Download Numerical modelling of powder interaction with plasma flow in tandem-type plasma torch

22nd International Symposium on Plasma Chemistry
July 5-10, 2015; Antwerp, Belgium
Numerical modelling of powder interaction with plasma flow in tandem-type
plasma torch
A.M. Essiptchouk1,3, F.R. Caliari2, F. Miranda1, D.A.P. Reis2, G. Petraconi1 and L.I. Charakhosvki3
1
2
Instituto Tecnológico Aeroespacial, São José dos Campos, SP, Brazil
Universidade Federal de São Paulo, São José dos Campos, SP, Brazil
3
Luikov Heat and Mass Transfer Institute, Minsk, Belarus
Abstract: A simplified numerical 3D model of tandem-type plasma torch for supersonic
plasma spraying at atmospheric pressure is presented. The computational model was
validated by comparing with experimental data. Peculiarities of the flow-flow and
flow-particles interaction was shown by distribution of principal thermal and kinetics
parameters.
Keywords: atmospheric plasma spray, plasma torch, coating with supersonic flow
1. Introduction
Plasma spraying processes is commonly used
worldwide. Thus, the numerical simulation of such a
process is currently well developed. The main goal of
those simulations is to study the impact of the working
parameters on particle acceleration and heating, i.e., the
heat transfer and drag force effect. More complex, from
the point of view of experimental study and mathematical
modelling, is the energy transfer process from electric arc
to gas and particles. The complexity of the numerical
simulation is due the strong couplings between many
physical processes involved besides of a large
temperature range (from 300 K to more than 20 kK).
Most of the existing numerical models describe the
plasma spraying torches of conventional linear type. In
the present work a simplified numerical model of a novel,
tandem-type plasma torch for supersonic coating is
presented.
2. Model formulation
The principal goal of this simulation effort was to
visualize the internal flow of plasma and its interaction
with solid particles with the purpose of prediction of
particle velocities and temperatures in the supersonic
plasma jet produced by tandem-type plasma torch. In
order to diminish computational resources and requiring
simulation time, the following simplifications were
assumed:
- The heat generated in the electric arc of plasma torch
due to Joule heating was approximated by a simple
volumetric heat source.
- The problem was treated as being stationary.
The computational domain is shown in Fig. 1. As seen
the model is made of two principal subdomains with axis
7 and 8. The subdomain for arc stabilization consist of
principal gas inlet 1 (with 6 tangential channel for the
vortex formation), mixer chamber 2 and reverse vortex
chamber 3. The subdomain for gas/powder acceleration
consist of powder inlet 4 in axis 7 direction, supersonic
P-II-12-4
nozzle 5 and calculation boundary 6 for ambient pressure
application.
In these fluid domains, the temperature, velocity and
pressure fields are calculated by using the classical
conservation equations for mass, momentum and energy.
Fig. 1. Computation domain.
Thermodynamic and transport properties of nitrogen,
assumed as working gas, were taken from [1] for
temperature range up to 24 kK. Normal conditions were
assumed as initial conditions, i.e. static pressure
131325 Pa and temperature 293 K.
Boundary conditions at model´s walls were assumed as
thermally insulated and heat generation rate of the
volumetric heat source were adjusted taking account of
heat losses by plasma torch refrigeration. Thus the total
enthalpy rate on the plasma torch outlet was 15 kW.
A principal mass flow rate at the inlet 1 (Fig. 1) of 5g/s
was equally divided between 6 symmetrical channels,
which formed strong vortex at axis 8 (Fig. 1) needed for
arc stabilization.
An additional flow for powder
transportation was 1.3 g/s and was assumed as fully
developed at cylindrical inlet 4 and directed along axis 7.
All flows are initially at normal thermodynamics
1
conditions like as conditions on the boundary 6 that
simulated surrounding environment. For diminishing
computational resources the model was divided by plane
9 at which was applied symmetry condition.
Alumina was used as material for powder particles
injected in 4 with initial velocity equal to velocity of
transporting gas. The particles were assumed cylindrical
with outer diameter of 60 µm, density 3970 kg/m3,
specific heat 765 J/(kg K) and thermal conductivity
16 W/(m K). The change in particle diameter and phase
transformation were not taken in account.
Fig. 2. Velocity pattern
3. Numerical Simulation Results
Validation of the presented model was made by
comparing predictions of the particle´s velocity and
temperature, pressure in discharge chamber with those
measured experimentally [2]. Table 1 presents the results
of validation. As it can be seen, for the preliminary study
the developed model can be validated.
Table 1. Validation of the simulation model.
Parameter
Particle diameter, μm
Particle velocity, m/s
Particle temperature, K
Total enthalpy rate, kW
Total enthalpy of jet,
MJ/kg
Jet velocity, m/s
Mach number
Jet temperature, K
* values estimated
Average
from model
60
300-350*
2000-2500*
28,2
4,417
Experimental
data
56±28
410±40
23400±330
27.8*
4,430*
1348
1.30
2830
1224*
1.5*
3780*
The velocity pattern inside the mixer chamber and
outlet nozzle is shown in Fig. 2. As it was expected,
existence of the sink disturbs the singular vortex structure
(formed in the closed electrode cavity) and induces two
counter-flow vortices. This phenomenon, probably, can
diminish the arc stability as well as displace the arc to
inner surface of the mixer chamber.
Other phenomenon observed in the Fig. 1 is the flow
redistribution in the outlet channel. The hot flow with
high dynamic pressure flows in bottom part of the outlet
whereas relatively cold gas is forced to top part of the
channel.
This effect is more clearly observed in Fig. 3, where the
trajectories of the particle carrier gas is shown. The
points of injected gas is situated at the left hemi circle of
inlet (4 from Fig. 1). The lines are spread at channel wall
and forced the top part of the channel.
2
Fig. 3. Flow trajectory of the transportation gas.
Unfortunately, the flow pattern of this kind affect the
particle pattern, as it shown in Fig. 4. Particles, obliquely
incident to hot wall of the channel may cause the wall
damage by abrasion or promote the nozzle clogging.
Nevertheless, this nonlinear trajectory can increase the
particle dwell time inside the plasma torch.
Fig. 4. Particles trajectory inside (top) and outside
(bottom) of the plasma torch.
P-II-12-4
accelerating) is situated on the beginning of the divergent
nozzle where a first inner shock wave occurs (see Fig. 6).
Presence of a shock waves structure is confirmed by
parameters variation in open space.
Fig. 5. Parameters distribution on the vertical diameter of
outlet nozzle.
Nonuniformity of the exit flow is shown in the Fig. 5,
where the temperature, velocity and Mach number
distribution along the vertical of the outlet nozzle are
presented. A temperature difference of 500 K is observed
between top and bottom part of the exit. However, the
other parameters have more uniformity and can be treated
as fully developed supersonic flow (see Fig. 6).
Fig. 7. Parameters of the flow along the axis (7 in Fig. 1).
For the study of the particle behavior in plasma torch, it
was simulated an injection of 10 particles uniformly
distributed on the inlet surface (4, Fig. 1). All particles
have the same dimensions and initial parameters.
A typical variation of the temperature and velocity of the
particles is shown in Fig. 8 and Fig. 9.
Fig. 8.
Particle’s temperature evolution along its
trajectories.
Fig. 6. Diamond-like structure of the external jet. Isolines
of the pressure (top) and Mach number (bottom)
Variation of the principal parameters (temperature,
velocity, Mach number and dynamic pressure) along the
axis of the nozzle channel are shown in Fig. 7. When the
temperature of the gas gradually diminishes, the other
parameters grow. The maximum of velocity is localized
directly on the nozzle exit, but the maximum of the
dynamic pressure (that is responsible for the particle
P-II-12-4
From Fig. 8 it is clear that the particles attain its
maximum temperature in the mixing chamber, before to
be accelerated by flow. On the contrary, the particle
velocity grows continually (quasi linearly) up to nozzle
exit (Fig. 9) where exponential growing with saturation is
observed.
3
Fig. 9. Particle’s velocity evolution along its trajectories.
4. Conclusion
A design of plasma torch of tandem type, developed for
plasma spray, has been modelled and discussed. The
numerical model was validated by comparing with
experimental data. The evolution of the gas flux and
injected particles parameters are studied during the
spraying process. The electric arc heating flux was
represented by volumetric heat source. The proposed
numerical model permits to study thermal and kinetic
characteristics of plasma spray.
5. Acknowledgements
The authors acknowledge FAPESP, CNPq and Capes
for financial support.
6. References
[1] M.I. Boulos,P.Fauchais, E.Pfender Thermal Plasmas
Fundamentals and Applications. Springer (1994)
[2] F.R. Caliari, et al. Proc. ISPC 22 (2015)
4
P-II-12-4