Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
S0924-0136(16)30033-4
http://dx.doi.org/doi:10.1016/j.jmatprotec.2016.01.032
PROTEC 14712
To appear in:
Received date:
Revised date:
Accepted date:
20-9-2015
19-1-2016
30-1-2016
Please cite this article as: Ahmed, Dewan Hasan, Naser, Jamal, Deam, Rowan
Thomas, Particles Impact Characteristics on Cutting Surface During the Abrasive
Water Jet Machining: Numerical Study.Journal of Materials Processing Technology
http://dx.doi.org/10.1016/j.jmatprotec.2016.01.032
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
Abstract
The cutting performance of an abrasive water jet mainly depends on the abrasive
particles velocities and impact angles as well as the physical properties of the particles
and the workpiece being cut. This ultimately causes the different modes of erosion
process at the time of cutting of the job specimen. During the cutting process the cutting
profile changes with the depth of the cut to form the kerf. Due to the change of radius of
curvature of the cutting surface with depth, the erosion process per unit depth decreases
because of the changing of the impact angle. Here comprehensive numerical studies
have been carried out to find the particle impact characteristics on the groove wall
(cutting surface) as well as side walls for different radii of curvature. The results
obtained from the simulations have indicated that the particle impact angles depend on
the radius of curvature. The study has shown that the particles have a tendency to slide
or stay close to the cutting surface for the large radius of curvature and have very small
impact angles. The particles primary impact velocity is decreased little, however, the
particles impact velocities are significantly decreased for the secondary, tertiary and
following impacts due to fluid drag. The numerical simulation results have been used to
calculate the particles distribution factor for both deformation wear and cutting wear.
The distribution factors indicate that the particles have tendency to slide on the groove
wall for higher radius of curvature. These findings are consistent with the literature.
Nomenclature
Symbol: Explanation
2
A: Surface area, m
B: Body force, N
C: Constant
d: Diameter, m
F: Force, N
FD: Drag force, N
FB: Buoyancy force, N
Gx: Impact flux
H: Total enthalpy, KJ/kg
h: Thermodynamic enthalpy, KJ/kg
h: Jet width, m
2 2
k: Turbulent kinetic energy, m /s
m: Mass, kg
2
P: Pressure or shear production, N/m
2
p: Thermodynamic pressure, N/m
r: Fraction of phase
R: Radius of curvature, m
S: Source or sink term of the variable
S: Arc length, m
T: Temperature,
t: Time, sec
u: Particle velocity, m/s
U: Velocity, m/s
v: Velocity, m/s
V: Velocity, m/s
w: Width of jet in the cutting face, m
X : Particle distribution factor
X C : Particle distribution factor for cutting wear
1 Introduction
Abrasive water jet process is mainly a high speed erosion process. Abrasive particles
with high impact speed cause the erosion of the cutting surface. Two modes of erosion
by abrasive particles in abrasive water jet machining processes have been identified by
Hashish (1991) they are; cutting wear mode, and deformation wear mode. Bitter (1963a,
1963b) developed a mathematical model for erosion which based on eroding particle
impact angle and showed that two things happen simultaneously in erosion process, a)
repeated deformation during collision or deformation wear and b) the cutting action of
the free moving particles or cutting wear. The deformation wear usually occurs due to
the larger particle impact angle and cutting wear occurs due to the lower particle impact
angle. Finnie (1972) studied the erosion process of ductile materials by solid particles in
a fluid stream and found that the important factors that influenced the erosion process of
ductile materials were: angles of impingement, particle rotation at impingement, particle
velocity at impingement, particle size, surface properties, shape of the surface, stress
level in the surface, particle shape and length, particle concentration in the fluid stream
and nature of carried fluid and its temperature.
Srinivasa and Ramulu (1994) identify two cutting region that occur during the abrasive
water jet machining process, one is Smooth Cutting region (SCR), and other is Rough
Cutting region (RCR). The SCR is mainly at the top of the kerf and the RCR mainly at
the bottom of the curve. El-Domaty et al. (1996) reported that the depth of cut for
abrasive water jet cutting consists of two regions: one is due to a cutting wear mode at a
shallow angle of impact, and the other is due to a deformation wear mode at a large
angle impact. Preece (1979) reported that the maximum erosion results from the impact
of particles at a shallow angle of impact (15o-20o) for ductile materials and normal
angles for brittle materials. During the abrasive water jet cutting process the kerf profile
always changes with time. As the particles traverse the cutting wear zone, the velocity
of the abrasive particles decreases and for this reason the material removal rate in the
deformation (or step) zone is lower compared with the cutting wear zone, which is
called the cutting lag. Kitamura et al. (1992) showed that the cutting lag appears on the
bottom side of the cut surface in a direction opposite to the cutting direction. They
reported from their experimental results that cutting lag increases linearly with an
increase in cutting traverse speed. This is often called time lag in abrasive water jet
cutting process. Due to the change of the kerf profile (i.e. radius of curvature) particle
impact angle is also varied.
In the cutting process we can consider the following considerations:
The primary i.e. first impact angle will be very shallow as the initial cutting
surface should have large radius of curvature.
Due to collisions with the wall, the abrasive particles velocities decrease so that
the secondary and further impacts result in a lower wear rate of the workpiece.
Due to the lower rate of wear rate with depth, the kerf profile is changed i.e. the
local radius of curvature decreases.
With the decrease of the radius of curvature the particle impact angles increase.
The waterjet process can also be used for removing coatings from a workpiece (cleaning
or de-coating). Mabrouki et al. (2000) investigated the de-coating process with
numerical simulation and experimental study for polyurethane-coated alloy Aluminium
sheets (A2024T3), but only for the plain waterjet (not for abrasive waterjet). They used
Ls-dyna3D code for the simulations and used Eularian/Lagrangian coupling for
Waterjet and target. The simulations did not use the k- turbulence model. Leu et al.
(1998) investigated the coating material removal from the substrate by both analytically
and experimentally. They reported that there is a critical standoff distance for optimum
cleaning and the water pressure and nozzle radius affect this critical standoff distance.
Meng et al. (1998) showed the erosion model due to the impact of water droplets with a
mathematical model. They derived the relation of the critical cleaning standoff distance
based on the travel speed, water pressure and nozzle radius. However, these papers deal
with the theory and simulation of de-coating but they do not consider the total erosion
process for abrasive water jet cutting process.
Lebar and Junkar (2004) carried out numerical investigation on unit event approach
which means the impact of particular abrasive particle. With the model it explained the
surface characteristics for abrasive water jet. However, the model didnt consider many
aspects like particle rebound, impact velocity and the impact velocity after the impact
etc. Later Junkar et al. (2006) simulated the impact of a single particle on the workpiece
by the finite element analysis. They calculated the craters sphericity and validated their
results with experiments. However, they did not consider the potential effects of
forthcoming particles and their related erosion consequences.
Ahmadi-Brooghani (2007) extended the work of Junkar et al. (2006) by applying the
Johnson-Cook model with the inclusion of strain rate approach for a single abrasive
particle. Maniadaki et al. (2007) carried out the numerical simulation to investigate the
workpiece material behaviour under pure water jet impingement. They clearly showed
the velocity profile both inside the nozzle and at the workpiece stand-off distances.
Moreover, they investigated impingement of a pure water jet on polyurethane-coated
aluminum showing erosion stages. Later, Maniadaki et al (2011) extended their work
for an abrasive water jet and investigated the crater circularity on different particle
impact angle and velocity for multi-particle simulation. From their simulation results
they claimed that the erosion mechanism is mainly due to the cutting deformation and
the ploughing which is the erosion process by spherical particles.
Turenne and Fiset (1993) modeled abrasive particle trajectories during the erosion
process by slurry jet. They found that particle trajectories were more important than
abrasive particle size and the abrasive concentration in slurry. They modelled
axisymmetric fluid flow and concluded that impact velocity increased (comparing with
fluid velocity) with an increase of particle size and also deflected near the surface
(wall). In the case of finer particles, the velocity of the particles was about the same as
that of the fluid. Mostofa et al. (2010) and Liu et al. (2003) investigate the abrasive
water jet by computational fluid dynamics. However, their studies mainly focus on the
particle behavior and distribution inside the nozzle. Mostofa et al. (2010) calculated the
erosion of the nozzle with the help of a simple erosion model. On the other hand, Liu et
al. (2003) carried out computation fluid dynamics analysis and discussed different
aspects of CFD simulation like: the selection of boundary conditions, initial guess,
solver control and convergence strategies. They observed that there is minimum
variation of the velocity and pressure along the jet while the variation across the jet is
not significant within 80% of the jet diameter.
In recent years, there are significant works have been done by Prof Wangs group and
others. Prof. Wangs group has brought a new dimension on erosion mechanism. For
example, Li et al. (2013) described the erosion process in a fundamental manner rather
than the cutting or deformation wear phenomenon. Three different material failure
modes are identified, which are (i) inertia-induced tension failure, (ii) elongationinduced tension failure and (iii) adiabatic shear banding-induced failure. At a larger
impact angle, more impact energy is transferred onto the target and converted into larger
inertia of the target material so that the inertia-induced tension failure generally prevails
at a larger impact angle. When a particle impacts with low angle, the damage is formed
on the crater wall as the particle passes by, as result of an elongation-induced tension.
Material removal by sharp particles is mainly due to the elongation-induced tension
failure. A considerable amount of material removal is made due to melting in the case of
conical particle impact. From the sequential study of Prof Wangs group [Li et al.
(2013), (2014)] they concluded that inertia-induced fracture is the primary material
removal mechanism at normal impact angles, while the thermal instability-driven
failure, or specifically the adiabatic shear banding (ASB) induced failure, as well as the
elongation-induced fractures are the two major material removal mechanisms at oblique
impact angles.
Takaffoli and Papini (2012a) carried out numerical study by coupling both smoothed
particle hydrodynamics (SPH)/finite element (FE) model to simulate 124 nonoverlapping particle impacts on an Al6061-T6 target, both at normal and oblique
incidence. With the model they generated a realistic particle geometries based on
measurements of the distribution of particle surface area, circular diameter, sphericity
and thickness. Later, Takaffoli and Papini (2012b) used the same methodology to
simulate overlapping impacts, and thus the material removal mechanisms associated
with the solid particle erosion of this material. They observed that overlapping particles
impact plays significant contribution on material loss and solid particle erosion
mechanism such as the micromachining of chips, the ploughing of craters, and the
formation, forging and knocking off of crater lips.
Anwar et al. (2011) used finite element to investigate the craters profile for different
particle impact angles by considering the effects of high-strain rate plastic deformation
and adiabatic heating. Considering the Ti-based superalloy (Ti-6Al-4V) as work
specimen, their predicted depth of cut gave good agreement with experiments when the
particle impingement angle is 90o. However, for the angled impingement the results
varied because of various regions including divergence of the jet and sliding effects.
Recently, Lv et al. (2015) carried out numerical simulation to investigate the abrasive
erosion process in ultrasonic-assisted abrasive waterjet (AWJ) machining. The results
highlighted that the erosion rates under partially overlapping conditions are lower than
those under entire-overlapping conditions. Also, the sharp particle can cause higher
erosion rate at the impact angle of 30. For all conditions, the erosion rates with the
workpiece vibration are higher than those without vibration.
Kumar and Shukla (2012) carried out finite element analysis to model the erosion
behavior in abrasive water jet machining. They investigated the crater sphericity and
depth and erosion rate based on particle impact angle (ranging from 30o to 90o) and
velocity (180 m/s to 220 m/s). They observed that for the impact angle of 90, the
sphericity evaluated at the top surface of workpiece remains equal to 1.0, irrespective of
the particle velocity. For impact angles less than 90, the sphericity value goes on
decreasing as the number of impacting particles increases. The stabilization of
sphericity depends on the impacting particle angle and the particle velocity. On the
other hand, Ahmadi-Brooghani et al. (2007) claimed that the crater's sphericity
increases approaching 1 when the impact angle increases and the crater's sphericity
decreases when the velocity of particle increases.
The major objective of this paper is to investigate the abrasive particles impact
characteristics like particle impact angle, impact velocity for different radii of curvature
of the kerf. Numerical simulation is carried out for by Lagrangian particle transport
model to investigate the abrasive particles impact angle on the groove surface and also
the impact flux along the arc length of the groove. Moreover, the particles impact flux is
calculated to understand the erosion mechanism for abrasive water jet machining.
2 Solution Method
The numerical simulations were carried out for two phases (Water and Abrasive
Particles) using Lagrangian particle transport model with the help of commercial
software CFX (1997). The governing equations were discritised with finite volume
approach. In general the average velocity of abrasive water jet is more than 200 m/s.
Here the geometry was created for 22 mm groove. The flow was considered as a
turbulent flow. Here the simulation was carried out with the jet velocity 200 m/s and the
jet was seeded with 1000 particles that enter the groove.
GOVERNING EQUATIONS:
Governing equations were solved for the time averaged values of the velocity
components, pressure and turbulence parameters. The equations, represented below,
were closed using k- turbulence model.
The continuity equation
. r U 0
(1)
T
. r U eff U U r B p F
where eff T
(2)
(3)
eff is the effective viscosity, is the molecular viscosity and T is the turbulent
viscosity.
And T C
k2
(4)
Here k is the turbulence kinetic energy and is the dissipation length scale.
The energy equation
. r UH T 0
(5)
. r U rS
(6)
T
and is the turbulent Prandtl number
The transport equation for k and takes the same form as the generic scalar advectionwhere
diffusion equation
. r Uk T k rSk
(7)
(8)
. r U T rS
The source terms were considered to be the same as their single-phase analogous and
thus:
Sk P
(9)
C1 P C2
(10)
u
(11)
F
t
where F is the sum of the forces acting on the particle, u the particle velocity vector and
mp the mass of the particle. The principle force is the Drag force:
mp
1
FD d 2 CD vR vR
8
(12)
where CD is the drag coefficient, d the particle diameter and vR the relative velocity
between the particle and the fluid.
The buoyancy force is defined as
1
FB d 3 g f g
6
where g is the gravitational acceleration.
(13)
In order to investigate the abrasive particles impact behavior during the cutting process
of the job specimen, the geometry was created with a curved surface. As the particles
impact behavior changes continuously during the cutting process downstream, the
whole cutting face or the cutting surface also changes with time. Here the geometries
were created for the curved surface with a constant radius of curvature. To investigate
the impact characteristics on the curved surface a number of simulations were carried
out with different radii of curvature. As the job specimens are usually placed at different
standoff distances, here in the geometry curve surface started at 2 mm standoff distance.
The groove (with curved surface) was created with 2 mm 2 mm with 900 bends has
shown in Fig. 1.
Here the simulation was carried out with the jet velocity 200 m/s and the jet was seeded
with 1000 particles that enter the groove. It has calculated the divergence of the abrasive
particles (jet) from three velocity components (u, v and w velocity) that the jet diverged
around 5.50 just after the focus tube. Particles velocity at the inlet was specified as 200
m/s with the divergence 5.50 and the mass flow rate 0.45 kg/min. All walls of the
groove (except the curve wall) were treated with zero shear stress. Johnson (1987)
reported that the values of co-efficient of restitution vary in wide range, however, for
higher velocities specially on supersonic conditions the value decreases much.
Gudimetla and Yarlagadda (2007) used 0.1285 as co-efficient of restitution for the
interaction between abrasive water jet particle and polycrystalline alumina ceramic. On
10
other hand, Zeng and Kim (1996) considered the values of 0.2 for their erosion model of
polycrystalline ceramics in abrasive water jet machining. In the present study the coefficient of restitution of the curve wall was considered 0.2. The exit was treated as
pressure boundary.The geometrical parameters and the boundary conditions are shown
in Table 1.
2.1.3 Meshing and Grid Independence
This study considers different radii of curvature, but a constant 2 mm standoff distance
of the workpiece from the focus tube. Simulations were carried out for 5 different radii
of curvature (0.5, 0.1, 0.05, 0.02 and 0.01m). The geometry is shown in Fig. 1. For the
simulations a grid was generated with the body fitted to a three-dimensional Cartesian
coordinate system. For the grid independence tests four different simulations (cases)
were continued (total grids 35,072, 79,488, 105,408 and 193,536 respectively) for the
radius of curvature 0.1m. The residual of mass goes down from 10-2 to 10-6 for 8000
iterations. The grid structures for the radius of curvature 0.01 with three different planes
are shown in Fig. 2. Fine grids are used on groove walls. The 900 bend groove wall was
divided into 180 segments and defined as the wall. The number of hits of the particles to
the wall (segment) for radius of curvature 0.1 m is shown in the following Fig. 3. And
for all cases the number of impacts on the groove wall (for radius of curvature 0.1 m)
was almost same and showed the same trend. So, for the radius of curvature 0.1 m,
88540 grids were chosen for numerical simulations. However, to reduce
computational time, different numbers of grids were used for the numerical simulation
for different radius of curvature. The variation of the number of grids in the Z direction
was varied from 360 to 720. The number of grids in X and Y directions were 1212.
2.1.4 Particle Independence Test
As the numerical simulations were carried out with a Lagrangian particle transport
model, each particle represents a spherical abrasive particle. Here the simulations were
conducted with different number of particles (1000, 2000 and 4000) in the fluid flow for
the radius of curvature 0.1 m. The following Fig. 4 shows the number of hits at different
angles for different number of particles. From the figure of number of hits and the
pattern indicates that the number of collisions of the particles with the curve wall
increases linearly with the number of particles. This means, that at these particle
concentrations, the interactions between particles is negligible. Therefore, to save the
time the present studies are carried out for the total number of particles 1000.
11
P
V12
V2
Z1 2 2 Z 2
2g
2g
F P2 * A
(4.1)
(4.2)
12
were measured for three different standoff distances (2, 4 and 6mm) and for three
different pressures.
To validate the numerical simulation result, the experiments were carried and compared
with the numerical simulation results. To calculate the force on solid surface, numerical
simulations were carried out for three different pressures 276 MPa, 300 MPa and 340
MPa. As mentioned before, to measure the jet striking force on a solid surface, three
different water pressures were used (276 MPa, 310 MPa and 345 MPa). Fig. 6 shows
the comparison of numerically calculated force and the experimental force
measurement.
The comparison shows the numerical results follow the same trend line of the
experimental results. From the comparison of forces between numerical and
experimental results, it is clear that the maximum jet striking force on workpiece is not
high, which is one of the main advantage of abrasive water jet machining compared
with other conventional cutting techniques.
13
particle to the wall) at certain angle, the impact fluxes were decreased and remained
almost constant for the rest of the path due to the secondary, tertiary or so on impact to
the wall. Another thing is that for the every radius of curvature a certain angular
position has found where no collisions of the particles to the wall were found from the
simulations results. For example, Fig. 7(a), for the radius of curvature 0.01 m, up to 13.5
degree initial angle there were no collision of the particles with the wall. This angle
decreases with the increase of radius of curvature. Therefore, it indicates that if the first
impact angle is shallow, then there is less chance of the abrasive particles to hit the kerf
(groove wall).
Fig. 8 shows the track of 10 particles out of 1000 particles for the radius of curvature
0.01m. The mean path of the particles is shown. The fluctuations in the path are not
shown here. It is clear from the visualization pictures that after the primary impact of
particles to the groove wall, particles stayed near the wall for the rest of the path. This
was mainly due to the particles centrifugal forces in the groove. It should be mentioned
that particles lost velocity after impacting the wall and, again, gained the velocity from
the fluid (water).
From the above figures there is a maximum impact flux for every radius of curvature,
and these maximum impact fluxes are found at certain angles of the 90 degree bend
kerfs. Figs. 9 and 10 show the relationship between the maximum impact flux and h/R
and the relationship between theta (where maximum impact flux was found for certain
radius of curvature) and h/R, where h is the jet width and R is the radius of curvature.
From these two figures it is clear that the impact flux (Gx) and theta (max) is a linear
function of (h/R)0.5 and the Pearson coefficient is more than 0.99.
Usually, at the initial stage of the cutting process, the radius of curvature of the job
specimen is almost infinity (for plane surface) and the jet goes down vertically. With a
large radius of curvature there is less chance of particles impacting the surface. Hashish
(1993) experimentally showed that jet angles other than 900 (jet goes down vertically)
are better for maximum depth of cut and kerf depth uniformity. Once particles impact
the solid surface, the radius of curvature of that portion of the surface decreases, leading
to an increase in impact flux, which causes more erosion. However, there were some
angular positions where there would be no impact of the particle. From Hashishs
(1993) experimental results and the simulation results has confined that there should
have a starting angle or initial angle of the oncoming particles towards the job
specimen. For these positions, erosion would have to depend on the erosion process on
the upstream portion through the oncoming jet. Nonetheless, there are many of other
parameters such as impact angle and impact velocity etc. that are also involved for the
erosion process for abrasive water jet.
14
15
Comparing the angle distribution between Fig. 11(c) and 12, the secondary, tertiary and
so on impact angles are comparatively smaller than the primary impact angles.
However, here the secondary, tertiary and so on impact occurs on the groove wall,
which was calculated for the same radius of curvature. In practice, the downstream
radius of curvature of the kerf becomes smaller and as a result the particles will impact
the cutting surface with much higher angles and also with smaller velocities.
4.4
In reality, the front wall (i.e. opposite of the cutting face) should be open in air. Here for
more realistic simulations, the front wall was modeled as a pressure boundary, as was
the walls from the inlet to the curve surface (the distance from inlet to curve surface was
the 2 mm standoff distance). This improved model was set up to find the particles
impact angle on the groove wall and side walls. Fig. 14 and Fig. 15 show the particles
impact angle distribution on the groove wall and the side walls respectively.
The particles impact angle distributions from the more realistic model displayed the
same trend, which was found from the previous set of simulations (all walls were
treated as zero shear stress except the groove wall i.e. Fig. 11). For smaller radii of
curvature the impact angle distributions between the two sets are almost same. However
for higher radii of curvature the predicted number of collisions with the walls decreased,
because particles had enough time to leave the domain through the pressure boundary. It
can be seen from the above improved simulation results that there is a great influence of
particles impact angle over the radii of curvature. As has been mentioned earlier,
particle impact angle is one of the prime parameters controlling the erosion mechanism.
Fig. 15 shows the particles impact angle distribution on side walls and also shows that
few particles hit the wall with large angles (very few particles impacted the wall at
around 900). The reason is that the fluid and particles velocity near the wall were small,
and as a result the fluctuating velocity component (due to turbulence) mean that few
16
particles strike the wall at larger angles. Occasionally the particles may be trapped in
turbulent eddies near the wall, which might make the particles hit the side wall with a
larger angle. Dutta and Nandi (2015) showed numerically that the mean flow behaviour
(for the curved pipe flow) is strongly influenced by the pressure gradient forces and as
well as the bend curvature. However, Anwer et al. (1989) found that, for flow in a
curved pipe, the normal stresses near the pipe wall increase rapidly as the flow enters
the bend. In their turbulence measurements they also found, the anisotropy between v 2
and u 2 becomes more significant near the outer bend than the inner bend, where v 2 is
the kinematic circumferential normal stress and u 2 is the kinematic radial normal stress
for the bend pipe. When particles move toward the wall they have to penetrate the
boundary layer and as a result the abrasive particles velocity near the wall decreases.
Due to turbulence, the velocity component of the fluid continuously changes when the
fluid goes toward the downstream.
17
It is clearly observed from the Fig. 17 is that the trend of the average velocity on
different radii of curvature are same. However, the average impact velocity of the
particles decreased with the increase of the radius of curvature. As mentioned earlier,
for a constant radius of curvature, the abrasive particles did not hit the groove wall up to
a certain angular position. As a result, the average velocity was found to be zero up to a
certain angle for different radii of curvature. The main point is that the particles impact
velocity decreased with the increase of the radius of curvature. It is expected that the
primary impact velocity of the particles should be same for all radii of curvature at least
at the first angular positions where the impact is obviously the primary one. There are
two important aspects for abrasive water jet machining: one is the cutting surface at the
initial state, which would have higher radius of curvature i.e. there are less chance of
particles to impact the cutting face; the other one is that with a larger radius of curvature
the particles impact velocity would be comparatively lower and these might reduced the
cutting efficiency of the abrasive water jet. For higher radii of curvature the particles
impact velocity was lower because particles have to travel a long distance to hit the wall
i.e. it takes a long time to penetrate the boundary layer in order to hit the solid wall.
Once the particles start to penetrate the boundary layer their velocity decreases and as a
result the particles impact velocity at the wall is smaller. However for smaller radii of
curvature particles penetrate the boundary layer with a higher angle (virtually for
smallest radius of curvature it should be 900) and as a result the particles do not get
enough time to decrease their velocity.
4.6
It is important to know the particle distribution over the solid surface during the
abrasive water jet machining to obtain the material removal rates from the job specimen.
Liu et al. (2004) carried out a similar study on particle velocity distribution, but without
consideration of the interaction with the solid surface of the groove. However, from the
above simulation results of the present study indicate that after primary impact at the
groove wall, particles slide along the wall or at least have a tendency to stay near the
wall region. This distribution of the particles in the groove channel caused more impact
flux than expected and, as a result, particles lost energy and tended to slide close to the
wall. This distribution of the particles in the groove channel is defined as the particle
distribution factor ( X ).
Deam et al. (2004) defined the particle distribution factor, which ultimately indicates
how dense the particles are near the cutting face. Two distinctive particle distribution
factors have been found for deformation and cutting wear. If the particles have a
normalized probability density distribution g(h) = 1, the particle distribution factor is
X D = 2.25 for deformation wear and X C = 2.44 for cutting wear (for details see Deam
18
et al. (2004). These values mean that the particles are evenly distributed in the groove
channel. Higher values of X D and X C mean that particle distribution is denser near the
cutting face than it is near the free surface. However, these simulation results do not
support the idea that particles are evenly distributed. Back in Fig. 7, it was shown that
particle impact flux changed rapidly along the wall for smaller radii of curvature when
compared with the higher radii of curvature. It might be that a certain arc length or
distance on the downstream (particles need to travel a distance in the groove) is required
for settling down the distributions of the particles.
Another set of simulations was carried out in order to reveal this distribution of the
particles in the groove wall. Simulations were conducted for three different radii of
curvature (0.01m, 0.05 m and 0.5 m) and for six different angular bends (150, 300, 450,
600, 750 and 900). As the fine grids (1212 grid configuration in X and Y direction)
were generated near the wall with geometric progression ration 1.31, and expected to
get a more accurate particle probability distribution in the groove channel. For every
angular bend and for every radius of curvature, particles distribution factor ( X ) has
calculated from their respective probability density ( g(h) ). Fig. 18 shows the variation
of the particles distribution factor for deformation and cutting wear respectively.
The above two figures indicate that the particles become denser near the wall for higher
angles (e.g. maximum X D and X C was found for 900 bend groove channel). The angle
increase indicates that particles travelled a longer length and had more chance to
become close to the wall. It should be mentioned that the values of X D and X C for the
radius of curvature 0.05 m and 0.5 m were close since the angle of the bend was 300.
Turbulence increased along the flow, especially in those cases where the flow direction
continuously changes (like bend pipe) as described by Anwer et al. (1989). The
following Table 2 shows the arc length that the jet has to travel for different angular
bends with different radius of curvature. Fig. 19 shows the variations of X D and X C
over S/h and indicates that the particle distribution factor increased sharply with S and,
after a certain distance, it remains constant.
The increased value of X D and X C over S/h reveals that the particles get closer to the
wall as they go downstream and, as a result, there are more collisions of the particles
with the groove wall that would cause more erosion of the surface. However,
subsequent collisions with the groove wall causes a loss of kinetic energy and, as a
result, particles may just touch or slide down the wall without any effective normal
velocity component. As a consequence, effective erosion may not occur.
19
5. Conclusion:
In the present study, we carried out numerical simulations of particle impact
characteristics on the groove wall during the abrasive water jet cutting process.
Different curvatures of the groove wall were also considered in the present simulation
as the curvature of the cutting surface consistently changes during the cutting process.
The simulation results showed that primary impact needed a certain arc length of the
groove wall especially when the radius of curvature was large. The calculated particle
impact flux revealed that the secondary, tertiary and following impacts on the groove
wall are the common feature of abrasive water jet machining. Particle impact angle on
the groove surface was much influenced by the radius of curvature of the groove wall.
Larger radii of curvature exhibited low particle impact angles and on other hand,
smaller radii of curvature resulted in large particle impact angles. In addition, the
velocities other than the primary impact are small enough and the particles are almost
sliding the groove wall with a high particle density flow near the groove wall. The
particle distribution factor was also calculated for both cutting wear and deformation
wear and showed that the particle distribution factors are comparatively large compared
to an analytical solution of Deam et al. (2004), which could be modified by the present
findings.
Acknowledgement
This work was supported by the Australian Research Council (ARC) of Australia.
20
References
Ahmadi-Brooghani, S. Y., Hassanzadeh, H., Kahhal, P., 2007. Modeling of singleparticle impact in abrasive water jet machining. International Journal of
Mechanical Systems Science and Engineering. 1 (4), 231236.
Anwar, S., Axinte, D. A., Becker, A. A., 2011. Finite element modelling of a singleparticle impact during abrasive waterjet milling. Proc. IMechE Vol. 225 Part J:
J. Engineering Tribology, 821-932.
Anwer, M., So, R. M. C., Y. G. Lai, Y. G., 1989. Perturbation By and Recovery
From Bend Curvature of a Fully Developed Turbulent Flow. Physics of Fluids
A. 1, 1387-1397.
Bitter, J. G. A., 1963a. A study of erosion phenomena Part . Wear. 6, 5-21.
Bitter, J. G. A., 1963b. A Study of Erosion Phenomena - Part II. Wear. 6, 169-190.
CFX, 1997, Solver Manual. CFX-4.2, AEA Technology plc.
Deam, R. T., Lemma, E., Ahmed, D. H., 2004. Modelling of the Abrasive Water Jet
Cutting Process. Wear. 257, 877-891.
Dutta, P., Nandi, N., 2015. Effect of Reynolds Number and Curvature Ratio on
Single Phase Turbulent Flow in Pipe Bends. Mechanics and Mechanical
Engineering. 19 (1), 516.
El-Domiaty, A. A., Shabara, M. A., Abdel-Rahman, A. A., Al-Sabeeh, A. K., 1996.
On the Modelling of Abrasive Waterjet cutting. International Journal of
Advance Manufacturing Technology. 12 (4), 255-265.
Finnie, I., 1972. Some Observations on the Erosion of Ductile Metals. Wear. 19,
81-90.
Gudimetla, P., Yarlagadda, P. K. D. V., 2007. Finite element analysis of the
interaction between an AWJ particle and a polycrystalline alumina ceramic,
Journal of Achievements in Materials and Manufacturing Engineering. 27 (1), 714.
Hashish, M., 1991. Wear Modes in Abrasive Waterjet Machining. PEDVol.54/TRIB-Vol.2, Tribological Aspects in Manufacturing, ASME, 141-153.
Hashish, M., 1993. The Effect of Beam Angle in Abrasive-Waterjet Machining.
Journal of Engineering for Industry, Transactions of the ASME. 115, 51-56.
Johnson, K. L., 1987. Contact Mechanics, Cambridge University Press, New York.
Junkar, M., Jurisevic, B., Fajdiga, M., Grah, M., 2006. Finite element analysis of singleparticle impact in abrasive water jet machining. International Journal of Impact
Engineering. 32 (7), 10951112.
Kitamura, M., Ishikawa, M., Sudo, K., Yamaguchi, Y., Ishimura, T., Tujita, K.,
1992. Cutting of Steam Turbine Components using Abrasive Waterjet. 11th
International Conference on Jet Cutting Technology, St Andrews, Scotland, 543554.
21
Kumar, N., Shukla, M., 2012. Finite element analysis of multi-particle impact on
erosion in abrasive water jet machining of titanium alloy. Journal of
Computational and Applied Mathematics. 236, 46004610.
Lebar, A., Junkar, M., 2004. Simulation of Abrasive Water Jet Cutting Process: Part
1. Unit Event Approach. Modeling and Simulation in Materials Science and
Engineering. 12, 1159-1170.
Leu, M. C., Meng, P., Geskin, E. S., Tismeneskiy, L., 1998. Mathematical Modeling
and Experimental Verification of Stationary Waterjets Cleaning Process, Trans.
ASME Journal of Manufacturing Science and Engineering. 120, 571-579.
Li, W. Y., Wang J., Zhu, H., Huang, C., 2014. On ultrahigh velocity micro-particle
impact on steels A multiple impact study. Wear, 309 (12), 52-64.
Li, W. Y., Wang J., Zhu, H., Huang, C., 2014. On ultrahigh velocity micro-particle
impact on steels A multiple impact study. Wear, 309 (12), 52-64.
Liu, H., Wang, J., Brown, R. J., Kelson, N., 2003. Computational Fluid Dynamics
(CFD) Simulation of Ultrahigh Velocity Abrasive Waterjet. Key Engineering
Materials. 233-236, 477-482.
Liu, H., Wang, J., Kelson, N., Brown, R.J., 2004. A study of abrasive water jet
characteristics by CFD simulation. Journal of Material Processing Technology,
153-154, 488-493.
Lv, Z., Huang, C., Zhu, H., Wang, J., Yao, P., Liu, Z., 2015. FEM analysis on the
abrasive erosion process in ultrasonic-assisted abrasive waterjet machining.
International Journal of Advanced Manufacturing Technology, 78, 16411649.
Mabrouki, T., Raissi, K., Cornier, A., 2000. Numerical Simulation and Experimental
Study of the Interaction between a Pure High-velocity Waterjet and Targets:
Contribution to Investigate the Decoating Process. Wear. 239, 260-273.
Maniadaki, K., Kestis, T., Bilalis, N., ntoniadis, A., 2007. Finite element based
model for pure waterjet process simulation. International Journal of Advanced
Manufacturing Technology. 31 (910), 933940.
Maniadaki, K., Bilalis, N., ntoniadis, A., 2011. Effect of impact angle and velocity
in crater circularity in abrasive water jet machining by means of multi-particle
impact simulation. International Journal of Machining and Machinability of
Materials. 10 (1-2), 34-47.
Meng, P., Geskin, E. S., Leu, M. C., Li, F., Tismenseskiy, L., 1998. An Analytical
and Experimental Study of Cleaning with Moving Waterjets. Trans. ASME Journal of Manufacturing Science and Engineering. 120, 580-589.
Mostofa, M. G., Kil, K. Y., Hwan, A. J., 2010. Computational Fluid Analysis of
Abrasive Waterjet Cutting Head. Journal of Mechanical Science and
Technology. 24, 249-252.
Preece, C., 1979. Teatise on material science and technology. Erossion, vol.16,
Academic press, New York, 69-126.
22
Srinivasa, P. R., Ramulu, M., 1994. Predicting Hydro-abrasive Erosive Wear During
Abrasive Water Jet Cutting: Part - A Mechanistic Formulation and Its Solution.
PED- Vol.68-1, Manufacturing Science and Engineering, ASME. 1, 339-351.
Takaffoli, M., Papini, M., 2012a. Numerical simulation of solid particle impacts on
Al6061-T6 part I: Three-dimensional representation of angular particles. Wear,
292-293, 100-110.
Takaffoli, M., Papini, M., 2012b. Numerical simulation of solid particle impacts on
Al6061-T6 Part II: Materials removal mechanisms for impact of multiple
angular particles. Wear, 296 (1-2), 648-655.
Turenne, S., Fiset, M., 1993. Modeling of Abrasive Particle Trajectories During
Erosion by a Slurry Jet. Wear. 162-164, 679-687.
Zeng, J., Kim, J. T., 1996. An erosion model of polycrystalline ceramics in abrasive
waterjet cutting, Wear. 193, 207-217.
23
Figure Captions
Right Wall
Left Wall
Free Surface (inner surface)
Groove Wall
(outer surface)
24
Fig. 2. Grids for the radius of curvature 0.01 m with three different planes
25
120
100
80
60
40
20
0
0
20
40
60
80
100
12*12*540
12*12*720
16*16*720
26
No. of Impacts/No. of
Particles used
0.12
0.1
0.08
0.06
0.04
0.02
0
0
20
40
60
80
100
P 2000
P 4000
Fig. 4. Number of impact distribution along the angle for particle independence test
27
a)
b)
Fig. 5. a) Multicomponent force measurement, b) Standard measuring chain
28
20
Force, N
18
16
14
12
10
250
270
290
310
330
350
Pressure, MPa
Force Experimental
Force (Numercal)
Fig. 6. Comparison of jet striking force between numerical and experimental results
29
200
(b)
120
120
Hit Flux
Hit flux
150
(a)
160
80
40
0
90
60
30
0
20
40
60
80
100
20
40
60
80
100
100
(d)
60
Hit Flux
Hit Flux
75
(c)
80
60
40
20
45
30
15
0
0
0
20
40
60
80
100
20
40
Hit Flux
60
80
30
25
20
15
10
5
0
100
(e)
20
40
60
80
100
Fig. 7. Impact flux along with the arc angle for different radius of curvature
(a) R = 0.01m, (b) R = 0.02 m. (c) R = 0.05 m, (d) R = 0.1 m and (e) R = 0.5 m
30
31
Impact Flux
200
160
120
Gx = 489.0(h/R)0.542
Pearson coefficient = 0.999
80
40
0
0
0.05
0.1
0.15
0.2
h/R
Collisions
Power (Collisions)
32
60
Theta (max)
50
40
30
20
10
0
0
0.05
0.1
0.15
0.2
h/R
theta (max)
33
(a)
(b)
(c)
(d)
(e)
Fig. 11. Particle impact angle distribution on groove wall for different radii
of curvature (a) R = 0.01 m, (b) R = 0.02 m, (c) R = 0.05 m, (d) R = 0.1 m and
(e) R = 0.5 m
34
Fig. 12. Particle impact angle (all i.e. primary, secondary, tertiary and so on)
distribution for radius of curvature 0.05 m
35
Average Impact
Angle, degree
20
16
12
8
4
0
0
20
40
60
80
100
r =0.02
r = 0.05
r =0.1
r =0.5
Fig. 13. Particles average impact angle with groove wall for different radius of
curvatures
36
(a)
(b)
(d)
(c)
(e)
Fig. 14. Particles impact angle distribution on groove wall for different
radii of curvature with real conditions
37
(a)
(b)
(c)
(d)
(e)
Fig. 15. Particles impact angle distribution on side walls for different radii
of curvature with real conditions
38
(a)
(c)
(b)
(d)
(e)
Fig. 16. Particles primary impact velocity distribution on the groove wall
39
Average velocity, m/s
180
150
120
90
60
30
0
0
20
40
60
80
100
R = 0.02
R = 0.05
R = 0.1
R = 0.5
Fig. 17. Average particles impact velocity on the groove wall for different radii of
curvature
40
75
60
60
45
45
XC
XD
75
30
15
30
15
0
0
20
40
60
80
Angle, degree
R = 0.01
R = 0.05
100
0
0
20
40
60
80
100
Angle, degree
R = 0.5
R = 0.01
R = 0.05
R = 0.5
41
75
60
60
45
45
XC
XD
75
30
30
15
15
0
0
0
100
200
300
400
100
r = 0.01
r = 0.05
200
300
400
S/h
S/h
r = 0.5
Fit
r = 0.01
r = 0.05
r = 0.5
Fit
Fig. 19. Variation of X D and X C over S/h for different radii of curvature
42
Tables
Table 1. Geometrical parameters and boundary conditions
Groove dimensions
2 mm 2 mm
2 mm
200 m/s
0.45 kg/min
Particle density
4100 kg/m3
Number of particles
1000
Particles diameter
100 m
43
Table 2. Arc length variations for different angular bend with different radii of curvature
R = 0.01 m
R = 0.05 m
R = 0.5 m
15
2.61667
13.08333
130.8333
30
5.23333
26.16667
261.6667
45
7.85
39.25
392.5
60
10.4667
52.33333
523.3333
75
13.0833
65.41667
654.1667
90
15.7
78.5
785
44