Professional Documents
Culture Documents
P. Vadasz
The fundamental theory of flow and thermal convection in porous media is reviewed and
a systematic classification and identification of the relevant problems is introduced. An
initial distinction between rotating flows in isothermal heterogeneous porous systems and
free convection in homogeneous non-isothermal porous systems provides the two major
classes of problems to be considered. The properties of flow in rotating porous media are
being demonstrated theoretically as well as experimentally and compared to the
corresponding properties of flow in rotating pure fluids (i.e. in non-porous domains).
Application of the latter to non-isothermal problems of thermal convection in rotating
porous layers raises a wide range of distinct problems that are being reviewed in light of
their theoretical, experimental as well as their applied impact. These include the effects of
centrifugal as well gravity induced buoyancy on convection as well as the impact of the
offset distance of the fluid saturated porous layer from the axis of rotation on the resulting
heat transfer. Other important effects include the Coriolis effect on gravity driven
convection in a porous layer heated from below. Examples of solutions to selected problems
are presented, highlighting the significant impact of rotation on the flow in porous media.
1. INTRODUCTION
The study of flow in rotating porous media is motivated by its practical applications in
geophysics and engineering. Among the applications of rotating flow in porous media to
engineering disciplines, one can find the food processing industry, chemical process
industry, centrifugal filtration processes, and rotating machinery. Detailed discussion on
particular applications was presented by Nield and Bejan (1999), Bejan (1995) and Vadasz
( 1997, 1998).
Limited research is available on isothermal flow in rotating porous media, while some
results are available for natural convection in rotating porous media, e.g. Palm and Tyvand
(1984), Rudraiah et al. (1986), Patil and Vaidyanathan (1983) and Jou and Liaw (1987).
Nield (1999, 1991a,b), while presenting comprehensive reviews of the stability of
convective flows in porous media found also that the effect of rotation on convection in a
porous medium attracted limited interest. The lack of experimental results was particularly
noticed. The lack of interest for this type of flow is the fact that isothermal flow in
homogeneous porous media following Darcy's law is irrotational (Bear, 1972). However, it
turns out that there are sufficient cases when the above conditions do not apply. This applies
for heterogeneous porous media and for the effect of free convection in porous media.
470
2. GOVERNING EQUATIONS
The dimensionless equations governing the flow and heat transfer in a rotating
heterogeneous (but isotropic) porous medium, following Darcy's law, but extended to
include the Coriolis and centrifugal terms are presented in the form
(i) Continuity equation
V'V-0 (1)
(ii) D a r c y ' s law (extended to include rotation effects)
q = -~:
[VPr+Ra~TV(i~"X)- Rao~T~o x (eo~ x X ) + -E-s
1 x q
] (2)
(iii) E n e r g y equation
Or
~+q'VT=V2T (3)
Ot
where it was assumed that local thermal equilibrium between the solid and fluid phases
applies. In equations (1)-(3) q is the dimensionless filtration velocity (Darcy's flux), Pr is
the dimensionless reduced pressure generalized to include the constant components of the
centrifugal as well as the gravity terms, T is the dimensionless temperature, k ( X ) is the
dimensionless permeability function, ~o is a unit vector in the direction of the imposed
angular velocity and X = x~ x + yey + ze z is the position vector. The values u c = Vo/l c and
Apc = l,toVo/k o were used to scale the filtration velocity and pressure, respectively, and k o
was used to scale the permeability function. For heat convection problems a rescaling of the
equations by using u c = aeo/lc and Apc = Poaeo/ko for scaling the filtration velocity and
pressure, respectively, was applied. The characteristic length lc, used to scale the space
variables x,, y, and z,, needs to be defined in the context of the particular problem under
consideration, and the characteristic time At c = lc / u c was used to scale the time variable. A
linear approximation was assumed for the relationship between the density and temperature
in the form p = 1 - fir T where fit = fl, ATc and fl, is the thermal expansion coefficient, A T c
is a characteristic temperature difference ( typically A T c = ( Tn - T c ) ) used to scale the
temperature in the equations. Furthermore, by using the Oberbeck-Boussinesq
approximation the density was assumed constant everywhere except in a body force term of
eq.(2). The dimensionless groups in eq.(2) are the gravity related Rayleigh number in porous
media, the porous media Rayleigh number related to the centrifugal body force, and the
porous media Ekman number, defined as
Rag = fl*ATcg*lck~ Ra~ = fl*ATcco212k~ E k = qbv~ (4)
V o l~Zeo V o O(,eo 2co, k o
where ~b is porosity, co, is the angular velocity of rotation, g, is the acceleration of gravity,
k o is a reference value of permeability, v o is the kinematic viscosity, ~eo is the effective
thermal diffusivity extended to account for the ratio between the heat capacity of the fluid
and the effective heat capacity of the porous domain, lc is a characteristic length scale, 13, is
the thermal expansion coefficient and A T c is a characteristic temperature difference. These
dimensionless groups control the significance of the different phenomena. Therefore, the
value of Ekman number (Ek) controls the significance of the Coriolis effect and the ratio
471
between the gravity related Rayleigh number (Rag) and the Rayleigh number related to the
centrifugal body force (Ra~o) controls the significance of gravity with respect to centrifugal
forces as far as free convection is concerned. This ratio is Rag/Ra~o = g./c021c .
A third category related to convection due to solutal buoyancy alone could have been
introduced but it would not bring any significant contribution to (i) above as the results
obtained for thermal buoyancy alone can be easily converted to the third case by analogy.
Three separate cases for each one of the above categories can be considered, depending
on the driving body force, i.e., convection driven by gravity, convection driven by the
centrifugal force and convection driven by both gravity and centrifugal force.
Given typical values of viscosity, porosity and permeability one can evaluate the range of
variation of Ekman number in some engineering applications. There, the angular velocity
may vary from 10 rpm to 10 000 rpm leading to Ekman numbers in the range from
Ek = 1 to Ek = 10-3 . The latter value is very small, pertaining to the conditions considered
in this section. Therefore, in the limit of Ek ~ O, say Ek = 0, eq.(6) takes the form
ez x q = -VpR (7)
and the effect of permeability variations disappears. Taking the "curl" of eq.(7) leads to
vx(zxq)=O
Evaluating the "curl" operator on the cross product of the left-hand side of eq.(8) gives
(ez'V)q = 0 (9)
Equation (9) is identical to the Taylor-Proudman condition for pure fluids (non-porous
domains); it thus represents the proof of the Taylor-Proudman theorem in porous media and
can be presented in the following simplified form
0__qq= 0 (10)
3z
Experimental confirmation of these theoretical results is necessary. However, because of the
inherent difficulty of visualizing the flow, i.e., the Taylor-Proudman column through a
granular porous matrix, an indirect method was adopted. The rationale behind the proposed
method was to place a porous layer in-between two pure fluid layers inside a cylinder (see
Fig. l(a)). A disturbance in the form of a small fixed obstacle was created on the bottom
wall in the fluid layer located below the porous layer causing the appearance of a Taylor-
Proudman column in this layer. As the fundamental property of a Taylor-Proudman column
is that it does not allow for variations of velocity in the vertical direction, the proposed
experimental setup should allow us to detect Taylor-Proudman columns in the top
undisturbed fluid layer located above the porous layer, if the column exists in the porous
473
layer as a result of the disturbance in the bottom fluid layer. This means that a small object
located at the bottom of the lower fluid layer should create a Taylor-Proudman column
which extends upwards through the porous layer into the upper fluid layer. It is because of
this expectation that the top fluid layer will "feel" through the buffer porous layer the small
object located at the bottom of the lower fluid layer, that some colleagues of the author
named the effect resulting from the experiment "The Princess and the pea". Of course, if
experiments will not allow us to detect Taylor-Proudman columns in the top undisturbed
fluid layer another possibility exists which is consistent with eq.(10). This could for example
imply that the filtration velocity is zero in a rotating porous layer, meaning that the whole
fluid in a rotating porous layer rotates as a solid body instantaneously (almost).
The experimental apparatus consists of a record player turntable adjusted to allow variable
angular velocity in the designed range and to provide a better dynamic balance when the
cylindrical container is placed securely on the rotating plate. Preliminary results confirm the
appearance of Taylor-Proudman columns in the top fluid layer as shown qualitatively in Fig.
1(b). A further significant consequence of eq.(10) is represented by a geostrophic type of
flow. Taking the z-component of eq.(10) yields Ow/Oz = 0, and the continuity equation (1)
becomes
Ou Ov
0x t-~-y 0 (ll)
Therefore the flow at high rotation rates has a tendency towards two-dimensionality and a
stream function, ~ , can be introduced for the flow in the x - y plane
u = - c ) v / / O y ; v = c ) ~ / O x , which satisfies identically the continuity equation (1).
Substituting u and v with their stream function representation in eq.(7) yields
3gt _ 3pg 3gt _ 3pR
c9---~-- O---~ 0--7- ~--7 (12)
As both the pressure and the stream function can be related to an arbitrary reference value,
the conclusion from equations (12) is that the stream function and the pressure are the same
in the limit of high rotation rates (Ek---> 0). This type of geostrophic flow means that
isobars represent streamlines at the leading order, for Ek ~ O.
~gT
0-7 + q" VT = VET (15)
474
z=l m G
Z••* ".:-.: : : ..
z
, 03,
'.'"".'-'".'".'.'.'4... . . .
z=0
(a) (b)
Figure 1: (a) A closed cylindrical container divided into a porous layer and two fluid layers
above and below the porous layer. A solid object is fixed at the bottom surface of the
container. (b) Qualitative description of the Taylor-Proudman column as observed in the
preliminary runs of the experiment.
Temperature gradients collinear with the centrifugal body force represent a stability
problem. The objective is then to establish the stability condition as well as the convection
pattern when this stability condition is not satisfied.
6. CONCLUDING REMARKS
The effect of rotation was shown to have a significant impact on the flow in porous
media. In isothermal systems this effect is limited to the Coriolis effect. In non-isothermal
systems the effect of rotation is expected in free convection.
475
z'A
ik z' 03, X,._
v-
z
" " " " " " " " " " " " " z - ' H
~ _...~ y'
Xt X
--" i .v.
Xo i
[ '
Figure 2: A rotating fluid saturated porous layer distant from the axis of rotation and subject
to different temperatures at the sidewalls.
~ . . i I t8 8 ~ ! ! ! ! ! I I
4~ ~ 4
3 3
.......iiiii~.iiiiiiiiiiii,!;.ii'
.. ~........4.....iiiiiiiiiiiii~,
,,:4........~...... ~........
~ 1 ~...... 2 21
0 0
-2 -1 0 1 2 0 1 2 3 4
l~ Ra~o/x2
(a) (b)
Figure 3: (a) The variation of the critical values of the centrifugal Rayleigh numbers as a
function of q; (b) The stability map on the Rao,- Rao,o plane showing the division of the
plane into stable and unstable zones.
REFERENCES
Bear J. Dynamics of Fluids in Porous Media, New York: Elsevier, 1972, pp. 131-132.
Bejan A. Convection Heat Transfer, 2nd ed., New York: Wiley, 1995.
Jou JJ, Liaw JS. Thermal convection in a porous medium subject to transient heating and
rotation. Int. J. Heat Mass Transfer 30:208-211, 1987.
476