WEAR

Wear186187(1995) 179-188

Sand erosion of wear-resistant materials: Erosion in choke valves

K. Haugen, 0. Kvernvold, A. Ronold, R. Sandberg
Det Nor& Veritas Industry A/S, Hovik, N-1322, Norway

Abstract

Sand particles from oil and gas production may cause considerable erosion damage in critical parts of transport and processing equipment
such as valves and chokes. Increased longevity of such components will lead to significant savings as offshore oil and gas production move
subsea, and it can be achieved in two ways; through selection of erosion-resistant materials or through design optimization.
A total of 28 different materials including standard steel grades, solid tungsten carbide materials, coatings and ceramics have been tested
under various impact angles and velocities of sand particles, and examined for erosion characteristics. The test results have been used to
determine empirical model equations and to determine the relative erosion resistance for the materials. The results show that a longevity gain
factor of up to more 10’ is possible to achieve by selection of proper material. Optimization of choke design for a needle choke has shown
that additional savings with a factor of more than 10’is obtainable.

Keywords: Sand; Erosion; Choke valves; Wear-resistant materials; Oil and gas production

1. Introduction

The production of sand from oil and gas wells may cause
extensive wear in hydrocarbon processing and transport
equipment. The components most exposed to erosion in this
context are choke valves, which often operate at sonic flow
velocities. In general, ‘normal’ production conditions from
sand stone reservoirs may lead to a choke replacement fre-
quency in the range from 3-4 months up to 18-24 months.
In extreme cases in the North Sea critical components have
been completely destroyed by erosion within a few hours.
Figs. 1 and 2 show some typical examples of erosion damages

Fig. 2. Erosion of needle in needle and seat choke.

observed on chokes used in the North Sea. Fig. 1 shows

erosion in the ports in the cage in a plug in cage choke having
been in operation 3-4 months at normal operation conditions,
while Fig. 2 shows erosion attacks on the needle for a needle
and seat choke having been in operation for only some few
hours at test production with a very high sand feed.
Choke replacements are associated with high costs and are
also undesirable from a safety point of view. Considering
future development of subsea fields, with consequent higher
replacement costs, an increase in the service life of a choke
valve would lead to significant cost reductions as well as
Fig. 1. Erosion in ports of a plug in cage choke. increased safety.

0043-1648/95/$09.50 0 1995 Elsevier Science S.A. All rights reserved

SSDIOO43-1648(95)07158-X
180 K. Haugen et al. /Wear 186-187 (1995) 179-188

I 1 I I I I flow, secondly modelling of the sand particle behaviour

within the flow, and thirdly specific modelling of the erosion
f\ (a) attack by particle impingement on choke internals.
- I ‘@tile Computational fluid dynamics (CFD) techniques are
applied to solve the Navier-Stokes equations governing the
I ‘1
fluid flow. The standard k-e model of turbulence is used to
- I \
close the equations, and local properties of the flow are thus
obtained throughout the choke. The sand particles are sub-
sequently treated in a Lagrangian manner, that is, the local
velocities and trajectory of each individual particle are pre-
dicted based on a force balance, taking into account the effects
of drag, gravity and added mass. In this way information
about sand particle speed and impact angle is obtained for
each particle impingement.
The general correlation of erosion rates has been estab-
0 30 60 90
lished empirically [ l-31 as
Impact angle
E=M,KF(cz)V; (1)
Fig. 3. Typical erosion behaviour of ductile and brittle materials as a function
of particle impact angle. with E being the weight loss of the target material, MI, the
mass of sand hitting the target material, VPthe particle impact
Subsea development of high pressure and high temperature
velocity, (Y the particle impact angle, K and n constants
reservoirs require typically choking from 700-800 bar pres-
assumed to depend on the physical characteristics of the mate-
sure to 300-400 bar, leading to sonic flow velocities in ‘vena
rials involved, and F(a) a functional relationship describing
contracta’ of the order 300-500 m s- ‘. Field experiences
the dependence of erosion on the particle impact angle. Two
from the North Sea indicate that sand production in the order
commonly observed forms of this latter empirically deter-
I-30 ppmW can be expected. Although erosion damage is
mined function are shown in Fig. 3 for ductile and brittle
considered unavoidable under such conditions, the degrada-
materials, respectively. Values of n range typically from 2 to
tion can be significantly reduced through proper selection of
3, and K then defines the erosion resistance of the material,
erosion-resistant materials and through choke design optim-
provided that F( (Y) is defined with a maximum value equal
ization. The present paper addresses both these topics, with
to 1. By adding up the erosion contribution from each parti-
particular emphasis on material testing and selection.
cle-wall collision, the total erosion rate can be computed at
all locations within the choke, given the initial flow condi-
tions.
2. Erosion modelling

The erosion modelling is performed in order to support 3. Material testing

laboratory experiments during the design optimization pro-
cedure. Modelling of the erosion process in the chokes con- The purpose of the experimental investigation were two-
sists of three steps. Firstly modelling of the hydrocarbon fluid fold: Firstly to establish a ranking of the erosion resistance
for various prospective materials, and secondly to establish
data which enable the development of correlations for esti-
mation of erosion rates for materials exposed to sand particle
impact.

3.1. Experimental procedure

A test rig was built in order to test the erosion behaviour

of different materials at different impact conditions. The rig
layout is shown in Fig. 4. Angular sand particles of diameter
200-250 p+m-typical of the North Sea fields-were injected
into an air stream at the inlet of a 2 m long pipe of ID 6 mm.
The sand particles, pictured in Fig. 5, were accelerated along
the pipe and directed towards a specimen holder positioned
20 mm from the pipe exit. The sand particles were subse-
quently collected in a settling tank. No sand was used for
Fig. 4. Experimental layout. more than one test due to the possibility of fracturing of the
 K. Haugen et al. /Wear 186-187 (1995) 179-188 181

on the recorded air flow. The uncertainty related to the com-

puted sand particle velocities is estimated to be of the order
15-20%. The impact angle and velocity were varied in order
to establish K, n and F( cu) of Eq. ( 1) for each material.
A total of 28 different materials were examined, covering
6 standard steel materials, 10 surface coatings, and 12 ceram-
ics and tungsten carbide materials. The list comprises the
most relevant materials presently in use in offshore valves/
chokes as well as some new materials for future considera-
tion. C-steel was used as base material for all coatings. The
coatings comprised chemical nickel, electrochemical nickel
of different hardness and WC coatings basically manufac-
tured in one of two ways-hot spraying or detonation gun
(Degun) deposition. The solid tungsten carbide materials
consisted of different binder materials; i.e. cobalt and nickel
binder, and grain size of the tungsten particles. A summary
Fig. 5. Sand particles used in the test
of the material properties are shown in Table 1.
All materials were tested at impact angles of 22.5” and 90”,
sand grains. Pressure and air flow rates were monitored in and some also at 7.5”, with test velocities (air) being kept at
order to ensure identical test conditions for the different mate- 22,55 and 320 m s- ‘. The corresponding sand impact veloc-
rials. Testing at each test condition was repeated until a stable ities were estimated to 18-20, 40-45 and 2Ofl-220 m s-‘,
erosion rate was achieved. The results were recorded by respectively. C-steel was selected as a reference material and
weight loss measurements using a scale with an accuracy of was, as such, subject to more extensive examination than the
0.2-0.3 mg. Scanning electron microscope (SEM) photos other materials, with additional tests also performed at impact
were taken of the different materials, to examine the erosion angles of 15”, 30” and 45”.
mechanisms. Impact angles 22.5” and 90” were selected since maximum
The sand particle velocities were estimated using CFD erosion is achieved at those angles for ductile and brittle
modelling and a Lagrangian particle tracking model, based materials, respectively.

Table 1
Density and hardness and fracture toughness for materials tested

Material Density (kg rne3) Hardness (HV) Fracture toughness (MN m- 3’2)

C-steel 7800 240 (70)

3 16L-steel 8000 300
Duplex stainless steel 7850 350
SMO 8000 330
Stellite 6 8400 400
Boronized Stellite 6 8400 1500
Electrochemical nickel, Ni-250 8850 250
Electrochemical nickel, Ni-500 8850 500
Chemical nickel 8850 1400
Hot sprayed WC-60INi 12000 8001
Hot sprayed WC-40%Ni 12000 700?
Degun WC, 0.25 mm thick 15000
Detonation gun WC, thin layer 13500 1200
Cobalt-based coating 7600 600
WC, DC-05 15250 1800
WC, cs-10 14800 1750 10
WC, CR-37 14600 1350 12
95 A1203 3700 1960 4-6
99.5 Al,O, 3700 1960 4-6
PSZ 5700 1120 12-15
zro*-Y, 6070 1330 5-10
Sic 3100 2100 5.2
Si,N, 3200 2000 6.3
TiB2 4520 2400 5
B,C 2500 2600 3.5
SiSiC 3100 2400 4
182 K. Haugen et al. /Wear 186-187 (1995) 179-188

3.2. Results Based on the relative erosion resistance, REF, defined as

Volume loss material

REF=
The reference C-steel material was found to behave in a Volume loss C-steel
ductile manner with maximum erosion attack at impact angles
of 15-30”, dependent on the impact velocity. The erosion rate the standard steel grades show nearly identical erosion resis-
shows, however, low dependence of impact angles in the tance, as seen in Fig. 6. For an impact angle of 22.5” the
range 15-45”. A weight loss of X&60% of the maximum variation was found to be less than lo%, whereas a variation
value is recorded at 90” impact. A strong dependence of the up to 60% was recorded at 90” impact. Similar results have
impact velocity is measured; i.e. an velocity exponent of 2.6. also been reported previously [ 71.
This is in accordance with results reported in the literature, Boronized Stellite, which has a hardness of 1500 HV com-
[ 1,2,4-61.
pared with 400 HV for standard Stellite, results in an
improvement of the erosion resistance by a factor of approx-
The results for the weight losses per kilogram of sand
imately 10 compared with steel at low impact velocities. The
impact for the various tests are shown in Table 2. All tests
hard cover layer was, however, observed to crack and fall off
have been repeated until a linear relation between weight loss
at high impact velocities, in particular at 90” impact. The
and sand feed has been obtained. Of the other standard mate-
resulting catastrophic erosion is shown in Fig. 7.
rials tested, 3 16L steel, duplex stainless steel, and SMO seem
The two electrochemical nickel coatings of hardness 250
to behave in a ductile manner, as do the two electrochemical
and 500 HV, showed erosion resistance similar to steel, as
nickel coatings and one chemical nickel coating, whereas
seen in Fig. 8. Chemical nickel which has a hardness of
Stellite and boronized Stellite show brittle behaviour. How-
1500 HV was, however, found to flake off from the base
ever, the variation of impact angles was not sufficiently exten-
sive to determine the exact ductility, and some of the results
may be interpreted as brittle behaviour at high impact veloc-
ities, such as for the one electrochemical nickel coating with 16

the highest hardness and also for SMO. 1.4

I 2

I
k
Table 2 = 0.6

Weight loss (mg per kg sand) as function of particle impact velocity and
06
impact angle
0.4

“P 45-50 ms-’ 200-220 m s-’ 02

0
90” 22.5” 90” 22.5” C-St.%3 316-steel Duplex SMO Stellite Boronlzed
Stelllte
C-steel 14 23 1085 1700
??45m/s. 90 deg ? ?45 m/s. 22.5 deg
3 16-s& 16 20 1770 1845
0 200m/s.90deg 0 2001-n/s 22 5 deg
Duplex 13 23 1400 1825
SMO 12 20 2170 1690 Fig. 6. Erosion rates of steel qualities relative to C-steel.
Stellite 17 13 1870 1030
Boronized Stellite 1.7 1.7 860 265
Electrochemical nickel Ni-250 15 27 1070 1930
Electrochemical nickel Ni-500 17 22 1460 1460
Hot sprayed WC--60%Ni 27 13 4670 1145
Hot sprayed WC_40%Ni 32 17 4070 1470
Detonation gun WC, thin layer 108 33 13000 2700
Cobalt based coating 70 50 7100 4200
Degun WC, 0.25 16 4 860 265
WC, DC 05 0.7 0.58 22 11
WC, cs 10 1.6 1.0 45 17
WC, CR 37 1.2 0.83 50 17
95% A&O, >130 >20 > 2500 >700
99.5% Al,O1 100 25 620 360
PSZ 48 6 1880 360
zro-Y, 1.2 0.8 68 30
SIC 8.9 1.5 150 22
S&N, 0.37 0.17 7 1.1
TiB, 15 2.4 270 48
BX 0.97 0.37 3.7 2.0
SiSiC 2.5 0.5 150 27 Fig. 7. Boronized stellite eroded at impact velocity 200 m s- ’and impact
angle 90”.
 K. Haugen et al. /Wear 186187 (1995) 179-188 183

lowest content of binder material and the smallest carbide

particles. Similar behaviour has also been reported in [7].
The results show that an improvement in the erosion resis-
tance of the order 50-300 compared with steel can be
achieved for the WC grades tested.
Testing of the ceramics revealed large variations in the
erosion resistance. All ceramic materials showed better ero-
sion resistance than C-steel at low impact angles. This effect
was particularly pronounced for S&N_, and B& which
showed an improvement in erosion resistance of the order of
0.01 ,
more than 100 as compared with C-steel at the highest impact
C-steel Ni-250 Ni-500 WC- WC- C&gun Degun Degun velocities and lowest impact angles. These two ceramics were
60%Ni 40XNi WC-CO co- Union
66-12% based Carbide also performing somewhat better than the solid tungsten car-
bide materials. At high impact angles, however, several of
m 45 m/s. 90 deg ??
45m/s. 22.5 deq
the ceramics revealed less good erosion resistance than steel.
0 200117/s. 90 deg ??
2OOm/s. 22.5 deg
This was most pronounced for A1203 and PSZ. It is, however,
Fig. 8. Erosion of surface coatings relative to C-steel. to be noticed that none of the ceramics shows significant
better erosion resistance than the WC materials. The relative
material during the test, with consequent erosion as illustrated erosion resistance for the solid tungsten carbide materials and
in Fig. 9 the ceramics is shown in Fig. 11.
The hot sprayed coatings with a nickel binder content var- SEM examination of the materials were carried out in order
ying from 40% to 60% showed almost identical erosion resis- to determine details of the erosion mechanisms. Typical
tance: better performance at low impact angles and low examples of SEM photos taken before and after the erosion
impact velocity, and less good at high impact angles as com- tests are shown in Figs. 12-15 for a thin layer hot sprayed
pared with C-steel. Two of the detonation gun coatings- tungsten carbide coating and solid tungsten carbide, respec-
with a coating layer of 0.05-0.1 mm-showed poor resis- tively. The eroded specimen of the coating (Fig. 13) is seen
tance compared with C-steel, most probably because the car- to contain less tungsten carbide particles, especially of the
bide particles was falling out of the substrate during the larger sizes, than the non-eroded sample (Fig. 12). This
erosion process. However, the third Degun coating-with a shows that the tungsten carbide particles are torn out of the
coating layer of approximately 0.25 mm-demonstrated the matrix during the erosion process. The non-uniformity of
very best performance of all the coatings tested, and was the both size and distribution of the WC particles for the hot
only one with a significant improvement in erosion resistance sprayed coating should be noticed. For the solid material, all
as compared with the reference material. It is not known signs of the manufacturing process, as evident on the non-
whether the large difference in the behaviour of the Degun eroded material (Fig. 14)) are removed during the erosion
coatings is due to the variation in coating thickness or other process, and a surface of uniform character is revealed (Fig.
aspects of the manufacturing process. The relative erosion 15).
resistance for the coating materials is presented in Fig. 8. Some examples of SEM photos for ceramics are shown in
The remaining 12 materials-9 ceramics and 3 solid tung- Figs. 16-20. For A&O,, which showed poor erosion behav-
sten carbide materials-all behave in a brittle manner. The
ceramics comprise 95% and 99.5% A1203, magnesium partly
stabilized zirconia (PSZ) , yttria stabilized zirconia (ZrO,
Y3), silicon nitride (S&N,), silicon carbide (Sic), titanium
boride ( TiB2), boron carbide (B,C) , and silicon-infiltrated
silicon carbide (SiSiC). Examples on the observed brittle
behaviour of some ceramics as compared with the ductile C-
steel is shown in Fig. 10.
In general, the ceramics and solid tungsten carbide mate-
rials showed large differences in erosion resistance, with few
specimens having less good erosion resistance than the ref-
erence C-steel.
The three solid tungsten carbide materials, DC-05
(95 wt.% WC, nickel + cobalt binder and extra fine grain
size), CS-10 (89.5 wt.% WC, cobalt binder and fine grain
size) and CR-37 (65 vol.% WC, nickel binder and medium
grain size), all showed good erosion resistance, with the best Fig. 9. Chemical nickel eroded at impact velocity 200 m s-’ and impact
performance being recorded for the one specimen with the angle 90”.
184 K. Haugen et al. / Wear 186-187 (I 995) 179-188

700 P
/’

- 95% A1203
600
----A---.
99.5% A1203
6
$ 500
u ----*--.p02_y3
ii
2 400
s3
5
2 300 _ d.- 0
P :-
_.- ,.>
___-
i! ----_m_ ..- *_e.c
1 200 .e.- ----___
$ .>,.__.,p...._.-
I.;‘.. ,__ ..-. .-7-...=----; --__--_f

100

0 -
0 30 60 so

Impact angle
Fig, 10. Erosion rate ( mm3 per kg sand impact) as function of impact angle for steel and some selected brittle materials at 200 m s- ’impact velocity.

m 45mfs. 90 deg ? ?45tn/s. 22.5 deg 2oom/s. 90 deg 0 ZOOm/s.22.5 deg 1

Fig. 11. Relative erosion resistance of solid WC and ceramics relative to C-steel

iour, SEM investigation revealed that the binding between during the manufacturing process, have been removed and
the A&O3 particles was broken and that the particles had been the surface has become smoother. However, both erosion pits
crushed and partly torn out of the matrix during the erosion and cracks are observed. The cracks have probably been
process (Figs. 16-18). For the B,C material porosities in the initiated at weak points at the grain boundaries.
surface can be observed, probably from carbide grains being In general, the surface structure of the materials may
torn out during the manufacturing process (Fig. 19). At the change significantly during the initial stage of the erosion
eroded surface (Fig. 20), the grains which had loosened process. This effect give rise to a transient behaviour in the
 K. Haugen et al. / Wear 186-l 87 (1995) 179-188 185

3.3. Model coejjkients

The results from the tests on C-steel compare well with

other data [ 1,2,6,8,9], albeit the present results are in the

Fig. 12. Non-eroded surface of hot sprayed tungsten carbide coating with
atomic number contrast. WC particles-bright and binder dark.

Fig. 15. Eroded surface of solid tungsten carbide.

Fig. 13. Eroded surface of hot sprayed tungsten carbide coating.

Fig. 16. Non-eroded surface, 99.5% Al,O,.

Fig. 14. Non-eroded surface of solid tungsten carbide.

erosion characteristics of each material, and is the reason why

each test had to be repeated until stationary erosion conditions
was obtained. Fig. 17. Eroded surface, 99.5% A1,O1
186 K. Haugen et al. /Wear 186-187 (1995) 179-188

F(u)=&-l)“+“Ai fg
i=l
(1 (2)

where the Ais are given in Table 3.

Based on both present and previous experiments, values
of K and n in Eq. ( 1) are proposed as given in Table 4 in
order to obtain conservative estimates of the erosion rate for
the materials tested.
Most other materials have been tested at two different
impact angles only, and F(a) is thus not possible to describe
satisfactory. However, typical behaviour of F( a) curves are
shown in Fig. 8. For the brittle materials, the dependence on
the impact angle shows almost linear behaviour. Using this
as a rough assumption, the model coefficients K and n have
been obtained accordingly, as shown in Table 4. The large
Fig 18.Eroded surface, 99.5% A1203.
variation in the exponent n should be noticed.

4. Design optimization

As a part of the project, investigations with respect to

optimization of choke design were also performed. Extensive
laboratory tests and numerical simulations have been per-
formed. However, only an extract of the main results will be
given.
Different steel choke designs were examined at laboratory
conditions of moderate pressure ( AP = 10 bar). Two needle
chokes-one traditional design and one modified-were
tested at identical flow conditions. The resulting erosion rates
were found to differ by a factor of more than 10’ (Fig. 21) .
Numerical modelling of the particle flow and subsequent
erosion of the two different chokes, using the method
described in Section 2 and model coefficients of Table 2, has
Fig. 19. Non-eroded surface, B4C.
reproduced experimental data with good accuracy both with
respect to location of erosion attack and order of magnitude
of erosion rate. This example firstly shows that large improve-
ments can be achieved through design optimization, and sec-
ondly that the erosion modelling is of sufficient accuracy to
assist costly high-pressure laboratory testing and thus repre-
sent a valuable tool in the design optimization process.

5. Conclusions

A total of 28 different material types have been examined

for erosion characteristics. These comprise 6 standard mate-
rials, 10 surface coatings, 3 solid tungsten carbide materials
and 9 ceramics. By varying the impact velocity and angle of
the eroding sand particles, the erosion behaviour as functions
of these variables have been investigated for each material,
Fig. 20. Eroded surface, B4C. and empirical coefficients have been determined for erosion
modelling purposes. The impact angle variation, however,
lower range of what others have reported. Based on the C- does not comprise a detailed screening, and for most materials
steel results from this study, the function F( CX)of Eq. ( 1) the resulting model coefficient should be used with caution.
has been determined using the relation One material, C-steel, was used as a reference material and,
 K. Haugen et al. / Wear 186-l 87 (1995) 179-188 187

Table 3
Constants to be used in Eq. (2)

AI A2 AS A, As A, A, As
9.370 42.295 110.864 175.804 170.137 98.298 31.211 4.170

Table 4 carbide particles. These observations were most often found

Coefficients recommended for use in model E@.( 1)
at materials of high binder content and/or large carbide par-
KX lo9 n
ticles. For the ceramics and tungsten carbide qualities with
Material
high erosion resistance, the binding material between the
Steel 2.0 2.6 grains are not broken to the same extent as for the less resistant
Hot sprayed WC-60%Ni 5.2X 1O-2 3.4 materials.
Hot sprayed WC-40%Ni 1.2x lo-* 3.2
The model coefficient achieved for the steel material were
Degun WC, 0.25 mm 6.1X10-’ 2.7
5.3x 10-l 3.2
used to carry out erosion modelling of two choke designs.
Detonation gun WC, thin layer
Cobalt based coating 5.3x 10-l 3.1 Numerical simulations of the flow by using the k--E model of
WC, DC-05 1.1 x 10-l 2.3 turbulence to solve the governing Navier-Stokes combined
WC, cs-10 3.2X 10-l 2.2 with Lagrangian tracking of the sand particles were carried
WC, CR-37 8.8X lo-’ 2.5
out, and subsequent erosion attack was modeled using the
95 A&O, 6.8 x 10’ 2.0
99.5 A&O3 9.5 x lo* 1.2
coefficient obtained from the testing procedure. Comparison
PSZ 4.1 2.5 of model results with laboratory experiments indicate that the
zrO,-Y 3 4.0x 1o-2 2.7 present modelling technique can predict the erosion rate and
SIC 6.5 1.9 location with sufficient accuracy to be applied in extrapola-
S&N4 2.0x 10-l 2.0 tion of test results and for design purposes.
TiB, 9.3 1.9
B,C 3.0x 10’ 0.9
SiSiC 7.2X lo-’ 2.7

Acknowledgements

This work was carried out as part of the Norwegian

Research Programme within Multiphase Flow, PROFF, dur-
ing 1991-92. The programme has been sponsored by the
Norwegian Research Council, Norsk Hydro a.s., Saga Petro-
leum a.s., Statoil, Elf Petroleum Norge a.s., Esso Norge a.s.,
AS. Norske Shell, Amoco Norway Oil Company, Total
Norge a.s., Aker Engineering, Framo Engineering, Kvaerner
Engineering and NAT.
The materials tested were supplied by Kvaerner Eureka
(N) , Bohler Stainless Foundry (D) , Diffusion Alloy Ltd.
(UK), Norsk Everit, AHC Oberllachentechnik (D) , Castolin
Norge/Eutectic Switzerland, Weartec AB (S), Union Car-
bide (UK), Sandvik Hard Materials (S) , ICI Advanced
Fig. 21. Two identical choke needles, from two differently designed hous- Ceramics (UK) and Sintec Inovation in Keramik (D)
ings, and subject to identical sand flow conditions. The erosion rates differ
by a factor of 103.

as such, subject to particularly detailed analyses, comprising

References
a wide range of impact angles and velocities.
Of the 28 materials, the most erosion resistant were found
[ 11 G.P. Tilly, Erosion by impact of solid particles, Treatise on Material
to be the three solid tungsten carbide materials and two of the
Science and Technology, Academic Press, New York, 1979
ceramics, S&N4 and B&. Only one coating, a Degun tungsten [2] E. Raask, Tube erosion by ash impaction, Wear, 13 (1969)301-315.
carbide layer, was found to give significantly improved ero- [31 W. F. Adler, Assessment of the state of knowledge pertaining to solid
sion characteristics as compared with the reference material, particle erosion, Final Rep. CR79-680, to the U. S. Army Research
C-steel. Office for contract DAAG29-77-C-0039.1979.
SEM examination of the materials prior to and after the [4] G.L. Sheldon, Similarities and differences in erosion behaviour of
materials, Trans. ASME, J. Basic Eng.. 92 (1970) 619630.
tests show that high erosion rates generally are associated [51 A. Magnee, A. Piccinin and D. Coutsouradis, Behaviour in impact
with fracturing of the carbide particles, combined with ero- erosion of ductile and brittle materials, Metall. Rep. (CRM), 37 ( 1973)
sion of the binder material, resulting in fall-out of individual 43-50.
188 K. Haugen et al. /Wear 186-187 (1995) 179-188

[6] L.K. Ives, Erosion of 310 stainless steel at 975 “C in combustion gas [ 81 M.E. Gulden, Influence of brittle and ductile transition on solid particle
atmosphere, Trans. ASME. J. Eng. Mater. Tech&, (1977) 126 behaviour, 5th Int. Conf: on Erosion by Liquid and Solid Impact,
132. Cambridge, 1979.
[ 71 J.S. Hansen, Relative Erosion Resistance of Several Materials, ASTM- [9] C.D. Wood, Erosion of metals by the speed impact of dust particles,
STP 664, ASTM, Philadelphia, PA, 1979. Proc. Annual Tech. Meet. Inst. EnvironmentalScience, 1966.