You are on page 1of 11

Wear 225–229 Ž1999.

1059–1069

A comparison of laboratory abrasion and field wear results


Joe H. Tylczak ) , Jeffery A. Hawk, Rick D. Wilson
US Department of Energy, Albany Research Center, 1450 Queen AÕenue SW, Albany, OR 97321, USA

Abstract

In the ongoing battle on wear, laboratory tests have been one tool used to evaluate and model the process of wear. A second, less
commonly used tool is field wear testing. Field wear testing, while being more time-consuming, has the advantage that the materials are
exposed to the actual environmental conditions and abrasives responsible for the wear loss. This paper examines four different abrasive
wear tests Žpin-on-drum, dry-sand rubber-wheel, jaw crusher, and impeller-in-drum., and compares the results obtained from these tests
with field wear tests using the Albany Research Center’s Planar Array Field wear test. A variety of ferrous-based alloys commonly used
to resist abrasion in the mineral processing industry were tested, including carbon steels, low alloy steels, austenitic steels, and white cast
iron. q 1999 Published by Elsevier Science S.A. All rights reserved.

Keywords: Abrasion; Field wear testing; Laboratory wear testing; Steel

1. Introduction specific sitertest conditions. Temperature, humidity, ve-


locity, abrasive particle size and type, ore characteristics,
Wear, and abrasive wear in particular, is an ongoing etc., can influence abrasive wear and limit the analytical
expense in many industries. Dozens of laboratory wear value of a specific field test. Despite these disadvantages,
tests have been developed to model specific wear prob- however, field wear tests have played a critical role in
lems, or conditions, with varying degrees of success. The wear research w4x.
problem with using laboratory wear tests tends to be the Because the goal of wear research is to understand and
inability of these tests to accurately and completely model reduce field wear, all test datarconclusions must reflect
the conditions found in the field, or the fear that the real world conditions. Considerable effort has been put
laboratory test does not adequately model the conditions into the design of laboratory wear tests, trying to make
found in the field. The lack of faith in laboratory wear tests them simulate some ‘real’ condition, such as the impact-
can lead to the field wear testing of materials, which are abrasion of ore impact hammers, the abrasion of wire
used to predict the component’s lifetime. A poorly de- ropes used in draglines and the wear associated with
signed field wear test is fraught with the hazards of scraper and bulldozer blades. Classifications such as two-
spending considerable time and expense in testing a mate- body and three-body, high- and low-stress conditions have
rial, only to find out that it is unsuitable for the task at been developed over the years to describe abrasion pro-
hand. However, a well-designed field wear test can be a cesses in order facilitate meaningful discussion w5x. Conse-
valuable source of data and insight. The Albany Research quently, laboratory wear tests have an advantage over field
Center ŽARC. has conducted numerous field and labora- wear tests relative to time and cost issues, but there is
tory wear tests w1–3x. always the risk that they and their resulting data will not
Field wear tests are notoriously expensive, time-con- correlate sufficiently well with the performance of a mate-
suming, invasive, and difficult to control andror quantify. rial in real world conditions. If the wear mechanisms
In addition, before generalizations can be made as to the documented in a laboratory wear test do not replicate those
wear rate and mechanisms, multiple tests need to be found under field wear conditions, the test would be
conducted at numerous sites to mitigate the impact of useless in predicting the wear behavior.
Finding statistical correlations between field and labora-
tory test data would have a significant impact on the
)
C orresponding author. Fax: q 1-5419675845; e-m ail: design of future research programs. Therefore, the Albany
tylczak@alrc.doe.gov Research Center has conducted two field tests and four

0043-1648r99r$ - see front matter q 1999 Published by Elsevier Science S.A. All rights reserved.
PII: S 0 0 4 3 - 1 6 4 8 Ž 9 9 . 0 0 0 4 3 - 5
1060 J.H. Tylczak et al.r Wear 225–229 (1999) 1059–1069

laboratory wear tests to determine if such corollary rela- 2. Experimental procedures


tionships exist. Thus, the purpose of this paper is to
compare the wear in field conditions for a suite of ferrous- 2.1. Planar array field wear test
based alloys to the wear of the same alloys tested in the
laboratory. For the purposes of this paper, the planar array The specimen wear test plate used in PAFWT is shown
field wear test ŽPAFWT., shown in Fig. 1, was used. The in schematic form in Fig. 1. The PAFWT steel plate,
PAFWT w1x has been successfully used in the past to typically 280 = 330 = 20 mm, with two dovetail grooves
screen and rank a standard suite of wear resistant alloys. for the individual wear test specimens to reside is inserted
The laboratory wear tests used in this study include ASTM into an ore stream in a crushing–grinding facility. The
G 65, dry-sand rubber-wheel abrasion test ŽDSRW. w6x; PAFWT plate replaces a solid metal wear plate somewhere
ARC developed pin-on-drum abrasive wear test ŽPOD. w7x; in the ore conveyance system and allows rock and mineral
ARC’s variation of the jaw crusher gouging wear test ŽJC. of different sizes to either pass directly over the plate, slide
w8x; and ARC version of the impeller-in-drum impact against it, or impact it at different velocities, angles, and
abrasive wear test ŽIW. w9x. stress levels.

Fig. 1. Specimen configuration used in the PAFWT.


J.H. Tylczak et al.r Wear 225–229 (1999) 1059–1069 1061

Test specimens of the wear resistant ferrous-based al- of the adjacent standard reference specimens is taken as
loys Ži.e., commercially available wear resistant steels. and the correction factor. The corrected wear depth d c is
standard reference specimens Ži.e., generic irons and steels calculated for each test specimen according to the follow-
like AISI 4340. are inserted into the grooves in an alternat- ing equation
ing fashion prior to installation of the plate in the
crushing-grinding facility ŽTable 1 contains a summary of dR
the alloys with chemistries, densities and hardnesses.. This dc s d
d RA
arrangement of test and standard specimens allows the
simultaneous testing of a large number of test specimens.
The specimen configuration also allows for the correction where d is the wear depth of the test specimen calculated
of any local gradients in wear intensity across the face of from mass change measurements, d R is the mean wear
the plate as a result of the flow of mineral and rock against depth of all standard reference specimens, and d RA is the
the plate. For example, if all of the standard reference mean wear depth of the two standard reference specimens
specimens Ži.e., hardened AISI 4340. exhibit identical adjacent to the test specimen.
wear, the wear of the plate would be uniform and no The wear depth d of each specimen was determined
corrections would need be made. However, in most cases, from the mass loss Žmeasured to "0.01 g., density Žmea-
given the orientation of the plate in the facility Že.g., as a sured to three significant figures., and mean wear surface
sidewall., wear is non-uniform from top to bottom, and area Ždimensions measured before and after the test in
some correction will have to be made at a local level. several places and averaged to "0.01 mm.. Prior to any
A previously developed equation w10x was used to cor- measurements, the specimens were degreased, scrubbed
rect for these local wear gradients. The wear on the with a bristle brush, ultrasonically cleaned in soapy water,
standard reference specimens adjacent to each test speci- rinsed with water, rinsed with alcohol, and blown dry with
men indicates the intensity of the wear gradient on that a warm air stream.
section of the plate. The ratio of the mean wear of all For the PAFWT experiments, specimens were nomi-
standard reference specimens to the observed mean wear nally 25 = 75 = 12 mm, with the 25 = 75 mm surface of

Table 1
Test materials composition and hardness
Specimen number Common designation Composition Density Hardness
C Cr Mn Mo Ni Žgrcm3 . ŽHB.

Carbon steels
W93 AISI 1060 0.50 – 0.53 – 0.13 7.76 673
W375 Commercial steel a 0.15 0.02 1.50 0.05 0.10 7.85 368
W376 Commercial steel 0.25 0.02 1.50 0.05 0.10 7.84 409
W377 Commercial steel 0.31 0.02 1.50 0.05 0.10 7.82 509
W378 Commercial steel 0.26 0.02 1.50 0.05 0.10 7.78 392
W379 Commercial steel 0.14 0.02 1.50 0.05 0.10 7.81 469

Low alloy steels


W184 A514 type B steel 0.14 0.50 1.33 0.23 0.20 7.83 264
W198 REM 500 0.33 0.98 0.61 0.20 0.14 7.81 473
W311 AISI 4340 steel 0.40 0.76 0.73 0.30 1.80 7.81 515
W314 AISI 52100 steel 1.04 1.47 0.36 0.04 0.13 7.81 635
W367 Commercial steel 0.32 1.40 0.26 0.35 3.42 7.85 415
W368 Commercial steel 0.34 1.40 0.26 0.35 3.42 7.83 514
W369 Commercial steel 0.26 1.48 0.90 0.30 3.44 7.84 461
W370 Commercial steel 0.22 1.48 0.90 0.30 3.44 7.85 425
W371 Commercial steel 0.20 1.24 1.11 0.49 1.43 7.84 444
W372 Commercial steel 0.25 1.24 1.11 0.49 1.43 7.84 448
W373 Commercial steel 0.33 0.29 2.47 0.16 0.80 7.84 437
W374 Commercial steel 0.19 0.53 0.97 0.20 0.10 7.85 387

High alloy steels and cast irons


W313 D2 tool steel 1.60 13.70 0.45 0.83 0.19 7.67 597
W299F 17% Cr white iron 3.19 15.30 0.76 1.02 0.50 7.64 709

Austenitic steels and stainless steels


W316 304 SS – 18.90 2.38 0.30 8.00 7.90 156
W315 13% Mn steel 1.12 0.44 12.80 0.10 0.20 7.86 224
a
Commercially available wear resistant alloy steels.
1062 J.H. Tylczak et al.r Wear 225–229 (1999) 1059–1069

the wear specimen exposed to the sliding and scratching of The second laboratory wear test was conducted using
the moving ore and rock Žsee Section 2.2 for further the dry-sand rubber-wheel abrasion test. This test simulates
details.. Test specimens were selected at random to fill low-stress, three-body abrasive wear. This type of wear
each dovetailed groove in the PAFWT plate. occurs in the mining industry in linkages, pivot pins, and
wire ropes, which suffer slow wear from the sliding and
2.2. Field wear test sites rolling action of abrasive fragments of rock and ore trapped
between metal surfaces. This test became an ASTM stan-
dard in 1980 and has become the most popular abrasive
The PAFWT plate was run at the Asarco Troy mine in
wear test in the United States.
Troy, Montana. This is a goldrsilver mine with the ore
This test consists of a rubber wheel f 228 mm = 12.7
consisting of chert. The test plate was installed for a period
mm thick, that turns at 200 rpm. A curtain of 50r70 mesh
of 52 days, during which time a total of 225,000 tons of
Ž300r200 mm nominal particle size. rounded silica sand
ore passed over its surface at a velocity of ; 0.1 mrs. The
flows between the rubber wheel and the rectangular test
ore from this mine was moist with a neutral pH. The ore
specimen, 25 = 75 = 12 mm. The specimen is held against
ranged in size from 50 to 1000 mm and was angular in
the rubber wheel using a lever arm with a force of 130 N.
geometry.
The duration of the test is equal to a distance of 1436 m
The second PAFWT plate was run at Vulcan Materials
ŽProcedure B.. A more detailed description of the test
in Helena, AL. In this particular case, a limestone quarry
apparatus is contained in ASTM G65-95 w6x and Labora-
was selected because its mildly abrasive environment would
tory Wear Testing Capabilities of the Bureau of Mines w8x.
give different wear rates compared to those of the Asarco
The wear rate is given as a volume loss per unit distance
Troy mine. This test plate was installed for a period of 274
Žmm3rm., where the abrasion distance is equal to the
days, during which time a total of 200,000 tons of ore
lineal distance traversed by the rubber wheel during the
passed over the specimens. The velocity of the ore passing
test.
over the samples was ; 3 mrs. The crushed limestone
The third laboratory wear test was the jaw crusher
was dry Žalthough the relative humidity was high., ranged
gouging-abrasion test w11x. This test simulates a type of
in size from 20 to 30 mm, and was angular in shape.
high-stress wear that may be produced by either two-body
or three-body conditions, where the mineral is pressed into
2.3. Laboratory wear tests the surface of the metal with stresses high enough to
remove large chunks of material. The jaws used to crush
Four laboratory wear tests were used to generate data the rock are taken as the test specimen. Several investiga-
on the abrasive wear rates of the tests specimens. This tors believe that the jaw crusher test gives the best correla-
information was gathered to see if any of the field wear tion to wear that occurs on earth-penetrating equipment,
test data followed trends which could be described by the such as excavator teeth, power shovel buckets, scoops, and
laboratory wear tests. grader blades, as well as real jaw crusher wear.
The first laboratory wear test was conducted using the The jaw crusher used in these tests is smaller than the
pin-on-drum w7x, which was specifically designed by Blick- one described in Ref. w9x, and consists of 50 mm wide jaws
ensderfer et al. to simulate two-body abrasive wear w8x. that has been stiffened and rebuilt to provide precision
The pin-on-drum abrasive wear test involves high-stress, wear results. This apparatus uses specimen similar in size
two-body abrasive wear. One end of a cylindrical pin and shape to the DSRW test, with the addition of tapered
specimen is moved over an abrasive paper, abrading mate- ends to hold them in place during the test. A more detailed
rial from the specimen and crushing the fixed abrasive description of the test apparatus is contained in Laboratory
grains of the abrasive paper in the process. The wear is Wear Testing Capabilities of the Bureau of Mines w8x.
believed to simulate grinding abrasion that occurs during Subsequently, the test has been modified to use rock of
the crushing and grinding of ore in ball mills. different types, varying in quantity, and to report wear as
In this test, a rotating 6.35 mm test pin is pressed the actual volume loss per unit mass of rock crushed
against f 500 mm drum with a load of 66.7 N. The drum instead of as a ratio Ži.e., in mm3rkg.. The test now uses a
is covered with an abrasive cloth, in this case a 150 mesh coarse, angular, high-silica quartzite rock with a diameter
garnet Žnominal particle size of 100 mm.. The drum is then of f 25.4 mm as the standard Žalthough other ores can and
rotated and the pin translated down the axis of the drum have been used—e.g., limestone, granite, river rock, etc..
Žin a helical fashion. so fresh abrasive is constantly en- For each jaw crusher run, two test materials, for a total of
countered. The test duration is selected to achieve at least four wear plates-two of each specimen, are run. The test is
40 mg of specimen mass loss. The mass loss is then set up so one specimen of each test material is positioned
converted into a volume loss per unit distance traversed on in the movable jaw and one is positioned in the stationary
the abrasive, i.e., mm3rm. A more detailed description of jaw, with test material ‘A’ facing test material ‘B’. Speci-
the test apparatus is contained in Laboratory Wear Testing mens A and B are selected so they are approximately of
Capabilities of the Bureau of Mines w8x. the same hardness. After running 11.4 kg of high-silica
J.H. Tylczak et al.r Wear 225–229 (1999) 1059–1069 1063

quartzite through the jaws, the A and B specimens are The last class of alloys were high-chromium high-carbon
weighed to determine the mass loss, then the positions of alloys, consisting of large chromium carbides in a marten-
the specimens are reversed Ži.e., the ones in the movable sitic matrix. These alloys included a 17% Cr white iron
jaw are placed in the stationary jaw and vice versa. for a and a D2 tool steel with hardnesses of HB 709 and 597,
second test. An additional 11.4 kg of high-silica quartzite respectively.
is crushed and the specimens are once again weighed. The Several of the tests depend on reference specimens to
wear rate is thus the average mass loss from the two tests. correct for otherwise uncontrollable changes in abrasive
The fourth laboratory test used was the ARC developed conditions, such as changes in wear intensity in the PAFWT
impeller-in-drum wear test w9x. This test develops high- as a result of position in the plate with respect to the
stresses from high velocity impacts against the surface of abrasive and differences in the abrasive type, and for
the specimen, and is considered an open two-body test. changes in the abrasiveness of the garnet paper used in the
This test simulates the movement and handling of loose POD that result from using different lots. The reference
ores and rock that are common during manipulation and specimens for the PAFWT were made of AISI 4340 steel,
transport. This laboratory test also has a great advantage in hardened to 515 " 6 HB. The reference specimens for the
that a variety of ores can be used in it. POD test were made of ASTM A514 type B steel with a
The general test procedure for the impeller-tumbler hardness of 269 HB.
starts with sizing the ore in the range of y25 mm to q19
mm. After this step, 600 g of ore is measured and the
number of particles that make up the charge is counted.
Typically, for 600 g of high silica quartzite, the number of 4. Results and discussion
particles range from 38 to 44. The 600 g charge is then
loaded into the impeller and the cover is closed. The drum Fig. 2 shows the results of POD wear testing with the
and impeller are then set into motion. This marks the wear rate plotted as a function of material hardness ŽHard-
beginning of the first of two 1-h tests. The speed of the ness in this case is the pre-test hardness of the alloys.. This
impeller is adjusted to 620 " 5 rpm for the first 15-min figure generally shows the classic correlation of decreasing
interval. After the first 15-min test interval has elapsed, the wear rate with increasing alloy hardness, with the general
test is stopped, the cover is removed and the ore is exceptions being the austenitic 13% Mn steel and the 304
collected. A fresh 600 g charge of ore is then placed in the SS, which showed lower wear rates than their hardnesses
drum and the procedure is repeated for a second 15-min would seem to indicate ŽPrevious studies have shown that
interval. This is done twice more, for a total running time work-hardening during wear testing increases the hardness
of 1 h. The amount of ore passed through the system is 2.4 of the wear surface w9x.. A line of best fit was constructed
kg. After the four 15-min tests, the samples are removed, for the alloys used in the POD test minus the two austenitic
thoroughly cleaned and then warm-air-dried. They are alloys. The equation for this line was: wear rate POD s
weighed to the nearest tenth of a milligram to determine 1.667 y 0.0018ŽHB.. The correlation Ž r 2 . between wear
the mass loss. The specimen is then reversed and a dupli- rate and Brinell hardness was 0.82. For the POD wear test
cate series of four 15-min tests is run. The results of the the highest wear rate was exhibited by the softest of the
two series of tests are averaged and the standard deviation low alloy steels, i.e., the A514 type B steel ŽW184.. The
is calculated. The wear rate is given in terms of the volume alloy with the lowest wear rate was also the hardest
of material removed in 1 h and for comminuting 2.4 kg of material, the 17% Cr white cast iron ŽW299F..
rock Žmm3rkg h.. Fig. 3 shows the results for the DSRW test, where the
wear rate is plotted as a function of Brinell hardness. A
general trend of decreasing wear rate with increasing
hardness is observed. A line of best fit was also con-
3. Materials structed for this data set minus the data point for the 13%
Mn steel ŽW315.. The equation for the line in this case
The materials used in this investigation are shown in was: wear rate DSRW s 0.148 y 0.0002ŽHB.. Correlation
Table 1. A total of 22 alloys were used is this study, but between the wear rate and Brinell hardness was 0.92. The
not all the materials were used in every test. The materials 13% Mn steel ŽW315. again had a much lower wear rate
belong to four general classes of iron-based alloys. The than its hardness would indicate. For the case of the
first class consisted of hardened, low to medium, carbon DSRW wear test, the lower stress was not enough to
steels. These alloys ranged in hardness from HB 368 to workharden the 304SS ŽW316. which had a wear rate that
673. The second class of alloys consisted of low alloy coincided well with the other data. The two alloys with the
steels. These ranged in hardness from HB 425 to 635. A large volume fractions of chromium-based carbides had
third class of alloys were austenitic steels, consisting of a the lowest wear rates, i.e., the 17% Cr white cast iron
13% Mn steel and 304 stainless steel. These two steels had ŽW299F. —lowest and the D2 tool steel ŽW313. —next
initial pre-test hardnesses of HB 224 and 156, respectively. lowest.
1064 J.H. Tylczak et al.r Wear 225–229 (1999) 1059–1069

Fig. 2. Graph of wear rate as a function of Brinell hardness for the POD laboratory wear test. 100 mm garnet particles were used as the abrasive. Normal
load was 66.7 N.

Fig. 4 shows the results for the JC gouging–abrasion these materials than for cases of pure abrasion described in
wear test where the wear rate is plotted vs. alloy hardness. Figs. 2 and 3. For the alloys tested in this study, there was
The JC wear data are not as well behaved as either the no linear relationship between wear rate and Brinell hard-
POD and DSRW, and the trend of decreasing wear rate ness. However, in fitting various equations to the data set,
with increasing hardness is not clearly defined, especially an exponential equation of the general form: y s a q b
for the alloys with carbide phases ŽW299F and W313.. expŽyxrc ., described the general trend of the data, al-
Microscopic observations of the alloys with carbide phases though correlation between wear rate and hardness was not
show that these phases are fractured during JC testing, and very good Ž0.64.. This may not be a statistically appropri-
as a result, are more prone to removal during the abrasion ate approach since only three data points are available at
part of the process. This leads to higher wear rates for the lower hardness values. More data points are needed to

Fig. 3. Graph of wear rate as a function of Brinell hardness for the DSRW laboratory wear test. 200–300 mm rounded silica particles were used as the
abrasive. Normal load was 130 N.
J.H. Tylczak et al.r Wear 225–229 (1999) 1059–1069 1065

Fig. 4. Graph of wear rate as a function of Brinell hardness for the JC laboratory wear test. 25–35 mm high silica quartzite were used as the abrasive.

verify the validity of the trend described by the exponen- steel’s surface to significantly workharden before material
tial equation, or to dismiss it as an artifact of the materials is worn away Žfor high silica quartzite only w9x.. In addi-
tested in this study. tion, the stresses appear to be too high for the chromium
In general, the low alloy steels Žsee Table 1. performed carbide phases to remain intact as a result of the gouging-
well relative to the austenitic steels and carbide reinforced abrasion. As a result of the test conditions, they fracture
alloys. Also, the 13% Mn steel ŽW315. performed well during the high-stress gouging phase of the wear process
given its relatively low hardness Žas it did in the other and are easily removed from the surrounding matrix during
wear tests., but not significantly better than the low alloy the abrasion phase, leading to a higher wear rate than
steels. It is suspected that the nature of the JC test, i.e., would be expected given their hardness. These carbide
high-stress, gouging–abrasion, does not allow the 13% Mn reinforced materials do not provide any real advantage in

Fig. 5. Graph of wear rate as a function of Brinell hardness for the Impeller-in-drum laboratory wear test. 19–25 mm high silica quartzite particles were
used as the abrasive.
1066 J.H. Tylczak et al.r Wear 225–229 (1999) 1059–1069

Fig. 6. Graph of wear rate as a function of Brinell hardness for Plate 22 used in the PAFWT at the Asarco Troy Mine in Troy, MT. A chert ore served as
the abrasive.

gouging–abrasion environments. For this test and this suite very minor trend because of the scatter in the data, with
of materials, the best wear rate was exhibited by alloys increasing hardness is seen. In general, the carbide rein-
W367 and W368 with 0.3% C and 1.4% Cr. forced materials perform well as do the 13% austenitic Mn
Fig. 5 shows the results for the impeller-in-drum wear. steel ŽW315. and the ASTM A514 low alloy steel ŽW184..
For the materials tested in this study, there is no correla- One way in which impeller-in-drum data can be used is in
tion between wear rate and Brinell hardness ŽOther data conjunction with test data from the other laboratory tests.
indicate that wear rate decreases with increasing material In this way, if a material performs well in another setting
hardness w9x.. With the exception of the two austenitic Žscratching or grinding abrasion., but does not perform
alloys ŽW315 and W316. which significantly work-harden well Žrelative to other alloys. in the impeller-in-drum test,
during impact-abrasion testing w9x, and have lower than this would be known and substitute materials could be
expected wear rates, a trend of decreasing wear, albeit a selected that do perform well in all environments. So this

Fig. 7. Graph of wear rate as a function of Brinell hardness for Plate 20 used in the PAFWT at Vulcan Materials in Helena, AL. A limestone ore served as
the abrasive.
J.H. Tylczak et al.r Wear 225–229 (1999) 1059–1069 1067

Fig. 8. Correlation between the normalized wear rate using the DSRW vs. the normalized wear rate for the PAFWT Plate 20.

particular test may not work well as a predictive instru- ness. The equation of the line that fit the data was: wear
ment, but may provide insight into how a specific material rate plate 22 s 7.167 y 0.0063 ŽHB.. Correlation Ž r 2 . be-
will do under the combined action of impact and abrasion. tween wear rate and hardness was equal to 0.76. Like the
Fig. 6 shows results for the first field test Žplate 22., DSRW wear test results, the austenitic 304 SS ŽW316. had
conducted with the hard chert ore. The chert was generally the highest wear rate while the 17% Cr white iron had the
non-rotating as it passed over the surface of the test plate, lowest one ŽW299F.. This feature suggests that any differ-
thereby providing a scratching action against the samples ence in the field wear test conditions, such as the ore
in the PAFWT plate. This figure possesses the general moisture, ore velocity Ži.e., the ore traveled at a much
appearance of both Figs. 2 and 3, i.e., the laboratory tests slower speed y0.1 msy1 . and ore size Ži.e., relatively
corresponding to pure abrasion. Fig. 6 shows that for the large chunks between 50 and 1000 mm., was not as
nine alloys exposed to this environment, the wear rate significant a factor as test similarities, such as a non-rotat-
decreased in a linear fashion with increasing alloy hard- ing abrasive and a unidirectional sliding motion.

Fig. 9. Correlation between the normalized wear rate using the POD vs. the normalized wear rate for the PAFWT Plate 20.
1068 J.H. Tylczak et al.r Wear 225–229 (1999) 1059–1069

Fig. 7 shows the result for the second field test Žplate ŽW315.. For example, whenever material removal was due
20., conducted with a much softer mineral Žlimestone. to the sliding action of abrasive particles, the 13% Mn
sliding and tumbling over the surface of the test plate. The steel performed better than expected Žwith respect to its
wear rate is plotted as a function of alloy hardness and a pre-test hardness.. In the case of JC gouging–abrasion
curve fit to the data is shown in Fig. 7 Žminus the values wear test, the gouging action of the abrasive removed
for the austenitic steels.. The best correlation was obtained material so aggressively that the alloy’s ability to
for a power law equation of the general form y s ax b. workharden did not help much in reducing the wear rate.
Correlation between wear rate and hardness was quite Second, testing showed that for laboratory tests with
good at 0.91. The two austenitic steels ŽW315 and W316. similar material removal mechanisms, e.g., sliding abrasive
once again exhibited wear rates lower than expected given wear, the DSRW and POD tests showed similar behavior
their hardness, with the 13% Mn steel ŽW315. having a in wear rate as a function of hardness for the materials
wear rate equivalent to alloys 250 HB higher. The carbide tested in this study. Increased hardness of the alloy kept
containing alloys ŽW299F and W313. had the lowest wear the abrasive particles from penetrating too deeply in to the
rates of the materials tested Žwith the D2 tool steel ŽW313. surface of the material and slowed the rate of material loss.
having the lowest.. Aside from the behavior of the The alloys with the refractory metal carbides were very
austenitic steels in sliding abrasion, it is apparent that good at protecting the matrix from abrasive particle pene-
hardness played a critical factor in the alloys resisting the tration.
wear of the limestone, with the harder alloys performing The field wear tests, while in different environments as
very well. In a general sense, this is what both the POD far as size and hardness of the abrasive ore is concerned,
and DSRW laboratory wear tests showed. showed results similar to that of the POD and DSRW. The
In order to see if a reasonable correlation exists between JC gouging–abrasion wear test gave much different re-
the POD and DSRW laboratory abrasion wear tests and the sults, primarily because this test produced a different
results from the PAFWT for Plate 20, the measured wear mechanism of material removal–gouging–abrasion. Much
rates were normalized to the hardened AISI 4340 steel the same can be said about the results from the impeller-
ŽW311., i.e., the measured wear rate of the test specimen in-drum wear test. This means that laboratory tests can
was divided by the measured wear rate of the AISI 4340 provide a good measure of the relative wear behavior of a
steel, and the results plotted against each other in Figs. 8 material, if the laboratory and field wear tests exhibit
and 9. For each set of data points, i.e., laboratory wear test similar wear mechanisms. For other environments, if you
corresponding to the y-variable vs. PAFWT corresponding want to predict wear behavior, a laboratory test that simu-
to the x-variable, a curve of the following form fit the data lates the environment must be used, e.g., in applications
well: y s a q be y x . Fig. 8 shows the relationship between where gouging–abrasion is a significant material removal
the normalized wear rate for the DSRW and the normal- mechanism, a jaw crusher-type laboratory test should be
ized wear rate for Plate 20 Žlimestone ore in unidirectional used instead of a POD or DSRW test. The same rational
sliding abrasion.. The 304 SS ŽW316. data point was hold for impact–abrasion environments.
omitted when determining the curve of best fit for this data The use of hardness as a method of predicting the wear
set. This curve describes the relationship between the two behavior of materials must be used with caution. Nor-
tests quite well Ž r 2 s 0.82.. Fig. 9 shows the relationship mally, hardness is a good first indication of a material’s
between the normalized wear rate for the POD and the resistance to abrasion. However, a cautionary note is in
normalized wear rate for Plate 20. The same equation also order, because some materials perform exceptionally well
fits the data well Ž r 2 s 0.75.. However, from examination in pure abrasion, but not so well when the wear mode
of the data in these figures, for materials of roughly the changes to impact or gouging. For these wear modes, some
same relative hardness, field wear tests will be needed in other property Že.g., toughness or the work-hardening abil-
order to discriminate differences in wear behavior. ity of the material. controls the wear rate. For example,
many high-Cr white cast irons Žwhere W299F is an exam-
ple. suffer damage to the carbide phase as a result of
5. Summary and conclusions fracture of the carbide through impact or high-stress point
loads Žgouging.. The fractured carbides are then more
The results of these tests show several things about susceptible to removal by subsequent abrasion. Another
laboratory tests and how they compare with field wear example is the 13% Mn steel ŽW315. which readily
tests. The first feature is the general relationship of the workhardens from its interaction with the wear environ-
abrasive wear rate decreasing with hardness in the labora- ment and so increases its near surface hardness and wear
tory for both the POD and the DSRW. The same sort of resistance as the extent of plastic deformation increases.
behavior was observed from the PAFWT. As with most So, while the laboratory wear tests can help eliminate
generalities there are exceptions, both for materials and for poorly performing materials, they will rarely be able to
tests. One of the material exceptions was the performance unambiguously identify the optimum material to use for a
of the austenitic steels—304 SS ŽW316. and 13% Mn steel particular wear application.
J.H. Tylczak et al.r Wear 225–229 (1999) 1059–1069 1069

References w6x ASTM G65-94, Standard test method for measuring abrasion using
the dry sandrrubber wheel apparatus, Annual Book of ASTM
w1x R. Blickensderfer, J.H. Tylczak, Evaluation of the planar array field Standards, Vol. 03.02, Amer. Soc. Testing Mater., Philadelphia, PA,
wear test, in: K.C. Ludema, R.G. Bayer ŽEds.., Wear of Materials 1996, pp. 232–243.
1991, ASME, New York, NY, 1991, pp. 339–344. w7x R. Blickensderfer, G. Laird II, A pin-on-drum abrasive wear test and
w2x R. Blickensderfer, J.H. Tylczak, B.W. Madsen, J.H. Hansen, Evalu- comparison with other pin tests, J. Test. Eval. 16 Ž1988. 56–526.
ating and improving materials for mineral processing, Tribology in w8x R. Blickensderfer, J.H. Tylczak, B.W. Madsen, Laboratory wear
Mineral Extraction: War on Wear, IMechE, London, UK, 1984, pp. testing capabilities of the Bureau of Mines, IC 9001, 1985, 36 pp.
15–21. w9x R.D. Wilson, J.A. Hawk, Impeller wear impact-abrasive wear test,
w3x R. Blickensderfer, J.H. Tylczak, Evaluation of commercial US this conference.
grinding balls by laboratory impact and abrasion tests, Minerals and w10x R. Blickensderfer, Design criteria and correction factors for field
Metallurgical Processing, May 1989, pp. 60–66. wear testing, in: K.C. Ludema ŽEd.., Wear of Materials 1987, Vol 2,
w4x I.R. Sare, A.G. Constantine, Development of methodologies for the ASME, New York, NY, 1987, pp. 877–884.
evaluation of wear-resistant materials for the mineral industry, Wear w11x ASTM G81-83 Žreapproved 1989., Standard practice for jaw crusher
203–204 Ž1997. 671–678. gouging abrasion test, Annual Book of ASTM Standards, Vol.
w5x J.D. Gates, Two-body and three-body abrasion: a critical discussion, 03.02, Amer. Soc. Testing Mater., Philadelphia, PA, 1996, pp.
Wear 214 Ž1998. 139–146. 325–330.

You might also like