materials

Article
Impact Load Behavior between Different Charge and
Lifter in a Laboratory-Scale Mill
Zixin Yin 1,2 , Yuxing Peng 1,2, * ID
, Zhencai Zhu 1,2 , Zhangfa Yu 3,4 and Tongqing Li 1,2
1 School of Mechatronic Engineering, China University of Mining & Technology, Xuzhou, 221116 China;
yinzixincumt@126.com (Z.Y.); zhuzhencai@vip.163.com (Z.Z.); litongqingcumt@163.com (T.L.)
2 Jiangsu Key Laboratory of Mine Mechanical and Electrical Equipment,
China University of Mining & Technology, Xuzhou 221116, China
3 Luoyang Mining Machinery Engineering Design Institute Co., Ltd., Luoyang 471039, China;
citichic006@126.com
4 State Key Laboratory of Mining Heavy Equipment, CITIC Heavy Industries Co., Ltd.,
Luoyang 471039, China
* Correspondence: pengyuxing@cumt.edu.cn; Tel.: +86-138-0520-9649

Received: 28 June 2017; Accepted: 29 July 2017; Published: 31 July 2017

Abstract: The impact behavior between the charge and lifter has significant effect to address the mill
processing, and is affected by various factors including mill speed, mill filling, lifter height and media
shape. To investigate the multi-body impact load behavior, a series of experiments and Discrete
Element Method (DEM) simulations were performed on a laboratory-scale mill, in order to improve
the grinding efficiency and prolong the life of the lifter. DEM simulation hitherto has been extensively
applied as a leading tool to describe diverse issues in granular processes. The research results shown as
follows: The semi-empirical power draw of Bond model in this paper does not apply very satisfactorily
for the ball mills, while the power draw determined by DEM simulation show a good approximation
for the measured power draw. Besides, the impact force on the lifter was affected by mill speed,
grinding media filling, lifter height and iron ore particle. The maximum percent of the impact force
between 600 and 1400 N is at 70–80% of critical speed. The impact force can be only above 1400 N at
the grinding media filling of 20%, and the maximum percent of impact force between 200 and 1400 N
is obtained at the grinding media filling of 20%. The percent of impact force ranging from 0 to 200 N
decreases with the increase of lifter height. However, this perfect will increase above 200 N. The impact
force will decrease when the iron ore particles are added. Additionally, for the 80% of critical speed,
the measured power draw has a maximum value. Increasing the grinding media filling increases the
power draw and increasing the lifter height does not lead to any variation in power draw.

Keywords: DEM simulation; load behavior; impact force; power draw; ball mill

1. Introduction
A ball mill (Figure 1) is the key piece of equipment for secondary grinding after crushing and is
suitable for grinding all types of ores and other materials. Ball mill are used in the mining, cement,
chemical and agricultural industries, particularly tumbling ball mills [1–4]. The comminution process
is dependent on the rotation of the mill to lift the grinding media for reducing particle size, including
abrasion, crushing and impacting. Due to the impact load behavior between charge and lifter direct
impact of milling efficiency, the efficient operation of mill working parameters is a critical issue.

Materials 2017, 10, 882; doi:10.3390/ma10080882 www.mdpi.com/journal/materials

Materials 2017, 10, 882 2 of 17
Materials 2017, 10, 882 2 of 17

Figure 1. Schematic diagram of ball mill.

Figure 1. Schematic diagram of ball mill.

For ball mills, it is dependent on the rotation of the mill shell for lifting the grinding media to
impact For theballparticle,
mills, it is and dependent
the mill on the rotation
shell, grindingofmedia, the millparticle
shell for andlifting
lifterthearegrinding
contents media
of theto
impact the particle, and the mill shell, grinding media, particle
grinding environment [5,6]. The grinding efficiency is affected by operating variables, such as the and lifter are contents of the grinding
environment
charge, lifter[5,6]. and The grindingparameter.
controlling efficiency isThus affectedfar, bymanyoperating
direct variables,
and indirect suchmethods
as the charge,have lifter
been
and controlling parameter. Thus far, many direct and indirect
employed to measure the impact load behavior, such as X-rays [7], vibration acceleration [8–10], methods have been employed to measure
the impact
acoustic load behavior,
emissions [11], strain such[12], as X-rays
force sensor [7], vibration acceleration
[13], conductivity [14][8–10], acoustic emissions
and inductivity [15]. [11],
strainThe [12],charge
force sensor [13], conductivity [14] and inductivity [15].
consists of particle and grinding media. Charge parameters include powder filling,
The charge consists
grinding media filling and of particle
chargeand size. grinding media. are
The particles Charge parameters
grinding material, include
which powder
are used filling,
to
grinding media filling and charge size. The particles
further comminute to obtain a suitable size distribution. The grinding media is used for are grinding material, which are used to further
comminute
impact-breakage to obtain anda abrasion.
suitable size distribution.
Moreover, the sizeThe and grinding
fillingmedia is used for
of grinding impact-breakage
media has a strict
and abrasion. Moreover, the size and filling of grinding
influence on the milling process [16–20]. Rezaeizadeh et al. [21] studied the effect of the media has a strict influence on thegrinding
milling
process [16–20]. Rezaeizadeh et al. [21] studied the effect of the
media filling on impact load, concluding that the impact load on the lifters increases as the filling grinding media filling on impact load,
concluding
level decreases. that the impact load
Rajrajamani [22]oninvestigated
the lifters increases
the effects as of
thegrinding
filling level mediadecreases.
size and Rajrajamani
filling level[22] on
investigated the effects of grinding media size and
the impact spectra, noting that impact spectra changes with filling level and cannot be filling level on the impact spectra, noting that
analyzed
impact
accurately.spectra changes with
Additionally, as filling
the steel levelball andsizecannot be analyzed
increases, accurately.
the impact Additionally,
frequency decreases, as thebutsteel
the
ball size increases, the impact frequency decreases, but the impact
impact forces increases. The purpose of the lifter is lifting charge and preventing slip, and lifter forces increases. The purpose of the
lifter is liftingsuch
parameters, charge and preventing
as size, shape and slip, number and lifter parameters,
of lifters such as size,
have a significant shape
effect onandload number
behavior of
lifters
[23,24]. have a significant
Rezaeizadeh et effect
al. [21] onmounted
load behavior [23,24].
a quartz force Rezaeizadeh
sensor to measure et al. [21] themounted
impacta quartz
load onforce the
sensor to measure the impact load on the lifter. The results showed
lifter. The results showed that increasing the lifter height and number of lifters results in larger that increasing the lifter height and
number of lifters
impact forces. results inetlarger
Djordjevic al. [25]impact
usedforces. Djordjevic
the Particle Flowet al. [25]
Code 3Dused(PFC3D) the Particle
to model FlowtheCodeeffect3Dof
(PFC3D) to model the effect of lifter number, height and
lifter number, height and mill speed on net power draw, concluding that the net power draw mill speed on net power draw, concluding
that the netaspower
decreases draw
the lifter decreases
height as theControlling
increases. lifter heightparameters,
increases. Controlling
such as mill parameters,
speed andsuch mill as mill
filling,
speed and mill filling, have been investigated to determine
have been investigated to determine a suitable parameter combination that requires higher a suitable parameter combination that
requires
grindinghigher grinding
efficiency. Hence, efficiency. Hence, many
many researchers have researchers
studied the have studied
effect the effect of
of controlling controlling
parameter on
parameter on milling efficiency [26–28]. For mill speed, impact
milling efficiency [26–28]. For mill speed, impact load increases with mill rotation speed, but higher load increases with mill rotation speed,
but higher
speeds speeds
result in theresult
wearing in theofwearing
lifters and of lifters
increasedand increased energy consumption.
energy consumption. For milla filling,
For mill filling, minor
afilling
minorlevelfilling
results in economic cost increases, and an excessive filling level result in mill doesdoes
level results in economic cost increases, and an excessive filling level result in mill not
not break the
break the particle. particle.
Currently,
Currently,there thereare aretwo twomethods
methodstotomeasure measurethe theimpact
impact force
force onon thethelifter. Method
lifter. Method 1 changes
1 changes in
accordance to the drop height of charge in the mill, but transforms
in accordance to the drop height of charge in the mill, but transforms into measuring the single into measuring the single charge
impact
charge on the lifter
impact on the material in free fall
lifter material [29–36].
in free Method Method
fall [29–36]. 2 consists of installing
2 consists a sensor aonsensor
of installing the millon
shell to measure the impact signal of the charge for replacing
the mill shell to measure the impact signal of the charge for replacing the impact force of the lifter the impact force of the lifter [20,22,37].
Based on recent
[20,22,37]. Basedresearch
on recent forresearch
improving formilling
improving efficiency
milling byefficiency
selecting an by appropriate
selecting ancombination
appropriate
of lifter design,
combination of mill
lifterspeeddesign, andmillfilling,
speed little
and is filling,
known little aboutisaknown
direct measure
about a of the impact
direct measureforce of
of the
the lifter and the impact breakage behavior between different
impact force of the lifter and the impact breakage behavior between different shapes of grinding shapes of grinding media and lifters.
Thus,
mediaitand is significant
lifters. Thus, to study the impacttobehavior
it is significant study the between
impact charge
behavior and lifter. charge and lifter.
between
To
To investigate
investigate the the multi-body
multi-body impact impact breakage
breakage behavior,
behavior, we we designed
designed aa drop drop ballball tester
tester and
and
laboratory-scale mill to measure the power draw and
laboratory-scale mill to measure the power draw and impact force on the lifter. In this paper, impact force on the lifter. In this paper,
we
primarily studied the effects of mill speed (50–100% of critical speed), grinding media filling
(5–30%), lifter height (10–20 mm) and grinding media shape (square, ball and cylinder) on impact
Materials 2017, 10, 882 3 of 17

we primarily studied the effects of mill speed (50–100% of critical speed), grinding media filling
(5–30%), lifter height (10–20 mm) and grinding media shape (square, ball and cylinder) on impact
breakage behavior. This improved understanding of the impact process in the ball mills and also
improved mill grinding efficiency.

2. Power Draw Model of Discrete Element Method

The DEM is a numerical method for simulating and analyzing the bulk performance of granular
materials and many geomaterials such as coal, ores and rocks. This method was designed to
simulate dynamic systems of particles where each element is considered to be an independent and
non-deformable entity that interacts with other particles by the laws of contact mechanics and moves
following Newton’s second law. This method applies solid-rigid mechanics at the particle level, and the
discrete elements are considered to be rigid, non-deformable elements. The constitutive model or
behavior of the material is established at the contact areas between particles.
This method has proven to be a useful tool in the milling industry. According to DEM theory,
the power draws of ball mills can be calculated as follows:

PDEM = mc × g × x × w (1)

where PDEM is the power in watt; mc is the effective mass of charge; g is the acceleration of gravity; x is
the distance between the center of the mill and the center of the gravity of load; and w is the rotational
angular velocity of ball mill.
In the last century, excluding the DEM model of mill power prediction, many models applied
to predict the power draw of tumbling mills were very similar [38]. Bond [39] applied the data of
industrial mills and laboratory test results to obtain a semi-empirical power draw model; the prediction
equation is as follows:

d − 12.5 × D
   
0.1
PBond = β × mc × 4.879 × D0.3 × (3.2 − 3 × ϕ) × ψ × 1 − 9 + 1.102 × (2)
2 − 104 50.8

where PBond is the power in watt; β is the mill coefficient (β = 1.08); D is mill diameter; ϕ is mill filling;
ψ is mill speed; and d is particle diameter in meter.

3. Materials and Methods

3.1. Experimental Setup

3.1.1. Drop Ball Experiments

In the ball mill, iron ore particles are broken by the grinding media acting on high-speed crushing
and grinding. Reduction of the size of iron ore particles is caused by a cumulative process of single
breakage behavior with a longer period of time. To understand the iron ore particle size reduction
process, the single breakage behavior must be investigated.
To measure the single breakage behavior, a drop ball tester was designed, as shown in Figure 2.
The drop ball tester mainly consists of an impact-force sensor, data acquisition equipment and
a supporting disk. Firstly, we moved the sliding block to adjust the drop ball at the right height.
Secondly, we dropped the ball to impact the sensor. Lastly, we collected the data of impact force using
data acquisition equipment. From the tester, we changed the steel ball size (∅20, ∅30, ∅40 mm) and
the drop height (50, 120, 220, 320, 420 and 520 mm) and then obtained a better understanding of the
relation between impact force and time.
 Materials 2017, 10, 882 4 of 17
Materials 2017, 10, 882 4 of 17
Materials 2017, 10, 882 4 of 17

Figure 2. Drop ball tester.

Figure
Figure 2. Drop balltester.
Drop ball tester.
3.1.2. Impact Breakage Experiments
3.1.2.3.1.2.
Impact
To Breakage
Impact Breakage
carefully Experiments
collectExperiments
the load behavior on the lifter, a laboratory-scale mill was designed. Figure 3
shows To the laboratory-scale
carefully collect theloadmillbehavior
load and data on acquisition
the system, and the relative facilitiesFigure
are as3
To carefully collect the behavior on the lifter,
lifter,aalaboratory-scale
laboratory-scale mill waswas
mill designed.
designed. Figure 3
follows:
shows (i)
the A mill shell
laboratory-scale withmill 520
and mm
data in diameter
acquisition and
system, 40 mm
and in
the length
relative and equipped
facilities are as with 12
follows:
shows the laboratory-scale mill and data acquisition system, and the relative facilities are as
equally
(i) A mill spaced with and different height rows of rectangle
40 mm inlifters;
length(ii) A 3.7-kW three-phase motor spaced
with a
follows: (i) Ashellmill shell 520 mm in
with 520 diameter
mm inand diameter and 40 and
mmequipped
in length with 12 equally
and equipped with 12
rated speed of 1480 rpm; (iii) An inverter was used to control the motor
and different height rows of rectangle lifters; (ii) A 3.7-kW three-phase motor with a rated speed of velocity; (iv) A torque
equally
sensorspaced
with and different height rows of rectangle lifters; (ii) A 3.7-kW three-phase motor with a
1480 rpm; (iii)a An
range of 500
inverter wasNm was
used to mounted
control thebetween the motor
motor velocity; (iv)and mill shell
A torque sensortowith
measure
a rangemill
of
rated speed torque
velocity, of 1480 and rpm;
power(iii)byAnthe inverter was used tosoftware; control theAmotor velocity; (iv) A ontorque
500 Nm was mounted between theM400
motordata
andacquisition
mill shell to measure(v) slip ring
mill velocity, was mounted
torque and power
sensor
the with
shaft
by the a range
to
M400 prevent ofwinding
500 Nm
data acquisition and was mounted
supply
software; A slipbetween
(v)power; (vi) An
ring the motor
wasimpact
mounted force and
onsensor mill shell
mounted
the shaft to to measure
through
prevent the mill
winding
velocity,
and supply power; (vi) An impact force sensor mounted through the mill lifter walls was employed to on
mill torque
lifter walls andwas power
employedby the
to M400
measure data
the acquisition
impact force software;
on the (v)
lifter, A
withslip
a ring
sensor was
range mounted
of 300
kg,
the shaft and
measure the
to prevent output
the impact signalonthrough
winding
force and slip
supply
the lifter, withring
power;to the
a sensor DH5981
(vi)range
An impact data
of 300 acquisition
force
kg, and sensor
the outputequipment
mounted with a the
through
signal through
sampling
mill slip
lifter
ring rate
walls
to the of
was 20
DH5981 kHz.
employed to measure
data acquisition the impact
equipment withforce on therate
a sampling lifter, with
of 20 kHz.a sensor range of 300
kg, and the output signal through slip ring to the DH5981 data acquisition equipment with a
sampling rate of 20 kHz.

Figure 3.
Figure Laboratory-scale mill
3. Laboratory-scale mill and
and data
data acquiring
acquiring system.
system.

In Figure3,3,thethe
In Figure experiments
experiments werewere performed
performed in dryin dry conditions,
conditions, the mill
the mill shell with shell with one
one transparent
transparent end (semicircle perspex) and a high-speed digital camera is used to record
end (semicircle perspex) and a high-speed digital camera is used to record the charge trajectory, and the charge
the
trajectory, and the Figure 3. Laboratory-scale mill and data acquiring system.
milling process wasmilling process
run 3 min and was run 3was
the data mincollected
and the for
data
10was collected
s. The summaryfor of10 the
s. The summary
experimental
of the experimental
conditions conditions
is given in the Tableis1 given
below.in the Table 1 below.
In Figure 3, the experiments were performed in dry conditions, the mill shell with one
transparent end (semicircle perspex) and a high-speed digital camera is used to record the charge
trajectory, and the milling process was run 3 min and the data was collected for 10 s. The summary
of the experimental conditions is given in the Table 1 below.
Materials 2017, 10, 882 5 of 17

Table 1. Experimental conditions.

Mill Speed/of
Test Filling Level Lifter Height (mm) Media-Shape Particle
Critical Speed
50%, 60%, 70%,
4.2 20% 20 Steel ball None
80%, 90%, 100%
5%, 10%, 15%,
4.3 75% 20 Steel ball None
20%, 25%, 30%
4.4 75% 20% 10, 15, 20 Steel ball None
4.5 75% 20% 15 Steel ball, Square, Cylinder None
4.5 75% 20% 15 Steel ball, Square, Cylinder Yes

3.2. Materials
Table 2 shows the material properties of the grinding media, iron ore particle and lifter.
The grinding media was forged in steel forging with round, square and cylinder shapes with equal
mass from Jinan, China. The raw iron ore particle from the Xuzhou iron ore factory was a magnetite
(Fe2 O3 67.46%) pellet. Since grinding fineness has a great effect on iron ore floatation, appropriate
grinding fineness can not only guarantee good monomer dissociation, but also help to avoid excessive
grinding [23,40]; we choose iron ore feed with 2–2.8 mm. Iron ore particles were broken into suitable
sizes with an industrial jaw crusher, then the products were sieved for 10 min by a vibrating screen
to acquire mono-sized particles (2–2.8 mm). The lifter was designed with different heights, and the
material was the same as the mill shell.

Table 2. Material properties.

Materials Size (L = Length; W = Width; H = Height)

Grinding media Steel ball (∅15 mm) Square (L: 12 × W: 12 × H: 12 mm) Cylinder (∅12 × H: 16 mm)
Iron ore particle 2–2.8 mm
Lifter L: 40 × W: 20 × H: 10 mm L: 40 × W: 20 × H: 15 mm L: 40 × W: 20 × H: 20 mm

Table 3 shows the input material parameter in the DEM simulations.

Table 3. The parameters of DEM simulations.

Material Parameters Value

Iron ore particle density (kg/m3 ) 3886
Iron ore particle shear modulus (Pa) 2.59 × 109
Iron ore particle Poisson’s ratio 0.28
Shell density (kg/m3 ) 7800
Shell shear modulus (Pa) 7 × 1010
Shell Poisson’s ratio 0.3
Particle-shell restitution coefficient 0.32
Particle-shell static friction coefficient 0.48
Particle-shell rolling friction coefficient 0.2
Particle-Particle restitution coefficient 0.49
Particle-Particle static friction coefficient 0.48
Particle-Particle rolling friction coefficient 0.16

4. Results and Discussion

4.1. Effect of Steel Ball Size and Drop Height on Impact Force
The impact force plotted against the time in this research is characterized using DH5981 data
acquisition equipment collected from the drop ball test performed on lifter. The drop test is performed
Materials 2017, 10, 882 6 of 17

Materials 2017, 10, 882 6 of 17

using different drop height and using three different ball sizes, 20, 30 and 40 mm. The results presented
drop height.
in Figure The variation
4 show in impact
that the impact force
force increases
shows up to
a strong a maximum
sensitive to theand
balldecreases
size and rapidly, and
drop height.
the impact duration time is approximately 1 millisecond. Additionally, the maximum
The variation in impact force increases up to a maximum and decreases rapidly, and the impact magnitude of
impact
durationforce
time increases with the
is approximately increase of Additionally,
1 millisecond. drop heightthe formaximum
the givenmagnitude
ball size offorimpact
the given
force
drop height.
increases with the increase of drop height for the given ball size for the given drop height.
Experimental
Experimental results
results indicate
indicate that
that each
each breakage
breakage event
event depends
depends onon the
the steel
steel ball
ball size
size and
and drop
drop
height, and increasing
height, and increasingthe thesteel
steel ball
ball sizesize
andand
dropdrop height
height increases
increases the gravitational
the gravitational potentialpotential
energy.
energy. Hence, increasing
Hence, increasing the steelthe steel
ball sizeball
and size
dropandheight
drop height can increase
can increase the impact
the impact force
force in ballin ball mills.
mills.

1000

50 mm
800 120 mm
Steel ball:20 mm 220 mm
320 mm
600 420 mm
520 mm
400

200

0
2000
0.0 0.2 0.4 0.6 0.8 1.0 1.2

1600
Steel ball:30 mm
Force / N

1200

800

400

0
4000
0.0 0.2 0.4 0.6 0.8 1.0 1.2

3200
Steel ball:40 mm

2400

1600

800

0.0 0.2 0.4 0.6 0.8 1.0 1.2

Time / ms

4. Impact
Figure 4.
Figure Impact force-time
force-timeanalysis
analysiscurve forfor
curve a steel ball with
a steel diameters
ball with of 20, 30
diameters of and
20, 40
30 mm
and dropped
40 mm
from different heights.
dropped from different heights.

4.2. Effect
4.2. Effect of
of Mill
Mill Speed
Speed on
on Impact
Impact Load
Load Behavior
Behavior

4.2.1. Steel
4.2.1. Steel Ball
Ball Trajectories
Trajectories of
of DEM
DEM Simulations
Simulations and
and Experiments
Experiments
In the
In the experiment,
experiment, the the mill
mill fitted
fitted with
with aa transparent
transparent end, end, and
and aa high-speed
high-speed digital
digital camera
camera is is
used to record steel ball trajectories. Figure 5 shows the steel ball trajectories
used to record steel ball trajectories. Figure 5 shows the steel ball trajectories of DEM simulations of DEM simulations
and experiments
and experiments withwith different
different mill
mill speeds,
speeds,the themill
millspeed
speedchanged
changedfrom from50%
50%toto100%,
100%, a lifter size
a lifter of
size
of × 20
40 40 × 20× 20 mm,
× 20 mm,andanda grinding
a grindingmedia filling
media of 20%
filling of by20% volume. The photograph
by volume. resultsresults
The photograph are shownare
the grinding information inside a mill, which includes the steel ball trajectories
shown the grinding information inside a mill, which includes the steel ball trajectories and the and the particular
charge position.
particular chargeFrom the photographs
position. of the DEM of
From the photographs simulation
the DEMand experiment,
simulation andmore steel balls
experiment, are
more
projected
steel ballsinto
areflight to impact
projected into the milltoshell
flight as thethe
impact rotating velocity
mill shell as of
thethe mill increases.
rotating velocityTheof steel ball
the mill
behaviors are highly consistent in the DEM simulation and experiment. At 50%
increases. The steel ball behaviors are highly consistent in the DEM simulation and experiment. At and 60% critical speed,
the steel
50% and ball
60%move from
critical the bottom
speed, the steelof ball
the shell
movetofromthe shoulder
the bottom position
of thebegin
shellto
toslide downward
the shoulder across
position
begin to slide downward across the free surface, and much higher part of steel balls follow a
cascading motion. At 70% and 80% critical speed, the shoulder position becomes high, the steel
Materials 2017, 10, 882 7 of 17

Materials 2017, 10, 882 7 of 17

the free surface, and much higher part of steel balls follow a cascading motion. At 70% and 80% critical
speed, the shoulder position becomes high, the steel balls projected into flight follow a parabolic path,
balls projected into flight follow a parabolic path, and much higher number of steel balls follow a
and much higher number of steel balls follow a cataracting motion. As the mill speed increases from
cataracting motion. As the mill speed increases from 90% to 100% critical speed, the steel balls are
90% to 100% critical speed, the steel balls are almost projected flight follow a centrifuging motion, large
almost projected flight follow a centrifuging motion, large number of steel ball has a high speed
number of steel ball has a high speed collide on the lifter and result in the wear rate of lifter increases.
collide on the lifter and result in the wear rate of lifter increases.
DEM software provides an accurate prediction of the steel ball trajectory in the ball mill.
DEM software provides an accurate prediction of the steel ball trajectory in the ball mill.
However,
However, theabove
the above discussion
discussion is
is dependent
dependent on observing the
on observing the steel
steel ball
balltrajectory
trajectoryininphotographs.
photographs.
Simply
Simplyobserving
observing thethe trajectory
trajectory is
is not
not enough
enough toto describe
describe what
what isis occurring
occurring inside
insidethe
theball
ballmill.
mill.
Hence, an impact force sensor was mounted in the lifter to investigate impact load
Hence, an impact force sensor was mounted in the lifter to investigate impact load behavior. behavior.

(a) (b) (c) (d) (e) (f)

Figure5.5.Photographs
Figure Photographsofofthe
theimpact
impactload
loadbehavior
behaviorwith
witha agrinding
grindingmedia
mediafilling
fillingofof20%
20%bybyvolume
volumeat
at 50%–100% of critical speed: (a) 50%; (b) 60%; (c) 70%; (d) 80%; (e) 90%; and
50–100% of critical speed: (a) 50%; (b) 60%; (c) 70%; (d) 80%; (e) 90%; and (f) 100%.(f) 100%.

4.2.2. Effect of Mill Speed on Impact Force

4.2.2. Effect of Mill Speed on Impact Force
Figure 6 shows the results of impact force at different mill speeds. Increasing the mill speed
Figure 6 shows the results of impact force at different mill speeds. Increasing the mill speed can
can increase the impact load behavior, but an excessive increase in mill speed results in higher
increase the impact load behavior, but an excessive increase in mill speed results in higher energy
energy consumption. As shown in Figure 6, the frequency of impacts in a force range of 0–200 N is
consumption. As shown in Figure 6, the frequency of impacts in a force range of 0–200 N is the
the maximum at lower speeds. In lower speeds of 50%–60% critical speed, the impact force is
maximum at lower speeds. In lower speeds of 50–60% critical speed, the impact force is obviously
obviously weak, a large number of steel ball move at a lower speed and has lower impact energy,
weak, a large number of steel ball move at a lower speed and has lower impact energy, resulting in
resulting in particle abrasive comminution. At 70%–80% critical speed, a force range of 600–1400 N
particle abrasive comminution. At 70–80% critical speed, a force range of 600–1400 N has the maximum
has the maximum percent compared with other mill speeds. Hence, the lifter has a higher impact
percent
force andcompared
results inwith
more other mill speeds.
crushing Hence,Additionally,
and grinding. the lifter hasat a higher
90%–100% impact force
critical and results
speed, the steelin
more crushing and grinding. Additionally, at 90–100% critical speed, the steel
ball move at a high speed and increasing the rate of the lifter, but forces ranging primarily from ball move at a high
0 to
speed
600 Nand increasing
results in more the rate of the lifter, but forces ranging primarily from 0 to 600 N results in
grinding.
more grinding.
From the above results, we can conclude that mill speed has a significant influence on load
From the
behavior. Theabove
impactresults,
force ofwethe
canlifter
conclude that mill speed
for investigating has a significant
multi-body influence
load behavior is a on load
useful
behavior. The impact force of the lifter for investigating multi-body load behavior
method. At lower speeds (50–60% of critical speed), steel balls from the toe position to shoulder is a useful method.
At lower speeds
position cascade(50–60%
down along of critical speed),surface,
the charge steel balls from
so the thebehavior
load toe position to shoulder
is mainly position
in a grinding
cascade
process.down along speeds
At higher the charge surface,
(70–80% so the load
of critical behavior
speed), is mainly
the steel in aprojected
balls are grinding process. At higher
to fall down and
speeds (70–80% of critical speed), the steel balls are projected to fall down and impact
impact the toe position, so the load behavior is mainly in a crushing process. However, an excessive the toe position,
sospeed
the load behavior
result in steel is mainly
balls in a crushing
impacting process.and
space position, However,
increasesan the
excessive speed result in steel balls
lifter wear.
impacting space position, and increases the lifter wear.
Materials 2017, 10, 882 8 of 17
Materials 2017, 10, 882 8 of 17

250000

200000

150000

100000

50000

0
0-200
1600

1400

1200

1000

800

600

400

200

0
200-400 400-600 600-800
60

50

40

30

20

10

0
800-1000 1000-1200 1200-1400 ≥1400

Figure
Figure 6.
6. Impact
Impact force
force at
at different
different mill
mill speeds.
speeds.

4.2.3.
4.2.3. Effect
Effect of
of Mill
Mill Speed
Speed on
on Power
Power Draw
Draw
Figure
Figure 77 shows
shows thethe results
results of
of mill
mill speed
speed onon power
power draw.
draw. The
The DEM
DEM simulation
simulation andand experimental
experimental
power
power draw increased and then decreased with mill speed, the results of the DEM simulation agree
draw increased and then decreased with mill speed, the results of the DEM simulation agree
well with the
well with the experiment
experiment results.
results. However,
However, the the Bond
Bond semi-empirical
semi-empirical power
power draw
draw model
model increases
increases
with
with mill
mill speed,
speed, andand the
the power
power draw
draw value
value isis larger
larger than
than inin the
the experiment
experiment andand DEM
DEM simulation
simulation by by
approximately
approximately 1.71–3.07 times. This result shows that the accuracy of Bond formula is theislowest
1.71–3.07 times. This result shows that the accuracy of Bond formula the lowest
when
when
the millthespeeds
mill speeds
at 100% at of
100% of critical
critical speed.speed.
In
In the
the DEM
DEM simulation
simulation and and experiment,
experiment, the the power
power drawdraw increases
increases with
with mill
mill speed
speed up up to
to aa peak
peak
value
value of 80% of critical speed. This means that the optimum grinding efficiency occurs at mill speeds
of 80% of critical speed. This means that the optimum grinding efficiency occurs at mill speeds
of
of near
near 80%
80% of critical speed
of critical speed [23],
[23], because
because moremore steel
steel balls
balls are
are projected
projected into
into the
the cataracting
cataracting region.
region.
After 80% of critical speed, the mill power draw decreases until 100% of critical
After 80% of critical speed, the mill power draw decreases until 100% of critical speed and more speed and more steel
balls begin centrifuging, leading to a power draw decrease to drive
steel balls begin centrifuging, leading to a power draw decrease to drive the mill. This outcome the mill. This outcome
demonstrates
demonstrates that thatthetheDEM
DEMsimulation
simulationis isanan accurate
accurate calculation
calculation method.
method. In the
In Bond
the Bondformula, this
formula,
result indicates that Bond formula is not suitable for predict the power accurately
this result indicates that Bond formula is not suitable for predict the power accurately in this condition, in this condition,
and
and the
the major
major difference
difference cancan bebe explained
explained by by the
the fact
fact that
that Bond
Bond formula
formula does
does not
not consider
consider thethe effect
effect
of the lifters. In addition, Bond formula applied data collected from industrial
of the lifters. In addition, Bond formula applied data collected from industrial mills and laboratory mills and laboratory
tests results,the
tests results, themodel
modelbasedbased on torque-arm
on both both torque-arm or energyor balance
energy principals
balance principals and some
and some coefficients,
coefficients, and the calculate results are affected by the mill length and diameter,
and the calculate results are affected by the mill length and diameter, ball density, ball filling and ball density, ball
filling
rotation and rotation speed.
speed.
Materials 2017, 10, 882 9 of 17
Materials 2017, 10, 882 9 of 17
Materials 2017, 10, 882 9 of 17
100
100
90 DEM simulation
Bondsimulation
DEM
90
80 Experiment
Bond
80
70 Experiment
70
60

/w/w
60
50

Power
50

Power
40
40
30
30
20
20
10
10
0
0 50 60 70 80 90 100
50 60Mill speed
70 / % of critical
80 speed
90 100
Mill speed / % of critical speed
Figure7.7.Power
Figure Powerdraws
drawsat
atdifferent
differentmill
millspeeds.
speeds.
Figure 7. Power draws at different mill speeds.
From the above results and discussions, the mill speed crucially affects the charge motion,
From
Fromthetheabove
aboveresults and
results discussions,
and the the
discussions, millmill
speed crucially
speed affects
crucially the charge
affects motion,
the charge impact
motion,
impact force and power draws. In order to obtain an efficient grinding, the mill speed between 70%
force and
impact power
force anddraws. In order
power draws. In to obtain
order an efficient
to obtain grinding,
an efficient the mill
grinding, speed
the mill between
speed 70%70%
between and
and 80% can be beneficial to ball mill operation. Furthermore, the ball mill with one transparent end
80% can be beneficial to ball mill operation. Furthermore, the ball mill with one transparent
and 80% can be beneficial to ball mill operation. Furthermore, the ball mill with one transparent end end and
and a photographed trajectory of charge can offer accurate input parameters for DEM simulation.
aand
photographed trajectory
a photographed of charge
trajectory can offer
of charge accurate
can offer inputinput
accurate parameters for DEM
parameters simulation.
for DEM simulation.
4.3.Effect
4.3. Effect of GrindingMedia
Media Fillingon
on ImpactLoad
Load Behavior
4.3. EffectofofGrinding
Grinding Media Filling
Filling on Impact
Impact Load Behavior
Behavior
4.3.1.Steel
4.3.1. SteelBall
BallTrajectories
TrajectoriesofofDEM
DEMSimulation
Simulation
4.3.1. Steel Ball Trajectories of DEM Simulation
Figure 88 shows
Figure shows the steel
steel balltrajectories
trajectories on different
different grinding
grinding media
media filling,
filling, with
with the
the filling
filling
Figure 8 shows the the steel ball
ball trajectories on on different grinding media filling, with the filling
changedfrom
changed from5% 5%toto30%,
30%,aalifter
liftersize
sizeofof4040××20
20× × 20 mm, and a mill speed of 75% [23] critical critical speed.
speed.
changed from 5% to 30%, a lifter size of 40 × 20 × 20 mm, and a mill speed of 75% [23] critical speed.
Increasingthe
Increasing the grindingmedia
media filling not only increases the fraction flying in cataracting
the cataracting zone
Increasing thegrinding
grinding mediafilling
fillingnot
notonly
onlyincreases thethe
increases fraction flying
fraction in the
flying in the zone
cataracting but
zone
but increases
also also increases
the theofrate
rate ball of
on ball
ball on ball contact
contact per per
unit. unit. Commonly,
Commonly, increasing increasing
the fillingthe filling
level level
increases
but also increases the rate of ball on ball contact per unit. Commonly, increasing the filling level
increases
the impactthe
increases the impact
energy,
impactand energy, and
effective
energy, and effective
impact impactincreases.
increases.
effective impact increases.

(a)
(a) (b)
(b) (c)
(c) (d)
(d) (e)
(e) (f)(f)
Figure
Figure 8.8. Photograph of the
Photograph of the impact
impactload
loadbehavior
behaviorwith
with75%
75%critical
criticalspeed
speedatat5%–30%
5%–30%grinding
grindingmedia
media
Figure 8. Photograph of the impact load behavior with 75% critical speed at 5–30% grinding media
filling:
filling: (a)
(a) 5%; (b) 10%;
5%; (b) 10%; (c)
(c) 15%;
15%; (d)
(d) 20%;
20%;(e)
(e)25%;
25%;and
and(f)
(f)30%.
30%.
filling: (a) 5%; (b) 10%; (c) 15%; (d) 20%; (e) 25%; and (f) 30%.

4.3.2.
4.3.2. Effect of Grinding Media Media Filling
Filling on onImpact
ImpactForceForce
4.3.2. Effect of Grinding Media Filling on Impact Force
Increasing the filling
filling not
not only
only increases
increases the theimpact
impactload loadbutbutalsoalsoresults
resultsininmore
morewearing
wearingand and
Increasing the filling not only increases the impact load but also results in more wearing and
breakage
breakage of the lifter. Hence, a suitable filling level is critical to improving efficiency. Figure 9 9
Hence, a suitable filling level is critical to improving efficiency. Figure
breakage of the lifter. Hence, a suitable filling level is critical to improving efficiency. Figure 9 shows
shows
shows the impact force on on the
the lifter
lifter at at different
different filling
fillinglevels.
levels.The Theimpact
impactforceforcecan
canbebeonly
onlyabove
above
the impact force on the lifter at different filling levels. The impact force can be only above 1400 N at
1400
1400 N at the ball filling of of 20%,
20%, andand the the maximum
maximumpercent percentofofimpact
impactforce forcebetween
between200 200and
and1400
1400NN
the ball filling of 20%, and the maximum
is
is obtained
obtained at the grinding
grinding media
media fillingpercent
filling 20%.ofIncreasing
of 20%.
of
impact force
Increasing thebetween
the
200 results
fillinglevel
filling level and 1400
resultsininN is obtained
more
more balltoto
ball
atball
the grinding butmedia filling
the frequencyof 20%. Increasing the filling level results in more ball to ball collisions,
ball collisions, frequency of of steel
steelball
balldirect
directimpacts
impactson onthe
thelifter
lifterdoes
doesnotnotalways
alwaysincrease.
increase.
but the frequency of steel ball direct impacts on the lifter does not always increase.
From the above investigation,
investigation, increasing
increasing the
the grinding
grinding media
media filling
filling does
doesnot notincrease
increasethe the
impact
impactFrom the
force above
on the investigation,
lifter. Due to increasing
the lifter the
profilegrinding
and media
rotation filling
speed does
at the
lifter. Due to the lifter profile and rotation speed at the same conditions, the not increase
same the
conditions,impact
the
force
number
number on the
of lifter.
impacts Due ontothethe lifter
the lifter does
lifter profile
does not and
not rotation
change
change speed
when
when theatfilling
the the same
filling levelconditions,
level up the number
uptotoa asuitable
suitable level. of
level.
Therefore, further increases
Therefore, increases the the filling
filling level
level results
resultsin inmore
moresteel
steelballs
ballsimpacting
impactingthe thespace
spaceregion
region
and
and energy consumption.
consumption.
Materials 2017, 10, 882 10 of 17

impacts on the lifter does not change when the filling level up to a suitable level. Therefore, further
Materials 2017, 10, 882 10 of 17
increases the filling level results in more steel balls impacting the space region and energy consumption.

250000

200000

150000

100000

50000

0
0-200
2000

1800

1600

1400

1200

1000

800

600

400

200

0
200-400 400-600 600-800
40

35

30

25

20

15

10

0
800-1000 1000-1200 1200-1400 ≥1400

Figure
Figure 9. Impact
Impact force
force at
at different
different grinding media filling.

4.3.3.
4.3.3. Effect
Effect of
of Grinding
Grinding Media
Media Filling
Filling on
on Power
Power Draw
Draw
Figure
Figure 10
10 shows
shows the the results
results of
of power
power draw
draw at
at different grinding media
different grinding media fillings.
fillings. This
This shows
shows that
that
increasing the filling level leads to an increase in the power draw. As the filling level increases, the
increasing the filling level leads to an increase in the power draw. As the filling level increases, the variation
variation between the Bond semi-empirical power draw and experiment increases.
between the Bond semi-empirical power draw and experiment increases. The Bond semi-empirical power The Bond
semi-empirical power times
draw model is 1.51–1.82 drawgreater
modelthan
is the
1.51–1.82 times result.
experimental greater than the the
Additionally, experimental
experiment result.
power
Additionally, the experiment power draw value at 5–10% and 25–30% grinding media
draw value at 5–10% and 25–30% grinding media filling does not significantly change, and the DEM filling does
not significantly change, and the DEM simulation result has a larger
simulation result has a larger difference at 10% and 15% grinding media filling. difference at 10% and 15%
grinding media filling.
100
10090 DEM simulation
9080 Bond
DEM simulation
Experiment
Bond
8070
Experiment
7060
Power / w

6050
Power / w

5040

4030

3020

2010

10 0
5 10 15 20 25 30
0 Grinding filling / %
5 10 15
Grinding 20 / % 25
filling 30
Grinding filling / %

Figure
Figure 10.
10. Power
Power draw
draw at
at different
different grinding
grinding media
media filling.
filling.
Materials 2017, 10, 882 11 of 17
Materials 2017, 10, 882 11 of 17

From the
From theabove
aboveresults,
results,this
this Bond
Bond formula
formula is only
is only suitable
suitable for afor a rough
rough estimation
estimation of the
of the mill mill
power
powerand
draw, draw,
the and
Bondthe Bond semi-empirical
semi-empirical power draw power
modeldraw model
is less is lessthan
accurate accurate
the DEM thansimulation.
the DEM
simulation. For the Bond semi-empirical power model, the main difference is
For the Bond semi-empirical power model, the main difference is that the calculation result does notthat the calculation
result does
consider the not consider
effect the effect
of the lifter. For theofDEM
the lifter. For the
simulation, thisDEM simulation,
difference this difference
can be explained by thecan be
effect
explained
of by the
the contact effect of the
parameter. Stillcontact parameter.
increasing Still increasing
the grinding the grinding
media filling could not media filling
increase thecould not
milling
increase the milling efficiency according the impact force on the lifter. As a result,
efficiency according the impact force on the lifter. As a result, a suitable grinding media filling has a suitable
agrinding media
beneficial filling
to ball mill has a beneficial
grinding to ball mill grinding process.
process.

4.4. Effect of Lifter Height on Impact Load Behavior

4.4.1. Steel Ball Trajectories of

of DEM
DEM Simulation
Simulation
Figure 1111 shows
showsthethesteel
steelball trajectories
ball at different
trajectories lifter
at different heights,
lifter grinding
heights, media
grinding filling
media of 20%,
filling of
and
20%,a and
mill aspeed
mill of 75% critical
speed of 75% speed.
criticalIncreasing the lifter height
speed. Increasing the lifter obviously increases the
height obviously numberthe
increases of
cataracting steel balls. steel balls.
number of cataracting
In the milling process, due to wear and impact, the lifter height gradually decreases until it is
too low and must be replaced. This condition condition causes
causes the
the steel
steel ball
ball not
not to
to reach a sufficient location,
and the
thenumber
numberofofhigh–speed
high–speed steel balls
steel decreases.
balls Finally,
decreases. if the iflifter
Finally, the height is too low,
lifter height the low,
is too workingthe
efficiency of the ballof
working efficiency mill
thedecreases.
ball mill Therefore,
decreases. the lifter height
Therefore, the is a significant
lifter height isfactor influences
a significant the
factor
steel ball trajectories.
influences the steel ball trajectories.

(a) (b) (c)

Figure 11.
Figure 11. Photograph
Photograph of
of the impact load
the impact load behavior
behavior with
with different
different lifter
lifter height:
height: (a)
(a) 10
10 mm;
mm; (b)
(b) 15
15 mm;
mm;
and (c)
and (c) 20
20 mm.
mm.

4.4.2. Effect
4.4.2. Effect of Lifter Height
of Lifter Height on
on Impact
Impact Force
Force
Figure 12
Figure 12 shows
shows thethe result
result ofof impact
impact force
force onon the
the lifter
lifter for
for different
different lifter
lifter heights.
heights. Increasing
Increasing thethe
lifter height lifts the steel balls to an increased height and results in higher impact
lifter height lifts the steel balls to an increased height and results in higher impact velocity. As the velocity. As the
lifter height
lifter height increases
increases from
from 10 10 to
to 2020 mm,
mm, the
the steel
steel balls
balls contact
contact thethe mill
mill shell
shell earlier
earlier and
and impact
impact force
force
increases. Increasing
increases. Increasingthe the lifter
lifter height
height increases
increases the shoulder
the shoulder position
position and
and lifts thelifts
steelthe
ballssteel balls to
to increased
increased heights the higher position. The percent of impact force ranging
heights the higher position. The percent of impact force ranging from 0 to 200 N decreases with from 0 to 200theN
decreasesofwith
increase lifterthe increase
height. of lifterthis
However, height. However,
perfect this perfect
will increase above will
200 N.increase above 200 N.
From the above results, it can be concluded that lifter height
From the above results, it can be concluded that lifter height has a significant has a significant
influenceinfluence
on workingon
working efficiency. Since the steel balls are lifted much higher when the lifter height
efficiency. Since the steel balls are lifted much higher when the lifter height increases, many high-speed increases, many
high-speed
steel steel balls
balls impact impactTherefore,
the lifter. the lifter.effectively
Therefore,changing
effectively thechanging the helps
lifter height lifter height helpsthe
to improve to
improve the milling
milling efficiency. efficiency.
Materials 2017, 10, 882 12 of 17
Materials 2017, 10, 882 12 of 17

10 mm 15 mm 20 mm
250000
250000

200000
200000

150000
150000

100000
100000

50000
50000

0 0
1800 0 -2 0 0
0-200
1800
1600
1600
1400
1400
No. of impacts per 10 s
1200
1200
1000
1000
800
800
600
600
400
400
200
200
0
0 2 0 0 -4 0 0 4 0 0 -6 0 0 6 0 0 -8 0 0
80 200-400 400-600 600-800
80
70
70
60
60
50
50
40
40
30
30
20
20
10
10
0
8 0 0 -1 0 0 0 1 0 0 0 -1 2 0 0 1 2 0 0 -1 4 0 0 ≥1400
0
800-1000 Force range1200-1400
1000-1200 /N ≥1400

Figure 12.
Figure Impact force
12. Impact force at
at different
different lifter
lifter heights.
heights.

4.4.3. Effect of Lifter Height on Power Draw

Figure 13 shows the effect of lifter height on power draw. Increasing the lifter height has a less
significant effect on power draw. For the Bond semi-empirical
semi-empirical power draw model, the power draw
values are still much
much greater
greater than
than the
the experimental
experimental values
values by
by approximately
approximately 1.58–1.9 times. For the
DEM simulation, the power draw values are greater than the experimental values by approximately
1.05–1.21 times.

100

90 DEM simulation
Bond
80 Experiment
70

60
Power / w

50

40

30

20

10

0
10 12 14 16 18 20
Lifter height / mm

Figure 13. Power draw at different lifter heights.

From the above results, the Bond semi-empirical power draw model is only suitable for
From the above results, the Bond semi-empirical power draw model is only suitable for estimating
estimating the power draw, and difference is influenced by the lifter. Besides, the difference
the power draw, and difference is influenced by the lifter. Besides, the difference between experiments
between experiments and DEM simulation results are determined by the distance between the
and DEM simulation results are determined by the distance between the center of the mill and the
center of the mill and the center of gravity of the load. Thus, the DEM simulation can predict the
load behavior at different lifter heights when the input parameters are accurate.

1
Materials 2017, 10, 882 13 of 17

center
Materialsof gravity
2017, 10, 882of the load. Thus, the DEM simulation can predict the load behavior at different13lifter of 17
heights when the input parameters are accurate.
From the
From theabove
aboveresults
results and
and discussions,
discussions, the the
lifterlifter height
height has ahas a significant
significant effect
effect on on motion
charge charge
motion and impact force, but little effect on power draw. Increasing the lifter height increases
and impact force, but little effect on power draw. Increasing the lifter height increases the lifting ability, the
lifting
and theability, and the
lifter height is lifter height
related to theis charge
related diameter.
to the charge diameter.
Thus, Thus, ittoischoose
it is important important to choose
a suitable liftera
suitable lifter height for a
height for a determined ball mill. determined ball mill.

4.5. Effect of Media-Shape on Impact Load Behavior

4.5.1. Effect of Particle on Load Behavior

Figure
Figure 1414shows
showsthe effect of particle
the effect on load behavior.
of particle on load The mill hasThe
behavior. a lifter of 40a ×lifter
sizehas
mill 20 × size
15 mm,of
mill
40 × speed of 75%
20 × 15 mm,critical speed,ofgrinding
mill speed media
75% critical fillinggrinding
speed, of 20%, and
mediapowder
fillinggrinding
of 20%,media ratio of
and powder
0.8. Adding
grinding iron ratio
media ore particles
of 0.8. increases
Adding ironthe height of charge,
ore particles and more
increases the charges
height are projected
of charge, into
and the
more
cataracting zone.
charges are projected into the cataracting zone.

(a) (b)

Figure 14.
Figure 14. Photograph
Photograph of the impact
of the impact load
load behavior:
behavior: (a)
(a) without
without particle and (b)
particle and (b) with
with particle.
particle.

4.5.2. Effect of Particle on Impact Force

4.5.2. Effect of Particle on Impact Force
The charge, the particle and lifter which collide together form a complex environment in the
The charge, the particle and lifter which collide together form a complex environment in the mill
mill shell. To investigate the real grinding process, a simulation was performed to obtain the
shell. To investigate the real grinding process, a simulation was performed to obtain the influence
influence of particles and different media shapes on the impact breakage behavior. The results are
of particles and different media shapes on the impact breakage behavior. The results are shown
shown in Figure 15. Adding iron ore particles significantly reduced the impact force from 1400 to
in Figure 15. Adding iron ore particles significantly reduced the impact force from 1400 to 500 N.
500 N. From Figure 15a,b, the number of impacts for a force range of 0–200 N in order are square,
From Figure 15a,b, the number of impacts for a force range of 0–200 N in order are square, steel ball
steel ball and cylinder shaped charge and for other force ranges are cylinder, steel ball and square
and cylinder shaped charge and for other force ranges are cylinder, steel ball and square shaped charge.
shaped charge.
From the above results, it can be determined that media shape has a significant influence on the
From the above results, it can be determined that media shape has a significant influence on
impact force in the lifter. The results occur because of the media shape of the steel ball, cylinder and
the impact force in the lifter. The results occur because of the media shape of the steel ball, cylinder
square is point, line and surface contact, respectively. Moreover, adding iron ore particles buffers the
and square is point, line and surface contact, respectively. Moreover, adding iron ore particles
impact force on the lifter, which decreases the impact force.
buffers the impact force on the lifter, which decreases the impact force.
Materials 2017, 10, 882 14 of 17

Materials 2017, 10, 882 14 of 17

Steel ball Cylinder Square
250000
250000

200000
200000

150000 Steel ball Cylinder Square

150000 250000
100000 250000
100000 200000
50000 200000
50000 150000
0 150000
1000 0 -2 0 0
0
0-200 100000
1000
No. of impacts per 10 s

100000
800
50000
800

No. of impacts per 10 s

50000
600
0
600 0 -1 0 0
1 0 00
400 0-100
100
90
400
890
0
200
780
0
200
670
0
0
2 0 0 -4 0 0 4 0 0 -6 0 0 6 0 0 -8 0 0 560
0
2 00
200-400 400-600 600-800 450
0
120
8

118
6 340
0

116
4 230
0

114
2 120
0

112
0 10
0
1 0 0 -2 0 0 2 0 0 -3 0 0 3 0 0 -4 0 0 4 0 0 -5 0 0 ≥500
10 0
8
100-200 Force range
200-300 300-400 / N400-500 ≥500
68

46

24

02
8 0 0 -1 0 0 0 1 0 0 0 -1 2 0 0 1 2 0 0 -1 4 0 0 ≥1400
0
800-1000 Force range1200-1400
1000-1200 /N ≥1400

(a) Without particle (b) With particle

(a) Without particle (b) With particle
Figure
Figure 15. Effect of
15. Effect of particle
particle on
on impact
impact force.
force. (a)
(a) Without
Without particle;
particle; (b)
(b) With
With particle.
particle.

4.5.3. Effect of Particle on

4.5.3. on Power
Power Draw
Draw
To further
To furthercompare
comparethethe effect
effect of adding
of adding ironiron ore particles
ore particles on theongrinding
the grinding
process, process,
the millthe mill
power
power
and and particle
particle size distribution
size distribution are investigated.
are investigated. As shown As in
shown
Figurein16,Figure
adding16,iron
adding iron ore
ore particles
particles increases
increases the powerthe power
draw, anddraw,
resultsand
in aresults
powerindrawa power draw
increase increase
from 36.10 from 36.10toincreases
increases to
45.82 W at
45.82 W at experimental
experimental measurement. measurement.
Due to the iron Due oretoincreases
the iron the
oremill
increases
filling,the millpower
so the filling,draw
so the power
increases.
draw
For theincreases. For the
Bond predict Bond
result, thepredict result, the DEM
DEM simulation simulation
results results is more accurately.
is more accurately.

100
DEM simulation
Bond
80 Experiment

60
Power / w

40

20

0
Particle Without particle

Figure 16. Power draw with and without particles.

Figure 16. Power draw with and without particles.

To study the difference of media shape on the milling process, an iron ore particle size
To study the difference of media shape on the milling process, an iron ore particle size distribution
distribution analysis is the most direct way to explain the effect of multi-body impact breakage
analysis
behaviorisofthedifferent
most direct wayshapes.
media to explain the effect
As shown inof multi-body
Figure 17, theimpact breakagepassing
mass fraction behavior
of of different
cylinder is
media shapes. As shown in Figure 17, the mass fraction passing of cylinder is greater than
greater than that of ball and square. The percent of −0.074 mm particle is 8.08% (cylinder), 6.42% that of ball
and
(ball)square. The percent
and 5.14% (square), −0.074 mm particle
ofrespectively. is 8.08%
The cylinder, (cylinder),
steel ball and6.42%
square(ball) andhave
shapes 5.14%line,
(square),
point
and surface contact, so the cylinder projected to impact the particle has the optimum effect. This
result also demonstrates the effect of different media 1 shapes on the impact force in the lifter.
 Materials 2017, 10, 882 15 of 17

respectively. The cylinder, steel ball and square shapes have line, point and surface contact, so the
cylinder projected to impact the particle has the optimum effect. This result also demonstrates the
effect2017,
Materials of different
10, 882 media shapes on the impact force in the lifter. 15 of 17

100

Cylinder
80 Steel ball

Mass fraction passing / %

Square

60

40

20

0.0 0.5 1.0 1.5 2.0

Iron ore particle size / mm

Figure 17. 17.

Figure IronIron
oreore
particle sizesize
particle distribution.
distribution.

From the above results and discussions, it can be concluded that different media shapes have a
From the above results and discussions, it can be concluded that different media shapes have
different influence on grinding efficiency. Besides, adding the particle in the ball mill can influence
a different influence on grinding efficiency. Besides, adding the particle in the ball mill can influence
the milling process. Therefore, choosing a suitable media shape based on the milling environment is
the milling process. Therefore, choosing a suitable media shape based on the milling environment is
an effective method for improving grinding efficiency.
an effective method for improving grinding efficiency.
5. Conclusions
5. Conclusions
In In
thisthis
paper, a laboratory-scale
paper, ballball
a laboratory-scale mill is employed
mill in in
is employed investigating thethe
investigating impact of of
impact different
different
operating variables on the impact behavior of lifter. The determinations of ball trajectory,
operating variables on the impact behavior of lifter. The determinations of ball trajectory, impactimpactforce
force and power draw are used to characterize the impact behavior between the charge and
and power draw are used to characterize the impact behavior between the charge and lifter. The main lifter.
Theresearch
main research conclusions
conclusions are summarized
are summarized as follows.
as follows.
 The magnitude of impact force has strong sensitive to ball size and drop height. The impact
• The magnitude of impact force has strong sensitive to ball size and drop height. The impact force
force increases with the increase of ball size and drop height.
increases with the increase of ball size and drop height.
 Mill speed has profound effect on the ball trajectories, impact force and power draw. The mill
• canMill
reachspeed
the has
bestprofound
performance effectaton thethemill
ballspeed
trajectories,
rangingimpact
from force70% to and 80%power draw. speed,
of critical The mill
and correspondingly the maximum percent of the impact force between 600 and 1400 Nspeed,
can reach the best performance at the mill speed ranging from 70% to 80% of critical is
and correspondingly the maximum percent of the impact
obtained. For the 80% of critical speed, the measured power draw has a maximum value. force between 600 and 1400 N is
 obtained. For the 80% of critical speed, the measured power draw
Commonly, increasing the grinding media filling not only increases the fraction flying in the has a maximum value.
• Commonly,
cataracting zone increasing the grinding
but also increases the ratemedia filling
of ball on not
ball only
contactincreases
per unit. theInfraction flying itinisthe
this research,
interesting to find that increasing the grinding media filling does not the number of ball on ballit is
cataracting zone but also increases the rate of ball on ball contact per unit. In this research,
interesting
contact. to find force
The impact that increasing
can be only theabove
grinding
1400media filling
N at the does not
grinding the number
media filling ofof20%,ball on
andball
thecontact.
maximum The impact
percentforce can beforce
of impact only above
between 1400 200N at
andthe1400
grinding
N is media
obtained filling of 20%,
at the and the
grinding
maximum percent of impact force between 200 and 1400 N is
media filling of 20%. Besides, increasing the grinding media filling increases the power draw. obtained at the grinding media
 filling
Lifter of 20%.
height hasBesides, increasing
a significantly the grinding
influence on the media
loadfilling increases
behavior. the power
Increasing the draw.
lifter height
• Lifter height has a significantly influence on the load behavior. Increasing
increases the shoulder position and lifts the steel balls to the higher position,. The percent the lifter height increases
of
the shoulder
impact position
force ranging fromand lifts
0 to 200theNsteel balls to
decreases the the
with higher position,.
increase Theheight.
of lifter percent of impactthis
However, force
ranging
perfect willfrom 0 to 200
increase above N decreases
200 N. In with the increasing
general, increase of the lifter height.
lifter heightHowever,
does notthis leadperfect
to any will
increase above
variation in power draw.200 N. In general, increasing the lifter height does not lead to any variation in
 Thepower
impactdraw.
force will decrease and the power draw will increase when the iron ore particles
• are added. Theforce
The impact mass will decrease
fraction and the
passing of power
cylinder draw will increase
is greater than that when of the
ballironandore particles
square. Theare
added.
percent of The mass
−0.074 mm fraction
particlepassing
is 8.08% of (cylinder),
cylinder is6.42%greater(ball)
thanandthat5.14%
of ball(square),
and square. The percent
respectively.
 The −0.074 mm particle
ofsemi-empirical poweris draw
8.08%of(cylinder),
Bond model 6.42% (ball)paper
in this and 5.14%
does not (square),
apply respectively.
very satisfactorily
for the ball mills, while the power draw determined by DEM simulation show a good
approximation for the measured power draw.

Acknowledgments：This research was supported by the National Natural Science Foundation of China (grant
No. 51475458, U1510205), the Jiangsu Postgraduate Scientific Research and Innovation Projects and the
Program for Changjiang Scholars and Innovative Research Team in University (grant No. IRT_16R68). The
authors also wish to thank the Priority Academic Program Development of Jiangsu Higher Education Institutions
Materials 2017, 10, 882 16 of 17

• The semi-empirical power draw of Bond model in this paper does not apply very satisfactorily for
the ball mills, while the power draw determined by DEM simulation show a good approximation
for the measured power draw.

Acknowledgments: This research was supported by the National Natural Science Foundation of China (grant No.
51475458, U1510205), the Jiangsu Postgraduate Scientific Research and Innovation Projects and the Program for
Changjiang Scholars and Innovative Research Team in University (grant No. IRT_16R68). The authors also wish
to thank the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD) and the
Top-notch Academic Programs Project of Jiangsu Higher Education Institutions (TATP).
Author Contributions: Zixin Yin and Yuxing Peng conceived and designed the experiments; Zixin Yin and
Tongqing Li performed the experiments; Zixin Yin and Yuxing Peng analyzed the data; Zhencai Zhu provided
useful comments for this study; Zhangfa Yu contributed material; Zixin Yin and Yuxing Peng wrote the paper.
All authors read and approved the manuscript.
Conflicts of Interest: The authors declare no conflicts of interest.

References
1. Huang, P.; Jia, M.P.; Zhong, B.L. Study on the method for Collecting vibration signals from mill shell based
on measuring the fill level of ball mill. Math. Probl. Eng. 2014, 11, 1–10. [CrossRef]
2. Si, G.Q.; Cao, H.; Zhang, Y.B.; Li, J. Experimental investigation of load behaviour of an industrial scale
tumbling mill using noise and vibration signature techniques. Miner. Eng. 2009, 22, 1289–1298. [CrossRef]
3. Delaney, G.W.; Cleary, P.W.; Morrison, R.D.; Cummins, S.; Loveday, B. Predicting breakage and the evolution
of rock size and shape distributions in AG and SAG mills using DEM. Miner. Eng. 2013, 50, 132–139.
[CrossRef]
4. Li, T.Q.; Peng, Y.X.; Zhu, Z.C.; Zou, S.Y.; Liu, S.Y.; Yin, Z.X.; Ni, X.; Chang, X.D. Multi-layer kinematics
and collision energy in a large-scale grinding mill-the largest semi-autogenous grinding mill in china.
Adv. Mech. Eng. 2016, 8, 1–10. [CrossRef]
5. Chen, B.C. Grinding Theory; Metall. Ind. Press: Beijing, China, 1989.
6. Tian, Q.J. Study on the Working Performances of Large Tumbling Ball Mill Based on the Discrete Element
Method. Master’s Thesis, Jilin University, Jilin, China, 2011.
7. Powell, M.S.; Nurick, G.N. A study of charge motion in rotary mills part 2-experimental work. Miner. Eng.
1996, 9, 343–350. [CrossRef]
8. Zeng, Y.G.; Forssberg, E. Effects of operating parameters on vibration signal under laboratory scale grinding
conditions. Int. J. Miner. Process. 1992, 35, 273–290. [CrossRef]
9. Behera, B.; Mishra, B.K.; Murty, C.V.R. Experimental analysis of charge dynamics in tumbling mills by
vibration signature technique. Miner. Eng. 2007, 20, 84–91. [CrossRef]
10. Tang, J.; Zhao, L.J.; Yue, H.; Yu, W.; Chai, T.Y. Vibration analysis based on empirical mode decomposition
and partial least square. Procedia Eng. 2011, 16, 646–652. [CrossRef]
11. Pax, R.A. Non-Contact Acoustic Measurement of in-Mill Variables of SAG Mills. In Proceedings of the
International Conference on Autogenous and Semi-Autogenous Grinding Technology; University of British
Columbia: Vancouver, BC, Canada, 2001.
12. Tano, K.T.; Pålsson, B.I.; Sellgren, A. Online lifter deflection measurements showing flow resistance effects in
grinding mills. Miner. Eng. 2005, 18, 1077–1085. [CrossRef]
13. Moys, M.H.; Skorupa, J. Measurement of the radial and tangential forces exerted by the load on a liner in
a ball mill as a function of load volume and mill speed. Int. J. Miner. Process. 1993, 37, 239–256. [CrossRef]
14. Moys, M.H. The measurements of parameters describing the dynamic behaviour of the load in a grinding
mill. Int. Conf. Miner. Sci. Technol. 1984, 1, 205–219.
15. Kiangi, K.K.; Moys, M.H. Measurement of load behaviour in a dry pilot mill using an inductive proximity
probe. Miner. Eng. 2006, 19, 1348–1356. [CrossRef]
16. Deniz, V. The effects of ball filling and ball diameter on kinetic breakage parameters of barite powder.
Adv. Powder Technol. 2012, 23, 640–646. [CrossRef]
17. Kiangi, K.K.; Moys, M.H. Particle filling and size effects on the ball load behaviour and power in a dry pilot
mill: Experimental study. Powder Technol. 2008, 187, 79–87. [CrossRef]
Materials 2017, 10, 882 17 of 17

18. Deniz, V.; Onur, T. Investigation of the breakage kinetics of pumice samples as dependent on powder filling
in a ball mill. Int. J. Miner. Process. 2002, 67, 71–78. [CrossRef]
19. Schnatz, R. Optimization of continuous ball mills used for finish-grinding of cement by varying the L/D
ratio, ball charge filling ratio, ball size and residence time. Int. J. Miner. Process. 2004, 74, S55–S63. [CrossRef]
20. Kiangi, K.K. Effect of Particle Filling and Size on the Behaviour of the Ball Load and Power in a Dry
Mill. Ph.D. Thesis, Faculty of Engineering and the Built Environment, University of the Witwatersrand,
Johannesburg, South Africa, 2011.
21. Rezaeizadeh, M.; Fooladi, M.; Powell, M.S.; Mansouri, S.H. Experimental observations of lifter parameters
and mill operation on power draw and liner impact loading. Miner. Eng. 2010, 23, 1182–1191. [CrossRef]
22. Rajamani, R.; Delgadillo, J.; Duriseti, V. Online SAG Mill Pulse Measurement and Optimization Semi-Annual
Technical Progress Report; The University Of Utah: Salt Lake City, UT, USA, 2006.
23. Cleary, P.W. Charge behaviour and power consumption in ball mills: Sensitivity to mill operating conditions,
liner geometry and charge composition. Int. J. Miner. Process. 2001, 63, 79–114. [CrossRef]
24. Hlungwani, O.; Rikhotso, J.; Dong, H.; Moys, H. Further validation of DEM modeling of milling: Effects of
liner profile and mill speed. Miner. Eng. 2003, 16, 993–998. [CrossRef]
25. Djordjevic, N.; Shi, F.N.; Morrison, R. Determination of lifter design, speed and filling effects in AG mills by
3D DEM. Miner. Eng. 2004, 17, 1135–1142. [CrossRef]
26. Moys, M.H.; Smit, I. The effect of discharge arrangement, load volume, slurry rheology and mill speed on
power and slurry hold-up in rotary mills. Chem. Technol.-Bedfordview 1998, 27–31.
27. Mulenga, F.K.; Moys, M.H. Effects of slurry filling and mill speed on the net power draw of a tumbling ball
mill. Miner. Eng. 2014, 56, 45–56. [CrossRef]
28. Owen, P.; Cleary, P.W. The relationship between charge shape characteristics and fill level and lifter height
for a SAG mill. Miner. Eng. 2015, 83, 19–32. [CrossRef]
29. Deniz, V. The effect of mill speed on kinetic breakage parameters of clinker and limestone. Cem. Concr. Res.
2004, 34, 1365–1371. [CrossRef]
30. Deniz, V. Effects of mill speed on kinetic breakage parameters of four different particulate pumices.
Part. Sci. Technol. 2013, 31, 101–108. [CrossRef]
31. Huang, H.; Dallimore, M.P.; Pan, J.; McCormick, P.G. An investigation of the effect of powder on the impact
characteristics between a ball and a plate using free falling experiments. Mater. Sci. Eng. A 1998, 241, 38–47.
[CrossRef]
32. Dong, H.; Moys, M.H. Measurement of impact behaviour between balls and walls in grinding mills.
Miner. Eng. 2003, 16, 543–550. [CrossRef]
33. Xu, G.H.; Liu, Y.J.; Shen, W.C. A method of measuring the impact stresses in ball mills. Phys. Exam. Test.
1995, 1, 23–26.
34. Xu, G.H.; Shen, W.C.; Liu, Y.J. Measurement and calculation of impact stress of ball mill. J. Tsinghua Univ.
(Sci. Tech.) 1996, 36, 87–92.
35. Xu, G.H.; Liu, Y.J.; Shen, W.C.; Fang, H.S. Impact and breakage simulation experiment of ball mill.
Min. Process. Equip. 1994, 2, 18–22.
36. Wu, Q.; Bai, Y.; Zhang, J.G. Characteristic analysis on process of grinding ball impacting charge in ball mill.
Min. Process. Equip. 2014, 42, 68–72.
37. Katubilwa, F.M. Effects of Pool Volume on Wet Milling Efficiency. Ph.D. Dissertation, Faculty of Engineering
and the Built Environment, University of the Witwatersrand, Johannesburg, South Africa, 2012.
38. Mohsen, Y. Analysis of Liner Wear Impact on Charge Behaviour in AG and SAG Mills. Ph.D. Thesis, Faculty
of Engineering, Shahid Bahonar University of Kerman, Kerman, Iran, 2010.
39. Bond, F.C. Crushing and grinding calculations. Can. Min. Metall. Bull. 1961, 47, 466–472.
40. Koleini, S.M.J.; Barani, K.; Rezaei, B. The effect of microwave treatment on dry grinding kinetics of iron ore.
Miner. Process. Extr. Metall. Rev. 2012, 33, 159–169. [CrossRef]

© 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).