Anda di halaman 1dari 7

Wear 266 (2009) 968974

Contents lists available at ScienceDirect


Wear
j our nal homepage: www. el sevi er . com/ l ocat e/ wear
Modelling the erosion rate in micro abrasive air jet machining of glasses
J.M. Fan
a
, C.Y. Wang
a,
, J. Wang
b
a
Faculty of Electromechanical Engineering, Guangdong University of Technology, Guangzhou 510006, China
b
School of Mechanical and Manufacturing Engineering, The University of New South Wales, Sydney, NSW 2052, Australia
a r t i c l e i n f o
Article history:
Received 24 May 2008
Received in revised form
22 November 2008
Accepted 18 December 2008
Available online 25 December 2008
Keywords:
Modeling
Micro abrasive air jet machining
Erosion rate
Dimensional analysis
a b s t r a c t
Micro abrasive jet machining (MAJM) is an economical and efcient technology for micro-machining of
brittle material like glasses. The erosion of brittle materials by solid micro-particles is a complex process
in which material is removed from the target surface by brittle fractures. The rate of material removal
is one of the most important quantities for a machining process. Predictive mathematical models for
the erosion rates in micro-hole drilling and micro-channel cutting on glasses with an abrasive air jet are
developed. A dimensional analysis technique is used to formulate the models as functions of the particle
impact parameters, target material properties and the major process parameters that are known to affect
the erosion process of brittle materials. The predictive capability of the models is assessed and veried by
an experimental investigation covering a range of the common process parameters such as air pressure,
abrasive mass ow rate, stand-off distance and machining time (for hole machining) or traverse speed
(for channel machining). It shows that model predictions are in good agreement with the experimental
results.
2008 Elsevier B.V. All rights reserved.
1. Introduction
The knowledge on solid particle erosion has been used for
the development of a new micro-machining technology, abra-
sive jet micro-machining or called micro abrasive jet machining
(MAJM). This is a high efcient and low cost machining technol-
ogy for brittle materials such as glasses, silicon and ceramics with
broad applications in semiconductors, electronic devices, MEMS,
at display panels, and micro uidics [14]. In particular, MAJM
has been an effective means for the fabrication of micro-channels
and holes on brittle glasses. For the production of at plasma
displays, MAJM was judged to be the best process to create thou-
sands of holes at once at low cost, high speed and high accuracy
[3,5].
In MAJM, ne particles are accelerated in an air or gas stream
and directed towards the target material. When a particle impacts
the surface of a brittle material, it causes an impact fracture to the
material and a material removal mostly by crack propagation. Such
an impact event is often compared to an indentation made with
a standard indenter. Over the last decades, several models have
been reported on this subject trying to relate the erosion rate to
the material properties [69]. Nevertheless, the practical applica-
bility of these theories is limited. At this stage of development, the
amount of material removedbya single particle impact is oftenpre-

Corresponding author. Tel.: +86 20 3932 2206; fax: +86 20 3932 2206.
E-mail address: cywang@gdut.edu.cn (C.Y. Wang).
dicted, while analytical models for the overall erosion rate in MAJM
are difcult to develop because of complex phenomena involved in
the erosionprocess, as a large number of parameters are involvedin
the process. Dimensional analysis is an effective and feasible tech-
nique of obtaining relationships betweenthe parameters that affect
erosion [10].
In the previous experimental study of MAJM for glasses by
the authors [11,12], the inuences of the technical parameters on
material erosionrate andthe shape of the eroding dimple are inves-
tigated; the inuences of the nozzle type/size and abrasive type on
the material erosion rate and structure prole are analyzed; and
the basic technical laws of these effects on machining performance
for MAJM are obtained.
The erosion of brittle materials by hard, solid micro-particles
is a complex process in which material is removed from the tar-
get surface by brittle fractures. The rate of material removal is
one of the most important quantities for a machining process.
To exploit the potentials of MAJM for applications of industrial
interest, the erosion rate of the blasting process has to become
more efcient and better predictable. In this paper, mathematical
models for the erosion rate in MAJM of micro-holes and channels
on glasses are developed using a dimensional analysis technique.
The predictive models are established as functions of the particle
projectile parameters, target material properties and some major
process parameters that are known to affect the erosion process
of brittle materials. The models are nally assessed by analyzing
the model predictions and by comparing with the experimental
results.
0043-1648/$ see front matter 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.wear.2008.12.019
J.M. Fan et al. / Wear 266 (2009) 968974 969
Nomenclature
C, C
0
, K
0
, K
1
, k
0
, k
1
, k
2
, a, b, c, a
1
, b
1
, c
1
, a
2
, b
2
, c
2
, l
1
, m
1
, n
1
, l
2
,
m
2
, n
2
, x
1
, y
1
, z
1
, x
2
, y
2
, z
2
constants
D nozzle diameter (mm)
E target material Youngs modulus (GPa)
E
rate
erosion rate (mg/g)
H Vickers hardness of target material (GPa)
K
c
target material fracture toughness (MPa m
1/2
)
m mass of removed material (g)
m
a
mass of impacting particles (g)
m
p
average mass of a particle (g)
n number of particles supplied to a jet
P air pressure (Mpa)
q
m
abrasive mass ow rate (g/s)
r average particle radius (m)
S stand-off distance (mm)
t machining time (s)
u traverse speed (mm/s)
v particle velocity (m/s)
V
p
volume removed by a single particle (mm
3
)
particle density (g/cm
3
)

t
target material density (g/cm
3
)
2. Model formulation
2.1. Erosion rate
Traditionally, the erosion rate may be expressed as the aspect
ratio of the mass of material removed to the mass of impacting
particles and is given as a dimensionless quantity by [9]
E
rate
=
m
m
a
(1)
where mis the mass of the removed material, and m
a
is the mass
of the impacting particles for achieving m.
Since the total material removal is the accumulationof the mate-
rial removed by individual particles, m may be given by
m = C
0
n V
p

t
(2)
where V
p
is the volume removed by a single particle,
t
is the tar-
get material density, C
0
is an efciency factor to allow for the fact
that not all particles are involved in the erosion process and some
particles do not have sufcient energy to cut the material, and n is
the number of particle supplied to the jet and is given by
n =
m
a
m
p
(3)
where m
a
is the mass of the impacting particles as dened earlier
and m
p
is the average mass of a particle, i.e. the mass of an average
sized particle. If assuming that the shape of particles is spherical,
the mass of a particle is given by
m
p
=
4
3
r
3
(4)
where r is the particle radius and is the particle density. Substi-
tuting Eqs. (2)(4) into Eq. (1) results in
E
rate
= C

V
p
r
3

(5)
where C is a constant generalising all the constants, i.e.
C=C
0
/(4/3).
2.2. Volume of material removed by a single particle
It is now essential to develop the volume of material removed
by a single particle V
p
. The erosion of brittle materials results from
the propagation and intersection of cracks produced by impact-
ing projectiles (see Fig. 1). It has been shown that the quasi-static
indentationtheory canbe usedto describe the impact of a solidpar-
ticle on brittle materials [13]. Based on the quasi-static indentation
theory for hard angular parities, the volume removed by a single
particle impact perpendicularly to a target surface is dependent on
the particle and impacting parameters (v, r, ) as well as the target
material properties (K
c
, H, E) [7,14]. Thus, the volume of material
removed by a single particle V
p
is a function of these parameters
and can be expressed as
V
p
= (v, r, , K
c
, H, E) (6)
where v, r and are respectively the particle velocity, radius, and
density, and K
c
, H and E are the target material fracture toughness,
hardness and Youngs modulus, respectively.
According to the dimensional analysis technique [15], all vari-
ables appearing in a problem can be assembled into a smaller
number of independent dimensionless products (or pi (
i
) groups)
using the constraint that all products formed must have the
same dimension (i.e. dimensionless). The relations connecting the
individual variables can be determined by an algebraic expres-
sion relating each dimensionless product. For the seven variables
involved in Eq. (6), three fundamental dimensions have been iden-
tied, i.e. length (L), mass (M) and time (T). Therefore, (73) =4
dimensionless groups need to be formed. In addition, three repeat-
ing variables, v, r, and H, were selected to form the four pi groups.
Accordingly, the four pi groups are given by

1
=
V
p
r
3
(7)
V
p
/r
3
represents the ratio of the volume removal during an impact
to the volume of the impacting particle;

2
=
E
H
(8)
E/Hcan be considered as the ratio of the elastic to the plastic energy
density;

3
=
K
2
c
rH
2
(9)
K
2
c
/(rH
2
) represents the ratio of the inverse of target material brit-
tleness to the size of impacting particle, where (K
c
/H)
2
is a measure
of the relative resistance of a target material to fracture during an
impact event, the inverse quantity represents the brittleness of the
target material;

4
=
v
2
H
(10)
Fig. 1. The erosion mechanism of a sharp particle.
970 J.M. Fan et al. / Wear 266 (2009) 968974
v
2
/H represents the ratio of the particle energy density, i.e. the
ratio of kinetic energy per particle volume, v
2
, to the target hard-
ness H, which can be considered as a deformation energy density
[7].
Based on the dimensional analysis technique, the functional
relation between these four dimensionless products can be given
by

1
= f (
2
,
3
,
4
) (11)
or
V
p
r
3
= f

E
H
,
K
2
c
rH
2
,
v
2
H

(12)
It is noted that a non-dimensional quantity is proportional to
the product of other dimensionless groups raised to rational power
[16]. Hence applying the power law formulation gives
V
p
r
3
= k
0

E
H

K
2
c
rH
2

v
2
H

c
(13)
where k
0
, a, b and c are coefcients and can be determined from
experiments.
2.3. Equations for erosion rate in hole and channel machining
2.3.1. Particle velocity
Before implementing Eq. (13), it is necessary to determine the
particle velocity, v, in the equation. The material erosion in MAJMis
based on the energy transmission of the potential pressure energy
containedinthe compressedair tothe kinetic energy of the blasting
particle in the jet. The particles are accelerated by the air jet, so the
particlesize, particleuxes andair velocityhaveagreat effect onthe
particle velocity. For a given particle material, the particle velocity
is proportional to the particle size and the air feed pressure [17].
Theoretical prediction of particle velocity is complicated by the
signicant inuence of collisions between particles and the nozzle
wall, so that the diameter of the nozzle has an effect on the parti-
cle velocity [18,19]. Particles often have a velocity lower than the
air velocity when just leaving the nozzle [20]. This means that the
particles are still accelerated outside the nozzle, so the distance
between the nozzle exit and target surface has an effect on the
particle velocity.
Some experimental studies have conrmedthat for brittle mate-
rials, the quantity of material removed by a single impacting
particle is primarily dependent on the normal component of the
particle velocity [13]. In MAJM, the material surface condition and
the local dynamic impact condition varies for each impact. The par-
ticle impact velocity varies during the machining process due to
the change in the local curvature of the impact point, so that the
velocity is also dependent on the machining time or traverse speed
[21,22] (see Fig. 2).
Consequently, the particle velocity can be expressed as a func-
tion of six parameters including air pressure P, stand-off distance
S, abrasive mass ow rate q
m
, nozzle diameter D, particle radius r,
andmachiningtimet (or traversespeedufor channel machining). In
addition, the hole and channel machining processes involve differ-
ent parameters and it is expected that the overall erosion action of
particles in the hole machining process is less efcient since some
particles may be trapped inside the hole preventing impacts on
the material by subsequent particles, so that the coefcient C
0
in
Eq. (2) is expected to be different for hole machining and channel
machining. Thus, these two processes are considered separately in
determining the particle velocity v in Eq. (13) as follows.
Fig. 2. Erosion condition.
2.3.2. Hole machining
For hole machining, the particle velocity can be given by
v = g
1
(P, S, D, q
m
, r, t) (14)
Eq. (14) involves three fundamental dimensions (length L, mass
Mand time T). Three repeating variables, t, q
m
, and S are selected to
form four pi groups, i.e.

1
=
vt
S
(15)

2
=
r
S
(16)

3
=
D
S
(17)

4
=
PSt
q
m
(18)
By employing the power function, the following functional rela-
tion is obtained
v = k
1

S
t

r
S

l
1

D
S

m
1

PSt
q
m

n
1
(19)
where k
1
is a dimensionless constant and l
1
, m
1
and n
1
are the
exponents of the dimensionless products.
Substituting Eqs. (19) and (13) into Eq. (5) yields the nal equa-
tion for erosion rate for MAJM hole machining, i.e.
E
rate
= K
1

E
H

a
1

K
2
c
rH
2

b
1

S
2
Ht
2

c
1
r
S

x
1

D
S

y
1

PSt
q
m

z
1
(20)
where K
1
= Ck
0
k
2c
1
1
, a
1
, b
1
, c
1
, x
1
(=2c
1
l
1
), y
1
(=2c
1
m
1
), and z
1
(=2c
1
n
1
) are constants that can be determined from experiments.
2.3.3. Channel machining
For channel machining, the particle velocity is given by
v = g
2
(P, S, D, q
m
, r, u) (21)
Since the number of fundamental dimensions in Eq. (21) is three
(length L, mass M and time T), three repeating variables, u, q
m
, and
S were selected to form four pi groups, i.e.

1
=
v
u
(22)

2
=
r
S
(23)

3
=
D
S
(24)
J.M. Fan et al. / Wear 266 (2009) 968974 971
Table 1
Material properties of soda-lime glass.
Elastic modulus E (GPa) 70
Vickers hardness H (GPa) 5.5
Fracture toughness Kc (MPa m
1/2
) 0.75
Density t (g/cm
3
) 2.5

4
=
PS
2
uq
m
(25)
Applying the power law formulation gives
v = k
2
u

r
S

l
2

D
S

m
2

PS
2
uq
m

n
2
(26)
By substituting Eqs. (26) and (13) into Eq. (5), the erosion rate
for MAJM channel machining can be found to be
E
rate
=K
2

E
H

a
2

K
2
c
rH
2

b
2

u
2
H

c
2
r
S

x
2

D
S

y
2

PS
2
uq
m

z
2
(27)
where K
2
= Ck
0
k
2
2c
2
, a
2
, b
2
, c
2
, x
2
(=2c
2
l
2
), y
2
(=2c
2
m
2
), and z
2
(=2c
2
n
2
) are constants to be determined from experiments.
3. Experimental procedure
In order to understand the effects of process parameters on the
material erosion in MAJM and to determine the coefcients in the
developed erosion models for hole and channel machining, a sta-
tistically designed experiment has been carried out. The erosion
tests were carried out on an SS-White model K Series II Airbra-
sive unit, using dehumidied compressed air with abrasives. The
nozzle was mounted on a XYZ table in a closed and exhausted
chamber, which allowed to position and move the nozzle accu-
rately in the XYZ directions. Adust collector was used to suck and
clean abrasive dusts from the closed blasting chamber. The widely
used soda-lime glass of 5mm thickness was used as the eroding
specimens whose properties are summarized in Table 1.
In the MAJM process, a large number of variables are involved
and virtually all these variables affect the cutting results. For the
purpose of the current investigation, only the major and easy-
to-control dynamic variables were considered in this experiment.
Thus, the erosion rate was studied under four major process
parameters which included air pressures, abrasive mass ow rate,
stand-off distance, and machining time (for hole machining) or tra-
verse speed (for channel machining). The ranges of the process
parameters for this experiment were selected according to the sys-
temlimitations andcommonpracticeas giveninTable2. Withthese
conditions, the depths of holes or channels produced were within
700m.
To reduce the number of tests, and yet to give the experimental
data for sensible analysis, the Taguchi orthogonal array was used
to plan the experiment. Two groups of tests were considered in
the experiment design, one for hole machining and the other for
channel machining. Each group of tests included two parts; the rst
was the four-level, four-factor design scheme in Taguchi orthogonal
array to study the inuence of process parameters on the erosion
rate, and the other was a full-factorial design scheme for some typ-
ical conditions to enable to study the single-factor inuence using
the as measured data.
After an erosion test, the sample was cleaned with pressured air
before its mass loss was measured using an analytical balance (A&D
GR-200) with a resolution of 0.1mg. At least three measurements
for each sample were made and the average was taken as the nal
reading.
4. Model assessment
The models in Eqs. (20) and (27) may be considered as the
general form of the erosion rate models for brittle materials, in
particular amorphous glasses. For a given work material, the con-
stants in the models need to be determined from experiments.
Based on the experiment data obtained in the last section, a regres-
sion analysis has been carried out to determine the constants at a
95% condence level. When carrying out the multi-variable regres-
sion analysis, K
c
, H, E and r are constants in this case, so that E/H
and K
2
c
/(rH
2
) are essentially constants and can be removed from
the regression equations. The nal erosion rate equation for the
machining of holes on the soda-lime glass is given by
E
rate
= 1.854

S
2
Ht
2

0.846

r
S

0.182

D
S

1.37

PSt
q
m

1.07
(28)
Re-arranging the equation gives
E
rate1
= 1.854

t
r
0.182
D
1.37
P
1.07
S
1.21

0.154
H
0.846
t
0.622
q
1.07
m
(29)
Similarly, the nal erosion rate equation for the machining of
channels on the soda-lime glass becomes
E
rate
= 0.0196

u
2
H

0.732

r
S

1.36

D
S

0.289

PS
2
uq
m

1.06
(30)
Re-arranging the equation gives
E
rate2
= 0.0196

t
r
1.36
D
0.289
P
1.06
S
0.471
u
0.404

0.268
H
0.732
q
1.06
m
(31)
The above equations are valid for the conditions specied in the
experimental work. In these equations, r is in m, D and S are in
mm, P is in MPa, H is in GPa, and
t
are in g/cm
3
, q
m
is in g/s, u is
in mm/s, t is in s, K
c
is in MPa m
1/2
.
According to the mathematical models, the erosion rate is pro-
portional to air pressure, stand-off distance, and nozzle diameter
and inversely proportional to abrasive mass ow rate. For hole
machining, the erosion rate is inversely proportional to machin-
ing time, while in channel machining it is proportional to traverse
speed. These characteristics are consistent to the ndings from the
experiments [11,12], which implies that the models have been cor-
rectly formulated.
Table 2
Process parameters for the MAJM of soda-lime glass.
Process parameters Level 1 Level 2 Level 3 Level 4
Air pressure P (MPa) 0.43 0.52 0.60 0.69
Abrasive mass ow rate qm (g/s) 0.1 0.117 0.133 0.15
Stand-off distance S (mm) 1 2 3 4
Machining time t (s) (for hole machining) 3 6 9 12
Traverse speed u (mm/s) (for channel machining) 2 4 6 8
Nozzle diameter D (mm) 0.46 0.36
Eroding specimen: soda-lime glass of 5mm thickness. Abrasive: aluminum oxide particles of 27m average particle size with a density of 3.95g/cm. Impact angle: 90

.
972 J.M. Fan et al. / Wear 266 (2009) 968974
Fig. 3. Predicted and experimental results for the erosion rate of hole machining.
Figs. 3 and 4 showthe predicted trends (in line) and experimen-
tal data (insymbols) for holeandchannel machining, respectively. It
can be noticed fromthe gures that the erosion rate increases with
an increasing in the air pressure (Figs. 3(a) and 4(a)) and stand-
off distance (Figs. 3(b) and 4(b)), but decreases with the abrasive
mass ow rate (Figs. 3(c) and 4(c)) for hole and channel machin-
ing. Figs. 3(d) and 4(d) show that the erosion rate decreases with
an increase in machining time, but increases with an increase in
Fig. 4. Predicted and experimental results for the erosion rate of channel machining.
J.M. Fan et al. / Wear 266 (2009) 968974 973
Fig. 5. Percentage deviations of the model predictions from the experimental data.
the traverse speed. These trends may be anticipated as an increase
in the air pressure will increase the particle velocity and energy
to increase the material removal. Similarly, the effect of stand-
off distance, abrasive mass ow rate, machining time (for holes)
and nozzle traverse speed (for channels) on the erosion rate is
attributed to the particle interference and particle deposition on
the target surface. An increase in the stand-off distance increases
the jet erosion area and reduces the impacting particle density
which reduces the particle interference and particle deposition on
the target surface so that the average erosion rate is increased. A
similar effect can be achieved when the abrasive mass ow rate or
machiningtime is reducedor the nozzle traverse speedis increased.
Although an increase in stand-off distance may reduce the par-
ticle energy due to the surrounding air effect on the ow, this
reduction is less signicant than its positive effect to increase the
overall material removal. While these trends may be anticipated,
it is heartening in that the models have correctly predicted these
effects.
It is further noticedthat the predictedtrends are consistent with
the experimental data as shown in the gures. It can be seen from
Figs. 3 and 4 that the predicted erosion rates correlate with the
experimental data very well, indicating again that the models have
beencorrectlyformulatedtorepresent theeffects of various process
variables.
A quantitative assessment has also been made based on the
percentage deviation of the model predicted results with respect
to the corresponding experimental data, as shown in Fig. 5(a)
and (b) for hole machining and channel machining, respectively.
It can be noticed that the average deviation for the erosion rate
in hole machining is 0.479% (0.479%) with a standard deviation
of 16.51%, while the corresponding average deviation for chan-
nel machining is 1.055% (1.055%) and the standard deviation is
15.96%. From these excellent agreements between the predicted
and experiment results, it is apparent that the erosion rate models
developed can give good predictions both qualitatively and quanti-
tatively for hole and channel machining of glass by an abrasive air
jet.
5. Conclusions
Predictive mathematical models for the erosion rate in hole and
channel machining on glasses by micro abrasive air jets have been
developedusing a dimensional analysis technique. Amulti-variable
regression analysis of the data from a statistically designed experi-
ment combining the Taguchi orthogonal array with a full-factorial
design scheme has been used to determine the constants in the
models. It has been shown that the air pressure, stand-off dis-
tance and nozzle diameter have a positive exponent, and abrasive
mass owrate carries a negative exponent. For hole machining, the
machining time has a negative exponent, while a positive exponent
is associated with the traverse speed in channel machining. These
characteristics correctly reect the effects of these variables. The
predictive capability of the model has been further assessed by the
qualitative and quantitative comparisons between the predicted
results and experimental data considering a range of common pro-
cess variables suchas air pressure, abrasivemass owrate, stand-off
distance andmachining time (for hole machining) or traverse speed
(for channel machining). It has been shown that model predic-
tions are in good agreement with the corresponding experimental
data with very small average deviations of 1% or less. These models
provideanessential basis for theprocess optimizationof this micro-
machining technology to achieve efcient and effective operations
in practice.
Acknowledgements
The authors wish to acknowledge the National Natural Sci-
ence Foundation of China (through grant #50475045), Guangdong
Technological Plan Project (through grant #2003C50115), and the
Australian Research Council (ARC) (through grant #DP0878285) for
nancial support to this work.
References
[1] A. Ghobeity, H. Getu, T. Krajac, J.K. Spelt, M. Papini, Process repeatability in
abrasive jet micro-machining, Journal of Materials Processing Technology 190
(13) (2007) 5160.
[2] E. Belloy, S. Thurre, E. Walckiers, A. Sayah, M.A.M. Gijs, The introduction of pow-
der blasting for sensor and microsystemapplications, Sensors and Actuators A:
Physical 84 (3) (2000) 330337.
[3] P.J. Slikkerveer, P.C.P. Bouten, F.C.M. de Haas, High quality mechanical etching
of brittle materials by powder blasting, Sensors and Actuators A: Physical 85
(13) (2000) 296303.
[4] J.M. Fan, C.Y. Wang, J. Wang, Development of Micro Abrasive Jet Machining
Technology, Diamond and Abrasives Engineering 145 (1) (2005) 2530.
[5] Z. Moktadir, H. Wensink, M. Kraft, Analytical model of micromachining of brit-
tle materials with sharp particles, Microelectronics Journal 36 (36) (2005)
608611.
[6] B.R. Lawn, Amodel for the wear of brittle solids under xedabrasive conditions,
Wear 33 (2) (1975) 369372.
[7] S.M. Wiederhorn, B.J. Hockey, Effect of material parameters on the erosion
resistance of brittle materials, Journal of Materials Science 18 (1983) 766780.
[8] C.T. Morrison, J.L. Routbort, R.O. Scattergood, Solid particle erosion of mullite,
Wear 105 (1) (1985) 1927.
[9] P.J. Slikkerveer, P.C.P. Bouten, F.H. int Veld, H. Scholten, Erosion and damage by
sharp particles, Wear 217 (2) (1998) 237250.
[10] J. Wang, Predictive models for the depth of jet penetration in abrasive waterjet
cutting of alumina ceramics, International Journal of Mechanical Sciences 49
(3) (2007) 306316.
[11] J.M. Fan, C.Y. Wang, J. Wang, Study on the machining performance of
micro abrasive jet machining, China Mechanical Engineering 19 (5) (2008)
584589.
[12] J.M. Fan, C.Y. Wang, J. Wang, G.S. Luo, Effect of nozzle type and abrasive on
machinablity in micro abrasive air jet machining of glass, Key Engineering
Materials 359360 (2008) 404408.
974 J.M. Fan et al. / Wear 266 (2009) 968974
[13] Q.L. Li, J. Wang, C.Z. Huang, Erosion mechanisms of monocrystalline silicon
under a microparticle laden air jet, Journal of Applied Physics 104 (3) (2008),
034903 (8 pages).
[14] M. Wakuda, Y. Yamauchi, S. Kanzaki, Material responsetoparticleimpact during
abrasive jet machining of alumina ceramics, Journal of Materials Processing
Technology 132 (13) (2003) 177183.
[15] E. Isaacson, M. Isaacson, Dimensional Methods in Engineering and Physics,
Edward Arnold, London, 1975.
[16] G.I. Barenblatt, Dimensional Analysis, Gordon and Breach, New York, 1987.
[17] P. Chevallier, A.B. Vannes, Effects on a sheet surface of an erosive particle jet
upon impact, Wear 184 (1) (1995) 8791.
[18] A.N.J. Stevenson, I.M. Hutchings, Scaling laws for particle velocity in the gas-
blast erosion test, Wear 181183 (1) (1995) 5662.
[19] A.N.J. Stevenson, I.M. Hutchings, The inuence of nozzle length on the diver-
gence of the erodent particle stream in a gas-blast erosion rig, Wear 189 (12)
(1995) 6669.
[20] M. Achtsnick, A. Holtsmark, A.M. Hoogstrate, B. Karpuschewski, Design and
testing of a laval nozzle for micro-abrasive-air-jetting, in: IMPLAST 2003, New
Delhi, India, 2003, pp. 952962.
[21] F.H. int Veld, P.J. Slikkerveer, Towards prediction of ux effects in powder blast-
ing nozzles, Wear 215 (12) (1998) 131136.
[22] K. Anand, S.K. Hovis, H. Conrad, R.O. Scattergood, Flux effects in solid particle
erosion, Wear 118 (2) (1987) 243257.

Anda mungkin juga menyukai