Prob-4 5

Download as pdf or txt
Download as pdf or txt
You are on page 1of 8

4.

5 COMPLEX CHEMICAL EQUILIBRIUM BY GIBBS ENERGY MINIMIZATION

445

4.5 COMPLEX C HEMICAL E QUILIBRIUM BY GIBBS ENERGY


M INIMIZATION
4.5.1

Concepts Demonstrated

Formulation of a chemical equilibrium problem as a Gibbs energy minimization


problem with atom balance constraints. Use of Lagrange multipliers to introduce
the constraints into the objective function. Conversion of the minimization problem into a system of nonlinear algebraic equations.
4.5.2

Numerical Methods Utilized

Solution of a system of nonlinear algebraic equations with constraints.


4.5.3

Excel Options and Functions Demonstrated

Use of the Excel Add-In Solver for constrained minimization.


4.5.4

Problem Definition

Ethane reacts with steam to form hydrogen over a cracking catalyst at a temperature of T = 1000 K and pressure of P = 1 atm. The feed contains 4 moles of H2O
per mole of CH4. Balzisher et al.1 suggest that only the compounds shown in
Table 410 are present in the equilibrium mixture (assuming that no carbon is
deposited). The Gibbs energies of formation of the various compounds at the temperature of the reaction (1000K) are also given in Table 410. The equilibrium
composition of the effluent mixture is to be calculated using these data.
Table 410 Compounds Present in Effluent of Steam Cracking Reactor1

No.
1
2
3
4
5
6
7
8
9

Component
CH4
C2H4
C2H2
CO2
CO
O2
H2
H2O
C2H6

Gibbs Energy
kcal/gm-mol
4.61
28.249
40.604
-94.61
-47.942
0
0
-46.03
26.13

Feed
gm-mol

4
1

Effluent
Initial Estimate
0.001
0.001
0.001
0.993
1
0.0001a
5.992
1
0.001

aThis initial estimate is more realistic and useful than the original

published estimate of 0.007.

446

CHAPTER 4

PROBLEM SOLVING WITH EXCEL

(a) Formulate the problem as a constrained minimization problem. Introduce the constraints into the objective function using Lagrange multipliers and differentiate this function to obtain a system of nonlinear
algebraic equations.
(b) Use the POLYMATH Constrained solution algorithm to find the
solution to this system of nonlinear equations. Start the iterations
from the initial estimates shown in Table 410.
(c) Use Excel's Solver to solve the problem as a constrained minimization problem without the use of Lagrange multipliers and without differentiation of the objective functions. Compare the results with those
obtained in (b).
4.5.5

Solution

The objective function to be minimized is the total Gibbs energy given by


min
ni

G
-------RT

G0

i
i + ln ---------
ni ------RTn
i
i=1

(4-49)

where ni is the number of moles of 0component i, c is the total number of compounds, R is the gas constant and Gi is the Gibbs energy of pure component i at
temperature T. The minimization of Equation (4-49) must be carried out subject
to atomic balance constraints
Oxygen Balance

g1

= 2 n4 + n 5 + 2 n6 + n7 4 = 0

(4-50)

Hydrogen Balance

g2

= 4 n 1 + 4 n 2 + 2 n 3 + 2 n 7 + 2 n 8 + 6 n 9 14 = 0

(4-51)

Carbon Balance

g3

n1 + 2 n 2 + 2 n3 + n4 + n 5 + 2 n9 2

= 0

(4-52)

The identification of the various components is given in Table 410.


These three constraints can be introduced into the objective functions using
Lagrange multipliers: 1, 2 and 3. The extended objective function is
min
n i, j

ni 3
G0i
+ ln ---------- +
n ------RT
n i

i=1

j=1

j gj

(4-53)

The condition for minimum of this function at a particular point is that all the
partial derivatives of F with respect to ni and j vanish at this point. The partial
derivative of F with respect to n1, for example, is
F
=
n1

G0

n
ni

1
1
+ ln ---------- + 4 2 + 3
-------RT

= 0

(4-54)

The other partial derivatives with respect to ni can be obtained similarly. If it is

4.5 COMPLEX CHEMICAL EQUILIBRIUM BY GIBBS ENERGY MINIMIZATION

447

expected that the amount of a particular compound at equilibrium is very close


to zero, it is preferable to rewrite the equation in a form that does not require
calculation of the logarithm of a very small number. Rearranging Equation (454), for example, yields
0
G

n1 n i exp --------1 + 4 2 + 3
RT

= 0

(4-55)

The partial derivatives of F with respect to 1, 2 and 3 are g1, g2 and g3 respectively.
The complete set of nonlinear equations, as entered into the POLYMATH Nonlinear Algebraic Equation Solver, is shown in Table 411. There are
12 implicit equations associated with the 12 unknowns. In the POLYMATH
input, the amount (moles) of a compound (ni) is represented by the formula of the
compound, for clarity. The equations associated with O2 and C2H2 are written in
the form of Equation (4-55) as preliminary tests have shown difficulty in convergence of the solution algorithm when the equations that contain logarithms of
the amount of those compounds are used.
(b)

Table 411 Equation Input to the POLYMATH Nonlinear Equation Solver Program - File P4-05B1.POL
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26

Equation, # Comment
R = 1.9872
sum = H2 + O2 + H2O + CO + CO2 + CH4 + C2H6 + C2H4 + C2H2
f(lamda1) = 2 * CO2 + CO + 2 * O2 + H2O - 4 # Oxygen balance
f(lamda2) = 4 * CH4 + 4 * C2H4 + 2 * C2H2 + 2 * H2 + 2 * H2O + 6 * C2H6 - 14 # Hydrogen balance
f(lamda3) = CH4 + 2 * C2H4 + 2 * C2H2 + CO2 + CO + 2 * C2H6 - 2 # Carbon balance
f(H2) = ln(H2 / sum) + 2 * lamda2
f(H2O) = -46.03 / R + ln(H2O / sum) + lamda1 + 2 * lamda2
f(CO) = -47.942 / R + ln(CO / sum) + lamda1 + lamda3
f(CO2) = -94.61 / R + ln(CO2 / sum) + 2 * lamda1 + lamda3
f(CH4) = 4.61 / R + ln(CH4 / sum) + 4 * lamda2 + lamda3
f(C2H6) = 26.13 / R + ln(C2H6 / sum) + 6 * lamda2 + 2 * lamda3
f(C2H4) = 28.249 / R + ln(C2H4 / sum) + 4 * lamda2 + 2 * lamda3
f(C2H2) = C2H2 - exp(-(40.604 / R + 2 * lamda2 + 2 * lamda3)) * sum
f(O2) = O2 - exp(-2 * lamda1) * sum
H2(0) = 5.992
O2(0) = 0.0001 > 0
H2O(0) = 1
CO(0) = 1
CH4(0) = 0.001 > 0
C2H4(0) = 0.001 > 0
C2H2(0) = 0.001 > 0
CO2(0) = 0.993
C2H6(0) = 0.001 > 0
lamda1(0) = 10
lamda2(0) = 10
lamda3(0) = 10

448

CHAPTER 4

PROBLEM SOLVING WITH EXCEL

The initial estimates suggested by Balzisher et al.1 (shown in Table 410)


are entered in lines 15 through 23 of the input to POLYMATH. Note that the initial estimate O2(0)=0.0001>0, for example, indicates that this variable is constrained to be always positive (absolutely positive) during the problem solution.
For this problem, only the constrained POLYMATH solution algorithm converges to the solution from the given initial estimates, while the other algorithms stop with the error message that the calculation of the logarithm of a
negative number is attempted. The constrained algorithm also allows definition
of some or all of the unknowns to be positive at the solution (physically positive). For O2 in this problem, the corresponding physically positive format as
specified in the initial estimate would O2(0)=0.0001>=0. This format was not
used in this problem.
The variables for which the initial estimates are specified in lines 16, 19,
20, 21 and 23 in Table 410 are marked as absolutely positive. The rest of the
variables do not approach zero so as to obtain negative values during the solution process (lines 15, 17, 18 and 19), or they are allowed to have both negative
and positive values (lines 24, 25 and 26).
The POLYMATH solution for the equilibrium composition of the effluent
mixture is shown in Table 412. These results indicate that the effluent contains
Table 412 POLYMATH Results for Equilibrium Composition of Effluent
Stream - File P4-05B1.POL

1
2
3
4
5
6
7
8
9
10
11
12
13
14

Variable
C2H2
C2H4
C2H6
CH4
CO
CO2
H2
H2O
lamda1
lamda2
lamda3
O2
R
sum

Value
3.157E-10
9.541E-08
1.671E-07
0.0665638
1.388517
0.5449182
5.345225
1.521646
24.41966
0.2530591
1.559832
5.459E-21
1.9872
8.866871

f(x)
7.238E-25
-2.58E-13
-1.688E-13
0
2.442E-15
-1.11E-15
1.11E-16
-1.665E-15
0
0
0
-5.687E-27

Initial Guess
0.001
0.001
0.001
0.001
1.
0.993
5.992
1.
10.
10.
10.
0.0001

significant amounts of H2 (5.345 moles per mole of C2H6 feed), H2O, CO, CO2
and CH4 and contains only trace amounts of C2H6, C2H4, C2H2 and O2. All the
function values are smaller by several orders of magnitude than the respective
variable values, indicating that the solution is correct. The same values were
obtained also by Balzisher et al.1, who used a dedicated FORTRAN program to
solve the same problem.

4.5 COMPLEX CHEMICAL EQUILIBRIUM BY GIBBS ENERGY MINIMIZATION

449

The value of the objective function at this point may be added to the POLYMATH problem by including Equation (4-49). The additional equations comprising Equation (4-49) are summarized in Table 413. The objective function value
at the solution is -104.34, and this can be used to compare the POLYMATH solution with the Excel solution in part (c).
Table 413 Equations for Calculation of the Gibbs Energy Function - File P4-05B2.POL
Line
1
2
3
4
5
6
7
8
9
10

Equation
G_O2 = O2 * ln(abs(O2 / sum))
G_H2 = H2 * ln(H2 / sum)
G_H2O = H2O * (-46.03 / R + ln(H2O / sum))
G_CO = CO * (-47.942 / R + ln(CO / sum))
G_CO2 = CO2 * (-94.61 / R + ln(CO2 / sum))
G_CH4 = CH4 * (4.61 / R + ln(abs(CH4 / sum)))
G_C2H6 = C2H6 * (26.13 / R + ln(abs(C2H6 / sum)))
G_C2H4 = C2H4 * (28.249 / R + ln(abs(C2H4 / sum)))
G_C2H2 = C2H2 * (40.604 / R + ln(abs(C2H2 / sum)))
ObjFun = G_H2 + G_H2O + G_CO + G_O2 + G_CO2 + G_CH4 + G_C2H6 + G_C2H4 + G_C2H2

The use of the Excel Add-In Solver* to solve this problem requires only
the Gibbs Energy objective function of Equation (4-49) as presented in Table 4
13 and the atomic material balance constraints given by Equations (4-50), (4-51)
and (4-52). It is convenient to enter these needed equations into Excel via POLYMATH. The resulting equations from POLYMATH as exported to Excel are given
in Table 414. Note that the Gibbs Energy objective function is in line 4, and the
atomic constraints are in lines 5, 6 and 7. The equations in line 4 and lines 9
through 18 are components of the objective function (Equation (4-49). Lines 19
through 27 provide the initial estimates for the nine problem variables.
(c)

Table 414 Gibbs Energy Minimization with Atom Balance Constraints as Exported from POLYMATH to
Excel - Files P4-05C.POL and P4-05C.XLS
A
B
C
D
E
1 POLYMATH NLE Migration Document
2
Variable Value
Polymath Equation
3 Explicit Eqs R
1.9872
R=1.9872
4
sum
8.996
sum=H2 + H2O + CO + O2 + CO2 + CH4 + C2H6 + C2H4 +
C2H2
5
OxBal
-4.441E-16
OxBal=2 * CO2 + CO + 2 * O2 + H2O - 4
6
HydBal 0
HydBal=4 * CH4 + 4 * C2H4 + 2 * C2H2 + 2 * H2 + 2 * H2O +
6 * C2H6 - 14
7
CarBal 0
CarBal=CH4 + 2 * C2H4 + 2 * C2H2 + CO2 + CO + 2 * C2H6 2
8
eps
1E-21
eps=0.1e-20
9
G_O2
-0.0501104
G_O2=O2 * ln(abs((O2 + eps) / sum))

* The Excel Add-In Solver may require special installation from Microsoft Excel or the

Microsoft Office. If this Add-In is not available from the drop-down Tools/Add-Ins menu in
Excel, please consult the Microsoft instructions to install this software.

450

CHAPTER 4

PROBLEM SOLVING WITH EXCEL

Table 414 Gibbs Energy Minimization with Atom Balance Constraints as Exported from POLYMATH to
Excel - Files P4-05C.POL and P4-05C.XLS

10
11
12
13
14
15
16
17
18

B
G_H2
G_H2O
G_CO
G_CO2
G_CH4
G_C2H6
G_C2H4
G_C2H2
ObjFun

19
20
21
22
23
24
25
26
27

O2
H2
H2O
CO
CH4
C2H4
C2H2
CO2
C2H6

C
D
E
-2.4348779
G_H2=H2 * ln(H2 / sum)
-25.360025
G_H2O=H2O * (-46.03 / R + ln(H2O / sum))
-26.322183
G_CO=CO * (-47.942 / R + ln(CO / sum))
-49.464812
G_CO2=CO2 * (-94.61 / R + ln(CO2 / sum))
-0.0067847
G_CH4=CH4 * (4.61 / R + ln(CH4 / sum))
0.00404462
G_C2H6=C2H6 * (26.13 / R + ln(abs((C2H6 + eps) / sum))
0.00511094
G_C2H4=C2H4 * (28.249 / R + ln(abs((C2H4 + eps) / sum))
0.01132823
G_C2H2=C2H2 * (40.604 / R + ln(abs((C2H2 + eps) / sum))
-103.61831
ObjFun=G_H2 + G_H2O + G_CO + G_O2 + G_CO2 + G_CH4
+ G_C2H6 + G_C2H4 + G_C2H2
0.007
O2=0.007
5.992
H2=5.992
1
H2O=1
1
CO=1
0.001
CH4=0.001
0.001
C2H4=0.001
0.001
C2H2=0.001
0.993
CO2=0.993
0.001
C2H6=0.001

It should be noted that the Solver Add-In cannot find the minimum when
Equation (4-49) is entered in its original form. Excel execution stops with error
messages indicating that there is an attempt to calculate logarithm of a negative
number. In order to prevent calculation of logarithm of a negative number or logarithm of zero, the expressions for calculating the partial Gibbs energy of some of
the compounds have been changed according to
Gi
-------RT

G 0i
n -------- + ln
RT

n i +
abs --------------
ni

(4-56)

where is a very small number ( = 1 1021 ). This equation is used only for the
compounds which present in trace amounts in the effluent: O2, C2H2, C2H4 and
C2H6.
The use of the Excel Solver Add-In for the solution of this minimization
problem is best explained with reference to Figures 451, 452 and 453. After
the Solver interface is requested within the Excel Tools drop-down menu as
shown in Figure 451, the Target Cell to be minimized should be entered as cell
C18 which contains the formula for the objective function (see Table 414). Next
the Equal to: option should be set to Min. The cells to be changed are the cells
which contain the numbers of moles of the various compounds which is indicated
in Excel by $C$19:$C$27. The three constraint equations from the atomic material balance constraints given by Equations (4-50), (4-51) and (4-52) are entered
by reference to the Excel cell containing each equation as shown in Figure 451.
These balances should all be equal to zero at the solution. Addition of the first

451

CHAPTER 4

PROBLEM SOLVING WITH EXCEL

constraint is shown in Figure 452.

Figure 451

Solver Parameters for the Gibbs Energy Minimization Problem


- File P4-05C.XLS

Figure 452

Addition of First Constraint for Solver Minimization


- File P4-05C.XLS

The constraints regarding the positive value of the compounds can be specified in the Options communication box shown in Figure 453. Two default
options must be changed with the Options button to obtain a solution. First,
Assume Non-Negative must be selected; otherwise the Solver converges to a
solution where some of the mole numbers are negative. Second, Use Automatic
Scaling must also be selected in order to obtain a feasible solution. The Excel
Solver solution is partially shown in Figure 454.

Figure 453

Solver Options for the Gibbs Energy Minimization Problem

452

CHAPTER 4

Figure 454

PROBLEM SOLVING WITH EXCEL

Partial Excel Spreadsheet Showing Solver Minimization Solution

The results obtained by the POLYMATH constrained Nonlinear Equation


Solver, the Excel Solver constrained minimization algorithm and the values
reported by Balzisher et al.1 are summarized in Table 415. Examination of the
various results indicates that the minimum Gibbs energy obtained by POLYMATH is slightly lower than the value obtained by the Solver. The composition
values obtained by POLYMATH are almost identical to those reported by Balzisher et al.1. There are considerable differences between those values from
POLYMATH and those obtained by the Excel Solver Add-In. There may be other
optimization routines that can be used with Excel to achieve a closer agreement
in the solution.
Table 415 Effluent Composition and Minimum of Gibbs Energy by Various Solution
Techniques
No.
1
2
3
4
5
6
7
8
9

Component
CH4
C2H4
C2H2
CO2
CO
O2
H2
H2O
C2H6
Gibbs
Energy

Initial Estimate
0.001
0.001
0.001
0.993
1
0.007
5.992
1
0.001
-103.61831

POLYMATH
0.066564
9.54E-08
3.16E-10
0.544918
1.388517
5.46E-21
5.345225
1.521646
1.67E-07
-104.34

Excel Solver
0.00149444
1.0112E-06
2.9847E-07
0.53441967
1.46396259
0
5.52962977
1.46719804
6.0332E-05
-104.27612

Balzisher et. al1


0.066456
9.41E-8
3.15E-10
0.544917
1.3886
3.704E-21
5.3455
1.5215
1.572E-7

The problem solution files are found in directory CHAPTER 4 and designated
,
,
, and
.
P4-05B1.POL P4-05B2.POL P4-05C.POL

P4-05C.XLS

You might also like