You are on page 1of 7

Tribology International 79 (2014) 17

Contents lists available at ScienceDirect

Tribology International
journal homepage: www.elsevier.com/locate/triboint

Slurry erosion characteristics and erosion mechanisms


of stainless steel
Q.B. Nguyen a,n, C.Y.H. Lim a,n, V.B. Nguyen b, Y.M. Wan b, B. Nai a, Y.W. Zhang b, M. Gupta a
a
b

Department of Mechanical Engineering, National University of Singapore, 9 Engineering Drive 1, Singapore 117576, Singapore
Institute of High Performance Computing, 1 Fusionopolis Way, Singapore 138632, Singapore

art ic l e i nf o

a b s t r a c t

Article history:
Received 5 March 2014
Received in revised form
21 April 2014
Accepted 16 May 2014
Available online 24 May 2014

In the present study, slurry erosion experiments of stainless steel SUS-304 were carried out using new
advanced erosion test rig, where the multiphase ows of alumina sand and water were utilized as the
erodents. The results show that erosion rate was initially high and gradually reduced over the testing
time. In addition, the erosion rate increased with an increase in impact velocity. The surface roughness
increased with either increasing testing time or impact velocity. Further, the surface proles of W
shape were observed for all eroded samples. Microstructural characterization reveals two different
erosion regimes: plastic deformation mechanism dominated at high impact angles, while plowing/
cutting mechanism dominated at low impact angles. A correlation between the erosion rate, erosion
prole and microstructure is also discussed.
& 2014 Elsevier Ltd. All rights reserved.

Keywords:
Microstructure
Atomic force microscopy
microscopy
Fracture
Surface properties

1. Introduction
Slurry erosion study is vitally important in many engineering
applications associated with the transportations of abrasive particles, particularly in oil and gas production systems, hydraulic
machines, pumps, hydropower, mine industry, etc. [14]. Failures
due to erosion can lead to detrimental economic consequences as
well as safety concerns. The demand for longer service span, which
eventually reduces the cost of maintenance, has motivated many
research laboratories and companies to study slurry erosion for
particular cases for specic applications [46]. Several standard
laboratory tests, such as ASTM G65/75, pipe loop, jet impingement,
toroid wheel, slurry pot or coriolis, have been built to understand
the actual erosion situations of different materials under various
environmental conditions [5,710]. Each test method has its own
advantages and disadvantages. Some test rigs used centrifugal
pumps, which are cost saving. However, the pump propellers are
very prone to abrasive particles and damaged after a short time of
use. This leads to a huge variation in ow velocity as well as the
actual amount of slurry carried in a unit of time, and thus reduces
the reliability of experimental results [5,1114]. In this paper, we
design a new slurry erosion test rig using a peristaltic pump,
which eliminates the disadvantages of centrifugal pumps as
abrasive particles transported through a rubber hose. The

Corresponding authors. Tel.: 65 6874 8082; fax: 65 6779 1459.


E-mail addresses: nguyenquybau@gmail.com (Q.B. Nguyen),
mpelimc@nus.edu.sg (C.Y.H. Lim).
http://dx.doi.org/10.1016/j.triboint.2014.05.014
0301-679X/& 2014 Elsevier Ltd. All rights reserved.

peristaltic pump gives accurate and consistent ow rates through


a digital speed regulator, as well as an adjustable pressure of
pulsation dampener.
Erosion of materials varies from material to material, testing
conditions, types of abrasives and environmental conditions. The
literature search shows that different types of materials, such as
ductile and brittle metals, ceramic, polymer and composites, have
been tested [1522]. The reported results indicated different
erosion characteristics such as U- and W-wear scars for brittle
and ductile materials, respectively. Several types of abrasives have
been used in erosion tests to simulate different situations occurred
in our real life at certain places [3,4,23,24]. Particle size, shape,
hardness and density as well as impact angles and impact velocity
also play a crucial role in the erosion process [25]. Further, erosion
becomes more serious when the testing environmental conditions
become hostile and corrosive due to the synergic effect of erosion
corrosion [3,5,9,13].
However, the literature search also shows that a careful and
detailed study on slurry erosion of stainless steel SUS304 using the
peristaltic pump has not been published yet. Thus, this paper
studies the erosion of stainless steel SUS304 at a representative
impact angle of 901 with variations of testing time and impact
velocity in the medium of angular alumina sand and tap water
using the newly developed test apparatus equipped with peristaltic pump. The erosion characteristics and mechanisms are changed
from location to location. Different characterization tools such as
surface prolometer, scanning electron microscopy and atomic
force microscopy were used to examine the nature of erosion,
microstructure and its erosion mechanisms.

Q.B. Nguyen et al. / Tribology International 79 (2014) 17

auto-lift up mechanism so that sand particles are well mixed in the


agitating tank at a low stirring speed. The medium of sand and
water is then pumped to the pulsation dampener using a powerful
and speed adjustable 40SPX Bredel peristaltic pump. The discharged ow is smoothened by regulating the pulsation dampener
pressure before shooting to the sample through a 6.4 mm diameter
stainless steel nozzle, and the discharged medium is then circulated
to the agitating tank. The pulsation dampener pressure is set
according to the discharged pressure at the nozzle for each ow
velocity. The sample holder is designed to rotate from 01 to 901 and
is adjustable in x and z directions. The distance from nozzle to the
sample and the impact angle was pre-xed at 12.7 mm and 901,
respectively. The mixing speed is optimized at 76 rpm. Flow velocity
is adjustable using frequency through the digital indication knob on
the controller.

2. Experimental procedure
2.1. Wet erosion test rig
The new advanced slurry erosion test rig using the peristaltic
pump was developed according to ASTM-G73 as shown in Fig. 1
[20,23,24]. It is a closed, circulated and automated system. Sand
particles and water are pre-mixed in the agitating tank using a
frequency controllable motor. The tank bed and stirring blade were
designed to incline 301 to the horizontal direction to achieve an

Control board
Pressure regulators
Mixer

Peristaltic pump

Nozzle
2.2. Sample preparation

Sample
holder

Time Velocity

The 25  25  5 mm3 stainless steel SUS304 samples with an


experimental density of 7929 kg/m3 were cut from a 500  500
 5 mm3 sheet using a precision laser-cutting machine. Prior to
laser cut, the stainless steel sheet was auto-polished to a mirror
level with an average surface roughness of 25 nm (Fig. 2). All the
furs at the sides of sample were removed and the samples were
alcohol washed in an ultrasonic bath and then hot dried and kept
inside the digital dry cabinet prior to the tests. The chemical
composition of the stainless steel SUS-304 is C  0.024%,
Si  0.55%, Mn  1.8%, P  0.03%, So0.001%, Cr  18.2%, Ni  8.2%,
N  0.049% and Fe balance. Its mechanical properties are as 0.2%
proof strength  296 MPa, 1% proof strength  327 MPa, Tensile
strength  616 MPa, Elongation at 5 mm  56%, elongation at
50 mm  55% and Brinell hardness 187.

Mixer

Mixing tank
Drainage
Fig. 1. Photograph of wet erosion test rig.

2.3. Sand material


The sand material contains mainly aluminum oxide with an
average size of 90 mesh with chemical composition of Al2O3
 95%, TiO2 3%, SiO2  1.3%, F2O3  0.16%, and CaO  0.5%
(supplied by Pan-abrasive, Singapore). The sand has a density of
2400 kg/m3 and a hardness of 9 (Moh). The microstructural
morphology and size distribution of sand particle are shown in
Fig. 3 with an average experimental angularity of 0.58. A constant
0.5 vol% of sand was used in this study.
2.4. Weight loss
Prior to the test, the samples were taken out from the dry cabinet
and their weight was measured immediately to avoid the humidity
effect, using a microelectronic balance with an accuracy of 0.00001 g.
The samples after test were rst water washed, air-blew and then
alcohol washed in the ultrasonic bath, hot dried and kept inside the

Fig. 2. AFM image showing a mirror sample surface condition prior to the test.

Weight retained (%)

70
60
50
40
30
20
10
0

50

70

90

110

130

Size (mesh)
Fig. 3. Morphology and distribution of sand particles.

150

170

190

Q.B. Nguyen et al. / Tribology International 79 (2014) 17

dry cabinet for at least two hours before measuring its weight to
ensure that all sand particles are ultrasonically removed and there
was no effect of differential humidity. The weight loss and thickness
loss were then calculated accordingly.
The erosion rate is calculated based on the nature of inhomogeneous erosion as follows [17]:
E

Sample_weight_loss
g=g
Sand_weight_consumed

2.5. Surface characterization


The surface characterization was carried out to examine the
surface erosion prole, surface topography and surface roughness
as well as erosion mechanisms using different characterization
tools including mechanical surface proler, stylus surface prolometer, eld emission scanning electron microscopy and atomic
force microscopy.

3. Results and discussion


3.1. Slurry erosion
Table 1 presents the results of slurry erosion rate with respect
to testing time. The results indicated that the erosion rate was
initially high and gradually reduced with increasing testing time.
The main reason is attributed to work hardening on the ductile
sample surface subjected to repeated impact of sand particles. As
the sample is considered very ductile due to its high ductility of
55%, the longer the sample exposes to impacts, the harder the
sample surface becomes and it results in less material removal and

thus lower erosion rates. In addition, the sand particles were


rounded during the testing time; therefore, they penetrated to
sample surface instead of cutting like angular particle. Further,
particles initially ew away from the sample surface, such that the
less particle interference occurred on sample surface. When the
W shape was formed, particles rebounded back and caused
the interference behavior beyond the sample surface. These result
in a lesser erosion rate. The results are consistent with the ndings
in [11,26].
Table 2 presents the erosion rates of stainless steel sample at
different impact velocities. The erosion rate was very minimal at
low impact velocity. However, it signicantly increased with
increasing impact velocity. In other words, the erosion becomes
more serious at high impact velocity. Many efforts have been done
to formulate the relationship between the erosion rate and the
impact velocity. A general equation was developed by Wood et al.
[20] as follows:
E mp KV np F

where E is the erosion rate (g/g), mp is the mass of sand particles


impacting on an area, K is the constant depending on types of
material and sand, V is the impact velocity of sand particle (m/s), n
is the material dependent index, and F() is the material dependent function of the impact angle and is in the range from 0 to 1.
Our experimental results show a dependent equation as
follows:
E 2  10  5 V 4:0923

3
2

with an average reliability R 0.9829. The result is, in principle,


consistent and in line with the nding mentioned in [3,17] for tube
steel AISI1015, in which the material dependent index n 4. It is

Table 1
Results of erosion rate with respect to testing time.
Testing time (min)

00

05

10

15

20

25

30

Erosion rate (  10  6)
Roughness Ra (mm)

0.025

28.3 7 1.7
0.4050

24.4 7 1.5
0.5862

22.2 7 1.3
0.7677

20.2 7 1.2
0.9566

18.17 1.1
1.0475

15.9 7 0.9
1.2725

Table 2
Results of erosion rate with respect to testing velocity.
Testing velocity (m/s)

10

15

20

25

30

Erosion rate (x10  6)


Roughness Ra (m)

0.2 70.03
0.1358

2.17 0.1
0.2039

8.7 7 0.5
0.4184

10.17 0.6
0.6668

15.9 7 0.9
1.2725

3 4 5 6

Fig. 4. A representative eroded surface of stainless steel sample: a) front view and b) cross section view.

Q.B. Nguyen et al. / Tribology International 79 (2014) 17

interesting to note that the distance from the center of wear scar
to the maximum depth was about 4.5 mm and almost independent with either testing time or impact velocity. It is believed that
the real impact velocity of sand particles reached maximum at this

also noted that the material dependent index is variable and


dependent on the characteristic of target material, sand particles
and testing conditions.
Tables 1 and 2 also present the results of surface roughness of
samples at different testing times and impact velocities. It is seen
that the sample surface becomes rougher with increasing testing
time and impact velocity. Rougher surface means that the sample
has deeper cuttings on the surface and leads to higher chance of
failure [27].

3.70e+05

2.41e+05

3.2. Erosion shape


Indeed the erosion rate does not reect the overall picture of
erosion without having its eroded surface prole as the erosion
rate is not uniform and varies from location to location.
A representative eroded surface, which is called a round wear
scar, is shown in Fig. 4 and it looks like a round bow with a
protruded island at the middle. Similar observations were
reported in [7,12,16]. The cross-section surface prole having a
W shape is shown in Fig. 5a and b for different testing times and
impact velocities. It is observed that symmetrical W shapes were
achieved for all samples. This reects homogeneous distribution of
sand particles in the ow. Minor erosion was seen at the center of
wear scar and the maximum depth of erosion increased with
increased testing time. Maximum depth of erosion was recorded at
550 mm when tested at 30 m/s for 30 min. The diameter of wear
scar was about 16 mm for all testing conditions. This nding is in
agreement with the result obtained by Zheng et al. [3]. It is

9.28e+04

Fig. 6. Simulation result of static back pressure on the sample surface.

Distance from center (mm)


-8

-7

-6

-5

-4

-3

-2

-1

-50

5min

-100

10min

Erosion depth (m)

-150
-200
-250
-300
-350
-400

15min
20min
25min

-450

30min

-500
-550

Distance from center (mm)


-8

-7

-6

-5

-4

-3

-2

-1

-50

15m/s

-150
-200
-250
-300
-350
-400
-450

Erosion depth (m)

-100

20m/s

25m/s

30m/s

-500
-550

Fig. 5. Results of surface erosion with respect to a) different testing times and b) different owing velocities.

Q.B. Nguyen et al. / Tribology International 79 (2014) 17

position due to back pressure effect. The back pressure imposed on


the sample surface reached its maximum at the center of wear
scar and gradually reduced from the center to the outer region
(Fig. 6). The differential pressure is caused by the jet ow and back
pressure leads to reduced impact velocity according to the
following equation:
v2 KP f  P b =

where v is the actual impact velocity of sand particles, K is the


discharge coefcient of ow, Pf is the pressure of ow and Pb is the
back pressure caused on the sample surface, is the average
density of the ow. Pb reaches maximum at the center of wear scar
and gradually reduces from the center to the outer region.
The back pressure does not only slow down the velocity of the
particle at the center, it also drives the particles ying away from
the center. The change in direction of particles depends on this
pressure distribution. This behavior only occurs with low Stoke
number particles; it is normally in the particle liquid ow. Therefore, the number of impact particle is reduced at the center. It also
leads to decrease in erosion rate at the center.

2.48e+01

1.32e+01

6.79e-03

Fig. 8. Simulation result of particle trajectory on the sample surface.

50 m

Fig. 7. Representative micrographs of eroded samples of W3090-30 showing dominated fracture mechanisms changing from plastic deformation to cutting from the center
wear scar to the outer region.

Q.B. Nguyen et al. / Tribology International 79 (2014) 17

Fig. 9. AFM images show a) plastic deformation mechanism of erosion at the center and b) micro-plowing/microcutting mechanism at the outer region.

According to this equation, the real velocity impacted on the


sample surface is the lowest at the center and reaches maximum
outside the jet nozzle and in this case was at about 4.5 mm
and thus leads to minor erosion at the center and maximum
erosion at 4.5 mm. On the other hand, higher jet ow leads to
increasing number of impact particles per unit time which lead to
more complicated and severe interaction of sand particles on the
sample surface at higher impact velocity and cause more serious
erosion.
3.3. Erosion mechanism
A careful study was carried out on the eroded samples to
understand erosion mechanisms. The FESEM images of eroded
samples were taken at 6 different locations (see Fig. 4) from the
center of wear scar to its outer region, and are shown in Fig. 7.
Erosion mechanisms changed from location to location due to the
sand particle sharpness, interaction trajectory (Fig. 8) and its
impact energy and thus led to varied erosion rate at different
locations on the sample surface.
Fig. 7 shows the characteristics of erosion surface at the center
and indicates evidence of plastic deformation mechanism due to
the normal indentation of sharp particles [22,25]. In this region,
particle trajectory is almost normal to the sample surface and the
real particle impact velocity was minimum due to high back
pressure (Figs. 6 and 8) and as a consequence, very less material
loss was observed.
In region 2 (2 to 3 mm away from the center) (Figs. 4, 5 and 7
(2)), most incoming particles with medium velocity (at real impact
angles of 60701) indented on the sample surface and some
interacted with the escaping particles from region 1 to change
the trajectory before hitting the sample surface (Fig. 8). The
surface material was removed after certain number of impacts
and thus the actual impact angle reduced gradually. Consequently,
plastic deformation and microplowing were dominated mechanisms in this region [7,23].
Region 3 (3.5 to 4.5 mm away from the center) is observed with
the maximum erosion and was believed to have the real highest
impact velocity and the number of impact particles (Figs. 4, 5
and 7(3)). The back pressure on the sample surface at this region
was negligible (Fig. 6) and thus the real impact particle velocity
reached its maximum and almost equaled the jet velocity. Most
incoming sand particles with highest velocity impacted on the sample
surface and re-bounced and led to severe erosion. Plastic deformation

and microplowing/microcutting mechanisms dominated in this region


[3,4].
Clearly re-bouncing effect was observed in region 4 (Fig. 7(4)).
This is about 56 mm away from the center (Figs. 4 and 5). The
escaping and re-bouncing particles from regions 2 and 3 dragged
and removed material in this region.
Fig. 7(5) shows the image at the edge of eroded circle (region
5 in Fig. 7). It obviously indicated the inter-section of two major
particle trajectories at the middle of image. One was from the
re-bouncing effect and another one was from escaping particles in
regions 1 and 2 which were mostly parallel to the sample surface
(Fig. 8). Purely microcutting/plowing effects were seen at two
sides of the image.
Fig. 7(6) shows the micrograph of the sample surface at the
outer region of round wear scar (region 6). Purely microcutting
mechanism is seen in this image. This was caused by the escaping
particles from regions 1, 2 and 3 at the initial stage of the test
when the escaping particle trajectory was almost parallel to the
sample surface [22]. Thus, material removal recorded in this region
was much lesser. Further, minor shallow micro-cracks were also
seen and this is attributed to repeated cutting actions.
The erosion mechanisms were re-examined using atomic force
microscopy. Fig. 9a is taken in region 1 and clearly indicates the
plastic deformation effect caused by particles indenting on the
sample surface at high impact angles. Fig. 9b taken at region
6 reveals purely microcutting effect due to cutting action of
particles impacting at very low angle [27].

4. Conclusions
The following conclusions are derived from this investigation.

 Slurry erosion experiments of stainless steel SUS-304 were


successfully carried out using the new advanced test rig.

 Erosion rate of stainless steel SUS304 was initially high and




reduced to a critical value over the testing time. In addition the


erosion rate increased with an increase in velocity. Further, the
sample surface roughness increased with either increasing
testing time or impact velocity.
The surface proles showed W shape for all eroded samples.
Erosion mechanisms included plastic deformation and microplowing/microcutting. Plastic deformation mechanism dominated at high impact angles which are around the center of

Q.B. Nguyen et al. / Tribology International 79 (2014) 17

wear scar while micro-plowing/cutting mechanisms dominated at lower impact angles.

Acknowledgments
The authors would like to acknowledge the National University
of Singapore and the Agency for Science, Technology and Research
(ASTAR) for providing facilities and nancial support under Grant
WBS no. R-265-000-412-305.
References
[1] Yu B, Li DY, Grondin A. Effects of the dissolved oxygen and slurry velocity on
erosioncorrosion of carbon steel in aqueous slurries with carbon dioxide and
silica sand. Wear 2013;302:160914.
[2] Castberg TS, Johnsen R, Berget J. Erosion of hardmetals: dependence of WC
grain size and distribution, and binder composition. Wear 2013;300:17.
[3] Zheng ZB, Zheng YG, Sun WH, Wang JQ. Erosioncorrosion of HVOF-sprayed
Fe-based amorphous metallic coating under impingement by a sandcontaining NaCl solution. Corros Sci 2013;76:33747.
[4] Arora HS, Grewal HS, Singh H, Mukherjee S. Zirconium based bulk metallic
glassbetter resistance to slurry erosion compared to hydroturbine steel.
Wear 2013;307:2834.
[5] Neville A, Wang C. Erosioncorrosion of engineering steelsCan it be managed
by use of chemicals? Wear 2009;267:201826.
[6] Llewellyn RJ, Yick SK, Dolman KF. Scouring erosion resistance of metallic
materials used in slurry pump service. Wear 2004;256:5929.
[7] Gnanavelu A, Kapur N, Neville A, Flores JF. An integrated methodology for
predicting material wear rates due to erosion. Wear 2009;267:193544.
[8] Rajahram SS, Harvey TJ, Walker JC, Wang SC, Wood RJK, Lalev G. A study on the
evolution of surface and subsurface wear of UNS S31603 during erosion
corrosion. Wear 2011;271:130213.
[9] Guanghong Z, Hongyan D, Yue Z, Nianlian L. Corrosionerosion wear behaviors
of 13Cr24Mn0.44N stainless steel in salinesand slurry. Tribol Int 2010;43:
8916.

[10] de Souza VA, Neville A. Corrosion and erosion damage mechanisms during
erosioncorrosion of WCCoCr cermet coatings. Wear 2003;255:14656.
[11] Ahmed SM, Al-Bukhaiti MA, Badran FMF, Emara KM. Effect of impingement
angle on slurry erosion behaviour and mechanisms of 1017 steel and highchromium white cast iron. Wear 2007;262:118798.
[12] Wheeler DW, Wood RJK. Erosion of hard surface coatings for use in offshore
gate valves. Wear 2005;258:52636.
[13] Shrestha S, Hodgkiess T, Neville A. Erosioncorrosion behaviour of highvelocity oxy-fuel NiCrMoSiB coatings under high-velocity seawater jet
impingement. Wear 2005;259:20818.
[14] Neville A, Reza F, Chiovelli S, Revega T. Erosioncorrosion behaviour of WCbased MMCs in liquidsolid slurries. Wear 2005;259:18195.
[15] Clark HMcI, Hartwich RB. A re-examination of the particle size effect in slurry
erosion. Wear 2001;248:14761.
[16] Speyer RJKWAJ, Stokes KR. Erosion of aluminium-based claddings on steel by
sand in water. Wear 2001;250:8028.
[17] Zhang VSL, Kent J, Dixon T, Novozhilov V. Analysis of boiler-tube erosion by
the technique of acoustic emission. Wear 2001;250:7629.
[18] Crnkovic SJ, Abbade Noelmar Pereira. Sandwater slurry erosion of API 5L X65
pipe steel as quenched from intercritical temperature. Tribol Int
2000;33:8116.
[19] Fang HXQ, Sidky PS, Hocking MG. Erosion of ceramic materials by a sandwater slurry jet. Wear 1999;224:18393.
[20] Wood YPRJK, Trethewey KR, Stokes K. The performance of marine coatings
and pipe materials under uid-borne sand erosion. Wear 1998;219:4659.
[21] Haugen KK, Ronold A, Sandberg R. Sand erosion of wear-resistant materialsErosion in choke valves. Wear 1995;186187:17988.
[22] Tan KS, Wood RJK, Stokes KR. The slurry erosion behaviour of high velocity
oxy-fuel (HVOF) sprayed aluminium bronze coatings. Wear 2003;255:
195205.
[23] Rajahram SS, Harvey TJ, Walker JC, Wang SC, Wood RJK. Investigation of
erosioncorrosion mechanisms of UNS S31603 using FIB and TEM. Tribol Int
2012;46:16173.
[24] Amarendra HJ, Chaudhari GP, Nath SK. Synergy of cavitation and slurry
erosion in the slurry pot tester. Wear 2012;290291:2531.
[25] Clark HM. A comparison of the erosion rate of casing steels by sandoil
suspensions. Wear 1991;151:21730.
[26] Hu X, Barker R, Neville A, Gnanavelu A. Case study on erosioncorrosion
degradation of pipework located on an offshore oil and gas facility. Wear
2011;271:1295301.
[27] Stachowiak GW, Batchelor AW. Engineering tribology. 3rd ed.Elsevier Inc.
Amsterdam; Boston: Elsevier Butterworth-Heinemann; 2005.

You might also like