Minerals Engineering 73 (2015) 39–43

Contents lists available at ScienceDirect

Minerals Engineering
journal homepage: www.elsevier.com/locate/mineng

Evaluation of grinding media wear-rate by a combined grinding method

Junfu King a, Qiang Li b, Alex Wang (Baikun Wang) a,⇑, Cathy He a, Joey Zhou a, Henry Deng a, Ray Xu a
a
King’s Ceramics & Chemicals Co., Ltd., No. 910, 9th Section, Jinsong, Chaoyang District, Beijing, China
b
East China Normal University, No. 500, Dongchuan Road, Minhang District, Shanghai, China

a r t i c l e i n f o a b s t r a c t

Article history: It is well-known that lab tests on wear-rate of grinding media cannot precisely represent its industrial
Received 20 May 2014 performance due to complex grinding conditions. Nevertheless the lab data provides reference to the
Revised 27 November 2014 industrial data. Therefore, a reproducible test method on wear-rate in lab is necessary.
Accepted 5 December 2014
By providing detailed data, this paper challenges the traditional wear-rate testing methods. Two com-
Available online 14 January 2015
monly used methods on wear-rate test, self-wear in water and grinding with mineral slurry, are respec-
tively employed. However, obvious fluctuation of wear-rate/time curves indicates neither of the two
Keywords:
normal methods is reliable.
Grinding media
Wear-rate
This paper introduces a wear-rate testing method on lab scale which combines self-wear in water and
Combined grinding method grinding with mineral slurry. By this method, some repeatable wear-rate/time curves are displayed after
a few hours. The mechanism will be discussed in detail in this paper.
Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction 2. Experimental

Wear-rate is one of the most important factors when evaluating
the overall performance of grinding media. Therefore, extensive
tests on media wear have been carried out by many media suppli-
ers, users and labs (Berthiaux et al., 1996; Blecher et al., 1996a,b;
Blickensderfer and Tylczak, 1989; Frances, 2004; Jensen et al.,
2010; Radziszewski, 2002). According to some papers (Chenje
et al., 2009; Morrow and Sepulveda, 2014) and many feedback
from customers of KING’S CERAMICS’s, there is a correlation on
media wear-rates between lab equipments and industrial mills.
Therefore, to get repeatable media wear-rates in lab is very
meaningful to predict the media performance in industrial scale.
However, two common approaches on media wear test, namely
self-wear rate test in water and wear rate test with mineral slurry,
are always poor in repeatability in lab. As a grinding media
supplier, KING’S CERAMICS has been making efforts in developing
a reproducible method in lab to evaluate the wear performance of
different ceramic media.
In this paper, we designed a rapid wear-rate testing method on
lab scale which took the advantages of the above two mentioned
common methods, and comparatively, repeatable results can be
drawn from this new method.
⇑ Corresponding author. Tel.: +86 10 87763540; fax: +86 10 67785715x3540.
E-mail addresses: alex.wang@kingsbeads.com, wangbaikun86@126.com (A.
Wang).
2.1. Test machines
Vertical stirred mill Fig. 1 (left) and rapid pot mill Fig. 1 (right)
were used in order to provide robust grinding and impacting
forces. Also some detailed parameters of both the two mills are
listed in Table 1.
2.2. Test methods
2.2.1. Self-wear in water
1200.0 g grinding media and 400.0 g water were added into the
vertical stirred mill to grind for some time. Grinding was con-
ducted at a speed of 900 rpm. The same ceramic media of 600.0 g
and water 200.0 g were filled into the rapid pot mill for wear test
at 450 rpm.
2.2.2. Grinding with mineral slurry
1200.0 g ceramic media, 400.0 g water and 400.0 g zircon sand
(Eucla Zircon Premium Grade with d50 = 92 lm provided by Iluka
Resources, 7–8 in Moh’s hardness) were added into the vertical
stirred mill. Grinding speed was also 900 rpm. 500.0 g identical
media, 100.0 g water and 100.0 g zircon sand (same with above)
were filled into the rapid pot mill for wear test at 450 rpm.
Both wear-rate test methods are applied in vertical stirred mill
and rapid pot mill, with consecutive grinding for 1 h each time. The
method was repeated many times to achieve stable wear-rates.

http://dx.doi.org/10.1016/j.mineng.2014.12.002
0892-6875/Ó 2014 Elsevier Ltd. All rights reserved.
40 J. King et al. / Minerals Engineering 73 (2015) 39–43

Fig. 1. Vertical stirred mill (left) and rapid pot mill (right).

Table 1 Fig. 2 shows the self-wear curves of different grinding media in

Parameters of vertical stirred mill and rapid pot mill. water versus time. The lines marked with diamonds (A1 and B1)
Parameters Vertical stirred mill Rapid pot mill are the results in rapid pot mill, and those marked with circles (A2
and B2) are in vertical stirred mill for media A and B, only the
Motor power (kw) 0.55 0.45
Grinding chamber volume (l) 1.87 1.0 wear-rate curve of media C in rapid pot mill is examined. For media
Impeller length/maximum 100 150 A, after grinding for 7 h, a relatively stable wear-rate could be
diameter (mm) obtained. For media B, it almost needs 8 h to observe a stable values.
Operation speed (rpm) 900 450 In case of media C, only a rising wave line is received as shown in
Fig. 3, which could not tell us at which time point is reliable to eval-
uate the wear-rate. The results mentioned above seem to indicate
The hardness of zircon sand is relatively higher among natural that self-wear is not an adaptable method for different grinding
minerals, so the wear-rate differences among different grinding media.
media can be maximized.

2.2.3. Combined grinding method 3.2. Wear-rate in mineral slurry

Combined grinding method, which was proceeded in the rapid
pot mill, includes two steps: (1) grinding with mineral slurry: the Grinding in mineral slurry is usually used as a relatively reason-
same ceramic media of 500.0 g, water 100.0 g and zircon sand able method for testing wear-rate.
100.0 g were ground at 450 rpm for 1 h, and then another 2 h in fresh The wear-rate curves of different ceramic media grinding with
zircon sand slurry after drying the media. (2) Self-wear in water: mineral slurry can be found in Figs. 4 and 5. However, the curves
after grinding with slurry, ceramic media were washed, dried, and of media A and C (only in rapid pot mill) are wave lines in wide
then filled into rapid pot mill again (media to water mass ratio is
3:1) to grind for 1 h at 450 rpm, followed by another 2 h grinding
in new water under the same condition after drying the media.

2.2.4. Grinding media

Three kinds of ceramic media with different densities of 3.32,
4.05 and 4.23 g/cm3 were selected from KING’S CERAMICS and
the market, which were named as media A, B and C, respectively.

2.2.5. Calculation on wear-rate

The media wear-rate can be calculated by equation ‘‘wear-
rate = a/b/t’’, where a is the media loss (g) after grinding, b is the
initial media charge (kg), and t is the grinding time (h). The unit
of wear-rate is g
kg 1 h 1.

3. Results and discussion

3.1. Self-wear test

Self-wear is a simple and easy method, so it is widely used. But

it cannot simulate the grinding conditions of practical operation. Fig. 2. Self-wear curves of media A and B in vertical stirred mill and rapid pot mill.
 J. King et al. / Minerals Engineering 73 (2015) 39–43
Fig. 3. Self-wear curve of media C in rapid pot mill.
Fig. 4. Wear rate curves of media A and B in mineral slurry in vertical stirred mill
and rapid pot mill.
Fig. 5. Wear rate curve of media C in mineral slurry in rapid pot mill.
range. So we cannot select the time point at which we can deter-
mine the reliable wear-rate.
3.3. Parallel wear-rate tests
Another way to examine the reliability of wear tests is to obtain
the reproducible curves by several parallel tests. To magnify the
wear-rate and minimize the systematical error, wear tests were
only deployed in rapid pot mill.
41
Fig. 6. Parallel tests on self-wear and wear rate in mineral slurry of media A in rapid
pot mill.
Fig. 6 gives the wear-rate curves of media A through parallel
tests by self-wear and grinding with mineral slurry in rapid pot
mill. The results indicate bad reproducibility. The wear-rate insta-
bility can be explained from the abrasive and abrasion processes
during the grinding. Different abrasive processes induced by differ-
ent mechanical forces play different roles in processes of self-wear
and grinding with mineral slurry.
3.4. Wear mechanism discussion
3.4.1. Self-wear mechanism
In the case of self-wear, impact force among different media
plays an important role to destroy the media surface structure.
Firstly impact induces defects on the surface due to plastic defor-
mation of crystal particles (Kang and Hadfield, 2005). The forma-
tion probability of defects is usually low, because the plastic
deformation can absorb some energy brought by impact to prevent
the media surface from being destroyed. But accumulated defects
would promptly lead to the structure collapse on the surface under
larger impact force (Torrance, 2005). When collapses take place,
the wear-rate rises. So, wave-like curves of wear-rate might be
observed after self-wear test, just like those in Fig. 2. Here the for-
mation of defects is a slow process while the collapse is a fast one.
After quick collapse, the surface can be cleaned and polished
quickly. In Fig. 7, collapses and holes on surface of media A can
be seen although the surface is smooth generally. This indicates
that the mechanical property of media A is not good enough to
resist the impact and friction when ground in the water. For media
B, holes with size around 1–2 lm are in presence as well. While for
media C, surface defects are less and micro-cuttings (slim
scratches) can also be observed thanks to its good balance between
hardness and fracture toughness.
3.4.2. Wear in mineral slurry
During grinding with mineral slurry, the frictional force from
slurry plays more important role (Becker and Schwedes, 1999;
Cleary and Morrison, 2011; Gers et al., 2010). It causes the corro-
sion on the surface of ceramic media. For most ceramic products,
including ceramic media, sintering aids are needed in the produc-
tion process, which will transform into grain boundary phase after
sintering. The corrosion preferentially takes place at the grain
boundary phase, and the defects are induced at the same place.
Fig. 8 shows typical SEM photos of media surface after grinding
in mineral slurry. Rough surfaces can be seen after grain bound-
aries are corroded. Compared with Fig. 7, the surfaces of media A
and B are more unsmoothed after grinding in mineral slurry. For
42 J. King et al. / Minerals Engineering 73 (2015) 39–43

Media A Media B Media C

structure collapse

micro-cutting

N x10k 10µm

Fig. 7. Typical SEM photos of media surfaces after self-wear.

Media A Media B Media C

N x10k 10µm

Fig. 8. Typical SEM photos of media surfaces after grinding with mineral slurry.

media C, more defects on the surface are observed and the surface
is not smooth anymore. When the defects accumulate to cause the
Table 2
structure collapse on the surface, wear-rate will rise suddenly Statistical analysis on wear-rates in mineral slurry of different media by combined
(Rezaeizadeh et al., 2010). Usually the formation of defects is a fast grinding method.
process while the collapse is a slower one. As the corrosion goes on
Media A Media B Media C
and the defects increase quickly, the collapse will increase in fre-
quency, and the curves of wear-rate will level off. This abrasive Wear-rate in mineral slurry (%) 0.400 0.409 0.413
0.395 0.408 0.402
process brings typical wear-rate curves in Figs. 4 and 5. 0.409 0.378 0.415
0.394 0.385 0.400
3.5. Combined grinding method 0.398 0.392 0.405
Lower confidence limit (%) 0.392 0.377 0.399
After analyzing and comparing the abrasive process of self-wear Upper confidence limit (%) 0.407 0.412 0.415
and grinding with mineral slurry, it can be proposed that the for-
mation of defects does not synchronize with structure collapses,
which makes the wear-rate unstable. If the fast process of defects
Table 3
formation when grinding with mineral slurry and the fast process Statistical analysis on self-wear rates of different media by combined grinding
of surface collapses in self-wear are combined, a fast and reliable method.
method for testing wear-rate may be found.
Media A Media B Media C
Basing on this assumption, the combined method is proposed as
Self-wear rate (%) 0.958 0.750 1.620
following steps (also in rapid pot mill): (1) grinding in mineral
0.950 0.758 1.595
slurry for 1 h; (2) grinding in fresh mineral slurry for another 0.946 0.812 1.578
2 h; (3) self-wear in water for 1 h; and (4) self-wear in fresh water 0.949 0.765 1.608
for another 2 h. After each step, the media will be washed clean 0.952 0.790 1.582
and dried to test the wear-rate before the next step. Step (1) is to Lower confidence limit (%) 0.945 0.743 1.575
simulate the media surface with that has grinding in mineral slurry Upper confidence limit (%) 0.957 0.807 1.618
for a period of time, and the wear-rate then obtained in step (2) can
be seen as the wear-rate grinding with mineral slurry. Step (3) is to
remove the defects and recover the media surface to a smooth sit- monitor the wear rate the daily production of ceramic media. For
uation, and the wear-rate obtained in step (4) can be treated as the further study, much work should be done to comprehensively
self-wear rate. research the wear mechanism of grinding behavior.
Tables 2 and 3 display the wear-rate values of different grinding
media and corresponding lower and upper confidence limits when
95% confidence level for the accuracy of wear rates are set. Taking 4. Conclusions
media A for example, if the wear rate in mineral slurry is between
0.392% and 0.407%, then the wear rate result is convincing. This 6 h Wear-rates of different grinding media are determined by two
combined grinding method is rapid and reliable to frequently commonly used methods of self-wear in water and grinding with
 J. King et al. / Minerals Engineering 73 (2015) 39–43
mineral slurry. After the analysis of all results, we design an
improved testing method on wear-rate, which combines the fast
process of defects formation when grinding with mineral slurry
with the fast process of surface collapses of self-wear in water.
The method proves to be reliable in terms of stability and objectiv-
ity. However, it is worthwhile mentioning that due to the compli-
cation of the grinding process, a further in-depth investigation on
wear mechanism is needed in the future.
Acknowledgements
The authors wish to thank our cooperative partner of East China
Normal University for the testing work during the course of these
experiments. This work has been carried out with the help of work-
ing staff of Q&C Center and R&D Center of KING’S CEREMICS’s.
Their support is also gratefully acknowledged.
References
Becker, M., Schwedes, J., 1999. Comminution of ceramics in stirred media mills and
wear of grinding beads. Powder Technol. 105 (1–3), 374–381.
Berthiaux, H., Heitzmann, D., Dodds, J.A., 1996. Validation of a model of a stirred
beads mill by comparing results obtained in batch and continuous mode
grinding. Int. J. Miner. Process. 44–45, 653–661.
43
Blecher, L., Kwade, A., Schwedes, J., 1996a. Motion and stress intensity of grinding
beads in a stirred media mill. Part 1: energy density distribution and motion of
single grinding beads. Powder Technol. 86 (1), 59–68.
Blecher, L., Kwade, A., Schwedes, J., 1996b. Motion and stress intensity of grinding
beads in a stirred media mill. Part 1: energy density distribution and motion of
single grinding beads. Powder Technol. 86 (1), 69–76.
Blickensderfer, R.A., Tylczak, J.H., 1989. Evaluation of commercial US grinding balls
by laboratory impact and abrasion tests. Miner. Metall. Process. 6, 60–66.
Chenje, T., Radziszewski, P., Hewitt, D., 2009. Steel media wear: experimentation,
simulation and validation. In: Proc. 2006, 41st Meeting of the Canadian Mineral
Processors, pp. 557–572.
Cleary, P.W., Morrison, R.D., 2011. Understanding fine ore breakage in a laboratory
scale ball mill using DEM. Miner. Eng. 24 (3–4), 352–366.
Frances, C., 2004. On modelling of submicronic wet milling processes in bead mills.
Powder Technol. 143–144, 253–263.
Gers, R., Climent, E., Anne-Archard, D., Frances, C., 2010. Numerical modelling of
grinding in a stirred media mill: hydrodynamics and collision characteristics.
Chem. Eng. Sci. 65 (6), 2052–2064.
Jensen, L.R.D., Fundal, E., Møller, , Jespersen, M, P., Jespersen, M., 2010. Prediction of
wear-rates in comminution equipment. Wear 269 (7–8), 525–533.
Kang, J., Hadfield, M., 2005. Examination of the material removal mechanisms
during the lapping process of advanced ceramic rolling elements. Wear 258 (1–
4), 2–12.
Morrow, R., Sepulveda, J.E., 2014. Moly-cop tools applications for the assessment of
grinding media performance at full industrial scale. In: 46th Annual Canadian
Mineral Processors Operators Conference. Ottawa, Canada.
Radziszewski, P., 2002. Exploring total media wear. Miner. Eng. 15 (12), 1073–1087.
Rezaeizadeh, M., Fooladi, M., Powell, M.S., Mansouri, S.H., Weerasekara, N.S., 2010. A
new predictive model of lifter bar wear in mills. Miner. Eng. 23 (15), 1174–
1181.
Torrance, A.A., 2005. Modelling abrasive wear. Wear 258 (1–4), 281–293.