You are on page 1of 11

石 油 学 会 誌 J. Japan Petrol. Inst.

, 26, (3), 163-173 (1983) 163

Calculations of Multicomponent Distillations

Mitsuho HIRATA*

The methods of calculation developed by the author and co-workersfor multicomponent


distillationsystems, especiallyfor nonideal systems,are reviewed. First, the degree of freedom
of multicomponent distillation is presented for a conventional distillation column and for a
system of extractive distillation columns. The method of group relaxation is described for a
complex distillation column with multi-feeds and multi-side streams. The method is further
applied to a system of extractive distillation columns. Finally, researches made by the author
and co-workerson the vapor-liquid equilibria, which are indispensable for calculations of dis-
tillation systems, are briefly reviewed.

1. Introduction

Distillation process is one of the most important


separation processes and it is widely used in many
chemical and petroleum industries. Distillation is
a unit operation to separate a liquid mixture into
its individual components. Generally, a mixture
contains a variety of components, therefore, mul-
ticomponent distillation is more of practical im-

portance.
In this paper, the researches made by the author
and co-workers on the multicomponent distillation
systems, especially on nonideal systems, are re-
viewed.

2. Degree of Freedom of Distillation4)

When distillation calculations are to be carried


out by use of a computer, we must consider how
many variables should be input. Namely, the num-
ber of variables, which can be set before and can
be assigned at start of the calculation, is called
the degree of freedom of distillation. Fig. 1 A Conventional Distillation Column

Concerning the researches on the degree of free-


dom of a distillation column, several reports are number of all variables involved and subtracting
found in literature, but all of them took the concept the number of all relations therefrom.
of the phase rule into consideration, and the pro- 2.1 Degree of Freedom of a Conventional
cedure for developing the relevant relations became Column
complicated and in some cases even misleading. Consider a conventional distillation column as
Moreover, no research has been made on the degree shown in Fig. 1. The column is assumed to be
of freedom for the system of extractive distillation adiabatic, that is, there are no heat losses from
columns. the walles of the column and pipeline.
In the following, a direct method is shown for 2.1.1 Number of Variables (Number of Com-
determining the degree of freedom by counting the ponents=m)

Received Dec. 25, 1982.


* Faculty of Technology, Tokyo Metropolitan University (Fukasawa 2-1-1, Setagaya-ku, Tokyo 158)

石 油 学 会 誌 J. Japan Petrol. Inst., Vol. 26, No. 3, 1983


164

Total number of variables=25+5m+(8+2m) (n1+n2)=NV

2.1.2 Number of Relations

石油学会誌 J. Japan Petrol. Inst., Vol. 26, No. 3, 1983


165

Total number of relations=15+4m+(7+2m)(n1+n2)=NR

2.1.3 Degree of Freedom The remaining degree of freedom becomes:


The degree of freedom of a conventional distilla- (10+m+n1+n2)-(m+2)-(n1+n2+4)=4.
tion column The simplest and, usually; the best way to fulfill
=NV-NR-25+5m+(8+2m)(n1+n2)-15-4m 4 degrees of freedom is to give the number of plates
-(7+2m)(n1+n2)=10+m+n1+n2 of each section of column(n1 and n2), reflux ratio
The rate, composition, temperature, and pressure (R), and flow rate of distillate (D).
of a feed should be defined, so 1+(m-1)+1+ 2.2 Degree of Freedom of a System of Ex-
1=m+2 should be reduced from the degree of tractive Distillation Columns
freedom. Further, the pressure of every part of Consider a system of extractive distillation col-
the column should be given, so n1+n2 for every umns as shown in Fig. 2. The first column is an
plate, 1 for distillate, 1 for reflux, 1 for bottoms, extractive distillation column and the second is
and 1 for reboiler vapor should be subtracted. a solvent recovery column.

Fig. 2 A System of Extractive Distillation Columns

石 油 学 会 誌 J. Japan Petrol. Inst., Vol. 26, No. 3, 1983


166

2.2.1 Number of Variables (Number of Components=m)

Subtotal number of variables=14+2m+(n1+n2+n3)(2m+8)=N1V

Subtotal number of variables=13+2m+(n4+n5)(2m+8)=N2V

石 油 学 会 誌 J. Japan Petrol. Inst., Vol. 26, No. 3, 1983


167

Streams (ST1=feed, ST2=distillate of 1st column, ST3=distillate of 2nd column, ST7=make-up


solvent, ST9=solvent feed)

Subtotal number of variables=38+9m=NSV


Total number of variables of a system of extractive distillation columns=N1V+N2V+NSV=65+13m
+(n1+n2+n3+n4+n5)(2m+8)=NV

2.2.2 Number of Relations

Subtotal number of relations=6+4m+(n1+n2+n3)(2m+7)=N1R

石 油 学 会誌 J. Japan Petrol. Inst., Vol. 26, No. 3, 1983


168

Subtotal number of relations=27+3m=NSR


Total number of relations=N1R+N2R+NSR=39+11m+(n1+n2+n3+n4+n5)(2m+7)=NR

2.2.3 Degree of Freedom column. The nonlinear material balance equations


The degree of freedom of a system of extrac- are solved simultaneously for each component,
tive distillation columns therefore, no matching relation is required. The
=NV-NR=(65+13m+(n1+n2+n3+n4 method may be employed only with the assumption
+n5)(2m+8))-(39+11m+(n1+n2+n3 that the liquid composition on each plate is equal
+n4+n5)(2m+7))=26+2m+n1+n2+n3 to the feed composition.
+n4+n5 For derivation of general working equations, a
The rate and composition of a feed should be hypothetical distillation column as shown in Fig.
defined, so m should be reduced from the degree 3 is considered as a convenient model column with
of freedom. Further, the pressure of every part any number of feeds and side streams. This col-
should be given, so n1+n2+n3+n4+n5 for every umn has n equilibrium plates and has a condenser
plate, 9 for every stream, 2 for both refluxes and (partial or total) at the top and a reboiler at the
2 for both reboiler vapors should be subtracted. bottom. The stages are numbered from top to
The number of degree of freedom then becomes bottom with the condenser as the zeroth stage
26+2m+n1+n2+n3+n4+n5-m-(n1+n2+n3+n4+ and the reboiler as the n+1-th stage. It is assumed
n5+9+2+2)=13+m. that at j-th plate one feed Fj, one vapor side stream
The simplest and usually the best way to fulfill SVj, and one liquid side stream SLj exist. This
these degrees of freedom is to give: model column can be reduce to a simpler one
Number of plates of each section in both by setting the undesired quantities to zero. Thus,
columns (n1,n2,n3,n4,n5) 5 for example, for a conventional column, all the
Reflux ratio (R1, R2) 2
quantities of the external streams except F, D, and
Flow rate of distillates (ST2, ST3) 2 W(=B) are zero.
Temperature of feeds (ST1, ST5) 2
Flow rate of solvent (ST9) 1
Temperature of solvent (ST9) 1
Temperature of make-up solvent (ST7) 1
Composition of make-up solvent (ST7) m-1

/13+m
3. Modified Relaxation Method1),2),3),5),6)

For calculation of distillation of multicomponent


systems, calculation of composition, of temperature,
and of mass flow of every part of the distillation
column or the system of distillation columns under
given conditions will be necessary. Many re-
searches have made the calculation by use of various
methods, but the relaxation method is considered
the most efficient for nonideal systems. In the
following sections, a new relaxation method modi-
fied by us is presented for calculating the com-
position and temperature profiles and the product
composition when the number of plates, the location
of feed plate, and the operating conditions are
specified. This method employs group relaxation
algorithm developed by us for the solution of
nonlinear material balance equations under un-
steady state conditions around each stage of the Fig. 3 A Model for the Complex Column

石 油 学会 誌 J. Japan Petrol. Inst., Vol. 26, No. 3, 1983


169

with respect to component, stage, and iteration.


Similarly, for the condenser and the reboiler, the
following equations are derived, respectively:
xk+1i0=xki0+ωki0(V1yki1-L0xki0-DxkiD) (7)

xk+1in+1=xkin+1+ωkin+1(Lnxk+1in-Vn+1ykin+1-WxkiW) (8)

3.2 Method of Computation


At first, one must specify the column pressure,
the number of plates; and the rate, the composition,
Fig. 4 An Ideal Equilibrium Stage and the thermal condition of each feed as well
as the locations of the feed plates and the side
Each stage in the model column is assumed to streams. Additional specifications made are the
be an ideal equilibrium stage, that is, the com- reflux ratio, the rate of distillate or bottoms, and
the rate of every side stream.
position of the liquid on the stage is identical
with that of the liquid leaving the stage, and the For the first trial, temperatures and liquid com-
vapor leaving the stage is in equilibrium with the positions for all stages of the column and also
liquid leaving the same stage. Fig. 4 shows such group relaxation factors for all components are
an ideal stage. assumed. The overall feed composition is usually
3.1 Derivation of Basic Equations taken as the initial liquid composition throughout
The material balance for component i around the column, but any other starting composition
may be used. As for the group relaxation factor,
j-th plate is given by the following differential
equation: we recommend it to be equal to the reciprocal

HLj(dxij/dt)=(FjZFij+Vj+1yij+1+Lj-1xij-1)
number of two to five times the sum of the
of feeds per unit time.
amounts

-(Vj+SVj)yij-(Lj+SLj)xij (1) The computational procedure may be written


The vapor holdup, HVj, in Fig. 4 is neglected in as follows:
the material balance, because it is usually small Step 1. Calculate the liquid and vapor flow rates
relative to the liquid holdup, HLj. For the con- of each stage from the following equations:
denser and the reboiler, similar but slightly different
Lj=RD+∑jr=1Frqr-∑jr=1SLr (9)
equations can be written respectively, as follows:

HL0(dxi0/dt)=V1yi1-L0xi0-DxiD (2) Vj=(R+1)D-∑j-1r=1Fr(1-qr)+∑jr=1SVr (10)

where qr is the liquid fraction in the feed to r-th


HLn+1(dxin+1/dt)=Lnxin-Vn+1yin+1-WxiW (3)
stage.
In the three equations above, the following finite Step 2. Assume temperature, liquid composition,
difference approximation is introduced. and group relaxation factors.
dxij/dt(xij)g+⊿t-(xij)t/⊿t (4) Step 3. Evaluate vapor composition, ykij, which
is in equilibrium with the liquid com-
For convenience, consider time t to be iteration position, xkij, by use of the vapor-liquid
k and t+⊿t to be iteration k+1, then Eq. (1) equilibrium relations.
becomes: Step 4. Solve the material balance equations for
xk+1ij=xkij+(⊿t/HLj)[(FjZFij+Vj+1yij+i+ xk+1ij
by Eqs. (6), (7) and (8).
Lj-1xij-1)-(Vj+SVj)yij-(Lj+SLj)xij]k (5) Step 5. Normalize the liquid composition ob-
The term (⊿t/HLj) is called the relaxation factor, tained.
but instead of it we use the group relaxation factor Step 6. If the convergence criteria, Eq. (14) and
ωij developed by us. Furthermore, when the cal- Eq. (15) are satisfied, a convergent solu-
culation begins from the top of the column and tion is obtained. If they are not satisfied
proceeds downward, instead of xkij-1in Eq. (5) go to Step 7.
the values calculated at k+1-th iteration, that is, Step 7. Calculate the following convergence cri-
xk+1ij-1
are used. Consequently, Eq. (5) may be teria:
written as follows:
xk+1ij=xkij+ωkij[FjZFij+Vj+1ykij+1+Lj-1xk+1ij-1 │∑N+1j=1(FjZFij-SLjxij-SVjyij)-DxiD-WxiW│
-(Vj+SVj)ykij-(Lj+SLj)xkij] (6) ξi= (11)
∑N+1j=1FjZFij
The group relaxation factor ωkij varies in value

石 油 学 会 誌 J. Japan Petrol. Inst., Vol. 26, No. 3, 1983


170

where ξi shows to what extent the mate-

rial balance of component i is satisfied.

The value of ξi differs greatly with each

component. Therefore, the correction


factor for group relaxation factor of each
component is calculated from the follow-
ing equation:

φi=│log ξi/log ξi│ (12)

where ξk is the smallest value of ξi. It

is obvious that the smaller is the value of

φi, the less satisfactory is the material

balance of component i. The new group

relaxation factor ωk+1ij is calculated from

the following relation:

ωk+1ij=βiωkij (13)

where βi is determined bascd on thc re-

suits of a large number of trials, that is,


when 1.0≧φ2≧0.7 βi=1.0
when 0.7>φi≧0.4 βi=1.1

when 0.4>φi≧0.0 βi=1.2

Using the new group relaxation factor, repeat


the steps from 3 through 7 until the desired ac-
curacy is obtained.
Convergence criterion is given by the following
two relations: Fig. 5 Liquid Composition Profile of Example 1 for a
ξi≦ε1 (14) Complex Column

∑N+1j=0(xk+1ij-xkij)2≦ε2 (15)
each stage of the column, a steady state solution
3.3 Numerical Example 1 was obtained in 1226 iterations, when ε1=10-4

A complex column with two feeds and three and ε2=10-6. Fig. 5 shows the liquid composition

side streams is considered. A mixture of five com- profile thus calculated.


ponents C1, C2, C3, C4, and C5 is to be separated. 4. Calculation of Extractive Distillation Col-
The column has 27 equilibrium stages, including
umns by Relaxation Method
a total condenser and a reboiler. The first feed,
whose composition is C1:0.50, C2:0.20, C3:0.15, Consider an extractive distillation system con-
C
4:0.10 and C5:0.05, consisting of 1.0 mole of sisting of two columns interlinked by one recycle
saturated vapor(q=0), is fed to the 10th plate stream as shown in Fig. 2. Both columns have
from top. The second feed, which has the com- a condenser and a reboiler. The bottoms of the
position:C1: 0.05, C2 :0.10, C3:0.15, C4: 0.20 and C5: first column is fed to the second column as feed
0.50, of 1.0 mole of saturated liquid(q=1) enters and the bottoms of the second column is recycled
at the 16th plate. Each liquid side stream of 0.4 and fed at near the top of the first column with
moles is to be drawn from the 5th, 13th, and 20th a make-up solvent stream.
plate. Relative volatilities of the components are To solve both columns individually using the
given hypothetically as follows: α15=10.0, α25= group relaxation method presented above, an initial
7.0, α35=5.0, α45=3.0 and α55=1.0. Two other assumption is made for the value of the recycle
conditions specified are the distillate rate D=0.4 stream composition, such as pure solvent. This
moles and the reflux ratio R=10.0. assumption makes it possible to solve the first
By assuming the initial group relaxation factors column and the composition of the bottom product
equal to the reciprocal number of five times the is then calculated. Using these results the second
sum of two feeds (1.0+1.0=2.0), that is 0.1, and column can be solved and the composition of the
starting with the average composition of two feeds bottoms can be obtained. The composition so
as the initial value of the liquid composition for determined is used in the correction of the com-

石油学会誌 J. Japan Petrol. Inst., Vol. 26, No. 3, 1983


171

position in the recycle stream. The amount of suming constant molar rates of liquid and vapor
the make-up solvent is determined by the overall in each section. For the first trial the liquid com-
material balance of the system. By using the cor- position of each stage in column 1 is equal to the
rected recycle stream composition, the first column average composition of feed and solvent and the
is reevaluated. This sequence is continued until composition of solvent is assumed to be equal to
the calculated composition of the recycle stream the composition of make-up solvent.
does not exceed the tolerance of errors. The calculated results of liquid composition pro-
When the changes in the vapor and liquid flow file in both columns are shown in Fig. 6.
rates in the column are to be considered, these
flow rates are corrected by using the enthalpy 5. Vapor-Liquid Equilibria
balance equation around each stage of the column. When calculations described above are carried
4.1 Numerical Example 2
out on multi-component distillation system, the
In this example, methanol and acetone are to most important factor to be considered is the vapor-
be separated by use of an extractive distillation
liquid equilibrium of the system to be distilled.
system consisting of two columns and using water It is most desirable that the observed data are
as a solvent. The system of methanol and acetone available for the whole system. Several methods
forms a minimum boiling azeotrope at 80mole% for measuring vapor-liquid equilibria and data ob-
acetone, but the addition of water as solvent in- served by us were presented in the following arti-
creases the relative volatility of acetone until the cles: (7, 9 to 15, 17, 19 to 21, 23 to 25, 32). In
azeotrope is destroyed. our book34), "Computer Aided Data Book of Vapor-
Column 1 (extractive distillation column) has Liquid Equilibria", we compiled data for about
15 stages including a total condenser and a reboiler. one thousand binary systems. Data for each system
A saturated liquid feed, whose composition is ace- are assembled onto one page. For 800 systems,
tone: 0.72, methanol: 0.18 and water: 0.10, is in- the tables include the original experimental data
troduced to the 10th stage at a rate of 100moles/ for vapor and liquid compositions and boiling tem-
unit time. A saturated liquid solvent of 200 moles
peratures, Antoine vapor pressure constants of the
is fed at the 5th stage of column 1. The top dis- components, optimized values of the Wilson's activ-
tillate rates of column 1 and column 2 are 75 ity coefficient equation, errors and x-y diagrams
moles and 28 moles, respectively. The reflux ratios for the systems. Wilson's equation was selected
are set at R1=3.0 and R2=4.0. 225 moles of the because of its accuracy and ease of application to
bottoms of column 1 is fed at the 5th stage of col- multicomponent systems. However, for other 133
umn 2 (solvent recovery column), which has 10 systems, whose vapor-liquid equilibria were mainly
stages including a total condenser and a reboiler. under high pressures, where Wilson's equation can-
The make-up solvent(3 moles from the total mate- not be used, the modified Redlich-Kwong equation
rial balance), whose composition is acetone: 0.02, of state was used for estimation of the vapor-liquid
methanol: 0.03 and water: 0.95, is added to the equilibria.
bottoms of column 2. Calculation is made as- If there are no experimental data of vapor-liquid
equilibria available, it is necessary to estimate them
by either using the equation of state or using the
atomic group solution theory. On these problems
we have discussed in the following articles: (8, 16,
18, 26 to 31, 33)

6. Conclusion

The methods of calculation developed by us for


multicomponent distillation systems were reviewed.
First, the degree of freedom of multicomponent
distillation was presented for a conventional col-
umn and for a system consisting of extractive
distillation columns. The method of group relax-
Fig. 6 Liquid Composition Profile in Column 1 and ation was described for a complex column with
Column 2 for Example 2 multi-feeds and multi-side streams. The method

石油学会誌 J. Japan Petrol. Inst., Vol. 26, No. 3, 1983


172

was applied to the system of extractive distillation 14) Hirata, M., J. Japan Petrol. Inst., 11, 613 (1968).
15) Hirata, M., Suda, S., Hakuta, T. Nagahama, K.,
columns. Finally, researches on the vapor-liquid
Mem. Facul. Tech. Tokyo Metropolitan Univ., 19, 103
equilibria, which are indispensable for calculation
(1969).
of multicomponent distillations, were reviewed 16) Nakahara, T., Hirata, M., J. Japan Petrol. Inst., 12,
briefly. 33 (1969).
Acknowledgment 17) Hirata, M., Suda, S., Hakuta, T., Nagahama, K., J.
Japan Petrol. Inst., 12, 773 (1969).
The author wishes to express his thanks to Takeshi
18) Hirata, M., Suda, S., Miyashita, R., Hoshino, T.,
Ishikawa, Hiromi Shono, Yoichi Kitamura, and Mem. Facul. Tech. Tokyo Metropolitan Univ., 20, 77
Reiko Miyashita for their collaboration in the de- (1970).
velopment of the methods presented in this paper. 19) Hakuta, T., Suda, S., Hirata, M., Mem. Facul. Tech.
Tokyo Metropolitan Univ., 20, 85 (1970).
References 20) Kato, M., Konishi, H., Hirata, M., J. Chem. Eng.
Data, 15, 435, 501 (1970).
1) Ishikawa, T., Hirata, M., Bull. Japan Petrol. Inst., 14, 21) Nagahama, K., Suda, S., Hakuta, T., Hirata, M., J.
18 (1972). Japan Petrol. Inst., 14, 252 (1971).
2) Ishikawa, T., Hirata, M., J. Chem. Eng. Japan, 5, 125 22) Hirata, M., Suda, S., Miyashita, R., Hoshino, T., J.
(1972). Japan Petrol. Inst., 14, 900 (1971).
3) Ishikawa, T., Hirata, M., Kagaku Kogaku, 36, 865 23) Hirata, M., J. Japan Petrol. Inst., 17, 76 (1974).
(1972). 24) Hoshino, D., Nagahama, K., Hirata, M., Bull. Japan
4) Hirata, M., Ishikawa, T., Shono, H., NIDIST Report, Petrol. Inst., 17, 9 (1975).
1, 7, 10 (1974). 25) Nagahama, K., Hirata, M., Bull. Japan Petrol. Inst.,
5) Hirata, M., Ishikawa, T., Shono, H., Kitamura, Y., 18, 79 (1976).
Miyashita, R., NIDIST Report, 2, 78 (1975). 26) Hoshino, D., Unno, Y., Nagahama, K., Hirata, M.,
6) Ishikawa, T., Shono, H., Kitamura, Y., Hirata, M., Bull. Japan Petrol. Inst., 19, 56 (1977).
Mem. Facul. Tech. Tokyo Metropolitan Univ., 25, 93 27) Hirata, M., J. Japan Petrol. Inst., 20, 435 (1977).
(1975). 28) Hoshino, D., Nagahama, K., Hirata, M., J. Japan
7) Hirata, M., Suda, S., Sagara, H., J. Japan Petrol. Petrol. Inst., 21, 128 (1978).
Inst., 9, 717 (1966). 29) Bae, H. K., Nagahama, K., Hirata, M., J. Japan
8) Hirata, M., Nakahara, T., J. Japan Petrol. Inst., 9, Petrol. Inst., 21, 249 (1978).
879 (1966). 30) Hoshino, D., Nagahama, K., Hirata, M., J. Chem.
9) Hirata, M., Sada, S., J. Japan Petrol. Inst., 9, 885 Eng. Japan, 12, 81 (1979).
(1966). 31) Bae, H. K., Nagahama, K., Hirata, M., Fluid Phase
10) Hirata, M., Hakuta, T., Onoda, T., J. Japan Petrol. Equilibria, 4, 45 (1980).
Inst., 10, 440 (1967). 32) Bae, H. K., Nagahama, K., Hirata, M., J. Chem. Eng.
11) Hirata, M., Hakuta, T., Mem. Facul. Tech. Tokyo Japan, 14, 1 (1981).
Metropolitan Univ., 18, 113 (1968). 33) Kato, K., Nagahama, K., Hirata, M., Fluid Phase
12) Hirata, M., Suda, S., Yokoyama, K., Mem. Facul. Equilibria, 7, 219 (1981).
Tech. Tokyo Metropolitan Univ., 18, 125 (1968). 34) Hirata, M., Ohe, S., Nagahama, K., Kodansha-Else-
13) Hirata, M., Suda, S., Bull. Japan Petrol. Inst. 10, 20 vier (Amsterdam) (1975).
(1968).

石油学会誌 J. Japan Petrol. Inst., Vol. 26, No. 3, 1983


173

要 旨

多 成 分 系 蒸 留 の 計 算

平 田 光 穂*

本 論 文 は多 成 分 系 の蒸 留 に 関 す る著 者 の 研 究 を ま とめ た もの 抽 出蒸 留 シ ス テ ム で は 各部 の 段数5個 と, 各塔 の還 流 比2個,
で あ る。 留 出量2個, 溶 剤 の量 (Fig. 2のST9) 1個 計10個 と, 各塔
第1章 で は普 通 の蒸 留塔 お よび 抽 出 蒸 留 シス テ ムの 自由 度 が へ の原 料 (ST1, ST5) の温 度2個, 溶 剤 (ST9) とメ イ クア ッ

も とめ られ て い る。 す なわ ち, 蒸 留 計算 を始 め るに あ た って 与 プ溶 剤 (ST7) の 温 度2個 総 計14個 を 与 え る。

え る こ とが で き る設 定 値 の数 (自由度 と呼 ば れ る) は1本 の 普 第2章 で は著 者 らに よる グル ー プ ・リ ラキ ゼ イ シ ョ ン法 に つ


通 の蒸 留 塔 (Fig. 1) に 対 して は いて 述 べ た。 この方 法 は 抽 出蒸 留 の よ うな非 理 想 性 の強 い系 に
10+m+n1+n2 は 最 も適 した方 法 で あ る。 基 礎 式 はEqs. (6), (7), (8) で,
で あ り, 2本 の蒸 留 塔 か らな る抽 出蒸 留 シス テ ム (Fig. 2) に Fig. 3お よび Fig. 4に しめ した よ うな 最 も一 般 化 され た 塔 に
対 して は 対 して 導 びか れ て い る。 これ らの式 で ωkijは グル ー プ ・リラ
26+2m+n1+n2+n3+n4+n5 キ ゼ イ シ ョ ン因 子 と呼 ば れ, そ の も とめ方 は2.2の Step 7に
とな る こ とを しめ した。 これ らに お い てmは 成 分 の数 で あ り, しめ され て い る。 例1に は5成 分 系 蒸 留 の例 を あげ て, 結 果 は
niは 蒸 留塔 の 各 部 に お け る 段 数 で あ る。 これ らの 自 由度 か ら Fig. 5に しめ した。
各 部 分 の 圧 力 や 原 料 の 組 成 な ど当 然 指 定 しな け れ ば な らな い も 第3章 で は グ ル ー プ ・リラキ ゼ イ シ ョ ン法 を抽 出蒸 留 系 に適
の を 差 し引 くと, 普 通 の 蒸 留 塔 では 残 りは4と な り, 抽 出蒸 留 用 す る方 法 に つ き述 べ, 例2と して は メ タ ノ ー ル とアセ トンを
シス テ ムで は14と な る。 これ らを 満 た す 最 も簡 単 で ま た最 も 水 を 溶 剤 と して 分 離 す る場 合 を あ げ, 結 果 は Fig. 6に しめ し
良 い 方 法 は 次 の 通 りで あ る。 た。
普 通 の 蒸 留塔 で は 濃 縮 部 と回 収 部 の 段 数2個 と還 流 比 と留 出 第4章 に は, これ らの 蒸 留 計 算 を お こな うに あ た って 不 可 欠
量 計4個 を 与 え る。 で あ る気 液 平 衡 関 係 に 関 す る著 者 らの 研 究 を 簡 単 に ま とめ た。
* 東京都立大学工学部 (158東 京 都 世 田 谷 区 深 沢2-1-1)

Keywords

Distillation, Extractive distillation, Freedom, Multicomponent distillation, Nonideal system, Relaxation method

石油学会誌 J. Japan Petrol. Inst., Vol. 26, No. 3, 1983

You might also like