Asia Pacific Journal of Chemical Engineering, 3 (2): 230-235, 2008

DOI: http://dx.doi.org/ 10.1002/apj.133

Correlations for the Grindability of the Ball Mill
As a Measure of Its Performance

By
A. Sahoo* and G. K. Roy
Chemical Engineering Department, National Institute of Technology, Rourkela-769008, India

Abstract

Ball mill is vital equipment in industries viz. mineral dressing, ore processing, fertilizers,
food and diary, pharmaceuticals and many others. The present work involves a meticulous
study of the effect of the various parameters on the performance of a ball mill. The
parameters studied in this work are particle size, number of balls, time of grinding, particle
density, and speed of the ball mill (rpm). An attempt has been made to develop correlation
for the performance of the ball mill by correlating these variables with the grindability on the
basis of dimensional analysis approach as well as fractional factorial design method. It is
observed that these parameters influence the performance of the ball mill significantly. The
performance of a ball mill is measured with reference to the quantity of undersize or fines
(amount of grounded material passing through120 mesh screen) per revolution of the mill,
collected for variation of each of the individual parameters.
Finally the calculated values of the fines in terms of the grindability of the mill obtained
through the correlations by both the above-mentioned methods are compared with the
experimental values thereby justifying the analysis with reasonable deviations. Thus the
developed correlations can be applied to know the grindability of the Ball Mill over a wide
range of parameters.
Keywords
Grindability, Ball mil, Dimensional Analysis approach, and Fractional Factorial Design.

…………………………………………………………………………………………………..

*Corresponding author; E-Mail: abantisahoo@gmail.com
Tele: +91-661-2462258(O), 2463258 (R);

1. Introduction
In almost all the metallurgical processes, crushing/grinding is considered to be an inseparable
unit operation which is adopted at the very first stage of a series of other unit operations
and/or unit processes. Size reduction of solids is done almost invariably. Ball mill finds its
application in many industries like abrasives, animal products, brewing industry, chemical,
confectionery, food processing, fuel preparation, metal powder, mineral preparation, paint
preparation, paper, pigments for color industry, abrasives for grinding, plastics, printing ink,
rubber, textiles, sintering, cement and limestone, powders for the detergent industry,
pulverized coal for power generation, refractory materials for investment casting, dry powder
opacifiers for ceramics industry, pharmaceuticals, mineral preparation, refractory materials
for investment casting, tungsten powder and dry lubricants, carbon black for rubber, charcoal
for briquetting and others.
Mainly because of their simple construction and application ball milling is a wide spread
milling technology, particularly in mining. Various parameters viz. particle size, number of
balls, time of grinding, particle density and speed of the ball mill (rpm) have been considered
for the present work to determine the performance of the ball mill.

2. Literature
Ball mills are cylindrical or conical shell rotating about a horizontal axis, partially filled with
a grinding medium such as natural flint pebbles, ceramic pellets or metallic balls. The
common classification of ball mill is any type of mill in which mild steel or iron balls are
used. In most cases the cylinder of such mill is made of alloy steel or some special type of
lining. The material to be ground is added so that it is slightly more than fills of the voids
between the balls (with a maximum limit of number of balls not exceeding 50% of the mill
capacity for proper grinding). The shell is rotated at a speed, which will cause the pellets to
cascade, thus reducing particle sizes by impact. It has been proposed that in the plastics
industry the term ball mill be reserved for metallic grinding media, and the term pebble mill

for non-metallic grinding media. There are two different methods of comminution:
autogeneous comminution where the material is pure and heterogeneous comminution where
the material is mixed with heavy spheres from steel of typical diameter of several centimeters
to increase efficiency. There exists much experimental knowledge on the operation
mechanism of ball mills and on the comminution of granular matter in these mills. To
increase efficiency of the mill one would have to tune the rotation velocity so that the
average collision velocity becomes maximum. In this context attempt has been made for a
meticulous study of the effect of the various system parameters on the performance of ball
mills.
2.1 Terms Associaed with Ball Mill Operation
2.1.1 Grindability (G):
Grindability is the number of net grams of screen undersized product per revolution [1]. The
chief purpose of study of the grindability is to evaluate the size and type of mill needed to
produce a specified tonnage and the power requirement for grinding. Detailed prediction of
grinding rate and product size distribution from mills await the development of a simulation
based on physics of fracture.
N. Magdalinovic [2] has stated simplified procedure for a rapid determination of the work
index by just two grinding tests. The applicability of the simplified procedure has been
proved on samples of Cu ore, andesite and limestone. The result by this method was not more
than 7% from the values obtained in the standard Bond test.
T. Yalcin et al. [3] investigated the effect of various parameters on the grindability of pure
Sulfur and used the obtained grinding data to establish mathematical models and set up a
computer simulation program. The established mathematical model is as shown below.
(1)

Where, y is cumulative percent passing size d, d
50
is the 50% passing size, n is distribution
constant, and k is a correction factor. The n values ranged from 0.84 to 1.84, and k values
from 0.95 to 1.00.
By using the Bond method of grindability, H. Ipek et al. [4] have observed that less specific
energy input is required in separate grinding of ceramic raw materials than grinding them in
admixtures. They have stated that the Bond work indices of the admixtures containing softer
component are greater than the weighted average of the work indices of the individual
components in the mixture.
2.1.2 Critical Speed:
If the peripheral speed of the mill is very high, it begins to act like a centrifuge and the balls
do not fall back, but stay on the perimeter of the mill and that point is called the "Critical
Speed” (n
c
).This phenomenon is called centrifuging. Ball mills usually operate at 65% to
75% of the critical speed. The critical speed is calculated as under [5].

r R
g
2
1
n
c
÷ t
= (2)

2.1.3 Work Index:
Work index is defined as the gross energy required in kilowatt-hours per ton of feed needed
to reduce a very large feed to such a size that 80% of the undersize passes through 100-µm
screen [5]. The expression for this is as given below
|
|
.
|

\
|
÷ × =
5 . 0
F
5 . 0
P
i
d
1
d
1
W 3162 . 0 W (3)
Deniz and Ozdag [6] have investigated the effect of elastic parameters on grinding and
examined the relationship between them. The most widely known measure of grindability is
Bond’s work index which is defined as the resistance of the material to grinding. The
standard equation used by them for the ball mill work index (Bond work index) is as follows.

( ) ( ) | |
80 80
82 . 0 23 . 0
/ 10 / 10
5 . 44
1 . 1
F P G P
W
bg i
i
÷
× =
(4)

In designing and optimizing a milling circuit using Bond Ball Mill Work Index[7], the
following equations are used (Bond 1961).
|
|
.
|

\
|
÷ =
80 80
1 1
10
F P
W W
i
And W T P * = (5)
Based on this equation it is possible to calculate, for example, the specific energy
requirement for a given grinding duty, BBMWI, feed size and required product size. It is then
possible to determine the size of mill required based on throughput and therefore the motor
power.
2.2 Factors Affecting Size of Product from Ball Mill
It is important to fix the point where the charge, as it is carried upward, breaks away from the
periphery of the Mill. This is called as “break point” or “angle of break” because it is
measured in degrees. It is measured up the periphery of the Mill from the horizontal [8].
There are four factors affecting the angle of break:
1. Speed of Mill
2. Amount of grinding media
3. Amount of material
4. Consistency or viscosity (for wet grinding)
2.3 Fractional Factorial Design [9]:
Full and fractional Factorial Design analysis is common in designed experiments for
engineering and scientific applications. In many cases, it is required to consider the factors
affecting the production process at two levels. The experimenter would like to determine
whether any of these changes affect the results of the production process. The most intuitive
approach to study these factors would be to vary the factors of interest in a full factorial
design, that is, to try all possible combinations of settings.
In statistics, fractional factorial designs are experimental designs consisting of a carefully
chosen subset (fraction) of the experimental runs of a full factorial design. The subset is
chosen so as to exploit the sparsity-of-effects principle to expose information about the most

important features of the problem studied, while using a fraction of the effort of a full
factorial design in terms of experimental runs and resources.
Fractional designs are expressed using the notation l
k − p
, where l is the number of levels of
each factor investigated, k is the number of factors investigated, and p describes the size of
the fraction of the full factorial used. Formally, p is the number of generators, assignments as
to which effects or interactions are confounded, i.e., cannot be estimated independently of
each other (see below). A design with p such generators is a 1/(l
p
) fraction of the full factorial
design. For example, a 2
5 − 2
design is 1/4 of a two level, five factor factorial design. Rather
than the 32 runs that would be required for the full 2
5
factorial experiment, this experiment
requires only eight runs.
If four or five factors are involved, the complete factorial might involve more than a practical
number of experiments. A 2
5
factorial would require 32 experiments. By careful selection of
the experimental conditions it is possible with only a fraction of the total experiments
required for the complete factorial to determine the main effects by aliasing them with the
higher order interactions which are usually not significant.
The eight experiments required for a complete three factor, two level factorial can be used to
determine the change required in four, five or under ideal conditions, even in seven
experimental variables to obtain the maximum change in the response variable. As (k-p)
factorial design is set up and the p factors not included in the complete 2
k-p
factorial are
aliased with one of the higher order interactions to form a generating contrast.
In practice, one rarely encounters l > 2 levels in fractional factorial designs, since response
surface methodology is a much more experimentally efficient way to determine the
relationship between the experimental response and factors at multiple levels. In addition, the
methodology to generate such designs for more than two levels is much more cumbersome.
The levels of a factor are commonly coded as +1 for the higher level, and -1 for the lower
level. For a three-level factor, the intermediate value is coded as 0.
With two cube (2
3
) Factorial Design Analysis, the correlation will be represented in the
following form.

ABC a BC a AC a AB a C a B a A a a Y
123 23 13 12 3 2 1 0 jv i
+ + + + + + + = (6)
3. Experimental
A ball mill of 36.6 cm diameter and 50 cm length has been used in the laboratory for
experimentation. The material of construction of the grinding media used is mild steel. The
steel balls each of size 5.41 cm in diameter and density 7.85kg/m
3
were selected for the
experiments. The mill was made to revolve at different speeds to grind various materials like
dolomite, manganese, iron ore, and limestone. Exhaustive study was carried out with the
dolomite material. The various system parameters (viz. particle size, material density, speed
of the mill, time of grinding and the number of balls) were considered to study their effects
on the performance of the ball mill. Scope of the experiment is given in Table 1 and the
experimental set up is shown in Fig. 1. The amounts of undersize or fine were found out by
sieving with the 120-mesh screen. Each time 1.0 kg of material was taken as feed material for
running the ball mill.
4. Results
4.1 Development of the Correlations
In the present work, attempt has been made to develop an expression correlating the
grindability of the ball mill with the various system parameters by means of dimensional
analysis and the fractional factorial design methods. The correlation plot for the former is
shown in Fig. 2 and the developed correlation is given as under.

( )

|
|
.
|

\
|
|
|
.
|

\
|
|
|
.
|

\
|
p
p
|
|
.
|

\
|
× =
÷ ÷
059 . 0
073 . 0
c
897 . 1
c
040 . 0
b
F
054 . 0
b
F
N
t
t
n
n
d
d
0402 . 1 G (7)
In case of the fractional factorial design method, a 2
5-2
fractional factorial design was
employed to determine the path of the steepest ascent for five factors with eight experiments.
The design involved the calculation of the grindability of the ball mill and the variables with

the upper and lower levels are shown in Table 2(A). The actual experiment design with the
calculation of the mean effects is shown in Table 2(B).
To study the effect of various parameters on the grindability of the mill the following
correlations has been developed using fractional factorial design by 2
5-2
approach.
BC 035 . 0 AC 09 . 0 E 11 . 0 D 19 . 0 C 115 . 0 B 165 . 0 A 19 . 0 0925 . 3 G ÷ ÷ ÷ ÷ ÷ + ÷ = (8)
5. Discussion
The critical speed of the ball mill was calculated to be 76rpm. It is observed that the
grindability of the ball mill or in other words the amount of fines increase with respect to the
overall effect of increase in number of balls, time of grinding, speed (rpm) of the ball mill.
But with increase in particle size and particle density the grindability of the mill decreases;
whereas it increases with the increase in speed of the ball mill, time of grinding and the
number of balls. Therefore the grinding is not so satisfactory for high density particles like
iron ores. The equations obtained are in accordance with the experimental observations.
From the above expressions it is clear that the factor, which affects the most, is the speed. So
for optimizing the grinding process, as compared to other parameters speed has to be
controlled to a maximum extent. The effect of various system parameters on grindability has
been shown in Fig. 2. The calculated values of the grindability from both the approaches
obtained through the developed correlations have been compared with the respective
experimental values with mean and standard deviations of -4.38 and 18.75 for D.A.-approach
and 3.54 and 11.56 for the F.F.D.-approach respectively as shown in Fig. 3. It is found that
the calculated values agree well with the experimental values.
6. Conclusion
In today’s industrial scenario, ball mill is widely used in multifarious industries as size
reduction process is energy inefficient, it is necessary to optimize the operation so as to
reduce cost to some extent. As it has been explicitly seen that the parameters influencing the
performance of ball mill cannot be ignored, the expression correlating all these variables can

be considerably used to optimize the operation of a ball mill in general over a wide range of
parameters. Based on the results one can comfortably determine the ranges of various
parameters to be used for a specific process. The future aspect of this work can be extended
to bond index calculation where the power consumption will indicate directly about the cost
benefit too.
Nomenclature
a
0
, a
1
, a
2
, a
3
,a
12
, a
13
, a
23
, a
123
: Mean effects for different factors
A, B, C,D,E : Factors for factorial design
d : Diameter (size) in mm
F
80 :
80% passing size of feed µm
g : Gravitational constant, 981gm/cm
2

G : Grindability of the mill g/rev
Gbg : Bond’s standard ball mill grindability g/rev
n : Speed of ball mill, rpm
N : Number of balls
P : Power draw kW
Pi : screen size for performing the test µm
P
80
: 80% passing size of product µm
r : radius of grinding balls cm
R : Radius of the ball mill cm
t : Time of grinding min
T : Throughput of new feed t/h
W : Work input kW-hr/t
Wi : Work index of the material kW-hr/t
p : Density of material kg/m
3

Subscripts:
b : for the grinding balls
c : for critical condition
F : for feed particles
p : for product particles
Abbreviations:
D.A. : Dimensional analysis
F.F.D. : Fractional Factorial Design

References
[1]. R. H. Perry and C. H. Chilton, Chemical Engineers’ Handbook , (5
th
edition),
McGraw-Hill, p 8-25.
[2] N. Magdalinovic, Calculation of energy required for grinding in a ball mill,
International Journal of Mineral Processing, Vol. 25, (1-2), January 1989, p 41-46
[3] T. Yalcin et al. A simulation study of sulphur grindability in a batch ball mill
Powder Technology, Vol. 146, Issue 3, 8 September 2004, p 193-199.
[4] H. Ipek et al. The bond work index of mixture of ceramic raw materials, Mineral
Engineering, Vol. 18, (2005), p 981-983.
[5] W. L. McCabe, J. C. Smith, and P. Harriot, Unit Operation in Chemical Engineering,
(Fifth Edition), McGraw-Hill, Inc, Singapore, 1993, p-980
[6] Deniz and Ozdag, A new approach to Bond grindability and work index: dynamic
elastic parameters, Mineral Engineering, Vol. 16, Issue-3, March-2003, p 211-217
[7] http://www.jktech.com.au/Product_Services/Comminution-Testing/Bond-Ball-Mill
[8]. W. L. Badger and J. T. Banchero, Introduction to Chemical Engineering, (28
th

Printing) McGraw-Hill, Inc, Singapore, 1985, p-677
[9]. W. Volk, (Second Edition), ‘Applied Statistics for Engineers’ McGraw-
Hill,Inc,USA, 1969, p-237 and p-336








Figure Caption:

Fig.-1: Experimental set-up
Fig.-2: Correlation plot for grindability against system parameters by D.A. approach
Fig.-3: Comparison plot of calculated values of grindability by both the approaches against
the experimental values


















Table-1: Scope of the experiment

Sl.No. Materials
Used
Density, p
F

(Kg/m
3
)
Particle
size, d
F
,
mm
No. of balls,
N
Speed of ball
mill, n, rpm
Time of
grinding, t,
min
1 Dolomite 2900 4.76 12 58 10
2 Manganese 4150 4.76 12 58 10
3 Iron ore 5200 4.76 12 58 10
4 Lime stone 2350 4.76 12 58 10
5 Dolomite 2900 9.525 12 58 10
6 Dolomite 2900 4.05 12 58 10
7 Dolomite 2900 1.85 12 58 10
8 Dolomite 2900 4.76 8 58 10
9 Dolomite 2900 4.76 16 58 10
10 Dolomite 2900 4.76 20 58 10
11 Dolomite 2900 4.76 12 52 10
12 Dolomite 2900 4.76 12 62 10
13 Dolomite 2900 4.76 12 66 10
14 Dolomite 2900 4.76 12 58 15
15 Dolomite 2900 4.76 12 58 20
16 Dolomite 2900 4.76 12 58 30
17 Dolomite 2900 4.76 12 58 60

























Table-2(A): A 2
5-2
Fractional Factorial Design for the actual experiment with the upper and
lower levels of the variables

Experiment A B C D=ABC -E=AB AC BC
1 - - - - + + +
a (de) + - - + - - +
b(de) - + - + - + -
ab + + - - + - -
c(d) - - + + + - -
ac(e) + - + - - + -
bc(e) - + + - - - +
abc(d) + + + + + + +







Table-2(B): Results and calculation of the mean effects for 2
5-2
Fractional Factorial Design

Experiment Experimental
Results
A-
effect
B-
effect
C-
effect
D-
effect
E-
effect
Main effects
1
3.25
-1 -1 -1 -1 +1 Overall mean 3.0925
Mean A-effect -0.19
Mean B-effect 0.165
Mean C-effect -0.119
Mean D-effect -0.19
Mean E-effect -0.11
Mean AC- effect -0.09
Mean BC- effect -0.035
a (de)
2.85
+1 -1 -1 +1 -1
b(de)
3.15
-1 +1 -1 +1 -1
ab
3.35
+1 +1 -1 -1 +1
c(d)
3.07
-1 -1 +1 +1 +1
ac(e)
2.87
+1 -1 +1 -1 -1
bc(e)
3.28
-1 +1 +1 -1 -1
abc(d)
2.92
+1 +1 +1 +1 +1













y = 4.4271x
0.837
1
10
0.1 1 10
(d
p
/d
b
)
-0.064
(rhos/rhob)
-0.0478
(n/n
c
)
2.2662
(T/T
c
)
0.0877
(N)
0.0703
G
-
e
x
p
Series1
Series2
Series3
Series4
Series5

Fig-2: Correlation plot for grindability of the ball mill against system parameters


0
1
2
3
4
5
6
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
G-exp
G
-
c
a
l
G - Exp
G-Cal_D.A.
G-cal_F.F.D.


Fig.-3: Comparison of calculated values of grindability by both the approaches against the
experimental values



Literature Ball mills are cylindrical or conical shell rotating about a horizontal axis. charcoal for briquetting and others. textiles. partially filled with a grinding medium such as natural flint pebbles. particle size. 1. which will cause the pellets to cascade. plastics. pulverized coal for power generation. carbon black for rubber. The common classification of ball mill is any type of mill in which mild steel or iron balls are used. cement and limestone. animal products. crushing/grinding is considered to be an inseparable unit operation which is adopted at the very first stage of a series of other unit operations and/or unit processes. mineral preparation. ceramic pellets or metallic balls. pigments for color industry. Introduction In almost all the metallurgical processes. powders for the detergent industry. metal powder. Ball mill finds its application in many industries like abrasives. number of balls. refractory materials for investment casting. refractory materials for investment casting. printing ink. The material to be ground is added so that it is slightly more than fills of the voids between the balls (with a maximum limit of number of balls not exceeding 50% of the mill capacity for proper grinding). E-Mail: abantisahoo@gmail.*Corresponding author. It has been proposed that in the plastics industry the term ball mill be reserved for metallic grinding media. Size reduction of solids is done almost invariably. rubber. paper.com Tele: +91-661-2462258(O). abrasives for grinding. food processing. The shell is rotated at a speed. time of grinding. particularly in mining. and the term pebble mill . In most cases the cylinder of such mill is made of alloy steel or some special type of lining. pharmaceuticals. brewing industry. Mainly because of their simple construction and application ball milling is a wide spread milling technology. tungsten powder and dry lubricants. chemical. confectionery. thus reducing particle sizes by impact. Various parameters viz. 2463258 (R). paint preparation. dry powder opacifiers for ceramics industry. 2. mineral preparation. particle density and speed of the ball mill (rpm) have been considered for the present work to determine the performance of the ball mill. fuel preparation. sintering.

1 Terms Associaed with Ball Mill Operation 2.for non-metallic grinding media. The applicability of the simplified procedure has been proved on samples of Cu ore. (1) . [3] investigated the effect of various parameters on the grindability of pure Sulfur and used the obtained grinding data to establish mathematical models and set up a computer simulation program. The result by this method was not more than 7% from the values obtained in the standard Bond test.1. There are two different methods of comminution: autogeneous comminution where the material is pure and heterogeneous comminution where the material is mixed with heavy spheres from steel of typical diameter of several centimeters to increase efficiency. 2. The established mathematical model is as shown below. The chief purpose of study of the grindability is to evaluate the size and type of mill needed to produce a specified tonnage and the power requirement for grinding. There exists much experimental knowledge on the operation mechanism of ball mills and on the comminution of granular matter in these mills. In this context attempt has been made for a meticulous study of the effect of the various system parameters on the performance of ball mills. andesite and limestone. To increase efficiency of the mill one would have to tune the rotation velocity so that the average collision velocity becomes maximum. Detailed prediction of grinding rate and product size distribution from mills await the development of a simulation based on physics of fracture. N. Magdalinovic [2] has stated simplified procedure for a rapid determination of the work index by just two grinding tests.1 Grindability (G): Grindability is the number of net grams of screen undersized product per revolution [1]. T. Yalcin et al.

1 0. They have stated that the Bond work indices of the admixtures containing softer component are greater than the weighted average of the work indices of the individual components in the mixture. 2.95 to 1.1. P0. By using the Bond method of grindability. and k is a correction factor.2 Critical Speed: If the peripheral speed of the mill is very high.84. The n values ranged from 0.3162  Wi  0. d50 is the 50% passing size. [4] have observed that less specific energy input is required in separate grinding of ceramic raw materials than grinding them in admixtures. y is cumulative percent passing size d. H. The critical speed is calculated as under [5]. but stay on the perimeter of the mill and that point is called the "Critical Speed” (nc). it begins to act like a centrifuge and the balls do not fall back.5 10 / P  10 / F80 80   (4) .This phenomenon is called centrifuging.5  0. nc  1 g 2 R  r (2) 2.Where. Wi  1.00. The expression for this is as given below  1 1  W  0. n is distribution constant. Ball mills usually operate at 65% to 75% of the critical speed.3 Work Index: Work index is defined as the gross energy required in kilowatt-hours per ton of feed needed to reduce a very large feed to such a size that 80% of the undersize passes through 100-µm screen [5].84 to 1.1. The standard equation used by them for the ball mill work index (Bond work index) is as follows.23Gbg82 i  44.5  d dF   P  (3) Deniz and Ozdag [6] have investigated the effect of elastic parameters on grinding and examined the relationship between them. and k values from 0. The most widely known measure of grindability is Bond’s work index which is defined as the resistance of the material to grinding. Ipek et al.

In statistics.2 Factors Affecting Size of Product from Ball Mill It is important to fix the point where the charge. fractional factorial designs are experimental designs consisting of a carefully chosen subset (fraction) of the experimental runs of a full factorial design. Amount of grinding media 3. BBMWI. The experimenter would like to determine whether any of these changes affect the results of the production process. Speed of Mill 2. to try all possible combinations of settings. breaks away from the periphery of the Mill. The most intuitive approach to study these factors would be to vary the factors of interest in a full factorial design. the specific energy requirement for a given grinding duty.3 Fractional Factorial Design [9]: Full and fractional Factorial Design analysis is common in designed experiments for engineering and scientific applications. feed size and required product size. that is. as it is carried upward. This is called as “break point” or “angle of break” because it is measured in degrees.In designing and optimizing a milling circuit using Bond Ball Mill Work Index[7]. It is then possible to determine the size of mill required based on throughput and therefore the motor power. Amount of material 4. There are four factors affecting the angle of break: 1. In many cases. Consistency or viscosity (for wet grinding) 2. The subset is chosen so as to exploit the sparsity-of-effects principle to expose information about the most . the following equations are used (Bond 1961). for example.  1 1   W  10Wi    P F80   80  And P  T *W (5) Based on this equation it is possible to calculate. it is required to consider the factors affecting the production process at two levels. It is measured up the periphery of the Mill from the horizontal [8]. 2.

i. where l is the number of levels of each factor investigated. As (k-p) factorial design is set up and the p factors not included in the complete 2k-p factorial are aliased with one of the higher order interactions to form a generating contrast. With two cube (23) Factorial Design Analysis. In practice. the complete factorial might involve more than a practical number of experiments. the methodology to generate such designs for more than two levels is much more cumbersome.important features of the problem studied. In addition. five or under ideal conditions. a 25 − 2 design is 1/4 of a two level. Rather than the 32 runs that would be required for the full 25 factorial experiment. The eight experiments required for a complete three factor. two level factorial can be used to determine the change required in four.e. cannot be estimated independently of each other (see below). The levels of a factor are commonly coded as +1 for the higher level. Fractional designs are expressed using the notation lk − p. assignments as to which effects or interactions are confounded. and p describes the size of the fraction of the full factorial used. while using a fraction of the effort of a full factorial design in terms of experimental runs and resources. k is the number of factors investigated. A 25 factorial would require 32 experiments. Formally. For example. and -1 for the lower level. By careful selection of the experimental conditions it is possible with only a fraction of the total experiments required for the complete factorial to determine the main effects by aliasing them with the higher order interactions which are usually not significant. A design with p such generators is a 1/(lp) fraction of the full factorial design. five factor factorial design. p is the number of generators. the correlation will be represented in the following form. even in seven experimental variables to obtain the maximum change in the response variable. one rarely encounters l > 2 levels in fractional factorial designs. For a three-level factor.. this experiment requires only eight runs. If four or five factors are involved. the intermediate value is coded as 0. . since response surface methodology is a much more experimentally efficient way to determine the relationship between the experimental response and factors at multiple levels.

a 25-2 fractional factorial design was employed to determine the path of the steepest ascent for five factors with eight experiments. The correlation plot for the former is shown in Fig. The design involved the calculation of the grindability of the ball mill and the variables with . 1.0 kg of material was taken as feed material for running the ball mill. attempt has been made to develop an expression correlating the grindability of the ball mill with the various system parameters by means of dimensional analysis and the fractional factorial design methods. 897  t  t  c     0 .  d G  1 .1 Development of the Correlations In the present work. The material of construction of the grinding media used is mild steel. 054  F    b      0 . material density. The amounts of undersize or fine were found out by sieving with the 120-mesh screen.Yi jv  a 0  a 1A  a 2 B  a 3 C  a 12 AB  a 13 AC  a 23 BC  a 123 ABC 3. Experimental (6) A ball mill of 36. The steel balls each of size 5. The mill was made to revolve at different speeds to grind various materials like dolomite. speed of the mill.0402    F   d b       0 . 040  n  n  c     1 . Exhaustive study was carried out with the dolomite material. and limestone.41 cm in diameter and density 7. 4. time of grinding and the number of balls) were considered to study their effects on the performance of the ball mill. The various system parameters (viz. manganese. 2 and the developed correlation is given as under. Scope of the experiment is given in Table 1 and the experimental set up is shown in Fig.059     (7) In case of the fractional factorial design method.85kg/m3 were selected for the experiments.073 N 0. Results 4. particle size.6 cm diameter and 50 cm length has been used in the laboratory for experimentation. iron ore. Each time 1.

0925 0. as compared to other parameters speed has to be controlled to a maximum extent. time of grinding.F. Conclusion In today’s industrial scenario.165B  0.-approach and 3.56 for the F. the expression correlating all these variables can . 6. The actual experiment design with the calculation of the mean effects is shown in Table 2(B).-approach respectively as shown in Fig.035BC 5. whereas it increases with the increase in speed of the ball mill. The effect of various system parameters on grindability has been shown in Fig.09AC  0.38 and 18. G  3. It is observed that the grindability of the ball mill or in other words the amount of fines increase with respect to the overall effect of increase in number of balls.the upper and lower levels are shown in Table 2(A). it is necessary to optimize the operation so as to reduce cost to some extent.A. But with increase in particle size and particle density the grindability of the mill decreases. Discussion (8) The critical speed of the ball mill was calculated to be 76rpm. is the speed. It is found that the calculated values agree well with the experimental values. So for optimizing the grinding process. Therefore the grinding is not so satisfactory for high density particles like iron ores.D. The equations obtained are in accordance with the experimental observations. time of grinding and the number of balls. ball mill is widely used in multifarious industries as size reduction process is energy inefficient. speed (rpm) of the ball mill. The calculated values of the grindability from both the approaches obtained through the developed correlations have been compared with the respective experimental values with mean and standard deviations of -4. To study the effect of various parameters on the grindability of the mill the following correlations has been developed using fractional factorial design by 2 5-2 approach. As it has been explicitly seen that the parameters influencing the performance of ball mill cannot be ignored.11E  0.19A  0.115C  0. 2.54 and 11. From the above expressions it is clear that the factor.19D  0.75 for D. 3. which affects the most.

a13. a23. : Dimensional analysis Fractional Factorial Design F. Based on the results one can comfortably determine the ranges of various parameters to be used for a specific process.D. a123 A. a2. rpm Number of balls Power draw screen size for performing the test 80% passing size of product radius of grinding balls Radius of the ball mill Time of grinding Throughput of new feed Work input Work index of the material Density of material µm : : : : : : : : : : : : : : : g/rev g/rev kW µm µm cm cm min t/h kW-hr/t kW-hr/t kg/m3 Abbreviations: D. B.A. : . The future aspect of this work can be extended to bond index calculation where the power consumption will indicate directly about the cost benefit too.E d F80 g G Gbg n N P Pi P80 r R t T W Wi  Subscripts: b c F p : : : : for the grinding balls for critical condition for feed particles for product particles : : : : Mean effects for different factors Factors for factorial design Diameter (size) in mm 80% passing size of feed Gravitational constant. a3. a1.a12. Nomenclature a0. C.F.D.be considerably used to optimize the operation of a ball mill in general over a wide range of parameters. 981gm/cm2 Grindability of the mill Bond’s standard ball mill grindability Speed of ball mill.

Mineral Engineering. Vol. Unit Operation in Chemical Engineering. L. Chemical Engineers’ Handbook . (1-2). The bond work index of mixture of ceramic raw materials. p-980 [6] Deniz and Ozdag. L. 1985. Banchero. Inc. Harriot. Badger and J.Inc. approach Fig. 16. H. Magdalinovic. W. p 8-25. (Fifth Edition). A new approach to Bond grindability and work index: dynamic elastic parameters.References [1]. Perry and C. Singapore. C. R. International Journal of Mineral Processing. 8 September 2004.-1: Experimental set-up Fig. ‘Applied Statistics for Engineers’ McGrawHill. Introduction to Chemical Engineering. p 193-199. http://www. (28th Printing) McGraw-Hill.USA. Issue-3. Vol.-3: Comparison plot of calculated values of grindability by both the approaches against the experimental values . Smith. March-2003. and P. Mineral Engineering. Calculation of energy required for grinding in a ball mill. H. (2005). Inc. 1993. [4] H. p 41-46 [3] T. Issue 3. p-237 and p-336 Figure Caption: Fig. January 1989. Vol. (Second Edition). 146.jktech. Yalcin et al. Volk.-2: Correlation plot for grindability against system parameters by D. Vol. McCabe. [2] N. [5] W. J. (5th edition).com. Ipek et al. Chilton. 18. A simulation study of sulphur grindability in a batch ball mill Powder Technology. McGraw-Hill. p-677 [9].A.au/Product_Services/Comminution-Testing/Bond-Ball-Mill W. 25. T. Singapore. 1969. McGraw-Hill. p 981-983. p 211-217 [7] [8].

dF.76 4.76 4.76 4. t.76 9.No. F (Kg/m3) 2900 4150 5200 2350 2900 2900 2900 2900 2900 2900 2900 2900 2900 2900 2900 2900 2900 Particle size.76 4. N 12 12 12 12 12 12 12 8 16 20 12 12 12 12 12 12 12 Speed of ball mill. rpm 58 58 58 58 58 58 58 58 58 58 52 62 66 58 58 58 58 Time of grinding. Materials Used Dolomite Manganese Iron ore Lime stone Dolomite Dolomite Dolomite Dolomite Dolomite Dolomite Dolomite Dolomite Dolomite Dolomite Dolomite Dolomite Dolomite Density. of balls.76 4. min 10 10 10 10 10 10 10 10 10 10 10 10 10 15 20 30 60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 .76 4.525 4.76 4.76 4. mm 4.76 4.76 4.05 1.76 No.Table-1: Scope of the experiment Sl.76 4.85 4.76 4. n.

87 3.119 Mean D-effect -0.Table-2(A): A 25-2 Fractional Factorial Design for the actual experiment with the upper and lower levels of the variables Experiment A 1 a (de) + b(de) ab + c(d) ac(e) + bc(e) abc(d) + B + + + + C + + + + D=ABC -E=AB + + + + + + + + AC + + + + BC + + + + Table-2(B): Results and calculation of the mean effects for 25-2 Fractional Factorial Design Experiment 1 a (de) b(de) ab c(d) ac(e) bc(e) abc(d) Experimental Results 3.92 Aeffect -1 +1 -1 +1 -1 +1 -1 +1 Beffect -1 -1 +1 +1 -1 -1 +1 +1 CDeffect effect -1 -1 -1 +1 -1 +1 -1 -1 +1 +1 +1 -1 +1 -1 +1 +1 Eeffect +1 -1 -1 +1 +1 -1 -1 +1 Main effects Overall mean 3.19 Mean B-effect 0.0925 Mean A-effect -0.35 3.11 Mean AC.165 Mean C-effect -0.15 3.effect -0.effect -0.09 Mean BC.28 2.85 3.07 2.19 Mean E-effect -0.25 2.035 .

10 Series1 Series2 Series3 Series4 Series5 G-exp y = 4.0703 10 Fig-2: Correlation plot for grindability of the ball mill against system parameters .4271x 0.2662 (T/T c ) 0.837 0.064 (rhos/rhob) -0.0877 (N) 0.1 (dp/db) -0.0478 1 (n/n1) c 2.

-3: Comparison of calculated values of grindability by both the approaches against the experimental values .F.6 G .D.5 G-exp Fig.5 2 2. 4 G-cal 3 2 1 0 0 0.Exp 5 G-Cal_D.5 4 4.5 3 3. G-cal_F.5 1 1.A.

Sign up to vote on this title
UsefulNot useful