<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Archiving and Interchange DTD v1.0 20120330//EN" "JATS-archivearticle1.dtd">
<article xmlns:xlink="http://www.w3.org/1999/xlink">
  <front>
    <journal-meta />
    <article-meta>
      <title-group>
        <article-title>Computer Modeling of Viscous Fluid Flow Based on the Regularized Lattice Boltzmann Model</article-title>
      </title-group>
      <contrib-group>
        <aff id="aff0">
          <label>0</label>
          <institution>Pryazovskyi State Technical University</institution>
          ,
          <addr-line>Universitetska str., 7, Mariupol, 87555</addr-line>
          ,
          <country country="UA">Ukraine</country>
        </aff>
      </contrib-group>
      <fpage>0000</fpage>
      <lpage>0003</lpage>
      <abstract>
        <p>The problem of the laminar fluid flow simulations with the lattice Boltzmann method at moderate and at large Reynolds numbers up to 20000 is considered. To avoid the fluctuations that appear in the numerical solutions with the LBM the regularization method that based on the median filtration is proposed. The regularization method was tested on the classical problem of the flow over the circular cylinder. The fluid flow modeling over the Nasa 0012 profile at different angles of attack is carried out at the Reynolds number 1000. The flow patterns, the drag and lift coefficients of the NASA profile studied in detail.</p>
      </abstract>
      <kwd-group>
        <kwd>viscous fluid</kwd>
        <kwd>Boltzmann equation</kwd>
        <kwd>regularization</kwd>
        <kwd>circular cylinder</kwd>
        <kwd>Nasa 0012 profile</kwd>
        <kwd>Reynolds numbers</kwd>
      </kwd-group>
    </article-meta>
  </front>
  <body>
    <sec id="sec-1">
      <title>-</title>
      <p>
        Methods of the computational fluid dynamics (CFD) are classified according to the
different approaches in continuous medium describing. These approaches are the
Euler and the Lagrange ones [
        <xref ref-type="bibr" rid="ref1">1</xref>
        ]. Besides them there are hybrid methods that combine
the advantages of the Euler and Lagrange approaches. The first hybrid CFD method
for the non-stationary flow calculating was developed in 1955 in the USA by Harlow
on the basis of the Euler equation. It was called Particles in Cell method (PiC) [
        <xref ref-type="bibr" rid="ref2">2</xref>
        ].
The computational domain divides by a stationary Euler grid into cells in which
particles are located. Fluid describes by a discrete model, as a collection of particles of
fixed mass, moving along the cells of a grid. The method of particles in cells is
convenient to study the dynamics of multicomponent flows and flows with free
boundaries. However, it has two significant disadvantages. The first one follows from the
fluid discrete model: the fluctuations appearing which is the irregularity of the
movement of particles. The second disadvantage is the high memory demands and the low
speed of calculations.
      </p>
      <p>
        In 1965 as a development of the PiC method was proposed the method of large
particles, developed by Davidov U. M. and Bilatserkovskiy O. M [
        <xref ref-type="bibr" rid="ref1">1</xref>
        ]. The
computational domain divides by an Eulerian grid which cells are considered as large
particles. The motion of such large particles is modeled on the basis of the non-stationary
Euler model. Thus, the method occupies an intermediate position between the
Harlow`s PiC method and the classical finite-difference methods.
      </p>
      <p>
        In 1973 was appeared another approach in the fluid flow modeling based on the
cellular automata - Lattice Gas Cellular Automata (LGA) [
        <xref ref-type="bibr" rid="ref3">3</xref>
        ]. In this method particles
can move to the next cells and collide in them according to the laws of gas dynamics
(Boltzmann equation). The disadvantage of this method is the stochastic noise that
appears when densities, impulse and velocities are calculating. As a means of sealing
such noise the integer number of particles replaces by the particles distribution
function. The evolution of the distribution function is described by the Boltzmann kinetic
equation. This approach was called the lattice Boltzmann method (LBM) [
        <xref ref-type="bibr" rid="ref3 ref4 ref5 ref6">3-6</xref>
        ].
      </p>
      <p>
        The LB method becomes widespread. A lot of scientists in Europe and America are
engaged in its development [
        <xref ref-type="bibr" rid="ref10 ref7 ref8 ref9">7-10</xref>
        ]. The popularity of the method is due to its
advantages:
      </p>
      <p>
        • Simplicity in description and programming, since all stages of the method are the
linear equations that solves by the explicit schemes [
        <xref ref-type="bibr" rid="ref3 ref6">3, 6</xref>
        ];
      </p>
      <p>
        • the simplicity in the boundary conditions settings, due to the fact that the method
is based on the kinetic approach in the continuous medium describing [
        <xref ref-type="bibr" rid="ref11 ref6">6, 11</xref>
        ];
• the ability to simulate flows in different difficult domains;
• ease in external forces adding [
        <xref ref-type="bibr" rid="ref12">12</xref>
        ];
• the possibility of using parallel computing technologies [
        <xref ref-type="bibr" rid="ref13 ref14">13, 14</xref>
        ];
• solving a wide class of tasks.
      </p>
      <p>
        However, despite the growing popularity, there are still the following problems:
1.Significant calculation time, which increases with Reynolds number increasing
[
        <xref ref-type="bibr" rid="ref15">15</xref>
        ];
2.Conditional stability [
        <xref ref-type="bibr" rid="ref15 ref16 ref3">3, 15, 16</xref>
        ].
      </p>
      <p>
        These problems make it difficult to obtain numerical solutions for the flows at
moderate Reynolds numbers Re~100 and prevent simulations at large numbers
Re&gt;100. For a partial solution of these problems at small Reynolds numbers Re&lt;100
schemes with several relaxation parameters [
        <xref ref-type="bibr" rid="ref17 ref18">17, 18</xref>
        ] or implicit schemes [
        <xref ref-type="bibr" rid="ref19">19</xref>
        ] are
used. However, the above problems are still not solved to the end and determine the
relevance of the article as well as its scientific and practical significance.
2
      </p>
    </sec>
    <sec id="sec-2">
      <title>Regularized lattice Boltzmann model</title>
      <p>
        According to the lattice Boltzmann method, the computational domain divides into a
grid which cells are treated as large particles [
        <xref ref-type="bibr" rid="ref3 ref4 ref5 ref6">3-6</xref>
        ]. These large particles can move to
one of the neighbors cells and collide in them according to the Boltzmann's kinetic
equation. Such large particles are describing statistically using the particle distribution
function [
        <xref ref-type="bibr" rid="ref20">20</xref>
        ].
      </p>
      <p>
        The mathematical formulation of the problem consists of the Boltzmann equation
(1), in which the particles collision integral is replaced by the Bhatnagar-Gross-Kruck
approximation (2) [
        <xref ref-type="bibr" rid="ref21">21</xref>
        ]:
t f  x f  mF
 f   1  2  f1' f2'  f f dd 2
      </p>
      <p>1 2 
Icoll    1  2  f1' f2'  f1 f2 dd 2 
f eq  f</p>
      <p>
 1 fk r,t  fkeq r, t 
fk rVk ts,ttreamt fkr,t  collision 
Numerical model is a system of the lattice equations (3), obtained by sampling the
mathematical model (1) - (2) with the absence of the external forces F  0 .
where fk r,t  – discrete particle distribution function;
V k – discrete series of particles possible velocities;
t – time step;
τ – dimensionless relaxation parameter;
fkeq r, t  – discrete approximation of the Maxwell-Boltzmann local equilibrium
distribution function.</p>
      <p>
        The justification of using the lattice Boltzmann method to model viscous flows is
given in [
        <xref ref-type="bibr" rid="ref22">22</xref>
        ]. It was shown that using the Chapman-Enskog expansion on the
equations (3) the Navier-Stokes equation and the continuity equation for a uncompression
isothermal fluid (4) can be obtained as well as the formulas that associate
macroscopic and lattice parameters (5) - (8).
      </p>
      <sec id="sec-2-1">
        <title>Kinematic of the liquid viscosity</title>
      </sec>
      <sec id="sec-2-2">
        <title>Lattice speed of the sound Lattice speed</title>
        <p>u  u  u  p  vu  ot ;
t
  u  0  ot   oM p .</p>
        <p>v  cs2t  0.5
cs 
c 
c
3
d
t
where d – cell`s size
(1)
(2)
(3)
(4)
(5)
(6)
(7)
Lattice Mach number</p>
        <p>M p  Umax
cs
(8)
where Umax – maximum velocity value in the domain</p>
        <p>
          The stability of the LBM numerical scheme was investigated analytically in [
          <xref ref-type="bibr" rid="ref16 ref22 ref3 ref5">3, 5,
16, 22</xref>
          ]. The instability of numerical solutions may be caused by one of the following
factors:
        </p>
        <p>
           lattice Mach number increasing. The Boltzmann kinetic equation approximates
the Navier-Stokes equation only at small Mach numbers M p  0.3 [
          <xref ref-type="bibr" rid="ref22">22</xref>
          ].
        </p>
        <p>
           relaxation parameter decreasing. The "safe" value (according to [
          <xref ref-type="bibr" rid="ref5">5</xref>
          ]) is τ=1. The
relaxation parameter decrease τ causes instability: the particles accumulate in some
cells, which results in pulsations in the velocity field (fig. 1);
        </p>
        <p> speed increasing. The numerical model assumes modeling flows only at low
speeds Umax  cs .</p>
        <p>
           Reynolds number increasing. As the Reynolds number increases, the flow
becomes turbulent (according to [
          <xref ref-type="bibr" rid="ref23 ref24">23-24</xref>
          ]).
        </p>
        <p>Macroscopic fluid parameters: density, velocity and pressure are defined as
moments of the distribution function according to the formulas:
 r,t    fk r, t ;ur, t  
8</p>
        <p>
          1 8Vk fk r,t ; pr, t   cs2 r,t .
 r, t  k 0
(9)
ux i, j  med ux i, j  2, ux i, j 1,ux i, j,ux i, j 1, ux i, j  2
ux i, j  ux i, j 1 ux i, j  ux i, j 1
(11)
The advantage of such filter is the nonlinearity with the following properties [
          <xref ref-type="bibr" rid="ref26">26</xref>
          ]:
 after smoothing, the sharp boundaries of the solution areas are preserving;
 non-correlated or weakly corrected obstacles are suppressing;
 fluctuations are reducing.
        </p>
        <p>Smoothing using (10-11) equations schematically depicted in fig. 2.
To verify the proposed method a series of numerical experiments of numerical
solutions regularizing was conducted. Considered the flow pattern at Re=500. Fig. 3a
illustrates velocity diagram got with the LBM algorithm at the moment t=2.5. The fig.
3a shows pulsations, which arise mainly on the boundary of the region. Further
solution is unstable. Figure 3b illustrates velocity diagram of smoothed solution.
a
b</p>
        <p>The regularization method has allowed to smooth the fluctuations and obtain a
stable solution over large time intervals, such as the velocity diagram (fig. 4 a) and the
streamlines (fig. 4 b).</p>
        <p>Fig. 4. Simulation with the LBM with smoothing at Re = 500 at time t = 50 a) velocity diagram
b) streamlines</p>
        <p>The resulting solution was compared with the solution of a similar problem that
was obtained in the Comsol Multiphysics package with the finite element method (fig.
5). Figures 4 and 5 shows the similarity of the obtained patterns and the streamlines.</p>
        <p>Fig. 5. Simulation in the Comsol Multiphysics package at Re = 500 at time t = 50 a) velocity
diagram b) streamlines</p>
        <p>A more detailed comparison was made in vertical sections near the streamlined
body x=0.8, 1.0 (fig. 6). The graphs of the velocity distribution, obtained by the LBM
method with smoothing and the finite elements method (FEM), were constructed.</p>
        <p>Consequently, the solutions of the test problem obtained by the lattice Boltzmann
method with smoothing showed a good correspondence with the results of other
numerical experiments at the Reynolds number Re = 500.</p>
        <p>
          Increase the Reynolds number up to Re = 10,000. The structuring of such flow
disappears and the turbulent is appearing [
          <xref ref-type="bibr" rid="ref23 ref24">23, 24</xref>
          ]. A photo of such flow is shown in fig.
7 [
          <xref ref-type="bibr" rid="ref27">27</xref>
          ].
The conducted numerical experiments give the following results. Fig. 8 illustrates the
simulation results got with the LBM with smoothing.
        </p>
        <p>Fig. 8. Simulation with the LBM with smoothing at Re = 10000 at time t = 100 a) velocity
diagram b) streamlines</p>
        <p>As the Reynolds number increases, the periodicity of vortex formation increases:
vortices descend from the edges of the cylinder at about t = 2. Such a rapid formation
of vortices led to a divergence of numerical solutions by the FEM and LBM methods
at about t = 2 (figures 8 and 9).</p>
        <p>The results of the fluid flow modeling at Reynolds number Re = 20000, are
depicted in fig. 10 with LBM method and fig.11 with the FEM method.</p>
        <p>a
a
b
b</p>
        <p>The results of numerical experiments for the viscous fluid flow modeling with the
lattice Boltzmann method and the finite element method in the Comsol package at
Reynolds numbers 500&lt;Re&lt;20000 have been studied in detail. The comparison
shows the good correspondence of the results, namely the process of the boundary
layer separation and the formation of vortices over the cylinder. The structure of the
streamlines also coincides.
4</p>
      </sec>
    </sec>
    <sec id="sec-3">
      <title>Experiments and results</title>
      <p>The fluid flow modeling over the Nasa 0012 profile at different attack angles was
carried out at the Reynolds number Re=1000. Figure 12 illustrates the flow scheme:
the velocity vector Vin, angle of attack α, drag and lift coefficients F drag , F lift .
The influence of the angle of attack on the flow pattern and on the drag and lift
coefficients was studied. The flow pattern with the zero of the angle of attack is depicted
in fig. 13.
b
Fig. 13. Flow over the Nasa 0012 profile with the absent angle a) velocity diagram b)
streamlines
As can be seen from the fig. 12, the flow around the profile is symmetric. Was
calculated the average values of the drag Cd and lift Cl coefficients at different times up to
t=100. The appropriate graphs are depicted in fig. 14.
The change of the flow pattern by the angle of attack increasing α=10°, 20°, 30° was
considered. The average values of the hydrodynamic coefficients shown in fig. 15.
The angle of attack increasing cause the drag and lift coefficients increasing too as
can be seen from the fig. 15. The numerical simulations with the angles of the Nasa
0012 profile more than 30° are unstable with the regularized lattice Boltzmann model.
Fig. 16 illustrates the velocity diagrams and streamlines of the flow over the profile
with the different angles of attack α=10°, 20°, 30°.</p>
      <p>а
b</p>
      <p>The conducted numerical experiments indicate the growth of the hydrodynamic
coefficients and the formation of vortices over the profile with the angle of attack
increasing.
5</p>
    </sec>
    <sec id="sec-4">
      <title>Conclusion</title>
      <p>The problem of the laminar fluid flow simulations with the lattice Boltzmann
method at moderate and at large Reynolds numbers up to 20000 was considered.
Noticed, that LBM is a new approach in CFD that used the Boltzmann equation to
simulate fluid behavior. Although this method becomes more popular there are some
problems that make fluid flow modeling at moderate and large Reynolds numbers
impossible. They are the significant calculation time, which increases with Reynolds
number increasing and conditional stability. The last one may causes fluctuations which
are the irregularity of the movement of particles. To avoid the fluctuations the
regularization method that based on the median filtration was proposed.</p>
      <p>The regularization method was tested on the classical problem of the flow over the
circular cylinder. The comparison of the numerical experiments with the lattice
Boltzmann method and the finite element method in Comsol Multiphysics package
showed the good correspondence of the results, namely the process of the boundary
layer separation and the formation of vortices over the cylinder. The structure of the
streamlines also coincides. So, the regularization method has allowed to smooth the
fluctuations and to obtain a stable solution. This allows to use the LBM for studying
the laminar patterns at large Reynolds numbers.</p>
      <p>The fluid flow modeling over the Nasa 0012 profile at different angles of attack
was carried out at the Reynolds number Re=1000. Declared the vortex formation at
the angles more than 7°. Size of the vortex is increasing with the angle increasing as
well as the hydrodynamics coefficients. The numerical simulations with the angles of
the Nasa 0012 profile more than 30° are unstable even with the regularized lattice
Boltzmann model so demands more detail study.</p>
    </sec>
  </body>
  <back>
    <ref-list>
      <ref id="ref1">
        <mixed-citation>
          1.
          <string-name>
            <surname>Bilatserkovskiy</surname>
            ,
            <given-names>O.M.:</given-names>
          </string-name>
          <article-title>Numerical modeling in continuum mechanics</article-title>
          .
          <source>Fizmatlit</source>
          , Moscow (
          <year>1994</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref2">
        <mixed-citation>
          2.
          <string-name>
            <surname>Harlow</surname>
            ,
            <given-names>F.X.</given-names>
          </string-name>
          :
          <article-title>Numerical method of particles in cells for problems of hydrodynamics</article-title>
          . Mir, Moscow (
          <year>1967</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref3">
        <mixed-citation>
          3.
          <string-name>
            <surname>Wolf-Gladrow</surname>
            ,
            <given-names>D.</given-names>
          </string-name>
          :
          <string-name>
            <surname>Lattice-Gas Cellular</surname>
          </string-name>
          Automata and
          <string-name>
            <surname>Lattice Boltzmann Models - An Introduction</surname>
          </string-name>
          .
          <article-title>Alfred Wegener Institute for Polar and Marine</article-title>
          ,
          <string-name>
            <surname>Bremerhaven</surname>
          </string-name>
          (
          <year>2005</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref4">
        <mixed-citation>
          4.
          <string-name>
            <surname>Succi</surname>
            ,
            <given-names>S.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Benzi</surname>
            <given-names>R.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Higuera</surname>
            ,
            <given-names>F.</given-names>
          </string-name>
          :
          <article-title>The Lattice Boltzmann Equation: A New Tool for Computational Fluid-Dynamics</article-title>
          . Physica D: Nonlinear Phenomena.
          <volume>47</volume>
          ,
          <fpage>219</fpage>
          -
          <lpage>230</lpage>
          (
          <year>1991</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref5">
        <mixed-citation>
          5.
          <string-name>
            <surname>Sucop</surname>
            ,
            <given-names>M.</given-names>
          </string-name>
          :
          <source>Lattice Boltzmann Modeling. An Introduction for Geoscientists and Engineers</source>
          . Springer, Miami (
          <year>2006</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref6">
        <mixed-citation>
          6.
          <string-name>
            <surname>Succi</surname>
            ,
            <given-names>S.:</given-names>
          </string-name>
          <article-title>The Lattice Boltzmann Equation for Fluid Dinamics and Beyond</article-title>
          . Univercity Press, Oxford (
          <year>2001</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref7">
        <mixed-citation>
          7.
          <string-name>
            <surname>Grazyna</surname>
            ,
            <given-names>K.</given-names>
          </string-name>
          :
          <article-title>The numerical solution of the transient heat conduction problem using the lattice Boltzmann method</article-title>
          .
          <source>Scientific Research of the Institute of Mathematic and Computer Science</source>
          .
          <volume>11</volume>
          ,
          <fpage>23</fpage>
          -
          <lpage>30</lpage>
          (
          <year>2006</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref8">
        <mixed-citation>
          8.
          <string-name>
            <surname>Leclaire</surname>
            ,
            <given-names>S.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Pellerin</surname>
            ,
            <given-names>N.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Reggio</surname>
            ,
            <given-names>M.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Trepanier</surname>
          </string-name>
          , J.-Y.:
          <article-title>Enhanced equilibrium distribution functions for simulating immiscible multiphase flows with variable density ratios in a class of lattice Boltzmann models</article-title>
          .
          <source>International Journal of Multiphase Flow</source>
          .
          <volume>57</volume>
          ,
          <fpage>159</fpage>
          -
          <lpage>168</lpage>
          (
          <year>2013</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref9">
        <mixed-citation>
          9.
          <string-name>
            <surname>Favier</surname>
            ,
            <given-names>J.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Revell</surname>
            ,
            <given-names>A.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Pinelli</surname>
            ,
            <given-names>A.</given-names>
          </string-name>
          :
          <article-title>A lattice Boltzmann-immersed boundary method to simulate the fluid interaction with moving and slender flexible objects</article-title>
          .
          <source>Journal of Computational Physics</source>
          .
          <volume>261</volume>
          ,
          <fpage>145</fpage>
          -
          <lpage>161</lpage>
          (
          <year>2014</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref10">
        <mixed-citation>
          10.
          <string-name>
            <surname>Anderl</surname>
            ,
            <given-names>D.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Bogner</surname>
            ,
            <given-names>S.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Rauh</surname>
            ,
            <given-names>C.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Rude</surname>
            ,
            <given-names>U.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Delgado</surname>
            ,
            <given-names>A.</given-names>
          </string-name>
          :
          <article-title>Free surface lattice Boltzmann with enhanced bubble model</article-title>
          .
          <source>Computers and Mathematics with Applications</source>
          .
          <volume>67</volume>
          (
          <issue>2</issue>
          ),
          <fpage>331</fpage>
          -
          <lpage>339</lpage>
          (
          <year>2014</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref11">
        <mixed-citation>
          11.
          <string-name>
            <surname>Latt</surname>
            ,
            <given-names>J.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Chopard</surname>
            ,
            <given-names>B.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Malaspinas</surname>
            ,
            <given-names>O.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Deville</surname>
            ,
            <given-names>M.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Michler</surname>
            ,
            <given-names>A.</given-names>
          </string-name>
          :
          <article-title>Straight velocity boundaries in the lattice Boltzmann method</article-title>
          .
          <source>Physical Review</source>
          .
          <volume>77</volume>
          ,
          <fpage>1</fpage>
          -
          <lpage>17</lpage>
          (
          <year>2008</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref12">
        <mixed-citation>
          12.
          <string-name>
            <surname>Kupershtokh</surname>
            ,
            <given-names>A.L.</given-names>
          </string-name>
          :
          <article-title>Accounting for the action of volume forces in the lattice Boltzmann equations</article-title>
          . Bulletin of the Novosibirsk State University. Series “Mathematics, mechanics, computer science”.
          <volume>6</volume>
          (
          <issue>2</issue>
          ),
          <fpage>75</fpage>
          -
          <lpage>96</lpage>
          (
          <year>2004</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref13">
        <mixed-citation>
          13.
          <string-name>
            <surname>Kupershtokh</surname>
            <given-names>A.L.:</given-names>
          </string-name>
          <article-title>3D modeling with the lattice Boltzmann Method on GPU. Numerical methods and programming</article-title>
          .
          <volume>13</volume>
          ,
          <fpage>130</fpage>
          -
          <lpage>138</lpage>
          (
          <year>2012</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref14">
        <mixed-citation>
          14.
          <string-name>
            <surname>Agrawal</surname>
            ,
            <given-names>T.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Lin</surname>
            ,
            <given-names>C.</given-names>
          </string-name>
          :
          <article-title>Implementation of an incompressible lattice Boltzmann model on GPU to simulate Poiseuille flow</article-title>
          .
          <source>Proceeding of the Forties National Conference on Fluid Mechanics and Fluid Power</source>
          .
          <volume>40</volume>
          ,
          <fpage>991</fpage>
          -
          <lpage>995</lpage>
          (
          <year>2013</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref15">
        <mixed-citation>
          15.
          <string-name>
            <surname>Bulanchuk</surname>
            ,
            <given-names>G.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Bulanchuk</surname>
            ,
            <given-names>O.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Ostapenko</surname>
            ,
            <given-names>A.</given-names>
          </string-name>
          :
          <article-title>Stability investigation of the twodimensional nine-vectors model of the lattice Boltzmann method for fluid flows in a square cavity</article-title>
          . Bulletin of Kharkov national University named after V.N.Karazin. Series: Mathematical modeling. Information technology.
          <source>Automated control system. 28</source>
          ,
          <fpage>113</fpage>
          -
          <lpage>125</lpage>
          (
          <year>2015</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref16">
        <mixed-citation>
          16.
          <string-name>
            <surname>Aslan</surname>
          </string-name>
          , E.:
          <article-title>Investigation of the Lattice Boltzmann SRT and MRT Stability for Lid Driven Cavity Flow</article-title>
          .
          <source>International Journal of Materials, Mechanics and Manufacturing</source>
          .
          <volume>2</volume>
          (
          <issue>4</issue>
          ),
          <fpage>317</fpage>
          -
          <lpage>324</lpage>
          (
          <year>2014</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref17">
        <mixed-citation>
          17.
          <string-name>
            <surname>Rettinger</surname>
            ,
            <given-names>C.</given-names>
          </string-name>
          :
          <article-title>Fluid Flow Simulation using the Lattice Boltzmann Method with multiple relaxation times</article-title>
          .
          <string-name>
            <surname>Friedrich-Alexander Universuty</surname>
          </string-name>
          Erlander-Nuremberg,
          <source>Erlander</source>
          (
          <year>2013</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref18">
        <mixed-citation>
          18.
          <string-name>
            <surname>Zaharov</surname>
            ,
            <given-names>A.M.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Senin</surname>
            ,
            <given-names>D.S.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Grachev</surname>
            ,
            <given-names>E.A.</given-names>
          </string-name>
          :
          <article-title>Lattice Boltzmann modeling with multirelaxationals times</article-title>
          .
          <source>Numerical methods and programming</source>
          .
          <volume>15</volume>
          ,
          <fpage>644</fpage>
          -
          <lpage>657</lpage>
          (
          <year>2014</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref19">
        <mixed-citation>
          19.
          <string-name>
            <surname>Krivovichev</surname>
            ,
            <given-names>G.V.</given-names>
          </string-name>
          :
          <article-title>On the computation of viscous fluid flows by the lattice Boltzmann method</article-title>
          .
          <source>Computer Reserch and Modeling</source>
          .
          <volume>5</volume>
          (
          <issue>2</issue>
          ),
          <fpage>165</fpage>
          -
          <lpage>178</lpage>
          (
          <year>2013</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref20">
        <mixed-citation>
          20.
          <string-name>
            <surname>Liapin</surname>
            ,
            <given-names>I.I.</given-names>
          </string-name>
          :
          <article-title>Introduction to the kinetic equations theory</article-title>
          .
          <source>UGTU</source>
          ,
          <string-name>
            <surname>Ekaterinburg</surname>
          </string-name>
          (
          <year>2003</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref21">
        <mixed-citation>
          21.
          <string-name>
            <surname>Mussa</surname>
            ,
            <given-names>M.</given-names>
          </string-name>
          :
          <article-title>Numerical Simulation of Lid-Driven Cavity Flow Using the Lattice Boltzmann Method</article-title>
          .
          <source>Applied Mathematics</source>
          .
          <volume>13</volume>
          ,
          <fpage>236</fpage>
          -
          <lpage>240</lpage>
          (
          <year>2008</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref22">
        <mixed-citation>
          22.
          <string-name>
            <surname>He</surname>
            ,
            <given-names>X.</given-names>
          </string-name>
          :
          <article-title>Lattice Boltzmann Model for the Incompressible Navier - Stokes Equation</article-title>
          .
          <source>Journal of statistical physics</source>
          .
          <volume>88</volume>
          ,
          <fpage>927</fpage>
          -
          <lpage>944</lpage>
          (
          <year>1997</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref23">
        <mixed-citation>
          23.
          <string-name>
            <surname>Schlichting</surname>
          </string-name>
          , H.:
          <article-title>Boundary layer theory</article-title>
          .
          <source>Nauka</source>
          ,
          <string-name>
            <surname>Moskow</surname>
          </string-name>
          (
          <year>1974</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref24">
        <mixed-citation>
          24.
          <string-name>
            <surname>Bilatserkovskiy</surname>
            ,
            <given-names>O.M.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Bilatserkovskiy</surname>
            ,
            <given-names>S.O.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Gushin</surname>
            ,
            <given-names>V.A.</given-names>
          </string-name>
          :
          <article-title>Numerical modeling of the unsteady periodic flow of the viscous fluid past a cylinder</article-title>
          .
          <source>Computational mathematics and mathematical physics</source>
          .
          <volume>24</volume>
          (
          <issue>8</issue>
          ),
          <fpage>1207</fpage>
          -
          <lpage>1216</lpage>
          (
          <year>1984</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref25">
        <mixed-citation>
          25.
          <string-name>
            <surname>Paskonov</surname>
            ,
            <given-names>V.M.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Polezhaev</surname>
            ,
            <given-names>V.I.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Chudov</surname>
            ,
            <given-names>L.A.</given-names>
          </string-name>
          :
          <article-title>Numerical simulation of heat and mass transfer process</article-title>
          .
          <source>Nauka</source>
          , Moscow (
          <year>1984</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref26">
        <mixed-citation>
          26.
          <string-name>
            <surname>Micek</surname>
            ,
            <given-names>J.</given-names>
          </string-name>
          ,
          <string-name>
            <surname>Kapitulok</surname>
            ,
            <given-names>J.: Median</given-names>
          </string-name>
          <string-name>
            <surname>Filter</surname>
          </string-name>
          .
          <source>Journal of Information, Control and Management Systems</source>
          .
          <volume>1</volume>
          (
          <issue>2</issue>
          ),
          <fpage>51</fpage>
          -
          <lpage>56</lpage>
          (
          <year>2003</year>
          )
        </mixed-citation>
      </ref>
      <ref id="ref27">
        <mixed-citation>
          27.
          <string-name>
            <surname>Van Dyke</surname>
            ,
            <given-names>M.:</given-names>
          </string-name>
          <article-title>An Album of Fluid Mechanics</article-title>
          . The Parabolic Press, Stanford (
          <year>1982</year>
          )
        </mixed-citation>
      </ref>
    </ref-list>
  </back>
</article>