) ( )
| | ( )

( ) ( )

( )

( ) ( )

( )

( )

( ) ( )

( ) ( )

( )
step 6

( )
( )
( )
( )
( )
( ) ( ) ( )
( ) ( )

( ) ( )
| | ( )

( )

( )
( )

( ) ( )

©2024, Cognizance Journal, ZAIN Publications, Fridhemsgatan 62, 112 46, Stockholm, Sweden, All Rights Reserved 387
 cognizancejournal.com
Adnan Hadžihasanović et al, Cognizance Journal of Multidisciplinary Studies, Vol.4, Issue.12, December 2024, pg. 381-394
(An Open Accessible, Multidisciplinary, Fully Refereed and Peer Reviewed Journal)
ISSN: 0976-7797
Impact Factor: 4.843
Index Copernicus Value (ICV) = 77.57

( )

( ) ( ) ( ) ( )
( )
̅ ( ) ( )

( )
| ̅ | ( )

III. RESULTS AND DISCUSSION

The computer implementation of the above algorithms is performed in the FORTRAN programming
language. For calculation of the variables by trial and error procedure (iteratively) the method of quadratic
interpolation is used [23]. Numerical integration is performed using the explicit Euler method with a fixed step
size of 0.01 h [24]. A cubic spline [25] is used for numerical interpolation and derivation.
The separation of a binary mixture whose components have relative volatility α1 = 2.0 and α2 = 1.0 in a batch
rectifier with N = 4 plates is considered. The vapor flowrate is fixed at V = 110 mol/h, and the total duration of
the process at T = 1 h. Table 1 gives seven different sets of input data, in terms of initial batch amount M(0)N+1,
initial batch composition x(0)N+1 and desired product purity x0*.
The amount of obtained distillate of the desired purity, in the given time of the process, at different operating
modes, is calculated. The obtained results are given in table 2. As can be seen from the table, the amount of
obtained distillate is the highest in the mode of operation with the optimal profile of the reflux ratio, for all
seven considered cases.
The results obtained for the operating mode with the optimal reflux ratio profile are also compared with the
results of other authors [4] who used the same model. As can be seen from table 3, the obtained results are in
good agreement with the results of Murty et al.
The obtained initial values of the additional variable p2(0), at which the distillate acquires the desired purity
*
x0 , are given in table 4. A large difference in the obtained values for different input data is evident.
The maximum values of the Hamiltonian function at the optimal reflux ratio are given in table 5, at the
beginning, middle and end of the process. As can be seen, these values are uniform and positive, which is in
agreement with Pontryagin's maximum principle. The small difference in values can be attributed to Euler's
method.
The obtained results are shown graphically in Figures 1 to 5, for case no. 1. Calculated reflux ratios as a
function of time, for three different operating modes, are given graphically in Figure 1. Current distillate
compositions as a function of time, for three different operating modes, are given graphically in Figure 2.
Dependences of the Hamiltonian function on the reflux ratio, for initial conditions and different values of p2(0)
and p1(0) = 1, are given graphically in Figure 3. Additional variables p1 and p2 as functions of time are given
graphically in Figure 4. Dependencies of the Hamiltonian function on the reflux ratio, at the beginning, in the
middle and at the end of the process, are given graphically in Figure 5.

IV. CONCLUSION
By analyzing the obtained results, the following conclusions can be drawn:
• The amount of distillate obtained is the highest at the operating mode with the optimal reflux ratio
profile, for the same initial conditions and the same restriction on the distillate composition;
• The results obtained using the semi-rigorous model are in good agreement with the corresponding
results of other authors, who used the same model;
• The maximum values of the Hamiltonian function at the optimal reflux ratio are uniform and positive,
which is in agreement with Pontryagin's maximum principle;
• With a semi-rigorous process model, which does not take into account either the holdup of the reflux
accumulator or that of the column plates, objective results of the dynamic optimization of batch distillation can
be obtained;

©2024, Cognizance Journal, ZAIN Publications, Fridhemsgatan 62, 112 46, Stockholm, Sweden, All Rights Reserved 388
 cognizancejournal.com
Adnan Hadžihasanović et al, Cognizance Journal of Multidisciplinary Studies, Vol.4, Issue.12, December 2024, pg. 381-394
(An Open Accessible, Multidisciplinary, Fully Refereed and Peer Reviewed Journal)
ISSN: 0976-7797
Impact Factor: 4.843
Index Copernicus Value (ICV) = 77.57

• Dynamic optimization of batch distillation, based on the maximum principle of Pontryagin, finds the
optimal solution even when other modes of operation give a result that is close to optimal;
• The use of the semi-rigorous model is limited to ideal systems and to the initial design of the column,
or to finding the initial operating conditions;
• The obtained results must be verified using a rigorous dynamic column model, accurate physical
properties of the mixture and rigorous integration methods.

REFERENCES
1. S. Domenech and M. Enjalbert, "Calcul du titre instantané du distillat d'une rectification discontinue de mélange complexe," The
Chemical Engineering Journal, vol. 15, no. 3, pp. 229-231, 1978/06/01/ 1978.
2. I. Coward, "The time-optimal problem in binary batch distillation," Chemical Engineering Science, vol. 22, no. 4, pp. 503-516, 1967.
3. A. O. Converse and G. D. Gross, "Optimal Distillate-Rate Policy in Batch Distillation," Industrial & Engineering Chemistry
Fundamentals, vol. 2, no. 3, pp. 217-221, 1963.
4. B. S. N. Murty, K. Gangiah, and A. Husain, "Performance of Various Methods in Computing Optimal Control Policies," The Chemical
Engineering Journal, vol. 19, no. 3, pp. 201-208, 1980.
5. U. M. Diwekar, R. K. Malik, and K. P. Madhavan, "Optimal reflux rate policy determination for multicomponent batch distillation
columns," Computers & Chemical Engineering, vol. 11, no. 6, pp. 629-637, 1987.
6. J. S. Logsdon, L. T. J. I. Biegler, and E. C. Research, "Accurate determination of optimal reflux policies for the maximum distillate
problem in batch distillation," vol. 32, pp. 692-700, 1993.
7. E. R. Robinson, "The optimisation of batch distillation operations," Chemical Engineering Science, vol. 24, no. 11, pp. 1661-1668,
1969/11/01/ 1969.
8. E. R. Robinson, "The optimal control of an industrial batch distillation column," Chemical Engineering Science, vol. 25, no. 6, pp.
921-928, 1970.
9. E. R. J. C. E. J. Robinson, "A note on the time optimal problem in binary batch distillation with a recycled waste cut," vol. 2, pp. 135-
136, 1971.
10. D. N. Mayur, R. A. May, and R. Jackson, "The time-optimal problem in binary batch distillation with a recycled waste-cut," The
Chemical Engineering Journal, vol. 1, no. 1, pp. 15-21, 1970.
11. T. T. Hansen and S. B. Jørgensen, "Optimal control of binary batch distillation in tray or packed columns," The Chemical Engineering
Journal, vol. 33, no. 3, pp. 151-155, 1986.
12. L. H. J. Kerkhof and H. J. M. Vissers, "On the profit of optimum control in batch distillation," Chemical Engineering Science, vol. 33,
no. 7, pp. 961-970, 1978/01/01/ 1978.
13. R. J. S. Bellman, "Dynamic Programming," vol. 153, pp. 34 - 37, 1957.
14. R. Weinstock, Calculus of variations, with applications to physics and engineering. New York (N.Y.) : McGraw-Hill, 1952.
15. ontriagin and D. E. Brown, The mathematical theory of optimal processes (International series of monographs in pure and
applied mathematics; v. 55). Oxford: Pergamon Press; [distributed in the Western Hemisphere by Macmillan, New York], 1964.
16. E. S. Lee, "Quasilinearization and invariant imbedding in optimization," AIChE Journal, vol. 14, no. 6, pp. 977-979, 1968/11/01 1968.
17. L. Lasdon, S. Mitter, and A. Waren, "The conjugate gradient method for optimal control problems," IEEE Transactions on Automatic
Control, vol. 12, no. 2, pp. 132-138, 1967.
18. J. C. Heideman and A. V. Levy, "Sequential conjugate-gradient-restoration algorithm for optimal control problems. Part 2. Examples,"
Journal of Optimization Theory and Applications, vol. 15, no. 2, pp. 223-244, 1975/02/01 1975.
19. K. R. Morison, "Optimal control of processes described by systems of differential-algebraic equations," Ph.D. , Imperial College,
University of London, 1984.
20. V. Vassiliadis, "Computational solution of dynamic optimization problems with general differential-algebraic constraints," PhD,
Imperial College London (University of London), 1993.
21. J. E. Cuthrell and L. T. J. A. J. Biegler, "On the optimization of differential-algebraic process systems," vol. 33, pp. 1257-1270, August
01, 1987 1987.
22. J. G. Renfro, A. M. Morshedi, and O. A. Asbjornsen, "Simultaneous optimization and solution of systems described by
differential/algebraic equations," Computers & Chemical Engineering, vol. 11, no. 5, pp. 503-517, 1987/10/01/ 1987.
23. D. E. Muller, "A Method for Solving Algebraic Equations Using an Automatic Computer," Mathematical Tables and Other Aids to
Computation, vol. 10, no. 56, pp. 208-215, 1956.
24. S. D. Conte and C. de Boor, Elementary Numerical Analysis publisher = {Society for Industrial and Applied Mathematics}. 1980.
25. C. de Boor, A Practical Guide to Spline. 1978.

APPENDIX
TABLE I.
INPUT DATA**
Case no. M(0)N+1, mol x(0)N+1, mol/mol x0*, mol/mol
1 116.1 0.715 0.950
2 74.3 0.420 0.900
3 116.1 0.580 0.930

©2024, Cognizance Journal, ZAIN Publications, Fridhemsgatan 62, 112 46, Stockholm, Sweden, All Rights Reserved 389
 cognizancejournal.com
Adnan Hadžihasanović et al, Cognizance Journal of Multidisciplinary Studies, Vol.4, Issue.12, December 2024, pg. 381-394
(An Open Accessible, Multidisciplinary, Fully Refereed and Peer Reviewed Journal)
ISSN: 0976-7797
Impact Factor: 4.843
Index Copernicus Value (ICV) = 77.57

4 116.1 0.487 0.930

5 74.3 0.715 0.950
6 116.1 0.715 0.930
7 116.1 0.715 0.900
** Relative volatility α = 2, number of plates N = 4, vapor flowrate V = 110 mol/h, process time T = 1 h.

TABLE II.
AMOUNT OF OBTAINED DISTILLATE AT DIFFERENT OPERATING MODES
Amount of distillate of required purity, mol

Case no. Optimum reflux

R = const. x0 = const.
ratio
1 26.131 26.062 25.846

2 6.136 6.132 5.820

3 16.380 16.351 16.125

4 4.106 4.105 4.018

5 22.963 22.856 22.289

6 37.985 37.786 37.661

7 51.411 50.925 51.163

TABLE III.
COMPARATIVE RESULTS
Amount of distillate of required purity, mol
Case no. Murty et al. This paper
1 25.979 26.131
2 6.082 6.136
3 16.397 16.380
4 4.379 4.106
5 22.827 22.963
6 38.752 37.985
7 46.204 51.411

TABLE IV.
VALUES OF P2(0) THAT SATISFY CONSTRAINT ON DISTILLATE PURITY
Case no. p2(0)
1 -547.99
2 -157.15
3 -351.41
4 -264.824
5 -340.15
6 -695.76
7 -1213.75

©2024, Cognizance Journal, ZAIN Publications, Fridhemsgatan 62, 112 46, Stockholm, Sweden, All Rights Reserved 390
 cognizancejournal.com
Adnan Hadžihasanović et al, Cognizance Journal of Multidisciplinary Studies, Vol.4, Issue.12, December 2024, pg. 381-394
(An Open Accessible, Multidisciplinary, Fully Refereed and Peer Reviewed Journal)
ISSN: 0976-7797
Impact Factor: 4.843
Index Copernicus Value (ICV) = 77.57

TABLE V.
MAXIMUM VALUES OF HAMILTONIAN FUNCTION AT OPTIMAL REFLUX RATIO
Hmax
Case
no. t = 0.0 h t = 0.5 h t = 1.0 h

1 3.74521 3.74898 3.75340

2 0.19755 0.19762 0.19771
3 1.34509 1.34591 1.34674
4 0.06288 0.06289 0.06290
5 2.82795 2.83439 2.84267
6 14.29615 14.31874 14.34854
7 62.77856 62.92354 63.14840

5.0

(a)

4.0
Reflux ratio

(b)

(c)

3.0

2.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Time, h
Fig. 1 Reflux ratio as a function of time for different operating modes:
(a) constant composition, (b) optimal reflux ratio and (c) constant reflux ratio.

©2024, Cognizance Journal, ZAIN Publications, Fridhemsgatan 62, 112 46, Stockholm, Sweden, All Rights Reserved 391
 cognizancejournal.com
Adnan Hadžihasanović et al, Cognizance Journal of Multidisciplinary Studies, Vol.4, Issue.12, December 2024, pg. 381-394
(An Open Accessible, Multidisciplinary, Fully Refereed and Peer Reviewed Journal)
ISSN: 0976-7797
Impact Factor: 4.843
Index Copernicus Value (ICV) = 77.57

0.958
0.957 (c)
0.956
Current distillate composition, mol/mol
0.955 (b)
0.954
0.953
0.952
0.951
(a)
0.950
0.949
0.948
0.947
0.946
0.945
0.944
0.943
0.942
0.941
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Time, h
Fig 2. Current distillate composition as a function of time for different operating modes:
(a) constant composition, (b) optimal reflux ratio and (c) constant reflux ratio.

3.7

3.6 P2(0) = - 545

P2(0) = - 544
3.5
P2(0) = - 543

3.4
Hamiltonian

3.3

3.2

3.1

3.0

2.9

2.8
2.0 2.5 3.0 3.5 4.0
Reflux ratio
Fig 3. Hamiltonian as a function of reflux ratio for
initial conditions and different values of p2(0) and p1(0) = 1.

©2024, Cognizance Journal, ZAIN Publications, Fridhemsgatan 62, 112 46, Stockholm, Sweden, All Rights Reserved 392
 cognizancejournal.com
Adnan Hadžihasanović et al, Cognizance Journal of Multidisciplinary Studies, Vol.4, Issue.12, December 2024, pg. 381-394
(An Open Accessible, Multidisciplinary, Fully Refereed and Peer Reviewed Journal)
ISSN: 0976-7797
Impact Factor: 4.843
Index Copernicus Value (ICV) = 77.57

-450 1.4

-475 1.3
P2

P1
-500 1.2

-525 1.1

-550 1.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Time, h

Fig 4. Additional variables p1 and p2 as functions of time.

4.0

(a)
3.5

(b)
Hamiltonian

3.0
(c)

2.5

2.0
2.0 2.5 3.0 3.5 4.0
Reflux ratio

Fig 5. Hamiltonian as a function of reflux ratio at: (a) t = 0.0 h, (b) t = 0.5 h and (c) t = 1.0 h.

©2024, Cognizance Journal, ZAIN Publications, Fridhemsgatan 62, 112 46, Stockholm, Sweden, All Rights Reserved 393
 cognizancejournal.com
Adnan Hadžihasanović et al, Cognizance Journal of Multidisciplinary Studies, Vol.4, Issue.12, December 2024, pg. 381-394
(An Open Accessible, Multidisciplinary, Fully Refereed and Peer Reviewed Journal)
ISSN: 0976-7797
Impact Factor: 4.843
Index Copernicus Value (ICV) = 77.57

NOMENCLATURE
H Hamiltonian function
K equilibrium ratio of evaporation
M holdup, mol
N number of plates in column
NC number of components in mixture
p additional variable
R reflux ratio
t current time, h
T total time, h
V vapor flowrate, mol/h
x fraction of component in liquid phase, mol/mol
y fraction of component in vapor phase, mol/mol

Greek letters
α relative volatility
ε accuracy

Indexes
i counter of component in mixture (i = 1, ..., C; i = 1, ..., NC)
j counter of plate in column (j = 0, ..., N+1; 0 – condenser, N+1 – reboiler)
n counter of plate in column (n = 0, ..., N+1; 0 – condenser, N+1 – reboiler)

Exponents
j counter of reflux ratio in k-th step of time
J total number of reflux ratios
k counter of step in time
K total number of steps in time

©2024, Cognizance Journal, ZAIN Publications, Fridhemsgatan 62, 112 46, Stockholm, Sweden, All Rights Reserved 394