PSM010 2

OPTIMIZATION OF COMPLEX CHEMICAL ENGINEERING
PROBLEMS USING MATHCAD AND LINGO PACKAGE PROGRAMS:

A COMPARISON OF PERFORMANCES OF THESE PACKAGES
Selahattin Gltekin
Dou University, Acbadem, Kadky 34722-Istanbul-TURKEY
INTRODUCTION
In chemical engineering, one faces nonlinear, highly complex problems in transport
phenomena, process dynamics and control, chemical reaction engineering, thermodynamics,
etc. Not only are the mere solutions of these problems extremely tedious and time consuming,
but the difficulties involved in finding optimum variables sometimes become unbearable.
Solution of these kinds of problems by classical numerical techniques takes hours if not days.
However, today in the market there are certain ready programs such as MATHCAD [1],
LINGO, LINDO, GINO [2] used for the solution of such sophisticated optimization
problems.
METHODOLOGY
This study considers some complicated problems in the fields of transport phenomena,
separation processes [3], and chemical reaction engineering [4],
. Accordingly some
examples were selected from heat transfer to minimize heat loss to find optimum insulation
thickness, from chemical reaction engineering and a multi-effect evaporator system has been
considered for optimum number of effects to minimize the cost of the process involved, and
so forth.
RESULTS AND DISCUSSIONS
It has been deduced that although MATHCAD is a general- purpose engineering problem
solving program (e.g. optimization, solution of simultaneous linear, nonlinear equations,
Laplace Transformations, Fourier Transformations, Z-transformations, and many more) , it
has some superiority as compared
to other available package programs in solving
optimization problems. Moreover, the statements written in MATHCAD are much shorter and
simpler as compared to those used in other programs. Selected examples from the above fields
are solved especially with MATHCAD and LINGO6, and the findings are discussed in detail
with advantages and disadvantages. The graphical representation of the results with
MATHCAD in 2-D or 3-D seemed to be much more versatile as compared to LINGO6
optimization program.
OPTIMIZATION
The objective in optimization is to find the most suitable solution among several alternatives.
Perhaps it is better to say that it is a procedure to choose the least worst one among others!
In optimization, objectives could be:
* to maximize profit
* to minimize loss
* to maximize chemical conversion in a chemical reactor
* to minimize cost, etc.
In an optimization problem one always has the following:
*An objective function, and
* Constraints (limitations, restrictions) in the form of
a)Equalities
b)Inequalities, or
c)Both
Depending on the objective functions and the constraints the following programming are
possible:
Linear Programming (Simplex Method and/or graphical method)
Non-linear Programming
Integer Programming
The objective of this humble work is to compile some well-known optimization problems
available in chemical engineering textbooks and solve them by using MATHCAD and
LINGO package programs and finally compare the performance and ease of use these
programs.
In the following pages some problem statements and their solutions are being given in both
packages:
SOLUTIONS OF OPTIMIZATION PROBLEMS with MATHCAD 2001 Professional [1]
1- Problem may be linear or non-linear. MATHCAD could handle either of them very
effectively.
2- For example, if
f (x1, x2, x3, , xn) is an objective function with constraints as
n
Cji xii = bj or bj or bj
i = 1,2, ..n
j = 1,2, ..m
3- One has to give a guess value to each of the variables, namely x1, x2, x3, xn. Then write
down the objective function. After that write down given block. After that either you should
i=1
maximize or minimize the function. If we apply the above information, the picture is as
follows
x1 = guess1 , x2 = guess2 , x3 = guess3 , xn = guessn
f (x1, x2, x3, , xn) := [ expression for objective function]
Given
n
Cji xii = bj or
i=1
bj or bj
Maximize (x1, x2, x3, , xn) =
Minimize (x1, x2, x3, , xn) =
LINGO PROGRAM [2]

This optimization program is suitable to solve Linear, Nonlinear and Integer optimization
problems. It works under windows and quite versatile.
Firstly the objective functions to be maximized or minimized is to be typed as
Min=f(x1,x2,xn); or
Max= f(x1,x2,xn);
Then constraints must be typed as
ajixikbj; or
bj;
where aji and bj are constants.
Then in windows tool bar click on solve to get the result.
Problem 1 [3]
A multiple-effect evaporator is to be used for evaporating 400,000 lb of water per day from a
salt solution. The total initial cost for the first effect is $18,000, and each additional effect
costs $15,000. The life period is estimated to be 10 years, and the salvage or scrap value at the
and of the life period may be assumed to be zero. The straight-line depreciation method is
used. Fixed charges minus depreciation are 15 percent yearly based on the first cost of the
equipment cost. All other costs are independent of the number of effects. The unit will operate
300 days per year. If the pounds of water evaporated per pound of steam equals 0.85 x
number of effects, determine the optimum number of effects for minimum annual cost.
MATHCAD
n := 10
Ct( n) := 4500 n + 900 +
Minimize( Ct , n) = 6.863
211976
n
LINGO
n>0;
Min=4500*n+900+211976/n;
Local optimal solution found
at step:
14
Objective value:62670.28
Variable
Value
n
6.863365
Problem 2 [3]
Determine the optimum economic thickness of insulation that should be used under the
following conditions: Saturated steam is being passed continuously through a steel pipe with
an outside diameter of 10.75 in. The temperature of the steam is 400F , and the steam is
valued at $1.80 per 1000 lb. The pipe is to be insulated with a material that has a thermal
conductivity of 0.03 Btu/(h)(ft2)( F/ft). The cost of the installed insulation per foot of pipe
length is $4.5 x It, where It is the thickness of the insulation in inches. Annual fixed charges
including maintenance amount to 20 percent of the initial installed cost. The total length of
the pipe is 1000 ft, and the average temperature of the surroundings may be taken as 70F.
Heat-transfer resistances due to the steam film, scale, and pipe wall are negligible. The air-
film coefficient at the outside of the insulation may be assumed constant at 2.0 Btu/(h)(ft2) (F)
for all insulation thicknesses.
3.14 1000 ( 10.75 + 2 I) ( 400 70)
12
24
C ( I) :=
1.8 365
+ 900 I
826 1000
10.75
+I
ln
10.75
1
2
24 3.14 100 ( 10.75 + 2 I) 2 0.03 1000 3.14
I := 1
Given
I>0
Minimize( C , I) =
9.19
No solution was found with LINGO
Problem 3 [4]
A refinery has available two crude oils that have the yields shown n the following table.
Because of equipment and storage limitations, productions of gasoline, kerosene, and fuel oil
must be limited as also shown in this table. There are no plant limitations on the products such
as gas oils.
The profit on processing crude #1 is $1.00/bbl and on crude # it is $0.70/bbl. Find the
approximate optimum daily feed rates of the two crudes to his plant via a graphical method.
Maximum allowable
Volume percent yields
product
rate
Crude #1
Crude #2
(bbl/day)
Gasoline
70
31
6,000
Kerosene
6
9
2,400
Fuel Oil
24
60
12,000
MATHCAD
LINGO
max=8.1*x1+10.8*x2;
0.8*x1+0.44*x2<24000;
0.05*x1+0.1*x2<2000;
0.1*x1+0.36*x2<6000;
x1>0;
x2>0;
x1 := 20
x2 := 30
f ( x1 , x2) := 8.1 x1 + 10.8 x2

Given
Global optimal solution found

at step:4
Objective value: 286758.6
Constraints
0.8 x1 + 0.44 x2 24000
0.05 x1 + 0.1 x2 2000
Variable
0.1 x1 + 0.36 x2 6000
2.621 104
3
6.897 10
Maximize( f , x1 , x2) =
Value
X1
26206.90
X2
6896.552
Problem 4 [4]
Feed to three units is split into three streams: FA, FB and FC. Two products are produces: P1
and P2 (see following figure), and the yield in weight percent by unit is
Yield
(weight %)
P1
P2
Unit A
Unit B
Unit C
40
60
30
70
50
50
Each stream has values in $/lb as follows:

Stream
F
Value($/lb) .40
P1
.60
P2
.30
Because of capacity limitations, certain constraints exist in the stream flows.

1. The total input feed must not exceed 10,000 lb/day.
2. The feed to each of the units A, B and C not exceed 5000 lb/day.
3. No more than 4000 lb/day of P1 can be used, and no more than $7000 lb/day of P2 can
be used.
In order to determine the values of FA, FB , and FC that maximize the daily profit, prepare a
mathematical statement of this problem as linear programming problem. Do not solve it.
MATHCAD
FC := 10
FA := 10 FB := 10
P ( FA , FB , FC ) := ( 0.4 FA + .3 FB + .5 FC ) 0.6 + ( FA .3 + FB .7 + FC .5 ) .3 F .4
F := FA + FB + FC
Given
FA + FB + FC < 10000
FA < 5000
FB < 5000
FC < 5000
( 0.4 FA + .3 FB + .5 FC ) < 4000
( FA .3 + FB .7 + FC .5 ) < 7000
3.754 10 13
3
Maximize ( P , FA , FB , FC ) =
5 10
3
5 10
No Solution with LINGO
CONCLUSIONS
The authors research experience in evaluating the experimental results for curve fitting to
data or finding optimization process variables in the chemical reaction engineering and
separation processes shows that the use of package programs not only saves time and money,
but also generates extremely impressive graphs in the chemical engineering field, especially
with MATHCAD
REFERENCES
1. MATHCAD 2001, Applesoft Co. Boston, MA (2001)
2. LINDO Systems Inc. (1988)
3. Peters, M. S., and Timmerhaus, K. D., Plant Design and Economics for Chemical
Engineers, 4th Ed. McGraw-Hill (1991)
4. Edgar, T.F., Himmelblau, D. M., and Lasdon, L. S. Optimization of Chemical
Processes 2nd Ed. McGraw-Hill (2001)

PSM010 2

Uploaded by

Copyright:

Available Formats

PSM010 2

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

PSM010 2

Uploaded by

Copyright:

Available Formats

OPTIMIZATION OF COMPLEX CHEMICAL ENGINEERING

PROBLEMS USING MATHCAD AND LINGO PACKAGE PROGRAMS:

to other available package programs in solving

Maximize (x1, x2, x3, , xn) =

Minimize (x1, x2, x3, , xn) =

LINGO PROGRAM [2]

where aji and bj are constants.

Then in windows tool bar click on solve to get the result.

Ct( n) := 4500 n + 900 +

3.14 1000 ( 10.75 + 2 I) ( 400 70)

24 3.14 100 ( 10.75 + 2 I) 2 0.03 1000 3.14

No solution was found with LINGO

f ( x1 , x2) := 8.1 x1 + 10.8 x2

Global optimal solution found

0.1 x1 + 0.36 x2 6000

Each stream has values in $/lb as follows:

Because of capacity limitations, certain constraints exist in the stream flows.

No Solution with LINGO

You might also like