A. J. Chamkha
Assistant Professor,
Department of Mechanical Engineering,
Kuwait University,
Safat 13060 Kuwait
Mem. ASME
J. Peddieson, Jr.
Professor,
Department of Mechanical Engineering,
Tennessee Technological University,
Cookeville, TN 38505
Boundary Layer Theory for a
Particulate Suspension
Order of magnitude considerations are employed to develop boundary layer equa-
tions for two phase particle/fluid suspension flows. It is demonstrated that a variety
of possibilities exist and three of these are examined in detail. Two are applied to
the problem of flow past a semi-infinite flat plate.
Introduction
This paper is concerned with boundary layer theory for a
particulate suspension. Given the importance of boundary lay-
ers in applications, this topic has received surprisingly little
attention. There is a history of work on the problem of the
steady laminar boundary layer on a semi-infinite flat plate with
recent contributions by Osiptsov (1980), Prabha and Jain
(1980), and Wang and Glass (1988). References to earlier work
can be found in these papers. In contrast to investigations of
the flat plate and a few other specific geometries, there appears
to have been little effort devoted to determining the general
form of boundary layer equations. The present paper deals
with this topic.
A boundary layer order of magnitude analysis is carried out
using a typical set of two fluid equations representative of the
current literature. It is found that a variety of outcomes are
possible depending on the order of magnitude assumptions
selected. Three of the most interesting cases are singled out
for explicit presentation. Some specific numerical results are
then given for the problem of steady laminar boundary layer
flow past a semi-infinite flat plate. It is shown that the bound-
ary layer model employed greatly influences predictions.
Governing Equations
The boundary layer analysis to be described in the present
work is based on the following typical set of two phase flow
equations.
W-V-((l-0)v
c
) = O, d,4.+ V-W>v
d
) = 0 (1)
represent respective balances of mass for the fluid and partic-
ulate phases (with the true densities of both phases assumed
constant). In Eqs. (1) V is the gradient operator, t is time, <j>
is the particulate volume fraction, v
c
is the fluid phase velocity
vector, and \
d
is the particulate phase velocity vector.
/o
c
(l-</>)(d,v
c
+ v
c
-W
c
)=V-g
c
-f
p
rf
<M«,v
d
+ v
d
-W
d
)=V-£
d
+f (2)
Contributed by the Fluids Engineering Division for publication in the JOURNAL
OF FLUIDS ENGINEERING. Manuscript received by the Fluids Engineering Division
May 7, 1992; revised manuscript received March 18, 1993. Associate Technical
Editor: M. W. Reeks.
represent respective balances of linear momentum for the fluid
and particulate phases (with external body forces neglected).
In Eqs. (2) p
c
is the fluid true density, p
d
is the particulate true
density, g
c
is the fluid phase stress tensor, g
d
is the particulate
phase stress tensor, and f is the interphase force per unit volume
acting on the particulate phase. The balance laws discussed
above will be supplemented by the constitutive equations
g
c
=-(\-M)pl +2
A
t
c
(l-<«D
c
; D
c
=(vv
c
+W
c
r
)/2
g
d
= -{\<f>p + q)l +2
Mrf
</.D
d
; D
d=
(
V
v
d
+ VvJ)/2
f = p
d
4>y
r
/r + \pV<t>;\
r
= \
d
-v
c
(3)
wherep is the indeterminate pressure, q is the particulate phase
dynamic pressure, X is a coefficient which determines the ap-
portionment of the indeterminate pressure gradient between
the phases (see below), ^
c
and fi
d
are dynamic viscosity coef-
ficients, T is the interphase relaxation time, 4 is the unit tensor,
and a superposed T indicates the transpose of a second order
tensor. In general q, X, LI
C
, \i.
d
, and T are functions of such
quantities as 4>, v
r
, and the invariants of D
c
and D
rf
.
The following brief comments about Eqs. (3) are in order.
The forms of Eqs. (3b,c) should be such that Eq. (2b) will
reduce to the Eulerian form of the equation of motion for a
single particle when the volume fraction is sufficiently small.
Angular momentum considerations do not require that g
c
and
g
d
be individually symmetric but do require that the combina-
tion a
c
+ g
d
be symmetric (in the absence of body moments).
Several particle phase stress mechanisms have been proposed.
These include direct contact between particles, wall effects,
local particle deformations, and the consequences of the av-
eraging required to model a system containing discrete particles
as a continuum. The precise forms of X, /n
c
, \s,
d
, and q are
model dependent. In the present work it is not necessary to
restrict attention to a specific model and this has not been
done.
The focus of the present paper is on laminar flow. The
inclusion of D
c
and D
rf
in the list of arguments given above
formally allows Eqs. (3) to include algebraic turbulence models.
It is highly doubtful, however, that such models are sufficiently
inclusive to simulate two phase turbulence under any but the
simplest conditions.
Journal of Fluids Engineering MARCH 1994, Vol. 116/147
Copyright © 1994 by ASME
Downloaded 01 Oct 2012 to 109.171.137.210. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm