Numerical Investigation of Nucleated Bubbles around a Heated
Square Obstacle using Thermal Hybrid Lattice Boltzmann Method
Salaheddine Channouf
1
a
, Mohammed Jami
1
b
and Ahmed Mezrhab
1
c
1
Mechanical and energetic laboratory, First Mohamed university, faculty of sciences, Oujda, Morocco
Keywords: Boiling phenomenon, nucleated bubbles, lattice Boltzmann method, 2-order Runge-Kutta finite difference
scheme, heat transfer, phase-change.
Abstract: The phenomenon of boiling has become one of the challenges of the LB community in recent years, as it is
frequently used in industry and deals with a complex natural phenomenon. For this purpose, we have devel-
oped a computational code that allows us to simulate this phenomenon around a square obstacle of variable
length in a rectangular cavity. We aim to study the behaviour of nucleated bubbles which appear around this
obstacle due to its heat exchange with the liquid. The pseudopotential multi-relaxation time lattice Boltzmann
method MRT-LBM proposed by (Li, 2013) is used as a solver of the fluid flows combined with the 2-order
Runge-Kutta finite difference scheme for modelling the phase change from a liquid to a vapour phase
(Zhao,2018) and the temperature field.
1 INTRODUCTION
Lattice Boltzmann method (LBM) has become a pre-
cision and efficient tool for simulating hydro-thermo-
dynamic phenomena because its simplicity of imple-
mentation and insertion of boundaries. This method
contains 2 main steps represented by a distribution
function denoted ๐‘“
๎ฏœ
๏ˆบ
๐’™,๐‘ก
๏ˆป
which presents a set of par-
ticles having precise velocities in a regular mesh
called lattice Boltzmann and follows their movements
in phase space. These two steps describe the collision
and streaming of particles over time on the lattice
(Meng, 2014) and allow to have a simple algorithm to
calculate the macroscopic variables as density, veloc-
ity components, temperature, etc. In this paper, the
collision and streaming steps are represented in mo-
mentum space, this space aims to shift each macro-
scopic quantity towards the equilibrium by the multi-
relaxation time (MRT) collision operator model
(Krรผger, 2017) as opposed to the simple relaxation
time operator known as the BGK operator model (An-
dries, 2002) which is used to relax the distribution
function around the equilibrium distribution function.
The MRT model is used due to its reliability and high
stability. In our work, the pseudopotential MRT-
a
https://orcid.org/0000-0002-8538-5629
b
https://orcid.org/0000-0002-5356-4729
c
https://orcid.org/0000-0001-6013-7496
LBM approach (Shan,2014) is adopted. This is the fa-
mous model of the LB community for the simulation
of a multiphase flow thanks to its simplicity to obtain
the separation of the phases without the need to track
at each step time the interface between the phases.
This model allows to successfully simulate several
numerical problems in multiphases such as wettabil-
ity, condensation, boiling, collapsing and rising bub-
bles, etc.
In the present work, this model is combined with
the celebrate Runge-Kutta finite difference scheme
(Zhao, 2018, Zheng, 2018) to simulate the thermal be-
havior of liquid chosen at saturated temperature dur-
ing phase change around a heated obstacle with a var-
iable characteristic length ๐ท.
2 METHODOLODY
The hybrid model proposed in this work is described
by the following equation
๐‘“
๎ฏœ
โˆ—
๏ˆบ
๐’™,๐‘ก
๏ˆป
๎ตŒ๐‘“
๎ฏœ
๏ˆบ
๐’™,๐‘ก
๏ˆป
๎ต†๐‘ด
๎ฌฟ๐Ÿ
๐œฆ๐‘ด๎ตซ๐‘“
๎ฏœ
๏ˆบ
๐’™,๐‘ก
๏ˆป
๎ต†๐‘“
๎ฏœ
๎ฏ˜๎ฏค
๏ˆบ
๐’™,๐‘ก
๏ˆป
๎ตฏ๎ต…๐›ฟ๐‘ก๐‘ด๐‘ญ
(1)
Channouf, S., Jami, M. and Mezrhab, A.
Numerical Investigation of Nucleated Bubbles around a Heated Square Obstacle using Thermal Hybrid Lattice Boltzmann Method.
DOI: 10.5220/0010730400003101
In Proceedings of the 2nd International Conference on Big Data, Modelling and Machine Learning (BML 2021), pages 171-174
ISBN: 978-989-758-559-3
Copyright
c
๎€ 2022 by SCITEPRESS โ€“ Science and Technology Publications, Lda. All rights reserved
171
The term in the left hand represents the post-collision
distribution function which can be calculated with the
previous distribution function ๐‘“
๎ฏœ
๏ˆบ
๐’™,๐‘ก
๏ˆป
by subtracting
its disturbance around the equilibrium distribution
function relaxing by nine relaxation times given by
the relaxation matrix ๐œฆ and by adding the forcing
term ๐‘ญ. The parameters ๐‘ด and ๐‘ด
๎ฌฟ๐Ÿ
describe the
transformation matrix and its inverse from the phase
space to the momentum space.
The 2-order Runge-Kutta finite difference scheme
is used as a solver of energy equation to simulate the
liquid-vapor phase-change transformation as (Zhao,
2018, Zheng, 2018)
๐‘‘๐‘‡
๐‘‘๐‘ก
๎ตŒ๎ต†๐ฏ.๐›๐‘‡๎ต†
๐‘‡
๐œŒ๐‘
๎ฏฉ
๎ตฌ
๐œ•๐‘ƒ
๎ฎพ๎ฏˆ๎ฏŒ
๐œ•๐‘ก
๎ตฐ
๎ฐ˜
๐›.๐ฏ ๎ต…
1
๐œŒ๐‘
๎ฏฉ
๐›.๏ˆบ๐œ†๐›๐‘‡๏ˆป
(2)
๐‘ƒ
๎ฎพ๎ฏˆ๎ฏŒ
is the pressure according to equation of state
proposed by Peng-Robinson (Yuan, 2006) which is
calculated from the pseudopotential function, ๐œ† is the
thermal conductivity and ๐‘
๎ฏฉ
is the specific heat coef-
ficient ๐ฏ is the macroscopic velocity of flows and ๐œŒ is
the macroscopic density. These macroscopic parame-
ters can be calculated as follows
๐œŒ๎ตŒ๎ท
๐‘“
๎ฏœ
๎ฏœ
; ๐œŒ๐ฎ๎ตŒ๎ท
๐‘“
๎ฏœ
๐‘
๎ฏœ
๎ฏœ
(3)
๐’„
๐’Š
is the lattice speed in i
th
direction.
3 VALIDATION MODEL
Figures 1 and 2 show the validation of our code using
the work of Zheng et al. (Zheng, 2018) to check its
accuracy and its efficiency. For this purpose, a ther-
mohydrodynamic effect is investigated to simulate
the evaporation behavior of a heated liquid droplet
placed in the center of a cavity and surrounded by a
saturated vapor by implementing the periodic bound-
ary for all cavity walls in which the thermal conduc-
tivity is set to be constant ๐œ†๎ตŒ1 which allows to ne-
glect the convective effect.
From Figure 2, the results show a reducing of ini-
tial radius of droplet over time ฮดt, this reduction is de-
scribed by the square rate of change of the radius with
a negative slope line. However, the vector field
shown in Figure 1 shows the direction of the loss of
liquid mass caused by the transformation of the liquid
into vapor. The same remarks are made from the ref-
erence work. Therefore, our numerical results show a
strong corresponding with those reported by Zheng et
al. (Zheng, 2018).
๏ˆบ
๐‘Ž
๏ˆป
Ou
r
resul
t
๏ˆบ
๐‘
๏ˆป
Zheng e
t
al. (Zheng, 2018)
Figure 1: Comparative results with reference (Zheng, 2018)
Figure 2: Square rate change of radius over time ๐›ฟ๐‘ก in com-
parison with reference (Zheng, 2018).
4 RESULTS AND DISCUSSIONS
4.1 Presentation of the Computational
Problem
Figure 3 shows the computational domain of a satu-
rated liquid confined in a rectangular cavity of height
๐ป and length ๐ฟ in which a heated square obstacle at
temperature ๐‘‡
๎ฏ›
๎ตŒ 1.16๐‘‡
๎ฏ–
๎ณ
is placed halfway between
the vertical walls at a distance of ๐ป/6 from the bot-
tom wall. The obstacle has a height h and a variable
characteristic length D. In our simulation, the critical
temperature is set to be ๐‘‡
๎ฏ–
๎ณ
๎ตŒ 0.07292. For the fluid
flow, the periodic boundaries are selected for both
right and left vertical walls. The bounce-back bound-
aries are adopted for the bottom wall, while the out-
flow boundary is applied on the top wall. For the ob-
stacle faces, the no-slip boundaries are selected. Their
density value is adjusted to be in a non-wetting phase.
All computational domain is set at saturated temper-
ature ๐‘‡
๎ฏฆ๎ฏ”๎ฏง
.
(Zheng, 2018)
BML 2021 - INTERNATIONAL CONFERENCE ON BIG DATA, MODELLING AND MACHINE LEARNING (BMLโ€™21)
172
Figure 3: Illustration of the computational problem.
4.2 Formation Pattern of Nucleated
Bubble around the Obstacle
Figure 4 illustrated the density behavior of both liquid
and vapor during phase change caused by the heated
obstacle. It can be seen that the nucleated bubble
formed around the obstacle grows progressively over
time until its volume becomes important, then, it
moves upward due to the buoyant force.
Figure 4: Steps of the nucleated bubble formation around
the heated obstacle.
Figure 5 describes the velocity at three different posi-
tions near to the obstacle at time ๐‘ก
โˆ—
๎ตŒ 3.66 corre-
sponding to the final step of formation of nucleated
bubble. At ๐‘ฆ๎ตŒ๐ป/6, the velocity is zero on the ob-
stacle and tends to increase in the surrounding area.
However, it remains weak due to friction with the
solid. At the obstacle-liquid interface ๏ˆบ๐‘ฆ ๎ตŒ ๐ป/6 ๎ต…
โ„Ž/2๏ˆป, the velocity increases but it keeps the same
form. finally, at ๐‘ฆ ๎ตŒ ๐ป/6 ๎ต… โ„Ž, an intense peak due
to the buoyancy effect during nucleation is noted.
Figure 5: The velocity component in y-direction at different
positions.
The heat flux exchanged between the obstacle and the
liquid is illustrated in Figure 6 at three different posi-
tions and at ๐‘ก
โˆ—
๎ตŒ 3.66. This heat leads to the for-
mation of the nucleated bubbles. For ๐‘ฆ ๎ตŒ ๐ป/6, the
exchange flux is greater at the vertical solid-liquid in-
terfaces then it decreases to be zero when moving
away from these interfaces. At position ๐‘ฆ ๎ตŒ ๐ป/6 ๎ต…
โ„Ž/2, the exchanged flux behaves in the same way as
in the first case but it is reduced. For the last position,
a decrease in flux at the center of the curve is noted.
This is due to the flux transferred to the bubbles.
Figure 6: The exchange heat flux between liquid and heated
obstacle at different positions.
4.3 Effect of Variation of
Characteristic Length D on
Nucleation
In this part, we adjusting the characteristic length ๐ท to
study its impact on the nucleated bubble for-mation.
The same parameters of fluid flows and temperature
field are chosen as the previous part.
๐‘‡
๎ฏ›
Liquid at ๐‘‡๎ตŒ๐‘‡
๎ฏฆ๎ฏ”๎ฏง
๐ท
โ„Ž
๐ป
๐ฟ
๐ป/6
g
x
y
๐‘ก
โˆ—
๎ตŒ 0.7 ๐‘ก
โˆ—
๎ตŒ 1.3
๐‘ก
โˆ—
๎ตŒ 2.45 ๐‘ก
โˆ—
๎ตŒ4
Numerical Investigation of Nucleated Bubbles around a Heated Square Obstacle using Thermal Hybrid Lattice Boltzmann Method
173
v
(
x
)
Figure 7 gives the behavior of the velocity com-
ponent in the y direction for three different values of
D and for the position ๐‘ฆ๎ตŒ ๐ป/6๎ต…โ„Ž. We notice that
the velocity decreases with the increase of D. Indeed,
when D increases, the number of nucleated bubbles
increases then they merge to give a large bubble
which is more slowed down by the liquid.
Figure 7: The velocity component in y-direction by varying
the length D of the obstacle at y = H/6+h.
Figure 8 shows the heat flux exchanged between the
upper face of the solid and the liquid to give more de-
tails on the formation of nucleated bubbles around the
obstacle. For this purpose, the heat flux exchanged is
studied for three different values of the characteristic
length D. The curves show that the increase in the
characteristic length leads to an increase in heat flux
at the right and left sides of the obstacle. Moreover,
in the middle of the cavity in the x direction, the heat
flux form changes with D and indicates that the heat
transferred to the liquid is important for small bubbles
than for large bubbles.
Figure 8: The exchange heat flux between liquid and heated
obstacle by varying the length D of the obstacle at y =
H/6+h.
5 CONCLUSIONS
In the light of this work, we can draw the following
conclusions:
- When the characteristic length D increases, the heat
exchange between the solid and the liquid increases
at the vertical faces of the obstacle. However, this flux
decreases in the area of nucleated bubbles.
- The large bubble formed by the nucleated bubbles
is all the more slowed down as its volume is im-
portant.
These results allow us to make the right choice of
the characteristic length D that gives a compatible
boiling shape. In future work, 3D will investigate and
validate that with experiment by applying this method
to the study of phase change materials.
REFERENCES
Li, Q., Luo, K. H., & Li, X. J. 2013. Lattice Boltzmann
modeling of multiphase flows at large density ratio with
an improved pseudopotential model. Physical Review
E, 87(5), 053301.
Zhao, W., Zhang, Y., Xu, B., Shang, W., & Jiang, S. 2018.
Pseudopotential multiple-relaxation-time lattice Boltz-
mann simulation of vapor condensation on vertical sub-
cooled walls. arXiv preprint arXiv:1808.04973.
Meng, J., & Zhang, Y. 2014. Diffuse reflection boundary
condition for high-order lattice Boltzmann models with
streamingโ€“collision mechanism. Journal of Computa-
tional Physics, 258, 601-612.
Krรผger, T., Kusumaatmaja, H., Kuzmin, A., Shardt, O.,
Silva, G., & Viggen, E. M. 2017. MRT and TRT Colli-
sion Operators. In the Lattice Boltzmann Method (pp.
407-431). Springer, Cham.
Andries, P., Aoki, K., & Perthame, B. 2002. A consistent
BGK-type model for gas mixtures. Journal of Statistical
Physics, 106(5), 993-1018.
Shan, X., & Chen, H. 1993. Lattice Boltzmann model for
simulating flows with multiple phases and components.
Physical review E, 47(3), 1815.
Zheng, S., Eimann, F., Fieback, T., Xie, G., & Gross, U.,
2018. Numerical investigation of convective dropwise
condensation flow by a hybrid thermal lattice Boltz-
mann method. Applied Thermal Engineering, 145, 590-
602.
Yuan, P., & Schaefer, L. (2006). Equations of state in a lat-
tice Boltzmann model. Physics of Fluids, 18(4),
042101.
Chen, L., Kang, Q., Mu, Y., He, Y. L., & Tao, W. Q. (2014).
A critical review of the pseudopotential multiphase lat-
tice Boltzmann model: Methods and applications. Inter-
national journal of heat and mass transfer, 76, 210-236.
BML 2021 - INTERNATIONAL CONFERENCE ON BIG DATA, MODELLING AND MACHINE LEARNING (BMLโ€™21)
174