You are on page 1of 9

It follows from analysis of the expressions (4.4)--(4.

7) that with increasing con-


centration of the packets in the bed the value of the parameter 6 c at which a circula-
tion region is formed inside or outside a packet increases. For both rising and sinking
packets, the cloud radius a c decreases with increasing Pd" For Pd = 0, the relations
(4.3)--(4.7) are identical to the corresponding expressions for an isolated packet [i].
Figure 3 shows the dependence of the ratio of the radius of the circulation region
formed for a rising system of bubbles to the radius of the region formed for an isolated
rising bubble on the concentration Pb of bubbles in the bed. Curves 1 and 2 correspond
to sets of parameters of the bed and bubble sizes for which the ratio of the rising ve-
locity U b of an isolated bubble to the velocity v 0 of the fluid phase in the gaps between
the particles in the homogeneous flow outside the bubble takes the values 2 and 3.

LITERATURE CITED
i. N. N. Bobkov and Yu. P. Gupalo, "Packet mechanism of mixing in a fluidized bed,"
Izv. Akad. Nauk SSSR, Mekh. Zhidk. Gaza, No. 5, 73 (1983).
2. N. N. Bobkov, L. M. Ga!ieva, and Yu. P. Gupalo, "Motion of inhomogeneities of a
developed fluidized bed at small Reynolds numbers," Izv. Akad. Nauk SSSR, Mekh. Zhidk.
Gaza, No. 4, 57 (1984).
3. M. Horio, K. Morishita, O. Tachibana, and N. Murata, "Solid distribution and movement
in circulating fluidized beds," in: Proc. Second Int. Conf. on Circulating Fluid.
Bed., Compiegne, France, 1988, Pergamon Press, Oxford (1988).
4. S. J. Happel and H. Brenner, Low Reynolds Number Hydrodynamics, Prentice-Hall, Engle-
wood Cliffs (1965).
5. V. A. Borodulya and Yu. P. Gupalo, Mathematical Models of Chemical Reactors with
Fluidized Bed [in Russian], Nauka i Tekhnika, Minsk (1976).
6. S. Ergun, "Fluid flow through packed columns," Chem. Eng. Prog., 48, 89 (1952).
7. L. M. Milne-Thomson, Theoretical Hydrodynamics, Macmillan, New York (1950).
8. J. F. Davidson and D. Harrison, Fluidized Particles, Cambridge (1963).
9. G. Marrucci, "Rising velocity of a swarm of spherical bubbles," Ind. Eng. Chem. Fundam.,
i, 224 (1965).
I0. Yu. G. Chesnokov and I. O. Protod'yakonov, "Motion of bubbles in a fluidized bed
under constrained conditions,': Zh~ Prikl. Khim., No. 4, 926 (1981).

MODELING OF THE PARTICLE DYNAMICS IN THE WALL REGION OF


TURBULENT GAS DISPERSION FLOW

I. N. Gusev and L. I. Zaichik UDC 532.529.532.517.4

INTRODUCTION
To determine the extent to which the particles are entrained in the pulsations of
the turbulent carrier flow, the correlation moments containing the velocity pulsations
and concentrations of the solid phase must be related to the properties of the gaseous
phase. In most studies, for example, [1--3], the corresponding expressions are obtained
in the framework of a locally homogeneous approximation - the second moments of the
pulsations of the variables of the solid phase are expressed directly in terms of the
Reynolds stresses for the coefficient of turbulent viscosity of the gas at the same point
Of space. The locally homogeneous approximation is valid for relatively fine particles
(whose relaxation time is less than the characteristic time scale of the turbulence)
provided there are no large velocity gradients in the flow. However, in the wall region
of the flow, the approach based on description of the pulsation properties of the par-
ticles in the locally homogeneous representation may lead to serious errors. This is
due to the circumstance that, on the one hand, the flow in the wall region is character-
ized by large gradients of the averaged velocity and of the turbulent energy of the gaseous
phase, and, on the other, the particles are relatively more substantial (inertial) due
to the decrease in the turbulence time scale near the wall.

Moscow. Translated from Izvestiya Akademii Nauk SSSR, Mekhanika Zhidkosti i Gaza,
Noo I, pp. 50-60, January-February, 1991. Original article submitted January 18, 1990.

0015-4628/91/2601-0041512.50 9 1991 Plenum Publishing Corporation 41


The nonlocal effects due to the inertia of the particles and the inhomogeneity of
the carrier flow can be taken into account fairly consistently by solving a system of
equations for the correlation moments of the solid-phase velocity analogous to the chain
of Fridman--Keller equations in the theory of single-phase turbulent flows [4]. The
first attempts to use the transport equations of the second moments of the particle ve-
locity pulsations to calculate the motion of an admixture in a jet or tube were made
in [5, 6]. To model the particle dynamics in the wall region in this paper, we use the
equations for the second and third moments of the solid-phase velocity; a chain of equa-
tions for the moments is obtained from the kinetic equation for the probability distribu-
tion density of the particles in a turbulent flow constructed in [7].
i. We consider the motion of solid spherical particles subject to the forces of
viscous interaction with a turbulent gas flow, an external body force (for example,
gravity), and a random force that gives rise to Brownian diffusion. The mass (and even
more the volume) concentration of the particles is assumed to be low, and therefore the
back reaction of the admixture on the properties of the carrier flow and the interaction
of the particles with one another as a result of collisions can be ignored. Assuming
a Gaussian random field of the velocities of the turbulent carrier flow and 5-correlated
Brownian force, we obtain from the Lagrangian equation of motion of an isolated particle
an equation for the particle probability density P in the phase space of the coordinates
x k and velocities v k [7]:

OP +v~ OP + 0 ( U~-v~ + Fk ) P = g<u~, u~, > 02P + ~---<ui'u.~'> OzP + D O~P (1.1)
Ot Oxk Ovk ~ T " Ox~ Ovk T Ov~ Ovk x 2 Ov~ Ovk

where U k and u~ are the averaged and pulsation components of the gas velocity, F k is the
external body force, 9 is the dynamical relaxation time of a particle, D is the coef-
ficient of the Brownian diffusion, and <uiu~> are the second single-point, single-time
moments of the velocity pulsations of the gas.
The coefficients of entrainment f and g of the particles in the pulsations of the
carrier flow are mainly determined by the structure of the energy-bearing turbulent eddies
and have the form

]= iS
T 0
W(s)ex p(+)ds,
- g=---f,
T
W(s)=<e/(t)a/(t+s)>/<a,'(t)a/(t)>

Here, ~(s) is the two-time correlation function of the velocity pulsations of the
gas along the particle trajectory. In the case of a step approximation of P(s), the
e x p r e s s i o n s f o r f and g t a k e t h e f o r m [7]

/=i-exp (- ' g= 7 - ~-

where T is the time of interaction of a particle with an energy-bearing eddy formation


(turbulent mole), which is taken in what follows to be equal to the integral time scale
of the turbulence.
In the absence of turbulent pulsations (<uiuj> = 0), Eq. (i.i) goes over into the
well-known Fokker--Planck equation for Brownian particles in a laminar flow.
We deduce equations for the moments from (I.i). Integrating Eq. (i.i) over the
complete volume in the velocity space, we obtain a balance equation for the mass of the
solid phase:
OC + 0 CVk=O, C = ~ P dv, V~ = - ~i ~vkPdv (1.3)
Ot Oxh
Here, C is the volume concentration of the particles, and V k is the averaged velocity
of the solid phase. Multiplying (i.i) by v i and integrating over v, we obtain the momentum
balance equation of the solid phase:

OCV~ { OC<v~v~>+g<u/ak, > O C = C ( U ~ - V ~ ~-F~) (1./4)


Ot Ox~ Oxh

42
or, taking into account (1.3),

OVi+_ OVi O<vfv,/> ~ U~--V~ t-Fi D~,, OlnC


-- Vt, ~
Ot Ox~ ~ x Ox~
(1.5)
<v~v~>----~-~--]v~v~Pdv=-ViV~+<vi'v/>, D,~=T(<ui'v~'>+g<u(a/>), Ui =U~--Vi

Here, <viva> is the stress tensor in the solid phase, and Dik is the diffusion tensor.
Multiplying (1.1) by viv j and integrating over v, we obtain an equation for the second
moments of the velocity:

OC<v~v~> ~ OC<v~vjv~> + g <u(u,>OCVj + < ~ , u,~, >:@:-:-}


OCV~\ =
Ot Oxh Ox~ Ox~ ~

(1.6)

or, taking into account (1.3) and (1.5), an equation for the second moments of the ve-
locity pulsations of the solid phase:

~ ,,,
a<v,'v/> + vka<vfv/>_ ~ Dik aVj F D~k aV, _ _1 _ _
at Ox~ ~: ax~ 9 Ox~ C Ox~\
(1.7)
> i
<v~v~vk -~ -C- ~ v~v~vkP dr,

The terms on the left-hand side of Eq. (1.7) describe, respectively, the variation
in time, the convective transport, generation from the averaged motion, and diffusion
of the turbulent stresses in the solid phase; the terms on the right-hand side of (1.7)
describe the generation and dissipation of pulsations of the particles as a result of
their interaction with the energy-bearing turbulent eddies and due to Brownian motion.
Under the assumption that the terms on the left-hand side of (1.7) are unimportant, we
deduce from it
<v+'v/>=/<u+'~'>+DSij~ (1.8)

which expresses the condition of local equilibrium between the generation of the pulsa-
tion energy of the particles from the turbulent and thermal energy of the carrier flow
and its dissipation.
With allowance for (1.8), the diffusion tensor of the particles in the locally homo-
geneous approximation is the sum of the turbulent diffusion tensor of the inertialess
mixture and the Brownian diffusion coefficient:

Allowance for the nonlocal effects due to the convective and diffusion transport
mechanisms, which have an important influence on the stress and diffusion tensors of
the particles in the wall region, can be made by solving Eq. (1.6) or (1.7). The equa-
tion for the third moments of the solid-phase velocity, which determine the diffusion
transport in the equation for the second moments, can be obtained from (i.i) and has
the form
aC<v~vy~> F aC<v~vjv~v.> ~ g(<a~'~,'>"aC<v~v~>+<u(~.'> aC<~vh> +
at Ox~ x Ox~ Ox.

T+F <vv>

<v+vjv+v+>= ~ v+vjv+v.P dv " (1.9)

43
2. In the wall region the flow is nearly one dimensional (the flow characteristics
vary mainly in the direction y perpendicular to the surface), and consequently the system
of equations simplifies appreciably, and relatively simple solutions can be obtained.
For the mass balance and correlations of the components normal to the wall, we obtain
from Eqs. (1.3), (1.4), (1.6), and (1.8) in the steady case

dCV~ = 0 (2.1)
dy
dC<v,,2> ~.g<a ,,> dC C
=--(U~-V~+~F~) (2.2)
dy ay
dC<v~> 2c +-v _

(2.3)
dy T

(2.4)
dy dy "~ '
Such a farmulation of the problem presupposes the existence of a constant external
flux Jw that compensates the loss of particles due to precipitation on the wall (Jw =
--CwVyw). Equation (2.1) has the integral CVy = --Jw, and with allowance for it Eq. (2.2)
takes the form

dC + ~C d<vu~>
D,--~y d~ C(Uy+~Fy)=J~, Dp=~(<vu~>+g<u/2>) (2.5)

I t can be s e e n from ( 2 . 5 ) t h a t t h e f l u x o f p a r t i c l e s i s due t o d i f f u s i o n as a r e s u l t


of the concentration gradient, migration due to inhomogeneity of the energy of the par-
ticles, the motion of the gas, and the effect of the external force.
To close Eq. (2.4), we determine the fourth moment in accordance with Millionshchikov's
quasinormal hypothesis: <v~> = 3<v9> 2. Then the following expression for the third
moment is obtained from (2.<) with allowance for (2.2):

dy

The first term on the right-hand side of (2.6) describes the diffusion mechanism
Of transport due to the gradient of the energy of the particles, while the second is
due to the convective transport mechanism. Taking into account (2.6), we obtain from
(2.3) a closed equation for the second moment:

d(DpC d<v~2>)+ 2C[ ,~ d 2


(2.7)
dy dy ~c -~
3. We consider the formulation of the boundary conditions on the wall. Under the
assumption that the probability X of a particle's rebounding and then returning to the
flow does not depend on the velocity and that there is no momentum loss of the particle
in the normal direction, the distribution density on the wall satisfies the condition

Such a f o r m o f t h e b o u n d a r y c o n d i t i o n d e s c r i b e s o n l y t h e a b s o r p t i o n p r o p e r t i e s o f
the surface and does not enable us to obtain conditions for the moments of the distribution
directly. In [7, 8] it is assumed that on the wall and in its neighborhood the distribu-
tion of the particles with respect to Vy is nearly normal, and the corresponding correc-
tions are found from perturbation-theory series. A similar approach makes it possible
to determine the boundary condition for Brownian particles [9]. Another approach involves
the introduction of a model distribution function [i0] and subsequent determination of
the necessary coefficients, either from the moment equations or directly from the equation
for P.
We use the hypothesis of a binormal distribution:

44
vu>O (3.1)
2<vuZ>,/'
Such a form of the distribution is close to the one obtained by numerical solution
of the one-dimensional Fokker--Planck equation [ii] and is a special case of quasinormal
distribution. Note that the variance of the distribution is not determined by the pulsa-
tions of the carrier medium but is an intrinsic characteristic of the particles. From
the following normalization, we have

f
C,+= j P , + d v ~ , n= V2 t _C~ ~ (3.2)
~ I+ X Y<vy~>~
In accordance with (3.1) and (3.2), the moments of the normal component of the ve-
locity are determined by
/2 I--%
v~ = . . . . <v~2>~'I~ (3.3)
a t+%

<v~>~ = - --'V~._8t - Z <v2>dh (3.4)


t+%
In accordance with (3.3), the precipitation velocity depends only on the reflection
coefficient X and on the energy <vg> w of the particles at the wall. The expression (3.3)
can be regarded as a boundary condition for Eq. (2.5). Taking into account (3.3) and
(3~ we obtain from (2.6) a boundary condition for Eq. (2.7):

(3.5)
dy ,~ ~ I+ z
4. In order to identify the characteristic effects of the dynamics of the inertial
particles in the transverse direction, we consider the solution of Eqs. (2.5) and (2.7)
with the boundary conditions (3.3) and (3.5) for Uy = Fy = 0 in individual zones of the
wall region of the flow.
Brownian Motion in Laminar Diffusion Sublayer. The system of equations (2.5) and
(2~ reduces for <u~2> = 0 and in dimensionless variables to the solution of the prob-
lem
f_!+ 21o d~+__2(i_~)= o (4.1)
d'q2 q)w+JoT]d~t q~

= =o, c=C~
(4.2)

-- ' . . . . . , 1] = - -
7o= D C= a t+Z ~="' ~ 'D 1/~D
To construct an analytic solution, we linearize Eq. (4.1), setting ~=~ in the
denominator of the last term. Then the solution of the problem is

, i+-= . (4.3)
~a Y~ t+Z'
In Fig. i, we compare the solution (4.2), (4.3) (continuous curves) with the results
of numerical solution of Eq. (4.1) (broken curves). It can be seen that there is good
agreement between the approximate analytic solution and the exact numerical solution.
At large distances from the wall, the concentration of the Brownian particles can
be determined from the solution of the ordinary diffusion equation diC/dy = = 0 with
boundary condition of the third kind (dC/dy) w = Cw/A, which effectively takes into account
the wall effects. The Milne extrapolation length A = aJ~-D can be obtained by matching
in the limit y § ~ to the solution (4.2), which gives a=~/J0. For a completely absorb-
ing wall (X = 0)

45
? C/Cu~

1.2

1.1
f
a~ j

Fig. i

a = V9 u(~+3/Y~)
_ -= t,40
2(i+2/ga)
which is quite close to the value a = 1.46 obtained by direct numerical integration of
the Fokker--Planck equation [ii].
Logarithmic Layer of Turbulent Flow. In this zone of turbulent wall flow (the re-
gion in which the "logarithmic wall law" holds [4]) the Brownian motion can be ignored,
the intensity of the turbulent pulsations of the gas is constant, <u~2> = ~u~, and the
time scale of the turbulence is proportional to the distance from the wall: T=~ylu,,
where u, is the dynamic velocity, $ = I.i, and • = 0.4 is the Prandtl--von K~rmgn constant.
The relation for T follows from the expression for the turbulent diffusion coefficient
of an inertialess admixture in the logarithmic layer:

D, . . . . lim D~,=T<u~'~>=
So, ~/r-+o Set

where ~,=• is the coefficient of turbulent viscosity, and Sc t = i/$ = 0.9 is the
turbulent Schmidt number.
With allowance for these relations and (1.2), Eqs. (2.5) and (2.7) with the bound-
ary conditions (3.3) and (3.5) take in the dimensionless variables the form

C0[q~+~l-I exp(-~l) an" eq~,= ~ - ~ - { T + 2 [ i - e x p ( - ~ l ) ] } + . 2 C 0 [ i - e x p ( - ~ l ) - c p ] = 0 (4.4)

dCo C dep 1/ g-l-z q) V=,


[v+n-l+~p(-n)]--~+ o~=~o', n=o: e'l, cp " = ~ . Co=i
(4.5)
~t=oo:_ q f = 0 , r [~u, ~ , Co-- C~' ~1 = - - , ~ a , e=~•

Since Eq. (4.4) contains a small parameter multiplying the highest derivative, the
problem can be solved by the method of matched asymptotic expansions. The exterior solu-
tion of Eq. (4.4), obtained in the limit s § 0 and valid far from the wall, will be

~ = 1 - e x p (--N) (4.6)

46
f
J
0.5

0 Z

Fig. 2 Fig. 3
The physical meaning of (4.6) is that in the locally homogeneous approximation it
expresses the energy of the particles in terms of the intensity of the gas pulsations:
<v~> = f<u~2>. To construct the interior solution ~ near the wall, we go over to the
new variab• ~=~/8, ~=N/e, in which the problem takes the form

dZ~ 2~' (d~) 2


- -d~
7 ' OJ~+O~ - ~ + i + -~-(~--0)=0
(4.7)

The b o u n d a r y c o n d i t i o n a s $ + ~ i s o b t a i n e d f r o m m a t c h i n g t o t h e e x t e r i o r solution
(4.6). To find an analytic solution, we linearize Eq. (4.7), setting % = %w in the de-
nominator of the last term; in this case,

@=%+--- } exp -- ~ , ~= 14 lq = (4.8)


' • ~'~+~ - ~ 1+%

The composite solution, combining the exterior, (4.6), and interior, (4.8), solu-
tions, has the form

(4.9)

I n F i g . 2, we c o m p a r e t h e d i s t r i b u t i o n (4.9) (continuous curves) with the results


Of n u m e r i c a l s o l u t i o n o f E q s . ( 4 . 4 ) and ( 4 . 5 ) ( b r o k e n c u r v e s ) f o r E = 0 . 1 8 ; the approx-
i m a t e analytic solution and the exact numerical solution agree rather well. With increas-
ing distance from the wall, the parameter ~/T = i/q, which characterizes the relative
extent to which the particles are inertial, decreases, i.e., the particles become less
inertial, and therefore their energy tends to the values determined by the locally homo-
geneous solution (4.6) (chain curve). However, as the wall is approached, the particles
become relatively more inertial, and therefore we observe an appreciable departure of
the particle energy from the solution (4.6). The nonzero energies of the particles at
the wall are explained by the action of the diffusion and convective mechanisms of trans-
port of the pulsations, these mechanisms being important for inertial particles.
In Fig. 3, we give the particle concentration distributions corresponding to re-
flection coefficients X = 0, 0.2, 0.4, 0.6, 0.8, 1.0 (curves 1--6) obtained by numerical
solution of Eqs. (4.4) and (4.5). The nonmonotonic nature of the distribution and the
enhanced particle concentrations near the wall for the cases of incomplete reflection
are due to the presence of two competing particle transport mechanisms - turbulent dif-
fusion and migration (the second term on the left-hand side of Eq. (4.5)), which leads
to accumulation of particles in regions with low level of the turbulent pulsations.
Wall Region with Viscous Sublayer. As an approximation of the pulsation structure
of the carrier flow, we take a very simple two-zone model consisting of a viscous sub-
layer with zero pulsation intensity and a turbulent region with constant pulsation in-
tensity, i.e., ro 2

where 5 i s t h e t h i c k n e s s of the viscous sublayer, and H ( y ) i s t h e H e a v i s i d e function.

47
0,#
l ]

(/,2 . , 0,0~

]!
to %

Fig. 4 Fig. 5

The time scale of the turbulence at a distance from the wall of order 6 can be taken
to be constant: T = ~6/u,, ~ ~ i. In this case, the problem of calculating the dis-
tributions of the characteristics of the particles in the transverse direction (without
allowance for the Brownian motion) takes the form

d dq) c/ [r (~l_ 1) l + 2Co (4.10)

dCo.~ dq~ , ]/, 2 1--% ~,/,


(Pw ~ g! C0=l, rl-'-~ ~'=0
t+% T0
(4.11)
<v~~> U Tu*~'/=
/=I--exp(- aiU~ ~,/~
), g = - - - -
C
q~= l~a.~ , ~1 6 ' To= 6 ' TO t
f, Co=.
T0 Cw

We restrict ourselves to solution of Eqs. (4.10) and (4.11) for the case of a wall
that reflects the particles (X = i). In the region 0 < ~ < I Eq. (4.10) takes the form

(d2(p 2 )=0 (4.12)


~ dB 2 ~02

With allowance for the boundary condition (p~'=O, Eq. (4.12) has two solutions:

q~=O, O<q<~l, , cp=(rl--~l,)Z,~o-2, ~1,<~<1 (4.13)


~=~+N2/T02 (4.14)
In the region i < ~ < ~, the problem (4.10)--(4.11) reduces to solution of the equa-
tion
d2qD 2(]--q~) 0 (4.15)
dq~ ~0~(~+g)

Setting in the denominator of the second term of (4.15) ~ = ~ (where (~, is the value
of ~ for q = i), we obtain an approximate solution in the form

r ((p~_/) exp[ _1 / T02(~t+g


2 ) (N--i) ] + ] (4.16)

From t h e e x p r e s s i o n s ( 4 . 1 3 ) and ( 4 . 1 6 ) , augmented by t h e f o l l o w i n g matching c o n d i -


t i o n s , we d e t e r m i n e
~1~+o=~1~-0--~, (~o,+g) ,§ ~-0 (4.17)

Y(~,+g) (l+~,)--~f2~?), t-~,=Y~-~0 (4.18)


Similarly, f o r t h e case of nonzero p u l s a t i o n s along t h e w a l l , we o b t a i n from ( 4 . 1 4 )
and (4.16) in conjunction with the conditions (4.17)

48
u (q%+g) (/--qot)---- qh, q~=q% -- - - (4.19)
To "[02

Figure 4 gives the distributions of the intensity of the particle velocity pulsa-
tions in accordance with the solutions (4.13), (4.16), (4.18) and (4.14), (4.16), (4.19)
for ~ u 2 = i (curves i--5 correspond to T o = 0.5, 1.5, 2.62, 5, I0). The dependence
of the pulsation intensity ~ at the wall on the parameter T o characterizing the inertia
of the particles is given in Fig. 5. The value T o = 2.62 is the point of transition
from the one type of solution to the other: the solution (4.13) holds for T o < 2.62
in the region 0 < q < I, and the solution (4.14) holds for ~0 > 2.62.
Thus, the pulsation energy of fine particles in the viscous sublayer is zero, while
the pulsation intensity of inertial particles in the viscous sublayer and at the wall
itself is nonzero. The presence of velocity pulsations of the solid phase in the viscous
sublayer could be obtained only on the basis of the nonlocal theory and is explained
by the diffusion mechanism of transport of pulsations from the turbulent region of the
flow due to the inertia of the particles.
In conclusion we note that under real conditions one of the three considered sub-
regions will be the most important, this depending on the size of the particles. Large
particles, for which the structure of the flow near the wall is unimportant, satisfy
the dependences noted above. Smaller particles, moving in the direction to the boundary,
do not preserve information about the core of the flow, and their behavior is deter-
mined by the manner in which the pulsations of the carrier phase vary in the wall region.
Finally, for the description of the finest particles the effects due to Brownian dif-
fusion must be taken into account.

LITERATURE CITED
I. A. A. Shraiber, L. B. Gavin, V. A. Naumov, and V. P. Yatsenko, Turbulent Flows of
Gas Suspensions [in Russian], Naukova Dumka, Kiev (1987).
2. C. P. Chen and P. E. Wood, "Turbulence closure modeling of the dilute gas-particle
axisymmetric jet," AIChE J., 32, 163 (1986).
3. A. A. Mostafa and H. C. Mongia, "On the modeling of turbulent evaporating sprays:
Eulerian versus Lagrangian approach," Int. J. Heat. Mass Transfer, 30, 2583 (1987).
4. A. S. Monin and A. M. Yaglom, Statistical Fluid Mechanics, Part i, MIT Press (1971).
5. D. Miloevich, O. Ts. Solonenko, and G. M. Krylov, "Comparative analysis of some
models of the turbulent transport of inertial particles," in: Transport Processes
in One- and Two-Phase Media [in Russian], Institute of Heat Physics, Siberian Branch,
USSR Academy of Sciences, Novosibirsk (1986), p. 70.
6 L. V. Kondrat'ev, "Structure of the turbulent flow of a gas suspension in the wall
region of a tube," Inzh.-Fiz. Zh., 55, 1029 (1988).
7 I. V. Derevich and L. I. Zaichik, "Precipitation of particles from a turbulent flow,"
Izv. Akad. Nauk SSSR, Mekh. Zhidk. Gaza, No. 5, 96 (1988).
8 I. V. Derevich and L. I. Zaichik, "Boundary condition for the equation of diffusion
of particles in an inhomogeneous flow," Inzh.-Fiz. Zh., 5_~5, 735 (1988).
9 S. V. G. Menon and D. C. Shani, "Derivation of the diffusion equation and radiation
boundary condition from the Fokker--Planck equation," Phys. Rev. A 32, 3832 (1985).
i0 S. Harris, "Steady one-dimensional Brownian motion with an absorbing boundary," J.
Chem. Phys. 75, 3103 (1981).
ii S. V. G. Menon, V. Kumar, and D. C. Sahni, "Green's function approach to the solution
of the time dependent Fokker--Planck equation with an absorbing boundary," Physica,
135A~ 63 (1986).

49

You might also like