7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30June 4, 2010
Effect of Powder Delivery System Parameters on a Powder Cladding System with one
Lateral Nozzle
F. Mohagheght, H. Najafian H. Bisadi and S. M. Hosseinalipour
Engineering Research Institute, Ministry of Agriculture
No. 68, North Saba, Tehran, Iran
mohagheghfi~alum. sharif. edu
Department of Mechanical Engineering, Iran University of Science and Technology
Narmak, Tehran, Iran
Email: bisadii~iust.ac.ir, hamed.naiafiani~imechenp.iust.ac.ir, alipouri~iiust.ac.ir
Keywords: laser cladding, gasparticle flow, Lagrangian tracking, deposition efficiency.
Abstract
Numerical simulation of laser cladding process has been implemented with the aim of finite volume base FLUENT CFD
software in order to study of some powder delivery system parameters. In the analysis, relations of continuum, momentum
and turbulence of air as the continuous phase were studied in Eulerian approach and the equations of continuous phase were
studied in Lagrangian approach. The study shows that the carrier gas flow deviates the streamlines against the nozzle.
Intense gradient pressures are seen in two regions: one in the lateral nozzle and the other in near the molten pool where gas
flow jet contacts the deposition surface. The maximum turbulence kinetic energy is just outside the lateral nozzle where the
high speed carrier gas impacts the stagnant air. Particles concentration distribution obeys the Gaussian distribution in the
nozzle. The Effective twophase flow parameters that have been studied include initial particles velocity, carrier gas
velocity and nozzle angle. The result show that any increase in initial particle velocity, carrier gas velocity and nozzle
direction angle leads to increase of deposition efficiency.
Introduction
In a Laser cladding process, deposition of thin layer of a
desired metal on a moving substrate is done by utilizing a
laser heat source to form the metallurgical bonding. This
postmanufacturing method is used to repair damaged or
womn components or to surfaceharden the materials that
are sensitive to corrosion and abrasion. The nature of the
laser beam causes a wellcontrolled heattreated zone
because a laser beam is well confined and tense and, as a
result, the rapid heating and cooling that occur in the
process have little effect of heat on the base material.
High diversity of materials can be deposited on a metal
basis by laser cladding with powder injection that can
provide a layer with thickness from 0.05 to 2 mm and
width as low as 0.4 mm. The process also has the ability of
multilayer deposition. The method is in a way that the
laser beam focuses on the surface of the metal basis and
produces a molten pool on the surface. The powder
particles with a diameter between 20 to 200 pm are carried
by a neutral carrier gas and injected towards the molten
pool and melt. The substrate moves with a velocity of
about few millimeters per second and the molten pool
becomes cold soon after formation and as a result the
surface covering would be created (Famia 2008).
Interaction of powder particles with the molten pool is a
significant parameter to reach the appropriate covering. The
impact of particles with the surface may take place in three
ways (Toyserkani 2004):
Contact of solid particles with solid surface results in
ricochet of particle and their waste.
Collision of solid or liquid particles with molten surface
causes formation of layer over surface.
Impact of liquid particles with solid surface brings about a
layer over surface.
Type of nozzle and its shape and also dispersion of powder
particles on the molten pool have significant effects on
interaction of particles with surface. The nozzle should be
designed in a way that the least amount of solid particles
contact the surface to increase the efficiency of powder
usage. Some other affecting parameters are laser power,
canonical distance of laser beam, material characteristics,
deposition rate and gasparticle two phase flow
characteristics which include powder feed rate, carrier gas
velocity, nozzle angle, nozzle diameter ... (Famia 2008).
Rajaratnam (1976), Antonia and Bilger (1973) and Fan et al
(1992) have studied the construction of turbulent free jets
and multiple jets theoretically and experimentally. Lin
(1998) was one of the pioneers in the study of gaspowder
flow in coaxial nozzles. He used FLUENT software in his
numerical analysis to explain the powder distribution in an
airpowder flow. A free jet was considered with no surface
collision in the study. Based on the results, the number
density of particles was declined by increase of gas
velocity. Steady state conditions and disregarding the laser
beam radiation were from the basic assumptions.
Toyserkani et al (2004) did different researches on laser
cladding process from them were numerical study of
coaxial nozzles in laminar and turbulence regimes. A
simple experiment was proposed to find the amount of fee
jet divergence after exiting the nozzle. Using the picture
study of particles in different nozzle diameters and
different flow rates of carrier gas, a function for jet flow
diameter was suggested which was based on the squared
distance from nozzle.
Zekovic et al (2007) studied the gasparticles flows
numerically and experimentally. They used four radially
symmetric nozzles to inject the particles to the laser
radiated area. In the numerical analysis with FLUENT, the
interaction between laser and particles were ignored. The
importance of using a shield gas in order to protect the
laser optical parts was emphasized, especially in turbulent
flows that contact a flat surface.
Lee (2008) studied the effects of different laser cladding
parameters on deposition efficiency (percentage of
particles trapping on the molten pool) when a low power
Nd:YAG laser is used with two symmetric lateral nozzles.
Based on the results, nozzle aiming position and angle,
flow rate of solid particles and the relative velocity
between laser and the substrate have significant effects on
deposition efficiency. In contrast, types of shield gas, laser
pulse shape and laser beam focal position do not have a
meaningful effect on the efficiency.
This article focuses on the twophase flow of a laser
cladding system with one nozzle. The aim is to adjust the
affecting parameters to have the maximum powder
efficiency.
TwoPhase Flows
There is not a common model for twophase flows and
selection of any model is based on the condition and
prevailing parameters. The analysis of gas particle flows
that contains a low volume of particles in relation to gas
volume is usually done by combining Eulerian and
Lagrangian approaches. The gas phase is analyzed as a
continuous phase in Eulerian approach and the following
equations are solved numerically:
1 Continuity
2 Momentum transfer
3 Turbulence (in the case of turbulence flows)
4 Energy (in the case of heat transfer)
Analysis of the solidphase is in Lagrangian approach and
for each particle momentum equations are considered.
Then the flow field is calculated for the entire domain. The
combination of two approaches leads to obtaining the
flight pass of particles statistically. The common
Lagrangian ways are deterministic trajectory approach
which neglects any turbulence fluctuation effect and
7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30June 4, 2010
stochastic trajectory approach that considers turbulent
fluctuations impacts on the particles.
During a particle motion within the air, different types of
force appear: drag, Magnus, Saffman, Brownian, Basset,
thermal gradient and buoyancy. Regarding the particle sizes
in the process and also physical conditions, only drag
coefficient is not negligible. If a particle moves in a carrier
fluid, the difference between droplet and the fluid results in
pressure gradient and viscosity tension. The resulted force
FD is the drag force:
18p CD Re
FD= p 2 24 (U Up) (1)
Where pu, ?,, d,, and Re are viscosity, particle density,
particle diameter and Reynolds number respectively.
If the volume fraction of solid phase were less than 10%,
the flow would be a dilute flow and the interaction of fluid
and particles become insignificant and the fluid phase
operates as the prevailing phase. In the present research, the
volume fraction is less than 1% and the effects of particles
on gas are not considered.
One of the essential parameters in identification of
twophase flows is the Stokes number:
St=" (2)
Where t, is particle response time and t, is the time scale
based on the characteristic length (L,) and the characteristic
velocity (V,) of the system under investigation:
t' V Y (3)
For St<<1.0, the particle will follow the flow closely and
for St > 1.0, the particles will move rather independent of
the flow (FLUENT 2006).
Physics and Dimensions
The analysis is done by FLUENT software. A lateral nozzle
injects the iron particles with mean diameter of 100 pum and
flow rate of 5 g/min towards the molten pool. Carrier gas
flows with a rate of 1.25 L/min. The shield gas flow rate is
1.5 L/min to protect the optical parts from soot deposition
and entrance. The distance between nozzle tip and the
molten pool is 10 mm and the molten pool diameter is 1.5
mm. The internal diameter of the nozzle is 1.4 mm.
The Reynolds number outside the nozzle is 1300 which is
higher than 1000, the criterion for transition from laminar
to turbulence regiments (Streeter 1961). Therefore, a proper
turbulence model should be used.
Modeling and Numerical Analysis
SSTk? turbulence model was utilized to analyze the
turbulent flow. The model was proposed by Menter (1994)
in order to mix the robust near wall k? formulation with
ke model which is appropriate for analysis far from the
wall. In other words, the model simultaneously has the
advantages of k? model in low Reynolds regions and k?
model in high Reynolds areas (Fluent 2006).
A cylindrical space was produced with GAMBIT 2.3
7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30June 4, 2010
software as the flow domain and discretized like Fig. 1.
Regarding the geometry, a three dimensional analysis is
required to solve the problem. The radius of the cylinder is
40mm and its height is 20mm. The domain is large enough
to study all of the phenomena around the nozzle and
molten pool (Zekovic et al. 2007). The produced grid
contains 143513 tetrahedral cells.
Results
To start the parametric study, a nozzle angle equal to 450 is
considered for the first run. In Fig. 2, pathlines of
continuous phase are displayed. Based on the picture, the
carrier gas flow is the dominant flow in the domain which
deviates the pathlines to the other side of the lateral nozzle
because of the stronger flow of lateral nozzle (which carries
the particles) in relation to the vertical, shield gas, nozzle.
Fig. 2: Flow pathlines
The velocity contours are shown in Fig. 3. The effects of
prevalent lateral nozzle flow could be recognized easily.
The velocity of the nozzle flow after entering the domain
and contacting the ambient gas reduces. After impacting the
substrate, the flow continues its motion in a direction
almost parallel to the surface. The reason is existence of
boundary layer near the wall which causes high reduction in
momentum of flow vertical component.
Domain
Domain
Limitations
flomainl
Moltenl Pool
Fig. 1: Calculation domain for the used laser cladding
system
The boundary conditions are velocity inlet for carrier and
shield gas nozzle entrances and constant pressure for top
and sides of the cylindrical domain. Other boundaries are
considered as wall.
The assumptions for the analysis were:
The effect of laser on the particle flow is neglected.
The carrier gas and the shield gas are nitrogen. However,
the ambient gas is air. Since the density and the viscosity
of nitrogen are near to the air, it is considered that the
entire domain is filled with air.
 Because the volume fraction of particles is 0. 05% which
is less than 10%, the discrete phase model of the FLUENT
could be used and the effects of particles on air flow
motion could be ignored (FLUENT 2006).
 The particles are spherical.
 Within the coating process, the substrate has a motion
with respect to the laser beam and causes an elliptical
molten pool. However, in order to simplify the problem,
the molten pool is considered as a circle with a diameter of
1.5 mm which is equal to laser beam diameter.
 Particlewall collision outside the molten pool is
completely elastic.
The initial velocity of the particles is equal to the
average carrier gas velocity in the nozzle feeding hose:
0.074 m/s.
 Regarding the physical conditions, the flow is considered
as steady state flow.
Because the flow is turbulence, a stochastic trajectory
approach should be used. However, since the Stokes
number is high, the particles are not affected by the
instantaneous variations of the continuous phase and thus
deterministic trajectory approach also could be used.
The equations discretized by QUICK method. The
SI1VPLE model was used in a pressurebased solver to
couple the pressure and velocity relations. The
convergence criterion was 10~' for all equations.
Static pressure contour is presented in Fig. 4. It shows that
the pressure gradient is intense just in two regions: one is
the region within the lateral nozzle and the other is the
region that belongs to lateral nozzle jet flow impact with
the substrate. The high pressure gradient in the former
region is because of high pressure drop and in the later one
is as a result of sudden decrease in dynamic pressure of
high velocity carrier gas flow. The other implication from
Fig. 4 is approving of necessary vastness of the calculation
domain. As the pressure gradients near the sides of
calculation domain are negligible, the assumption of
constant pressure boundary conditions is valid.
rjic
2lezoa
lusrm
~''"",,
aooroa
Fig. 3: Contour ofvelocity
SlooC
i3?c
ni ~i
I ii
F Yig.4 oturo ttc rsue
. . 1
Fig.5 ilsrte ublnc mtcenry nth aea
nozzcletefo sams a iawietruec
regimeis oenn usd h oze hsapoe h
asupio ftrblne eie Tefowrn it ih
eloityotietenzzeadfutaesi otc ih
low eloiyflwotid h nzl.
Toyserkani et al. (2005) consider the particle distribution
at the exiting point of the nozzle as Gaussian distribution.
Fig. 6 displays the obtained concentration of particles at
the end of nozzle which is clearly Gaussian.
. unexvo.os
,
e
*
"
,
.*
250e+01
200..01 i
150e*01 i
100e+01
so.00oo
000e+00 . .
00008 00006 00004 0.0002 0 00002 00004 00006 0.0008
Position (m)
Fig. 6: Particle concentration at the end of nozzle
In the following, effect of twophase flows parameters
such as particles initial velocity, carrier gas velocity and
nozzle direction angle on the percentage of particles
reaching the molten pool, namely, deposition efficiency
(E,) are studied.
7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30June 4, 2010
Initial velocity of Particles
With increase of particle initial velocity (by decreasing
diameter of nozzle feeding hose) from 0. 74 m/s to 1.5 m/s,
the percentage of particles reaching the molten pool
decreases from 29% to 19%. Increase of particle initial
velocity from 1.5 m/s to 8 m/s leads to an increase from
19% to 80% in deposition efficiency. However, higher
velocities do not have a significant effect on the results. The
variations in the deposition efficiency with of particles
initial velocity are presented in Fig. 7. In overall, any
increase in initial velocity of particles, causes more
deposition of particles on the molten pool in a way when
particles initial velocity becomes equal to carrier gas initial
velocity, 82% of particles reach the molten pool.
100
0
~0 5 10
Fig. 7: deposition efficiency vs. particles initial velocity
Effect of initial velocity on deposition efficiency could be
expressed as the following: Based on Eq. 1, the drag force
is proportional to relative velocity between two phases.
When the continuous phase velocity is fixed and the initial
particle velocity is decreased, the relative velocity between
two phases (discrete and continuous phase) increases. For
the Stokes number higher than one, the differences between
two phase's velocities last for longer time. When the
relative velocities between two phases are high, the
particles are more affected by the drag force and under the
influence of gas flow they diverge more and reach the
molten pool in less percentage. On the other hand, when the
particles initial velocities are high, the impact of gas
momentum decreases and the particles move mainly based
on their initial momentum and the deposition efficiency
increases.
Carrier Gas Velocity
Any change in velocity of carrier gas, causes variation in
the deposition efficiency. Fig. 8 shows this variation in the
case of identical velocities for two phases. With mecrease of
carrier gas velocity from 3.4 m/s to 13.5 m/s, the percentage
of particles reaching the molten pool rises from 19% to
82%. There is not any significant change above 13.5 m/s.
11 9:00
Fig. 5: Contour of kinetic energy
DPM
Concentration
(kg/m3)
4
2
?
a 
Fig. 8: Variation of deposition efficiency with initial
carrier gas velocity
Contact of carrier gas jet flow with the wall causes
changes in gas flow direction. In subsonic flows, the
changes would transfer to upstream and alters the
divergence path of flow stream and therefore particles
before the contact. It means that as a result of high
momentum, increase of carrier gas velocity leads to
delayed effect of downstream on upstream and having less
divergence. Consequently, the divergence of particles
directions reduces and more particles reach the molten
pool. On the other side, reduction in gas flow momentum
intensifies the divergence of particles path and the
deposition efficiency diminishes.
Noule Angle
Table 1 displays the effects of nozzle angle on percentage
of particles that reach the molten pool. Initial velocities of
particles and carrier gas velocity are equal to 13.5 m/s. It is
clear the increase of angle, increases the deposition
efficiency.
Table 1 Variation of deposition efficiency
with nozzle angle
Anle (Degrees) 30 35 40 45 50 551 60
Dpstion
60 71 77 82 89 91 95
Efficiency (%)
With increase of nozzle direction angle with horizontal
coordinate, two main advantages appear:
The more nozzle direction angle, the more deposition
efficiency.
 The more nozzle direction angle, the more exposure to
laser beam.
Each item is expressed in the following:
As depicted in Fig. 9, when an inclined cylinder crosses
the horizontal plate, the formed shape in the intersection is
an ellipse with a minor diameter equal to diameter of the
cylinder, d. The major diameter is obtained from
7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30June 4, 2010
d'
Where a and b are major and minor diameters of the ellipse,
respectively. Increase of angle a, declines the major
diameter of the ellipse and therefore, higher percentage of
ellipse would be inside the molten pool. It is necessary to
mention that the formed ellipse is almost equal to the
surface where particles contact the substrate. The injected
particles have a slight divergence, thus some of the particles
do not contact the formed ellipse.
O 10 20 30 40] 5
Fig. 9: Schematic of the formed ellipse on the surface
The increase of angle a may have another benefit that is not
considered in the numerical analysis. It causes more
particle exposure to laser beam and consequently, the
particles have more time to be heated and melt. It is also
required to mention that the particles velocity in the domain
is usually more than 5 m/s and the laser beam diameter is
restricted to few millimeters. Thus the particles pass the
exposure area in less than milliseconds. To heat up the
particles there are two ways: one is using a more powerful
laser to melt the particles in the short exposure time which
is costly and may have side effects on the substrate. The
Other is letting the particles to have more exposure time.
Increasing the nozzle direction horizontal angle helps in the
second way. Regarding Fig. 10:
L, 2o CO ,
(5)
L, cos a2
Where L is the distance a particle passes under the exposure
of laser beam.
Fig. 10: Exposure length distance for two particles with two
angles
~OD
50
40
PD
S10
If aI=450,(~ I;li and tl and to2 are the exposure
times for two paths of La, and Lt2:
t 2 COS 81 = (6)
tl cos a2
Conclusions
Use of FLUENT CFD software was successful in the
analysis of a laser claddmng process. The study shows that
the carrier gas flow is the dominant flow in the system and
deviates the streamlines against the nozzle. Intense
pressure gradients are seen in two regions: one in the
lateral nozzle and the other in near the molten pool where
gas flow jet contacts the substrate. The maximum
turbulence kinetic energy is just outside the lateral nozzle
where the high speed carrier gas impacts the stagnant air.
Particle concentration distribution in the nozzle obeys the
Gaussian distribution.
The Effective twophase flow parameters that have been
studied include initial particle velocity, carrier gas velocity
and nozzle angle. Initial particle velocity has a significant
effect on number of particles reaching molten pool. If the
initial velocity rises from 0.74 m/s to 13.5 m/s, the
deposition efficiency increases from 29% to 82%. Any
growth in carrier gas velocity results in increase of
deposition efficiency in a way that changing the carrier gas
velocity from 3.4 m/s to 20 m/s leads to a jump from 19%
to 80% in deposition efficiency. Increase of nozzle
direction angle also has a positive effect on deposition
efficiency. Increase of the angle from 450 to 600 raises the
deposition efficiency from 82% to 95%.
References
Antonia R.A., Bilger R.W., "An Experimental
Investigation of An Axisymmetric Jet on A Co?owing
Air Stream", Journal of Fluid Mechanics, vol. 61, pp. 805
822, 1973.
Fan J., Zhao H., Chen K., "An Experimental Study of
TwoPhase Turbulent Coaxial Jets", Experiments in
Fluids, vol. 13 pp. 279287, 1992.
Farnia, A. "Surveying about Controlling Of Laser
Cladding Processing", Doctoral Seminar in Material
Engineering, Tarbiat Modarres University, Page 1745,
2008.
Fluent 6.3 User' s Guide, Chapter 12: Modeling
Turbulence, Fluent Inc. September 2006.
Fluent 6.3 User' s Guide, Chapter 22: Modeling Discrete
Phase, Fluent Inc. September 2006.
Lee H. K. "Effects of the Cladding Parameters on The
Deposition Efficiency in Pulsed Nd:YAG laser cladding",
Journal of Materials Processing Technology, vol. 202., pp.
321327, 2008.
7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30June 4, 2010
Lin J., "Numerical Simulation of the Focused Powder
Streams In Coaxial Laser Cladding" Journal of Materials
Processing Technology, vol. 105, pp. 1723, 2000.
Rajaratnam N., "Turbulent Jets", Elsevier, Amsterdam,
1976.
Menter. TwoEquation EddyViscosity Turbulence Models
for Engineering Applications. AIAA Journal, vol. 32 no. 8,
pp. 15981605, August 1994.
Streeter V.L., "Handbook of Fluid Dynamics", McGraw
Hill, New York, 1961.
Toyserkani E., Khajepour A., Stephen F. Corbin, "Laser
cladding", CRC Press, 2004.
Zekovic S., Dwivedi R., Kovacevic R., "Numerical
Simulation and Experimental Investigation of GasPowder
Flow from Radially Symmetrical Nozzles in LaserBased
Direct Metal Deposition" Int. Journal of Machine Tools &
Manufacture, vol. 47, pp. 112123, 2007.
