See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/238771365

OPTIMIZATION OF MilLING PERFORMANCE OF A SUGAR Mill BY USING

LINEAR PROGRAMMING TECHNIQUE

Article · January 2003

CITATIONS READS
5 3,430

3 authors, including:

Anjum Munir Rasmiati Tahir

University of Agriculture Faisalabad Stisip Veteran Palopo
89 PUBLICATIONS 654 CITATIONS 6 PUBLICATIONS 68 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

International Center for Development and Decent Work (ICDD) and German Academic Exchange Service (DAAD) View project

Income Generation using Solar Based Food Processing Technologies for Rural Community Funded by ICDD (DAAD), Germany View project

All content following this page was uploaded by Anjum Munir on 20 March 2016.

The user has requested enhancement of the downloaded file.

 Pak. J. Agri. sci, Vol. 40(1-2), 2003

OPTIMIZATION OF MilLING PERFORMANCE OF A SUGAR Mill BY USING

LINEAR PROGRAMMING TECHNIQUE
Anjum Munir, A.R.Tahir, M.Shafi Sabir
Department of Farm Machinery and Power, University of Agriculture, Faisalabad

The proposed study was carried out at Pahrianwali Sugar Mills, Lalian District Jhang. Data was collected for
sugar lost and moisture content in bagasse at different milling parameters like mills settings, cane imbibition,
hydraulic pressure on top rollers of the mills and index of cane preparation separately. This data was used to
develop constraint equations for linear programming model. An objective function of linear programming (LP)
model was also developed to minimize the cost of steam production and sugar lost in bagasse. Quantitative
System Business (QSB) software was used to solve the linear programming model. The results of the LP model
indicated an optimum cane crushing capacity of 5000 tons/day for mill setting of 58.3 mm against a hydraulic
3 2
pressure of 0.225 x 10 kg/cm to earn maximum profit of Rs. 58312/hr.
Key words: Sugar mill, linear programming, mill setting, imbibition, bagasse

INTRODUCTION Development of objective function

Sugar industry is one of the major agro-based The following objective function was developed to
industries that is playing a vital role in the development minimize the milling losses in the form of sugar (pol)
of our country. The cane-sugar industry is one of the lost to bagasse, furnace oil consumed in the boilers
oldest industries in Pakistan and is running with due to moisture in bagasse and extra steam
traditional management and technology. Milling consumption due to changing levels of imbibition,
performance of majority of the sugar mills in Pakistan hydraulic pressure and index of cane preparation at
is much below the national and international standards. constant crushing rate.
Normal sugar extraction of about 90 % appears to be
Min Z = C1 X + C2 I + C3 P + C4 D
satisfactory, however, this reflects that 8 % sugar in the
Where
cane is lost to bagasse that is burnt as a fuel in the
boilers. Hence, updating is absolutely necessary in I, P and D are decision variables and C1, C2, C3 and C4
every section of sugar industry especially in are the co-efficient of objective function
milling/crushing section which is responsible to control X = 1
all the major losses of a sugar mill (Azeem, 1976). I = Index of cane preparation
It therefore, becomes extremely important to reduce P = Hydraulic pressure on top rollers of the mills
3
this loss by optimally using the available resources. (10 x kg/hr)
There are many important factors that determine the D = Weight of imbibition per unit weight of cane.(kg/kg)
overall performance of the milling train. An optimal mill Mathematically, this may be expanded to include.
setting under prevailing resources was the main object Min Z = 1068.94 (w-wm) QsPl/Rl + Wb S P2
of this study. To achieve this objective, the study was + 0.01639 C1 L d n (P-Pm) N P3/Rl
designed to construct linear programming model for + We (D-Dm) R2 R3 P3/Rl + (I + ml) P3/Rl (A)
Pahrianwali Sugar Mills Limited. The model aimed at Where
minimizing the milling losses and cost. w = Weight of moisture per unit weight of bagasse.
Wm = Maximum limit of weight of moisture per unit
MATERIALS AND METHODS weight of bagasse above which furnace oil has
to used to maintain the steam pressure constant
The Pahrianwali Sugar Mills consists of mainly cane at given flow rate of steam
preparation unit, milling train, juice processing unit, O, =Steam flow rate (tons/hr)
centrifugation and sugar drying units. Two bagasse p, = price of furnace oil (Rs/lb)
cum furnace oil boilers are used to generate super R1 = Steam produced per unit weight of bagasse.
heated steam at the rate of 120 tons/hr to run 500- Wb = Bagasse formation rate (tons/hr)
1500 kW steam turbines installed for cane preparation s = Weight of sugar (pol) per unit weight of bagasse
and milling units. The rate of steam consumption P2 = Net price of sugar (Rs/kg)
mainly depends upon the cane preparation index, C1 ;;; Ratio of actual steam consumption to the
pressure on top roller on milling train and imbibition theoretical steam consumption.
(addition of water to prepared cane). L =Length of rollers of the mills (ern).
d = Average diameter of rollers of mills (rn)
n == Angular speed of rollers of mills (rpm).

87
 Munir, Tahir & Sabir

Pm = Minimum value of hydraulic pressure on top

3
of "We" at different mills settings, 1300 rpm of turbine
Rollers of the mills (10 x kg/hr) speed and 14.25 fiber in cane is given in Table1. The
N =No. of mills in the tandem. value of "Dm" was taken as 0.07 as choking of juice
=
P3 Price of bagasse (Rs/ton) gutters and tanks were observed below this limit. The
=
We Cane crushing rates (tons/hr) value of "R2" was taken as 0.25. One ton of exhaust
2
steam at 120°C temperature and 1.0 Kg /cm pressure
Dm = Minimum level of weight of imbibition per unit
is required to evaporate one ton of water in each
weight of cane under operational conditions.
evaporator (Eisner, 1988) and there were four
=
R2 Weight of exhaust steam required to evaporate
evaporators in process house. The value of " R3" was
unit weight of water.
taken as 0.85
=
R3 Weight of live steam required to convert into unit By substituting the values of the co-efficient in the
weight of exhaust steam. objective function, it was formed in terms of I, P, D, w
=
1& m constants of the linear equation for index of and s. Then by substituting values of wand s against
cane preparation. each crushing rate from Table 1 and by simplifying the
An experiment was conducted using different crushing same, the following objective function equation in term
rates to derive relationship/equations for moisture and
of I, P and D was established.
pol in bagasse as a function of D, P and I.

Table 1: Values of Os, We' Wb, 1,,12 and C1 at different crushing rates

Mills Setting Crushing Rate Os We Wb 11 12 C1

(mm) (tcd) (Tons/hr) (Tons/hr) (Tons/hr)
3800 85 158.00 48.00 0.86 0.92 1.15
41.91
47.37 4200 92 175.00 53.05 0.86 0.90 1.30

5000 100 208.33 63.16 0.83 0.87 1.45

58.30
63.77 5400 106 225.00 68.21 0.80 0.85 1.60

For each crushing rate, data of moisture per unit =

Min Z 61383 X - 27387 1-69388 P - 309 D
weight of bagasse "w" and pol per unit weight of Development of Constraint Equations
bagasse "s" at different imbibition per unit weight of The following constraint equations were developed for
cane "D", hydraulic pressure on top rollers of the mills the prevailing condition of Pahrianwali Sugar Mill.
"P" and index of cane preparation "I" was recorded. A =
1) 0.1321 + 0.489 P - 0.423 D + w 0.670
software was used to develop linear equations for "w" =
2) 0.07131 + 0.100 P +0.055 D + s 0.124
and "s" in term of "I", " P" and "D".
3) s ~ 0.0212
The steam flow rates (Os) were measured for different
4) s::;0.0411
cane crushing rates at 0.07 imbibition per unit weight of
2 5) w ~ 0.4711
cane, 0.175 x 103 Kg ftm hydraulic pressure on top
6) w::; 0.54
rollers of the mills and least value of index of cane
preparation (see Table 1). These values of steam flow =
7) X 1
rates were taken at 14.25 % fiber in cane and 8) P ~ 0.175
2 9) P::; 0.225
maintaining 1.0 Kg /cm steam pressure at process
house. 10) I ~ 0.86
The value of "wm "was taken as 0.49 at each crushing 11) I ::;0.92
rate. The value of "P1"was taken as 1.36 rupees per 12) D ~ 0.07
pound. The value of "R/' was taken as 2.20. Detailed 13) D::; 0.27
explanation may be seen in Munir (1995). The values =
14) 41275 w- F 20226
of Wb & C1 for different cane crushing rates are given =
15) 92.82 1- Oi 74.25
in Table 1. The value of "P2" (after subtracting Excise =
16) 68.93 P - Op 12
Duty and other Levies by Government) was found to
be Rs.12.00/kg. Values of "L" and "d " were 204 cm
=
17) 41.96 D - Od 2.93
18) F::; 7168
and 0.9652 m respectively. The experiment was
19) Oi + o, + Od::; 43
conducted at 1300 rpm of turbine speed. As such, the
The constraint equation NO.1 and NO.2 at different
value of "n" is taken as 4.64. The value of "Pm was
2 crushing rates showed the combined effect of
taken as 0.175 x 103 kg/cm . The value of "N" was
taken as 4 as there were four mills in the tandem. The imbibition, hydraulic pressure and index of cane
value of" P3"was taken as RS.500 per ton. The values preparation on bagasse moisture and bagasse pol.

88

r.
 Optimization of milling performance

Constraint equation No 14,15,16 and 17 were the

Constraint equation NO.3 and constraint equation NO.4
relation developed for excessive steam consumption
are the lower and upper limits of pol per unit weight of
due to increasing levels of imbibition, hydraulic
bagasse respectively. The lower and upper limits of pol
pressure, index of cane preparation and imbibition at
per unit weight of bagasse were the minimum and
different crushing rates. The detail may be seen in
maximum values obtained against different crushing
Munir (1995).
rates and mills settings.
Constraint equation No. 18 was a limit for furnace oil.
Constraint equation No. 5 was the lower limit of
All the four burners of boilers can spray maximum
moisture per unit weight of bagasse. The lower limit of
moisture per unit weight of bagasse was the least 7168.1bs/hr.
Constraint equation NO.19 was the maximum steam
value obtained against different crushing rates as given
generating capacity of the boiler. The maximum steam
in Munir (1995).
generating capacity of both the boilers was 120 tons/hr
Constraint equation No. 6 was the upper limit of
with bagasse. It was observed during the session
bagasse moisture. It was observed that when moisture
1994-95 that the steam generating capacity could be
per unit weight of bagasse increases beyond 0.54,
enhanced to 130 tons/hr by using furnace oil under the
dumping of bagasse on the furnace grates in the form
existing condition of the plant. So the excessive steam
of heaps occurred due to an increase in bulk density of
required for level of imbibition, hydraulic pressure on
bagasse under the existing condition of air distribution
top rollers of the mills and index of cane preparation
system. The heap formation prevents the primary air
should be less than 130 - Qs. In mathematical form, it
supply through the dumping grates that ultimately
retarded combustion due to lack of oxygen. So the may be written as
upper limit for bagasse moisture was taken as 0.54.
o, + + o, o, ::;130 - o,
Where Qs is the steam consumption at imbibition 0.07
Constraint equation NO.8 was the lower limit for 3
weight of water per unit weight of bagasse, 0.175 x 10
hydraulic pressure. Its lower limit was taken as 0.175 x
Kg Icm2 of hydraulic pressure and least value of index
103 Kg Icm2for linear programming model.
Constraint equation NO.9 was the upper limit for of cane preparation (11) as given in Table 1.
hydraulic pressure. It was observed that by increasing
3 2 RESULTS AND DISCUSSION
hydraulic pressure more than 0.225 x 10 kg/cm , the
temperature of outlet water of brass bearings cooling
The objective function and all the constraint equations
system of the rollers of the mills was found to be very
were in linear form. As such, optimization of milling
high (850 C). The safe working levels of hydraulic
performance was done using linear programming
pressure on the top rollers of the mills was taken as
3 2 technique. A computer software "QSB" was used to
0.225 x 10 kg/cm .
solve LP model called "Quantitative System Business".
Constraint equation No. 10 was the lower limit of index
The output of LP model output of the QSB program at
of cane preparation. The values of lower limit of index
different crushing rate is given in Table 2.
of cane preparation (11) are given Table 1.
Constraint equation No 11 was the upper limit of index Table 2. Output of LP model for different cane
of cane preparation. The upper limit of index of cane crushing rates
preparation (12) was taken up to the temperature limit of Crushing Rates (tcd)
shredder bearings (850 C). So the upper value of index Variable 5400
3800 4200 5000
of cane preparation in this safe working level is given in 1.0000
X 1.0000 10000 1.0000
Table 1.
Constraint equation NO.12 was the lower limit of 0.9200 0.9000 0.8700 0.8500
I
imbibition per unit weight of cane. Its value was taken 0.2250 0.2250 0.2250 0.2250
P
as 0.07 0 0.2399 0.2221 0.1950 0.1377
Constraint equation No.13 was the upper limit of
0.0227 0.0292 0.0374 0.0492
imbibition. Its was observed that under the mill S
resource, imbibition more that 27% not only affected 0.5400 0.5400 0.5400 0.5400
W
the cane crushing rate but also created chocking 11.1444 10.4240 8.4525 6.6360
Q;
problems of mills Donnelley chutes due to slippage of
Qp 3.5092 3.8920 4.3470 4.7997
bagasse at the mill. As the variation in imbibition was
done at constant crushing rate in the experiment, so 7.1349 5.6589 6.9173 3.2415
Qd
the upper limit of imbibition was taken as 0.27 per unit F 2140 2221 2460 2540
weight of cane.
Constraint equation No.14 was the quantity of furnace Optimum 20500 26156 35790 47061
Value (Rs/hr)
oil used (F) in term of bagasse moisture.

89
 Munir, Tahir & Sabir

The optimum value of linear programming model for restricted the imbibition level at a point at which
different crushing rates (Table 2) indicate the minimum inefficient performance of boilers was faced due to
cost at each crushing rate for selected mill. The milling increase in moisture. It was concluded that the present

50000

-I-
-...c:
~'"
40000
'-'
.•.. 30000
'"0
U
20000
...~=
-...
~
10000

0
3800 4200 5000 5400
Crushing Rate (ted)

Fig. 1. Milling losses (Rs/hr) at different crushing rates

0.25

0.2

.•..= 0.15
.9
:E
:E 0.1
S
•••
0.05

0
3800 4200 5000 5400
Crushing Rate (ted)

Fig. 2. Imibition at different crushing rates.

cost (Rs/hr) increased with the increase in cane resources of the plant permits maximum possible
crushing rates as shown in Fig. 1. The optimum values imbibition on cane for better milling results.
of imbibition at different crushing rates reflected that The optimum value of hydraulic pressure at each
3 2
bagasse pol might be reduced even at the cost of crushing rate was found to be 0.225 x 10 Kg/cm . This
furnace oil used in boilers as well as extra steam was the maximum value of hydraulic pressure that can
consumption to evaporate the additional water is mixed be set under available resources of the plant. The
in evaporators under available resources of the plant. optimization model permitted the maximum pressure
It was also noted that the value of imbibition decreased on the layer of prepared cane during operation in safe
as the crushing rate increased under available working levels to reduce the sugar and moisture in
resources of the plant as shown in Fig. 2. It was due to bagasse to a minimum level.
the reason that the constraint equation (w:o; 0.54)

90
 Optimization of milling performance

60000

I:' 57000
.c
--'"
~
~-
E
54000

51000
-
~

Z
~
48000

45000
3800 4200 5000 5400
Crushing Rate(tcd)

Fig. 3. Net profit (Rs/hr) against different crushing rates

High values of index of cane preparation obtained from in Fig. 3. It may be found that as the crushing rates
LP model showed that cost of pol lost in bagasse and increased, the net profit (Pn) also increased up to a
cost of steam used due to increase in bagasse crushing rate of 5000 tcd and then decreased with the
moisture were much greater as compared to the cost increase in cane crushing rate. The crushing rate
of steam required in increasing the level of index of (5000 tcd) gave maximum profit and was considered
cane preparation. So maximum possible disintegration the optimum crushing rate under available resources of
of the rind of cane was suggested at milling section. the mill.

Table 3. Cost of Cane, Cost of Sugar, Total Losses, and Net Profit against different crushing Rates.

Mills Cost of Cost of Total Cost L* Ln Net Profit

Crushing (Pn)
settings Cane Sugar (Lt)
Rate (tcd) (Rs/hr) (Rs/hr) (Rs/hr) (Rs/hr) (Rs/hr)
(mm) (Rs/hr)
3800 95446 152000 20500 14968 5532 51022
41.91
4200 105493 168000 26156 16544 9612 52895
47.37
5000 125587 200000 35796 19695 16101 58312
58.30
5400 135634 216000 47061 21271 25790 54576
63.77

CONCLUSIONS
Economic Analysis
It was observed that as the crushing rate increased 1. As the crushing rate increased from 3800 to
from 3800 to 5800 tons/day by increasing the mills 5400 tcd, the minimum possible milling losses
settings, the sugar production increased but at the also increased from 20500 to 47061 Rs/hr.
same time total cost (minimum value of objective under available resources of the sugar mill.
function) also increased (see Table 3). The net losses
2. The optimum value of imbibition on cane
(Ln) were calculated by subtracting the losses (L*) of
decreased as the crushing rate increased
2.60% bagasse pol (Average figure of bagasse pol at
under available resources of the mill.
8.0% recovery) from the total losses "U" against each
crushing rate. Optimum crushing capacity was 3. The optimum value of index of cane
calculated by subtracting the cost of sugar cane at the preparation and hydraulic pressure on top
rate of 22.50 rupees per maund and net losses from rollers of the mills were found to be 0.87 and
2
the total sugar production at 8.00% recovery. The 0.225 x1 03 kg/cm respectively.
graph between crushing rates and net profit is shown

91

I
 Munir, Tahir & Sabir

4. The optimum value of cane crushing rate was Eisner, L.R. 1988. Mathematical calculation of a cane
found to be 5000 tcd under existing condition sugar factory. Fletcher and Stewart Limited,
of the ~Iant which gave maximum profit of Rs. England.
58312/hr.
5. Results of linear programming model showed Hugot, E. 1986. Hand book of cane sugar engineering,
rd
that the bagasse pol is a major milling loss that 3 edition. St. Lucia, Brisbane ,Australia.
must be reduced even at the cost of other
Munir, A 1995. Optimization of a sugar plant by using
milling losses.
linear programming technique M. Sc. Thesis,
Faculty of Agri. Engineering & Technolgy, UAF.
REFERENCES

Azeem, AA 1976. A critical Look in the milling

performance of Pakistan sugar industry.
th
Proceedings of 13 annual convention of Pakistan
Society of Cane Sugar Technologists.

92

View publication stats