TRANSIENT EFFECTS IN

GEOTHERMAL CONVECTIVE SYSTEMS

A thesis submitted in partial fulfilment of

the requirements of the degree of Doctor of Philosophy
at the University of Auckland

Roland N. Horne

School of Engineering
Department of Theoretical and Applied Mechanics

March 1975
(blank)
ABSTRACT

This work is a detailed analysis of the transient behaviour of geothermal

convective systems. The flow in these systems is found to be fluctuating or regular
oscillatory in a simplified two-dimensional model and these unsteady effects persist
when the model is refined to include the concepts of temperature dependent viscosity
and fluid withdrawal and recharge. The analysis is extended into three dimensions
to verify this behaviour. The supplementary exploration of added salinity gradients
indicates transient effects of a different kind in this case. The examination of
the porous insulator problem confirms the results of previous authors and verifies
the viability of the numerical methods that are used throughout the investigation.

vi
ACKNOWLEDGEMENTS

During the last three years the research that has lead to the presentation
of this thesis has been assisted by many people. Most credit is due to my supervisor
Dr M.J. O'Sullivan who originally inspired the work and has given much advice since.

The Hele-Shaw apparatus used for the experimental work was borrowed from
the Department of Civil Engineering and I am grateful for the assistance of technicians
Mr C. Raymond and Mr D. Browne, who helped set it up. I am also indebted to Dr John
Meikle and the Auckland Industrial Development Division of the D.S.I.R. for the use of
their AGA Thermovision camera. The many photographs taken of these experiments were
processed by the Photographic departments of the University of Auckland and the A.I.D.D.

Referencing of earlier publications has been facilitated by the library staff

of the School of Engineering, particularly Mrs Alison Reid who helped with several
translations from French.

A large portion of this work has depended on the extensive use of computer
facilities and the aid of the University of Auckland Computer Centre is appreciated.
Most of the calculations were performed on the Burroughs B6700 although early develop-
ment of the programs was done on the IBM 1130 in 1972.

Parts of the work have appeared previously in Horne and O'Sullivan ( 1 9 7 4 a

and b) and I am indebted to Ms Kitty Young for the preparation of these two papers.

I would also like to thank Professor C.M. Segedin of the Department of

Theoretical and Applied Mechanics, Associate Professor M.P. Hochstein of the Department
of Geology, University of Auckland, and Dr Ian Donaldson and Dr Robin Wooding of the
D.S.I.R., Wellington for the various helpful discussions I have had with each of them.

The text has been typed by Ms Mamie Long who was fortunately persuaded out
of her retirement from such tasks.
CONTENTS

Notation ii

CHAPTER 1 -
INTRODUCTION
1.1 The P r e c e p t
1.2 The W a i r a k e i Geothermal Region
1.3 P a s t Work
1.4 The Scope of t h i s Work

CHAPTER 2 - MATHEMATICAL FORMULATION

2.1 The E q u a t i o n s of Motion 8
2.2 Boundary C o n d i t i o n s 14
2.3 Two-Dimensional R e g i o n s 16
2.4 Three- Dimensional Regions 20
2.5 F l u i d S i n k s and Sources 22

CHAPTER 3 - EXPERIMENTAL SOLUTION I N TWO-DIMENSIONAL REGIONS 26

CHAPTER 4 - NUMERICAL METHODS

4.1 P a s t Numerical S o l u t i o n s 35
4.2 F i n i t e D i f f e r e n c e Methods 35
4.3 V a r i a t i o n a l Techniques 44

CHAPTER 5 - RESULTS FOR THE TWO-DIMENSIONAL PROBLEMS

5.1 The Uniformly Heated Model f = 1 - 47
5.2 The Non-Uniformly Heated Model - f < 1 52
5.3 Review o f C o n s t a n t V i s c o s i t y S o l u t i o n s 58
5.4 The Recharge - Discharge Solutions 61
5.5 The V a r i a b l e V i s c o s i t y Model 71

CHAPTER 6 - THREE-DIMENSIONAL TRANSIENT FLOW

6.1 Introduction 76
6.2 The Range o f S o l u t i o n s 76
6.3 The Numerical R e s u l t s 77

CHAPTER 7 - THERMOHALINE CONVECTION I N POROUS MEDIA

7.1 Introduction 81
7.2 The E q u a t i o n o f Motion 82
7.3 The Numerical R e s u l t s 81,

CHAPTER 8 - THE POROUS INSULATOR PROBLEM

8.1 Introduction 87
8.2 The Numerical S o l u t i o n s E8

iv
CHAPTER 9 - SUMMARY AND CONCLUSIONS
9.1 Evaluation of the Results 91
9.2 The Thermal Boundary Layer 91
9.3 The Preferred Solution 92
9.4 Disturbance Interactions 92
9.5 Temperature/Salinity Effects 93
9.6 The Presence of Boundaries 93
9.7 The Physical Implications 93
9.8 Exploitation 94
9.9 Future Modelling of Geothermal Fields 94

APPENDICES A - The Arakawa Differencing Schemes 96

B - An Extension of the Buneman Algorithm to Fourth-Order Accuracy g7
C - The Nusselt Number 101
D - An Extension of the Buneman Algorithm to Neumann-type
Boundary Conditions in Two and Three Dimensions 10 3
E - Spectral Representation of Equations of Motion 105
F - Evolution of the Thermal Boundary Layer 10 7
G - Streamline Representation in Three-Dimensional Flows 108

REFERENCES 110

...
Ill
NOTATION

' All variables and operators used are defined when they first appear in

the text, however the commonly used ones are summarised here.

Dimensional Variables

bl, b 2 , b 3 - Coefficients of variation of viscosity v with temperature T.

g - The gravitational acceleration
k - The permeability of the medium
m - The porosity of the medium
P - The dynamic pressure
s - The mean flux velocity
t - Time
-.
X - Spatial dimensions
C - Concentration of dissolved mineral salts
c1 - Minimum and maximum values of C

f - Buoyancy force
'ij - Thermal dispersion tensor
T - Temperature

To' T1 - Maximum and minimum values of T

a - Thermal expansion coefficient

al' O2
- Linear and quadratic thermal expansion coefficient
a' - Solutal expansion coefficient
B1, B 2 , B 3 - Viscosity variation coefficients
K - Thermal. diffusivity
K' - Solutal diffusivity
x - Ratio of volumetric heat capacities

u - Dynamic viscosity of fluid

V - Kinematic viscosity of fluid

vO
- Low temperature value of v
P - Density of fluid

PO
- Low temperature value of p

Non-Dimensional Variables

f - Fraction of lower boundary heated

9 - Strength of fluid sink

u - Velocity due to flow into sink

-
U' - Velocity not due to flow into sink
C - Concentration
Nu - Nusselt number
P - Pressure
R - Rayleigh number
S - Solutal Rayleigh number

u - Fluid velocity

x - Spatial dimensions
Y - Buoyancy ratio
AX, AY - Spatial increments
AT - Time increment
V' - Viscosity
P' - Density
e - Temperature
JI - Stream function
e - Vector potential

T - Time

Operators

6 - Jacobian
VZ - Laplacian

I
Chapter 1 - INTRODUCTION

1.1 THE PRECEPT

Geothermal Regions

This work is an attempt to determine the various ways fluid may flow through
the permeable material of a geothermal system under the influence of thermally or
solutallyy produced buoyancy effects. If an enclosed porous region is saturated with
a fluid that has a temperature dependent density, then by introducing a high
temperature anomaly at the base of the system it is possible, under certain conditions,
to create movement of the fluid as it becomes buoyant and rises through the material.
This is the process of natural convection in a porous medium and is the driving
mechanism in many geophysical situations where heat is convected to the earth's
surface from an energy source deep below ground. In transporting heat from an
inaccessible base formation these hydrothermal systems provide a valuable source of
energy which in small ventures may be exploited for home heating or factory processes,
or in larger projects for the generation of electrical power. Hydrothermal fields
have been developed for power production in Larderello, Italy, in the Geysers, U.S.A.,
in Kamchatka, U.S.S.R., and in Wairakei and Kawerau, New Zealand, and it is expected
that this energy resource will become increasingly important in future with develop-
ment of fields in Iceland, Japan, Hungary, Mexico, El Salvadore, Kenya, Ethiopia, the
Phillipines, Indonesia, Chile, Guadeloupe, Taiwan, Turkey and Communist China.
Therefore the modelling of convective flow through porous media is important from a
practical viewpoint since it is required to enable engineers to predict the
behaviour of a geothermal system under exploitation. This type of flow is also of
general interest to theoreticians since it is one of the simplest examples of hydro-
dynamic instability, and furnishes parallels with more complex fluid problems, for
example the natural convection of a Newtonian fluid in an open region. Furthermore
there are additional practical applications, not related to geothermal flows, to the
cooling of nuclear reactor cores and the efficiency of porous insulators, for
instance fibreglass house insulation or refrigerator walls. It is not surprising
then that this problem has, in its various forms, received attention for some time
and this study surveys and extends the current state of research in the field and
reviews the interrelations between the many solutions previously obtained.

Unsteady Solutions

This work supersedes many

y earlier studies in that it pays strict attention
to the variation of conditions in geothermal regions as
s time passes. The investiga-
tion examines the evolution of a geothermal system, with a particular view to
ascertaining the implications of artificial development and exploitation.
Furthermore it demonstrates that under certain conditions convective flow in porous

1
media c o u l d remain t r a n s i e n t i n d e f i n i t e l y due t o i n h e r e n t f l u i d i n s t a b i l i t i e s ,
The c o n d i t i o n s u n d e r which t h e s e u n s t e a d y flows can o c c u r have been m o d e l l e d by
e a r l i e r s t u d i e s b u t t h e s e were o f t e n a p p a r e n t l y c o n t r a d i c t o r y a n d t h e p r o c e s s e s
i n v o l v e d were n o t f u l l y e x p l a i n e d . I t s t i l l remained f o r t h e p r e s e n t s t u d y t o
v e r i f y t h e e x i s t e n c e o f f l u c t u a t i n g c o n v e c t i v e i n s t a b i l i t i e s i n p o r o u s media and t o
d e t e r m i n e w h e t h e r s u c h e f f e c t s may have a n u n e x p e c t e d and major i n f l u e n c e o n a
g e o t h e r m a l s y s t e m , f o r example by c a u s i n g a change from a h o t water t o a steam
dominated f o r m a t i o n .

1.2 THE WAIRAKEI GEOTHERMAL R E G I O N

The P h y s i c a l System

The New Z e a l a n d g e o t h e r m a l r e g i o n s l i e i n a l o n g t r o u g h e x t e n d i n g r o u g h l y
NE-SW a c r o s s t h e N o r t h I s l a n d , a p p r o x i m a t e l y f i v e k i l o m e t r e s wide and d e e p , f i l l e d
w i t h v o l c a n i c d e b r i s and c o n t a i n e d a t t h e s i d e s betwaen r , e l a t i v e l y impermeable walls
of non- fragmented i g n i m b r i t e as i n f i g u r e 1 . 2 . 1 .

FIGURE 1.2.1 - Sketch of problem configuration - New Zealand geothermal system. ,

I t i s n o t known w h e t h e r t h e b a s e of t h e f o r m a t i o n i s impermeable o r n o t , b u t
h e a t i s i n t r o d u c e d i n some way t h r o u g h t h i s boundary and i s t r a n s p o r t e d upwards by
c o n v e c t i o n o f t h e water t h a t f i l l s t h e i n t e r s t i c e s o f t h e p o r o u s r o c k , or by c o n d u c t i o n
through t h e m a t e r i a l i t s e l f . The o r i g i n i s u n c l e a r b u t i t h a s been s u g g e s t e d by
Mercer (1973) t h a t t h i s h e a t i s d e r i v e d from t h e g l o b a l t e c t o n i c s a s s o c i a t e d w i t h t h e
t r e n c h system. On r e a c h i n g t h e s u r f a c e o f t h e e a r t h t h i s h e a t p r o d u c e s h y d r o t h e r m a l
phenomena s u c h a s s t e a m i n g g r o u n d , f u m a r o l e s and h o t s p r i n g s .

2
 The fluid which saturates porous geothermal beds may be water, steam, brine
or a mixture of each. The Wairakei system is generally known as a hot water system
(Kruger 1973)) since the water in it holds only a small amount ((<
< ( 1 2 0 0 0 ppm)
) of
dissolved mineral salts and the steam phase is restricted to the surface region down
to approximately 4000 m below ground. Above this region there are periodic surface
phenomena such as geysers and hot lakes which occur due to interactions between the
steam and water r phases, but the overall convective pattern of the system is mainly
influenced by the larger scale motion which originates at greater depth between 2 and
5 km
m at which the fluid is wholly liquid. Below the shallow two phase region are
the artificial flows produced by withdrawal of fluid through boreholes and it is
expected that these effect the overall flow more significantly.

Model Systems

When creating an experimental or numerical model of a physical system it is

necessary to be able to accurately represent all the essential features of the actual
prototype. Unfortunately this is not usually possible in geothermal applications
since the region of interest extends so far below ground that its exact nature is not
directly determinable. The Wairakei geothermal field is the most significantly
developed hot water system in the world and as such has been the subject of various
theories which differ in their explanation of the structure of the system. Some
studies restrict the region of interest to the shallow surface system and use averaged
parameters to represent the lower convecting flows - Mercer ( 1 9 7 3 ) is an example of
such a "reservoir"" model. This kind of approach may be used to represent, in plan
detail only, the response of the upper geothermal reservoir under a given set of
lower reservoir conditions. The behaviour of the lower system, which may also be
effected, is not modelled at all (Mercer 1 9 7 4 ) . A better representation of the
system requires the modelling of the convection throughout the region, right down to
the source of the geothermal energy. Donaldson ( 1 9 6 2 ) initially proposed a two layer
model system consisting of a homogeneous porous layer overlying a homogeneous imper-
meable layer heated from below. Later ( 1 9 6 8 and 1 9 7 0 ) he refined this model to a
permeable channel representation which was then more specifically relevant to the
upper reservoir. Such a representation presupposes the existence of a bed of
material that is anisotropically permeable, with greater permeability in the horizontal
direction. Such beds are evident in the geological studies of Healy and Hochstein
( 1 9 7 3 ); ; these are at shallow levels and are fed with preheated water from a deeper
system. A layer representation has also been used for a larger scale model by
McNabb ( 1 9 6 5 ) who envisaged the deep system as having its base above a lava lake or
similar magmatic heat source. Elder ( 1 9 6 6 a ) has presented a comprehensive review
of several possible configurations for both the reservoir and deep systems. The
current investigation is concerned with the modelling of the deep system.

The Current Model

Bearing in mind the configuration of figure 1.2.1, the model region is

chosen initially to represent an infinitely long trench of rectangular cross-section,
completely filled with a homogeneous and isotropic porous medum which is wholly
saturated with water. The heat source at the base of the region could be of
arbitrary shape, as has been considered by Taunton and Lightfoot ( 1 9 7 0 ), but because
of the lack of information about the heat source this is an unnecessary generality
and the source is more usually taken to be flat. The output of geothermal heat is

3
either by withdrawal of hot fluid through wells or by l o s s at the surface to the
surrounding atmosphere. A s has generally been the case these upper and lower
boundaries are taken to be horizontal, constituting a type of system labelled by
Elder (1966a)) as a “wet-convector”. Inclined systems have been studied by Bories,
Combarnous and Jaffrenou (1972),, Kaneko, Mohtadi and Aziz (1974) and Bories and
Combarnous (1973) but are not considered here.

Since it is unknown whether or not fluid flows in through the base of the
formation, both possibilities are modelled. If a magmatic intrusion or lava lake is
viewed as being covered by a supercritical gas sheath that constitutes a phreatic
surface, then liquid transfer across the boundary is prohibited and the lower boundary
may be assumed to be impermeable. The production of magmatic water or the large
scale introduction of meteoric water through major horizontal fissures as is proposed
by Donaldson (1974) constitutes the alternative configuration in which the base is
permeable. Following Wooding (1957),, the enclosing vertical boundaries are assumed
to be impervious and also relatively non-conducting (Elder 1967a).

1.3 PAST WORK

Convection in a Porous Medium

The classic configuration which has been the most commonly modelled is the
box or infinite slab of porous material heated uniformly from below. The initial
onset of convection in such a region was the subject of early analytical studies by
Horton and Rogers (191r5) and by Lapwood (1948)) and later experimental and numerical
studies by
y Wooding (19571, Donaldson (1962), McNabb (19651, Katto and Masuoka (1967),
Elder (1966a, 1967a & b), Bories and Thirriot (19691, Combarnous (1970), , Palm, Weber
and Kvernvold (1972),
, Holst and Aziz (1972a) and Yen (1974)) who all derived steady
solutions to the problem. However Combarnous and Le Fur (1969) discovered in their
experiments that there is a second mode of convection at higher values of the defining
dimensionless parameter - the Rayleigh number (see section 2.1) - and this second
mode was later reported by Caltagirone, Cloupeau and Combarnous (19711, and
Combarnous and Bia (1971),, to be a fluctuating pattern of two dimensional rolls.

This second mode has yet to be observed in numerical solutions, for one of
three reasons. Firstly, the earlier numerical studies (for example Donaldson (1962),
Elder (1966a, 1967a)) presupposed a steady solution and did not include the transient
terms in their equations of motion. In view of the fluctuating solutions which were
discovered later, and of Elder’s own admission (Elder 1967a) that unsteady convective
instabilities cannot be examined under the assumption of quasi-steadiness, these
analyses may seem inappropriate. However this is not so as due to limitations of
the numerical methods used the only flows studied were under the benevolent
conditions of a low Rayleigh number for which steady flows do occur. When the flows
are moree vigorous the numerical procedures considered up to the present time become
unstable due to the false production of kinetic energy and this is the second reason
why the range of numerical solutions has been s o restricted, Thus the transient
representations of Elder (1967b and 1968)) and Holst and A z i z (1972b1 which use
similarnumerical methods could not have produced solutions in the unsteady mode,
Thirdly, despite the limitations of the techniques, the algorithms and computa-
tional machinery
y still would not have been fast enough to develop the solutions
for a duration sufficient for the nature of the unsteady effects to become clearly
apparent.

Various linear and non-linear stability analyses have been performed by

Westbrook (1969),, Beck (1972),
, Busse and Joseph (1972)) and Gupta and Joseph (1973)
but once again these authors either ignored transient effects or considered a range
of conditions in which they could not occur. The more recent analysis by Straus
(1974) attempted to explain the transition observed by Combarnous and Le Fur (1969)
as a transition from two-dimensional to three-dimensional flows, however it seems
apparent that Straus
s (1974) did not consider the results of Caltagirone et a1 (1969)
since he too specifically precludes oscillatory effects. Similarly the experimen-
tal conditions of Yen (1974) duplicate those of Caltagirone et a1 (1969) ) but no
reference is made to this earlier investigation and no mention of the form of the
solution is reported.

Therefore to avoid the limited applicability of these earlier solutions,

this study uses more sophisticatedd numerical procedures that are based on the
transient equations of motion and which are stable enough and fast enough to allow
lengthy solutions under a range of conditions that may be expected to occur in
geothermal regions (including those under which unsteady flows are evident).

Convection of a Newtonian Fluid

Since a unified appraisal of the set of closely interrelated problems

considered in this work has not previously appeared it is expedient to draw
parallels with convection in a fluid layer (the Benardd problem) which has been more
widely reported. Corresponding steady solutions have been obtained by Fromm (1965),
Veronis (1966),
, Plows (1968) and Brown (1973) and the special considerations of non-
uniform heating, temperature dependent viscosity and horizontal temperature gradient
have all been investigated in the Benard problem by Weber (1973),, by Torrance and
Turcotte (1971) and by Wirtz, Briggs and Chen (1972)) and Gill and Kirkham (1970)
respectively. Of these three situations only the last has previously been fully
considered in the porous medium problem.

The stability of this and closely related problems has been considered
for layers by Chandrasekhar (1961),, Gill and Davey (1969),
, Willis and Deardorff
(1967 E 1970),
, Krishnamurti (1970 a&b),, Busse and Whitehead (1971 & 1974)) and
Homsy (1973 E 19741, for channels by Birikh, Gershuni, Zhukhovitski and Rudakov
, for arbitrary containers by Joseph (1971) and generally by Gebhart (1973).
(1972),
These investigators are all notable in that collectively they observed that the mode
of convection may change completely with variation of the defining parameters, that
one or more of the higher modes may be unsteady or oscillatory, and that under
certain circumstances the oscillation may take the form of paired disturbances rising
in the convective plane.

Numerical studies of the Benardd problem have been somewhat more successful
than their porous medium counterparts. A forerunner of instabilities that were
discovered later was perceived in the numerical study of Elder (1966b)) in the form
of “secondary flows” which he speculated might have been due to disturbances

5
inherent in the numerical method exciting other modes of the system. This was
confirmed some time later by the comprehensive numerical investigation by Moore
and Weiss (1973) who used more advanced numerical techniques and obtained an
oscillatory solution in a square two-dimensional region heated from below. The
generation of disturbances that they observed is very similar to that obtained in
this work and their explanation of the mechanism emphasises the same processes,
Similar explanations of analogous oscillatory behaviour in simple loop models have
been proposed, first by Keller (1966) then more fully by Welander (1967) who later
(1971) extended the concept to include similar effects in a stratified fluid layer.

The numerical analysis of the convection equations has more commonly been
considered for the Benard rather than the porous medium problem, although these
studies provide the basis of the numerical methods that are used here. The
representation of the advection terms in the heat transport equations (see section
4.2) has been studied variously by Crowley (1967), Wilkes and Churchill (1966),
Torrance (1968), and Fromm (1969 a & b) and these analyses are directly applicable
to the problem at hand since the same terms appear (and are equally difficult to
represent).

Although results for the Bbnard problem are not directly applicable to
convective flow in a porous medium it is interesting to note the similarities and
differences as they arise, since suggested mechanisms for processes in either
situation can often indicate lines of approach to be followed in the other.

1.4 THE SCOPE O F THIS WORK

A mathematical formulation of the modelling of flow in geothermal regions

is proposed in chapter 2 , including transient flows, non-uniform heat input at the
base of a system, temperature dependent viscosity and a variety of fluid withdrawal
conditions. A single, exact model which encompasses all the relevant features of
a geothermal region is impractical due to the difficulty in determining the details
of real system. Therefore a r a n g e of models is proposed to include more than one
possible configuration.

It is demonstrated experimentally in chapter 3 that the flow in the two-

dimensional model r e g i o n described in section 2 . 2 may undergo irregular fluctuations
or regular oscillations under certain conditions. Although the first of these
effects has been observed previously in the experiments of Caltagirone et el. (1969)
there is a later and seemingly contradictory experimental study by Yen (1974) who
reported steady solutions. To resolve this uncertainty a different approach is
taken and a numerical representation used. To investigate the unsteady effects
it is necessary to use the transient form of the equations (derived in section 2.1)
and examine the development of the flow over a long period in time, which most
earlier studies have neglected to do. For this purpose a set of numerical
techniques is derived in chapter 4 that is sufficiently rapid to allow lengthy
solutions. These techniques rely on the energy conservation properties of the
Arakawa differencing schemes (Arakawa 1966) to avoid the numerical instabilities
that restricted earlier studies into investigating cnly relatively slow flows.

b
 In chapter 5 the numerical verification of the previous experimental
results is produced - the apparent contradiction in the earlier works is due to
the existence of alternative steady and unsteady solutions under the same boundary
conditions (but different initial conditions). The regular oscillatory solutions
of chapter 3 are also confirmed and explained,

The regeneration of water in a geothermal system by seepage of ground-

water has a prominent influence on the flow and this concept is included in the
numerical models in section 5 . 4 (for the first time in a transient analysis). The
inclusion of these recharge conditions is necessary for the representation of
artificial withdrawal and reinjection in an exploited geothermal system, and
simulations of developed systems are performed in section 5.4.

Past models of geothermal regions usually considered the porous rock

formation to be saturated with pure water of constant viscosity. To examine the
advisability of making this assumption the effects of temperature dependent
viscosity is investigated in section 5.5, and the possibility of dissolved mineral
salts is allowed in the separate analysis of chapter 7 .

The flow equations may be solved in either two or three dimensions. The
two-dimensional approach permits beneficial economies in the usage of computational
resources but it is considered essential to extend the analysis to three dimensions
for at least part of the work (see chapter 6 ) to confirm the significanceof the
solutions generated in two dimensions.

The effects that these several conditions have on the overall flow in a
geothermal region are various and complex and is preferable to consider them
separately. The direct and precise modelling of the Wairakei geothermal region
lies slightly beyond this work, but the contiguous consideration of all the
individual effects represented here could simulate the system quite fully. By
presenting a range of solutions it is possible to obtain a more general understand-
ing of the fundamental mechanisms, before attempting to construct a single,
tenaciously accurate model in which separate effects may interact and mask
observation of the basic processes governing the behaviour of the system.

Having derived a solution technique with a specific application to

geothermal modelling it is interesting to apply it to a different situation also.
In chapter 8 the methods are modified slightly to allow representation of the
porous insulator problem mentioned in section 1.1. The results of this simpler
investigation are readily comparable with previous work on this problem, thus
permitting further evaluation of the viability, of the techniques. Furthermore
the flow mechanisms are related to those in the geothermal case and furnish some
interesting parallels (see chapter 9 ) .

7
Chapter 2 - MATHEMATICAL FORMULATION

2.1 THE EQUATIONS OF MOTION

The e q u a t i o n s g o v e r n i n g t h e c o n v e c t i v e f l o w of f l u i d t h r o u g h a p e r m e a b l e
material have v a r i o u s forms d e p e n d i n g on which of t h e r a n g e o f s i m p l i f y i n g assump-
t i o n s are i n v o k e d and which d e p e n d e n t v a r i a b l e s are used. Several d i f f e r e n t
r e p r e s e n t a t i o n s are p r e s e n t e d h e r e . B a s i c a l l y t h e system o f e q u a t i o n s r e p r e s e n t
four physical r e l a t i o n s - c o n s e r v a t i o n o f mass, momentum and e n e r g y a n d t h e
empirical equations of state. The working f l u i d i s assumed t o b e i n c o m p r e s s i b l e
a n d Newtonian, r e p r e s e n t i n g t h e f l o w o f p u r e water o r a t w o r s t a d i l u t e a q u e o u s
s o l u t i o n o f mineral salts (see c h a p t e r 7 ) .

C o n s e r v a t i o n o f Mass

The form of t h e c o n t i n u i t y e q u a t i o n r e l e v a n t t o f l o w t h r o u g h p o r o u s media

is t h a t g i v e n by A r a v i n and Numerov (1965), namely

where qi , as d e f i n e d by E l d e r ( 1 9 6 6 a ) and Wooding (1957), i s t h e mean flux v e l o c i t y

o f t h e f l u i d t h r o u g h t h e medium which f i l l s s p a c e xi a t time t .The mean f l u x
v e l o c i t y s h o u l d n o t b e c o n f u s e d w i t h t h e mean v e l o c i t y o f t h e f l u i d t h r o u g h t h e
i n t e r s t i c e s o f t h e p o r o u s medium. The f l u x v e l o c i t y i s s m a l l e r by a f a c t o r m , t h e
p o r o s i t y of t h e m a t e r i a l .

C o n s e r v a t i o n o f Momentum

The u s u a l form o f t h e N a v i e r - S t o k e s e q u a t i o n i s m o d i f i e d by t h e i n c o r p o -
r a t i o n o f t h e experimental Darcy's L a w (Yih 1 9 6 9 , p.379), replacing t h e viscous
terms w i t h t e r m s t h a t a c c o u n t f o r t h e f l o w r e s i s t a n c e o f t h e p o r o u s material:

Here Fi i s t h e buoyancy f o r c e , p i s t h e dynamic p r e s s u r e i n t h e f l u i d , p i s t h e

dynamic v i s c o s i t y o f t h e f l u i d and k i s t h e p e r m e a b i l i t y o f t h e medium.

8
 The Wairakei geothermal formation has significantly anisotropic
permeabilities with the horizontal permeability larger than the vertical
permeability by a factor of approximately 10. However thrbughout most of the
region these two properties are uniform so it is possible to use the same value
for both and adjust the horizontal and vertical length scales accordingly.
Beck (1972) has observed that the form (2.1.2) of the momentum equation is not
always satisfactory due to the disparity between the order of the space derivatives
in the equation and the usual number of boundary conditions, however for natural
convective problems where inertia effects may be ignored this difficulty does not
arise.

Conservation of Energy

Since the flow is slow and therefore predominantly determined by

differences between boundary and fluid temperatures and not by energy dissipation
effects, the appropriate form of the energy equation as given by Rubin (1974) is

=
at
+ xqi

where X is the ratio of the volumetric heat capacity of the fluid to that of the
saturated formation, and K I - is the thermal dispersion tensor. The thermal
LJ
dispersion tensor incorporates two properties of the system- firstly the molecular
diffusivity and secondly the intrinsic dispersivity of the porous medium which is
effectively a measure of the uncertainty that a particular small parcel of fluid
will keep to a theoretically averaged mean flow path as it moves in a tortuous
manner through the interstices of the material. The behaviour of a particle of
fluid moving through a porous material is similar to that of a particle in a
turbulent flow and the dispersivity may be envisaged as analogous to an eddy
diffusivity. Dybbs and Schweitzer (1973) have produced a more rigorous derivation
of the set of conservation equations for flow in a porous medium by defining the
flow in the interstices and then averaging throughout the material, thus avoiding
the use of the empirical Darcy’s Law (although this is a consequence in some cases).
They arrive at a concept which they term the “convective diffusivity” but which is
another representation of the intrinsic dispersivity. In a non-homogeneous
porous medium the flow may be vigorous in certain areas and consequently dispersion
effects surpass those of diffusion, so the thermal dispersion tensor must be
represented in its complete form, including both longitudinal and lateral fluid
dispersion. However for the very slow flows of natural convection in geothermal
areas, dispersion effects are small and the thermal dispersion tensor reduces to
the scalar molecular diffusivity K of the saturated porous material (Rubin 1974).
Elder (1966a) and Combarnous (1972) both propose that an estimate of this term may
be obtained from

where m is the porosity of the rock. In general the thermal diffusivity is

neither constant nor isotropic in geothermal areas, but for simplicity it is taken
to be so. At Wairakei the difference in horizontal and vertical permeabilities
means that diffusion may also be less in the vertical direction however as was
observed earlier this difficulty may be avoided by vertically ”stretching” the

9
solutions obtained using an isotropic diffusivity. The variation of thermal
diffusivity throughout the various geological formations which make up the system
is difficult to determine and since the exact geophysical structure is unknown
there is no other choice but to assume a constant value. Previous authors have
taken K to represent the thermal diffusivity of the medium alone, but Katto and
Masuoka (1967) showed that the form (2.1.4) is required to obtain agreement between
theoretical and experimental results. Combarnous (1972) has found experimentally
that the heat transfer coefficients are characteristic of the particular fluid and
medium, and it is hoped that the incorporation of both solid and liquid terms takes
account of this difficulty. The effect of the thermal conductivity of the satura-
ting fluid on the heat transfer in a porous medium has also been observed by
Maksimov and Stradomskii (1971).

The Equations of State

The fluid properties which may vary with temperature are the viscosity
and the density, and it is necessary to use empirical formulae to represent the
relationships between these quantities. Different representations have been used
in the past, most authors -
Horton and Rogers (1945), Lapwood (1948), Wooding (1957),
McNabb (1965), Donaldson (1962), Elder (1967 a & b) etc. - used a linear relation-
ship for density and assumed that the viscosity was constant. Elder ( 1 9 6 6 a ) has
considered temperature dependent viscosity but without recording what effect this
would have on the flow. For a temperature differential of 250°C the viscosity of
water varies by a factor of approximately 10, so the assumption of constant
viscosity is seemingly inappropriate in a geothermal context, and furthermore the
variation of density in this range is more closely quadratic. Thus the fluid
density is better described by

and the viscosity by

"0
v z - , (2.1.6)
l+bl(T-To)+b2(T-To)2+b3(T-To)3

where al is the more commonly used linear coefficient of thermal expansion of the
fluid.

The Boussinesq Approximation

It is usual in convection problems, as for example in Yih (1969 p.441),

Torrance (1968), Wooding (1957) and Nield (1968), to invoke the Boussinesq approxi-
mation by assuming that the variation of fluid density need only be considered in
buoyancy terms; in the inertia and continuity terms the density is taken to be
constant. For steady state analyses (e.g. Donaldson 1962) the Boussinesq
approximation is unnecessary as all the density terms may be combined into the
fluid velocity terms to create "mass transfer" equations. I n such a case the
kinematic viscosity replaces the dynamic viscosity and the mass flux replaces the
flux velocity, but otherwise the equations are essentially the same as they would

10
be if the Boussinesq approximation were used, Thus the net effect of the
Boussinesq approximation on the complete transient equations is an unimportant
change in the viscosity term (for water the dynamic and kinematic viscosities are
almost the same at most temperatures) and a misrepresentation of the terms

in the continuity equation.

If the change in temperature relative to the absolute temperature in any

part of the fluid is small, the change in density is an order of magnitude less and
can be neglected as far as continuity is concerned.

Since the thermal diffusivity is a function of thermal conductivity,

density and heat capacity, then it is a further consequence of the Boussinesq
approximation that this quantity is also constant.

Thus, invoking Boussinesq and substituting for the buoyancy terms, the
equations become

(2.1.8)

(2.1.9)

where the Laplacian Vz 2 x

2 + q 2 + 2 2 .

Order of Magnitude Analysis

At this stage the momentum equations may be simplified further after

comparison of the relative magnitudes of the separate terms. If U is a
representative velocity of the fluid passing through the interstices of the porous
material, and d is a length representative of those interstices, then for the
inertia, viscous, pressure and buoyancy terms respectively,

11
 The ratio of inertial to viscous terms is then of order !!?$! , and for
flows in which this parameter (the Reynolds number) is much less than unity, the
inertia terms may be neglected. This condition is satisfied for the slow percola-
tion of water through soil (Batchelor 1967 p.223) and also for natural convective
flows. Elder (1966a) observes that with a maximum flux velocity of IO-’ cm/sec
and a gap size of 1 mm, the Reynolds number is typically of the order It
should also be remembered that a low Reynolds number was a prerequisite f o r the
assumptions made earlier, that dispersion effects may be ignored and that Darcy’s
Law is valid.

As a consequence of Darcy‘s Law, the pressure and viscous terms are of

similar order of magnitude, and the appearance of convective flows in this problem
indicates that the buoyancy effects are capable of overcoming viscous dissipation
and therefore both pressure and buoyancy terms must remain in the momentum equation.
By this argument, equations (2.1.8) reduce to

P-Po
g.(--)
Po
+ 1
Po
gi ; + qi = 0 (2.1.10)

Non-Dimensionalisation

The equations may now be expressed in non-dimensional form, following a

similar analysis to that of Torrance (1968), by introducing the new variables

xi
xi -- --
a ’

T-T
e = O - -T-To
T1-To AT ,

after which the equations become

-
au
axi = 0

a2 a2 a*
where now V2 f
m2 + x? + E 2 .

12
 Introducing the relations

and expressing the density and viscosity variations in full,

pa = 1 - al(9.AT) - a2(9.AT)' ,

and V* = 3 - 6 3 ,

then the momentum equations become,

giala2AT v0aZ ui
ap
Ri
= (8+ae2) --
k~ '
(2.1.16)
K2

,
Redefining P* =
w P

and rewriting (2.1.16), dropping the stars immediately for convenience, then

where R is known as the Rayleigh number and is defined by

This definition differs to that of earlier authors (e.g. Elder 1966a) due to the
inclusion of the term h which is necessary for the correct representation of the
heat content of the solid/liquid combination. Equations (2.1.11), (2.1.17) and
(2.1.13) are now the governing equations of the flow,

The Rayleigh Number

As R increases from below a critical value of Rc ( 4 r 2 for the region

considered by Lapwood 1948), the flow process changes from conduction to convection.
This transition has already received some attention from Katto and Masuoka (1967)
Westbrook (1969), Beck (1972) and Busse and Joseph (1972) and is not of particular
interest in this investigation. However the second transition discovered by
Combarnous and Le Fur (1969) in the range 240-280 has been shown by Caltagirone
et al. (1971) to mark the first appearance of the fluctuating convective state.

This transition is below the range of values of R that is considered

likely in geothermal regions (103-104). The uncertainty in this figure is due to
the difficulty in determining the medium properties at depth. A range of values
between 0 and 2000 is considered here as this includes both known regime transitions.

13
 I t s h o u l d be n o t e d t h a t t h e R a y l e i g h d e f i n e d by (2.1.18) i s b a s e d upon
t h e c o l d w a t e r v i s c o s i t y , a s h a s g e n e r a l l y been t h e case i n m o d e l l i n g o f h y d r o -
thermal systems. However t h e added c o m p l i c a t i o n o f v a r i a b l e v i s c o s i t y g i v e s t h e
R a y l e i g h number a new meaning; w i t h a t e m p e r a t u r e d i f f e r e n t i a l of 250°C t h e
e f f e c t i v e l o c a l R a y l e i g h number close t o t h e h e a t e r may be t e n times t h e a p p a r e n t
" c o l d water" v a l u e . Therefore i n solutions generated using a variable v i s c o s i t y
model t h e R a y l e i g h number c o n s i d e r e d must be much s m a l l e r t o m a i n t a i n t h e
c h a r a c t e r o f t h e f l o w s ( t o m a i n t a i n c o n s i s t e n c y w i t h p r e v i o u s work a l l t h e v a l u e s
c i t e d above a r e f o r t h e c o n s t a n t v i s c o s i t y c a s e ) .

2.2 BOUNDARY CONDITIONS

T h e r e a r e two t y p e s of boundary c o n d i t i o n s t o be employed h e r e , t h e r m a l

c o n d i t i o n s ( t e m p e r a t u r e ) and f l u i d c o n d i t i o n s ( v e l o c i t y and p r e s s u r e ) . The f l u i d
b o u n d a r i e s a r e a l s o d i v i d e d i n t o two s u b - t y p e s , e n c l o s e d b o u n d a r i e s and t h o s e
t h r o u g h which f l u i d may p a s s f r e e l y ( h e r e a f t e r kncwn as recharge boundaries).

Temperature Conditions

Heat i n p u t a t t h e b a s e i s a c h i e v e d by u s i n g a n i s o t h e r m a l c o n d i t i o n

A t ground l e v e l h e a t i s l o s t by c o n v e c t i v e t r a n s f e r t o t h e a t m o s p h e r e
a t a r a t e p r o p o r t i o n a l t o t h e d i f f e r e n c e between t h e ground and a i r t e m p e r a t u r e s ,

where kc i s t h e c o n d u c t i v i t y of t h e f l u i d s a t u r a t e d material a n d h is the

c o n v e c t i v e h e a t loss c o e f f i c i e n t . Now f o r a s a t u r a t e d medium kc is typically
3 x cal/cm/sec/°C , f o r a h o r i z o n t a l ground s u r f a c e h is of t h e order of
2 x lo2 cal/cm2/sec/OC and i n g e o t h e r m a l r e g i o n s t h e a v e r a g e t e m p e r a t u r e g r a d i e n t
is o f o r d e r 5 x OC/cm. T h e r e f o r e t o m a i n t a i n a b a l a n c e between t h e two
t y p e s of h e a t t r a n s f e r i n ( 2 . 2 . 2 ) , t h e t e m p e r a t u r e d i f f e r e n c e between ground and
a i r must be e x t r e m e l y small. I t i s found i n c a l c u l a t i o n s u s i n g t h e e x a c t
1
c o n d i t i o n t h a t f o r a c l o s e d s u r f a c e t h e maximum t e m p e r a t u r e g r a d i e n t a T y = l i s
s u c h t h a t (T-To) i s a t most o f o r d e r lo-' OC w h i l e f o r r e c h a r g e s u r f a c e s it may
reach OC. T h e r e f o r e it i s s a t i s f a c t o r y t o employ a s i m p l e r boundary
c o n d i t i o n T=To or e = O , as i f t h e a t m o s p h e r e were a n i n f i n i t e h e a t s i n k .

The impermeable s i d e walls c o n f i n i n g t h e porous r e g i o n a r e o f r e l a t i v e l y

low t h e r m a l c o n d u c t i v i t y , p a r t i c u l a r l y f o r f l o w s i n which c o n v e c t i v e h e a t t r a n s f e r
may be e x p e c t e d t o d o m i n a t e , as i s t h e c a s e f o r R a y l e i g h numbers above Rc .

14
 T h e r e f o r e t h e s e v e r t i c a l walls a c t a p p r o x i m a t e l y l i k e i n s u l a t e d b o u n d a r i e s
and t h e t e m p e r a t u r e g r a d i e n t normal t o t h e boundary i s t a k e n t o be z e r o . This zero
g r a d i e n t c o n d i t i o n a l l o w s t h e s e b o u n d a r i e s t o be c o n s i d e r e d l i n e s o f symmetry i n
which c a s e t h e r e g i o n may be imagined t o be e x t e n d e d i n e i t h e r d i r e c t i o n normal t o
t h e boundary by t h e a d d i t i o n o f " m i r r o r - i m a g e s " o f t h e f l o w w i t h i n t h e s e b o u n d a r i e s .

Enclosed Boundaries

The s i m p l e s t and most commonly c o n s i d e r e d model i s t h a t h a v i n g c l o s e d

boundaries. I n t h i s c a s e t h e flow of fluid i n t o o r o u t o f t h e problem r e g i o n i s
p r o h i b i t e d by r e q u i r i n g t h a t t h e v e l o c i t y normal t o a l l b o u n d a r i e s be z e r o . This
c o n f i g u r a t i o n i s n o t a n a c c u r a t e r e p r e s e n t a t i o n o f W a i r a k e i o r o f t h e many o t h e r
r e a l s y s t e m s where g r o u n d w a t e r f l o w d o e s o c c u r n e a r t o t h e s u r f a c e and p o s s i b l y
also a t t h e b a s e o f t h e f o r m a t i o n . However below t h e s u r f a c e t h e Waiora a q u i f e r
a t W a i r a k e i i s o v e r l a i n by t h e less p e r m e a b l e Huka F a l l s a q u i t a r d , so t h e c l o s e d
boundary model d o e s have some p r a c t i c a l f o u n d a t i o n . I t s main u s e however i s i n
p r o v i d i n g a comprehensive a r r a y o f s o l u t i o n s t h a t may b e compared t o t h e many
e a r l i e r e x p e r i m e n t a l a n a l y s e s for s u c h a r e g i o n ( i n p a r t i c u l a r t h e u n s t e a d y
s o l u t i o n s o f C a l t a g i r o n e e t a l . 1971), and which p r o v i d e an i n s i g h t i n t o t h e f l o w
processes without unnecessary complications.

Recharge B o u n d a r i e s

F l u i d may en t e r or l e a v e g eothermla1 arseas by e i t h e r n. a t u r'a1 or a r t i f i c i a l

processes. A t W a i r a k e i t h e r e i s s i g n i f i c a n t n a t u r a l d i s c h a r g e and r e c h a r g e o f
s u r f a c e w a t e r , a t a r a t e o f 440 k g / s e c ( F i s h e r 1964). The a r t i f i c i a l removal o f
f l u i d t h r o u g h b o r e h o l e s t a k e s p l a c e a t r a t e 4-5 t i m e s t h a t o f t h e n a t u r a l d i s c h a r g e
a n d p r o v i d e s t h e f u n d a m e n t a l means o f e n e r g y r e t r i e v a l from t h e s y s t e m . With s u c h
p r o m i n e n t f l o w s as t h e s e it i s d e s i r a b l e t o a v o i d t h e r e s t r i c t i o n o f a c l o s e d
boundary model and a l l o w f l u i d t o f l o w f r e e l y i n t o o r o u t o f t h e u p p e r boundary o f
t h e model. T h i s r e p r e s e n t s t h e n a t u r a l c o n d i t i o n s and p r o v i d e s a r e c h a r g e s o u r c e
f o r t h e f l u i d s i n k s i n t h e models o f t h e e x p l o i t e d f i e l d t h a t a r e c o n s i d e r e d l a t e r .
Such a n a t u r a l r e c h a r g e c o n d i t i o n h a s b e e n modelled p r e v i o u s l y i n t h e q u a s i - s t e a d y ,
low R a y l e i g h number a n a l y s i s by E l d e r ( 1 9 6 7 a ) and a more a r t i f i c i a l c o n d i t i o n by
Donaldson (1962) who c o n s i d e r e d a s p e c i f i c s u r f a c e o u t f l o w .

The r e c h a r g e c o n d i t i o n i s i n c o r p o r a t e d by s p e c i f y i n g t h a t t h e p r e s s u r e
a t t h e s u r f a c e i s u n i f o r m ( i . e . a t m o s p h e r i c ) and it i s t h e n a c o n s e q u e n c e o f
D a r c y ' s L a w ( e q u a t i o n 2.1.2) t h a t t h e v e l o c i t y t a n g e n t i a l t o t h e h o r i z o n t a l s u r f a c e
is zero.

15
2.3 TWO-DIMENSIONAL REGIONS

Justification

As has been described in section 1.2, the Wairakei geothermal system is

situated in a trench of porous rock that is much longer than it is wide or deep.
Taking the length of the rectangular trench model to be infinite, and assuming that
the flow is unvarying in this dimension, allows the governing equations to be
redefined in only two dimensions, although this must be done with caution. Beck
(1972) has used an energy method to show that at low Rayleigh numbers just above the
critical value, the most favoured mode of flow in a box of porous material would be
two-dimensional r o l l s . This has been confirmed for long regions by Holst and Aziz
(1972b) at higher Rayleigh numbers, but it is not certain that this would always be
so. For wider porous layers Bories, Combarnous and Jaffrenou. (1972) have found
experimentally that as the Rayleigh number is increased the preferred mode is
initially polyhedral cells, and only later as the Rayleigh number increases above 2 8 0
are purely two-dimensional effects (unsteady rolls) observed. These fluctuating
two-dimensional rolls have been frequently reported in three-dimensional experimental
regions by various French authors - Caltagirone et al. (1971), Combarnous and Bia
(1971), Bories, Combarnous and Jaffrenou (1972) and Bories and Combarnous (1973), and
as these unsteady solutions are of particular interest in this work it is legitimate,
at least for the initial appraisal of the problem, to use a simplified two-
dimensional form of the governing equations. Three-dimensional effects undoubtedly
do occur in both the real situation and in theoretical models, as has been shown
numerically by Holst and Aziz (1972b), experimentally by Bories and Thirriot (1969)
and analytically by Straus (1974), therefore for completeness a three-dimensional
exploration of the two-dimensional solutions is necessary. However, the two-
dimensional flows are more simply and more economically simulated and therefore have
been more extensively studied here than the three-dimensional flows.

The Governing Equations

The equations of motion written in two-dimensional form are

E
ax
, aay
-v = o ,

a p -_ U
-
ax -m ’

-
ap
ay
= (e+aeZ) - B
V
o ’

and = V28 - A(U + V ae ) ,

m
aT ax

a2 a’
where now Q2 E
R2,z2 *

The pressure may be eliminated from (2.3.2) and (2.3.3) by cross-differentiation and
subtraction, reducing the equations to

16
 Now, the continuity equation ( 2 . 3 . 1 ) may be satisfied identically in the
usual way (Yih 1969 p . 1 4 ) by defining a stream function JI , such that

2.3.6)

after which (2.3.5) becomes

Stream Function Equation Reformulation

Although in this form (2.3.7) and (2.3.8) are a suitable pair of governing
equations for the flow, it turns out that the products of first derivatives in
(2.3.7) are in a form that is difficult to represent satisfactorily numerically (see
section 4.2) so it is beneficial to reformulate them replacing (2.3.7) by coupled
equations in and P .
The non-linear terms

in (2.3.7) originate from the term

in (2.3.5) which may be rewritten in Jacobian form as

by substitution of U and V from (2.3.2) and (2.3.3). If (2.3.2) and (2.3.3) are
differentiated and added then, by using (2.3.1) also, a single equation for the
pressure is obtained,

If P is now scaled by a factor A/R then (2.3.7) finally becomes

17
 The pressure is defined by

and the same heat transport equation (2.3.8) completes the governing set of three
equations.

The Constant Viscosity Model

A special case of these equations is the simple linearly dependent density

and constant viscosity model which has a2 zero and bl,b2 and b3 all zero. This
means that a-0 , B(81-1 and B ’ ( 8 ) = 0 , therefore equation (2.3.7) reduces to

ae
v2JI = - a (2.3.11)

and there are just two governing equations again since the heat transport equation
(2.3.8) is unchanged and the pressure equation (2.3.10) is now no longer needed since
the reformulation of the stream function equation (2.3.7) is unnecessary.
Alternatively the stream function may be eliminated in favour of a pressure formula-
tion although this is harder to represent numerically. Whichever formulation is
used, this simpler model produces results that are qualitatively similar to the more
precise variable viscosity model (see section 5.5) which requires a more complicated
numerical solution technique.

Thermal Boundary Conditions

The heat input distribution at the lower boundary is represented as a

fraction f of the base held at an elevated temperature 8=1 and the remainder at
the lower temperature 8 3 0 .Thus the form of the thermal boundary conditions
introduced in section 2.2 that is relevant to two-dimensional regions is

(2.3.12)

(2.3.13)

on the horizontal boundaries and

(2.3.14)

on the insulated vertical boundaries.

Fluid Boundary Conditions

As a consequence of (2.3.6), the zero normal velocity condition on

impermeable boundaries implies that the stream function is constant (arbitrarily
zero) ,
$ = 0 on all closed boundaries.

Furthermore the zero tangential velocity on recharge boundaries implies that the
gradient normal to these boundaries is zero,

3 = 0 on (horizontal) open boundaries.

Pressure Boundary Conditions

When using the variable viscosity model a set of pressure boundary

conditions is also required, for the solution of (2.3.10). These conditions are
derived from the substitution of the fluid boundary conditions into (2.3.2) and
(2.3.3) and are

g=o on vertical boundaries,

and P = O on recharge horizontal boundaries.

Boundary conditions for a simple, enclosed, two-dimensional, constant

viscosity model are illustrated in figure 2.3.1.

t
st.0

$4.0

19
 Taking the divergence of (2.4.2) and substituting continuity (2.4.1)

v2p = R ae
x aY,
o r scaling the pressure by ,

The velocity vector may be eliminated between (2.4.2) and (2.4.3) such that

These last two equations are then the governing pair for the problem, in a form
described here as the pressure f o r m u t a t i o n . F o r reasons which are explained later
(section 4.2) this form is not always satisfactory for numerical representations.

The Vector Potential Formulation

Due to the solenoidal form of I! in (2.4.1), there is an alternative

formulation described by Holst and A z i z (1972b), in which a vector potential 9
may be introduced such that

x
K v = v.9 (2.4.6)

Then, taking the curl of equation (2.4.2),

Now, from its definition 9 is arbitrary to the gradient of a scalar,

hence it is possible to specify

V.? = 0

in the identify VxVxt = v(v.t) - v*? ,

whence VZf
ae ,
= (E 0, - ae
=I. (2.4.7)

Boundary Conditions

As in the two-dimensional formulation the boundary conditions on the

pressure are, for closed boundaries

where 5 i s an outward pointing normal, and for recharge boundaries

P = O . (2.4.9)

21
The Heated Fraction f

The v a l u e o f t h e p a r a m e t e r f is s i g n i f i c a n t i n t h i s i n v e s t i g a t i o n as
t h e v a r i a t i o n of t h e l o w e r t e m p e r a t u r e d i s t r i b u t i o n h i g h l i g h t s t h e i n f l u e n c e s t h a t
boundary c o n d i t i o n s have on t h i s p a r t i c u l a r t y p e o f flow problem, These e f f e c t s
h a v e n o t so f a r been r e p o r t e d by o t h e r a u t h o r s . When f=l t h e boundary i s u n i f o r m l y
h e a t e d , which i s t h e c l a s s i c c o n f i g u r a t i o n c o n s i d e r e d f o r e i t h e r t h e e n c l o s e d o r
t h e s e m i - i n f i n i t e p o r o u s l a y e r by Horton and Rogers ( 1 9 4 5 ) , Lapwood ( 1 9 4 8 ) , Wooding
( 1 9 5 7 ) , Donaldson ( 1 9 6 2 ) , McNabb (1965), E l d e r (1966a, 1967 a & b , 1968), K a t t o a n d
Masuoka (1967), Westbrook (1959), Combarnous and Le Fur (1969), Combarnous (1970),
Bories a n d T h i r r i o t (1969), Combarnous and Bia ( 1 9 7 1 ) , C a l t a g i r o n e e t al. (1971),
Palm, Weber and Kvernvold (1972), Beck (1972), Busse and J o s e p h ( 1 9 7 2 ) , Holst and
A z i z (1972 a & b ) , Gupta and J o s e p h ( 1 9 7 3 ) , Combarnous and Bories (1974), Yen ( 1 9 7 4 )
and Straus (1974). I n t h i s f o r m t h e problem i s s y m m e t r i c a l and l e n d s i t s e l f w e l l
t o s e v e r a l d i f f e r e n t methods o f s o l u t i o n , and t h e a u t h o r s r e f e r e n c e d above have u s e d
a v a r i e t y o f e x p e r i m e n t a l , n u m e r i c a l and a n a l y t i c a l t e c h n i q u e s . Although t h i s
u n i f o r m l y h e a t e d problem h a s been e x t e n s i v e l y i n v e s t i g a t e d , t h e r a n g e o f s o l u t i o n s
i s not y e t exhaustive. In p a r t i c u l a r t h e contradictory experimental r e s u l t s of
C a l t a g i r o n e e t a l . ( 1 9 6 9 ) and Yen ( 1 9 7 4 ) s u g g e s t f u r t h e r e x a m i n a t i o n o f t h e problem.

D e s p i t e t h e a t t e n t i o n p a i d t o i t t h e u n i f o r m h e a t i n p u t boundary
c o n d i t i o n may be o f l i m i t e d p r a c t i c a l s i g n i f i c a n c e s i n c e t h e n a t u r e o f t h e p h y s i c a l
h e a t s o u r c e i s unknown, b u t may n o t be so b e n e v o l e n t l y u n i f o r m . I f t h e lower
boundary i s h e a t e d non- uniformly ( i . e . f < 1) t h e symmetry of t h e u n i f o r m problem
is l o s t and t h e r a n g e of s o l u t i o n t e c h n i q u e s i s r e s t r i c t e d t o n u m e r i c a l o r e x p e r i -
mental simulations.

2.4 THREE-DIMENSIONAL REGIONS

The Governing E q u a t i o n s - P r e s s u r e Formulation

The s o l u t i o n o f t h e f u l l t h r e e - d i m e n s i o n a l e q u a t i o n s n u m e r i c a l l y i s a
v e r y t i m e consuming p r o c e d u r e and it i s t h e r e f o r e e x p e d i e n t t o u s e t h e s i m p l e s t
p o s s i b l e form o f t h e g o v e r n i n g e q u a t i o n s . T h e r e f o r e t h e assumption o f c o n s t a n t
v i s c o s i t y i s made and t h e e q u a t i o n s (2.1.11), ( 2 . 1 . 1 7 ) and (2.1.13), rewritten i n
v e c t o r f o r m , become

0 . u-. = 0 ,

where i i s t h e u n i t v e c t o r p o i n t i n g v e r t i c a l l y upwards,

20
 I n t h e e n c l o s e d r e g i o n t h e boundary c o n d i t i o n s on t h e f l o w f i e l d a r e , as
b e f o r e , t h a t t h e v e l o c i t y normal t o t h e boundary i s z e r o . For i n s t a n c e on t h e
planes X=O,l t h e v e l o c i t y component U i s z e r o , which from e q u a t i o n ( 2 . 4 . 6 )
i m p l i e s t h a t p o t e n t i a l components o2 and $ 3 are l i n e a r f u n c t i o n s o f z and Y
respectively. F o l l o w i n g a r o u n d t h e r e m a i n i n g boundary p l a n e s a n d a v o i d i n g i n c o n s i s -
t e n c i e s on t h e e d g e s it t u r n s o u t t h a t $2 and b3 must b o t h be c o n s t a n t
( a r b i t r a r i l y z e r o ) on X=O,1 . Due t o t h e s o l e n o i d a l p r o p e r t y of it i s t h e n a
c o n s e q u e n c e t h a t t h e normal g r a d i e n t of must a l s o be z e r o . Thus f o r a c u b i c
r e g i o n t h e boundary c o n d i t i o n s on are,

ao
5x1 = o2 = o3 0 x = 0,1,

Y = 0,1, (2.4.10)

z = 0,1,

f o r c l o s e d b o u n d a r i e s , w h i l e on a h o r i z o n t a l r e c h a r g e boundary

As a consequence of e q u a t i o n ( 2 . 4 . 7 ) and t h e boundary c o n d i t i o n s ( 2 . 4 . 1 0 ) the

component o2of t h e v e c t o r p o t e n t i a l i s z e r o everywhere w i t h i n t h e r e g i o n . Thus
t h e f i n a l f o r m of t h e v e c t o r p o t e n t i a l e q u a t i o n s may now be w r i t t e n

2.5 FLUID SINKS AND SOURCES

S i n k s on V e r t i c a l B o u n d a r i e s

Should a f l u i d s i n k (or s o u r c e ) be p r e s e n t i n a t w o - d i m e n s i o n a l r e g i o n ,
f u r t h e r c o n d i t i o n s a r e n e c e s s a r y t o i n c l u d e it i n t h e model.The s i m p l e s t case i s
a s i n k o f s t r e n g t h q o n o n e of t h e v e r t i c a l b o u n d a r i e s , s a y a t a p o i n t (O,Yo), t h e n
t h e boundary c o n d i t i o n i s

$lx=o = - 9 Y o < Y ( l , (2.5.1)

r e s u l t i n g i n a normal v e l o c i t y o f

22
a n d a volume d i s c h a r g e o f

Clearly if JI i s s t i l l z e r o a t t h e t o p o f t h e v e r t i c a l boundary on which

t h e r e i s no s i n k , t h e n t h e c o n d i t i o n (2.5.1) i s o n l y v a l i d i f JI i s v a r i a b l e a l o n g
t h e t o p h o r i z o n t a l boundary. T h e r e f o r e it i s always n e c e s s a r y t o i n c l u d e a
r e c h a r g e boundary a t t h e s u r f a c e i f a s i n k i s p r e s e n t , i n order t o allow t h e
s a t i s f a c t i o n o f o v e r a l l c o n s e r v a t i o n o f mass.

S i n k s or S o u r c e s W i t h i n t h e B o u n d a r i e s

If t h e s i n k l i e s s t r i c t l y i n s i d e t h e r e g i o n , s a y a t a p o i n t (Xo,Yo), then
an a l t e r n a t i v e formulation i s required. Using t h e c o n s t a n t v i s c o s i t y model
e q u a t i o n s as a b a s e , b u t w i t h o u t i n t r o d u c i n g t h e stream f u n c t i o n , t h e n t h e g o v e r n i n g
e q u a t i o n s become:

where g(X,Y) r e p r e s e n t s flow t o t h e s i n k ,

m
ap = - u , (2.5.5)

ap = *R e - v , (2.5.6)

and ae = v z e
E - a [U ae +
x v =aeI . (2.5.7)

The boundary c o n d i t i o n s a r e s t i l l t h a t U i s z e r o on t h e v e r t i c a l
boundaries, V i s z e r o on t h e l o w e r h o r i z o n t a l boundary and P is z e r o on t h e
u p p e r ( r e c h a r g e ) h o r i z o n t a l boundary.

I n o r d e r t o accommodate t h e s i n g u l a r n a t u r e o f t h e f l o w t h e v e l o c i t y
f i e l d i s s e p a r a t e d i n t o p a r t s , one which d e f i n e s t h e f l o w o f f l u i d i n t o a p o i n t s i n k
i n a n i n f i n i t e s p a c e and a n o t h e r which b o t h r e p r e s e n t s t h e c o n v e c t i v e f l o w and
s a t i s f i e s t h e f i n i t e p h y s i c a l boundary c o n d i t i o n s . These two f i e l d s a r e s u p e r -
posed such t h a t

where u and v a r e t h e s i n k v e l o c i t i e s and obey t h e r e l a t i o n s h i p

23
The v e l o c i t i e s u‘ and yc must t h e n s a t i s f y

and be c o n s i s t e n t w i t h t h e m o d i f i e d boundary c o n d i t i o n s t h a t a r e a s u p e r p o s i t i o n of
t h e r e a l c o n d i t i o n s and t h e f a l s e boundary f l u x due t o u and v .
The stream f u n c t i o n may t h e n be d e f i n e d as b e f o r e , e x c e p t t h a t now

u* : EA 9
ay 3

and = - A
?
k
!
ax ,

a n d t h e f l o w e q u a t i o n becomes

The h e a t t r a n s p o r t e q u a t i o n ( 2 . 5 . 7 ) is then

The stream f u n c t i o n boundary c o n d i t i o n s a r e , on t h e t h r e e c l o s e d

boundaries

$I = g(X,Y) , (2.5.10)

and o n t h e t o p r e c h a r g e boundary

For a p o i n t s i n k o f s t r e n g t h q a t l o c a t i o n (Xo,Yo) the sink flow is

d e f i n e d by

Y-Yo
where tan0 = - .,
x-xo rd = (X-X~)’ + ( Y - Y ~ ) ’ , (2.5.13)

and g(X,Y) = -& ( -2

2
rO
- 2) . (2.5.14)

A t t h e s i n k p o i n t (Xo,Yo) , u and v a r e u n d e f i n e d and g(X,Y) h a s a s i n g u l a r i t y ,

so it is n e c e s s a r y t o t a k e s p e c i a l p r e c a u t i o n s when r e p r e s e n t i n g t h i s p o i n t . This
may be done by s p e c i f y i n g
at the singularity. These two relationships are consequences of the zero net
production of thermal energy and zero net increase in vorticity at the sink.
Alternatively this difficulty may be evaded, where finite difference methods are
used, by placing the sink at a location that lies strictly between nodes of the
grid.

25
Chapter 3 - EXPERIMENTAL SOLUTION I N
TWO-DIMENSIONAL REGIONS

The E x p e r i m e n t a l Analogy

The two- dimensional flow of f l u i d t h r o u g h a p o r o u s medium i s d i r e c t l y

a n a l o g o u s t o t h e f l o w i n a Hele-Shaw c e l l (Wooding 1960). I t i s shown by Yih
( 1 9 6 9 p . 3 8 2 ) t h a t t h e mean f l o w i n a Hele-Shaw c e l l w i t h p l a t e s e p a r a t i o n b p r o v i d e s
a s i m p l e a n a l o g y t o t h e mean s e e p a g e f l o w t h r o u g h a material w i t h p e r m e a b i l i t y
k = b /12. Given then. t h a t a f l u i d moves s i m i l a r l y t h r o u g h t h e two s y s t e m s , it
r e m a i n s t o e s t a b l i s h t h a t t h e y a r e a l s o t h e r m a l l y comparable. This i s n o t s t r i c t l y
so as h e a t i s i n e v i t a b l y l o s t t o t h e s u r r o u n d i n g s t h r o u g h t h e g l a s s p l a t e s of t h e
Hele-Shaw c e l l , i n t r o d u c i n g t e m p e r a t u r e g r a d i e n t s i n t h e t h i r d s p a t i a l d i r e c t i o n ,
a n d a l s o t h e h e a t t r a n s f e r between s o l i d and l i q u i d p h a s e s i s d i f f e r e n t i n t h e
p o r o u s medium. However, t h e v o l u m e t r i c r a t i o of s o l i d t o l i q u i d i s similar, and
by s c a n n i n g t h e Bele-Shaw c e l l w i t h a n i n f r a - r e d " Thermovision" camera it was
found t h a t t h e mean t e m p e r a t u r e t h r o u g h t h e t h i c k n e s s e s o f t h e g l a s s and t h e water
i s comparable w i t h t h e mean t e m p e r a t u r e a t a similar p o i n t i n a f l u i d s a t u r a t e d
p o r o u s medium a s r e p o r t e d by C a l t a g i r o n e e t a l . (1971). Thus a l t h o u g h t h e p l a t e
i n s u l a t i o n i s n o t c o m p l e t e and t h e a n a l o g y t h e r e f o r e o n l y a n a p p r o x i m a t e o n e , t h e
a p p a r a t u s i s u s e f u l f o r i n d i c a t i n g s a l i e n t f e a t u r e s o f t h e flow.

Experimental D e t a i l s

The c e l l i s h e a t e d a l o n g i t s b a s e by a c o p p e r h e a t i n g j a c k e t t h r o u g h which
h o t water i s p a s s e d a t a c o n s t a n t r a t e f o r t h e d u r a t i o n o f t h e e x p e r i m e n t . Two
i n p u t c o n f i g u r a t i o n s a r e c o n s i d e r e d , a f u l l y - h e a t e d and a h a l f - h e a t e d l o w e r b o u n d a r y ,
or f = 1 . 0 and f = 0 . 5 r e s p e c t i v e l y .

Ti-.e u p p e r boundary is a f r e e s u r f a c e and i s t h e r e f o r e a s t r e a m l i n e o f t h e

flow. S i n c e t h i s s u r f a c e d o e s n o t n o t i c e a b l y a l t e r s h a p e from i t s i n i t i a l r e s t
c o n d i t i o n d u r i n g t h e e x p e r i m e n t s , t h e u p p e r boundary i s a n a l o g o u s to t h e c l o s e d
h o r i z o n t a l s u r f a c e i n t h e models p r o p o s e d i n s e c t i o n 2 . 2 . Heat i s l o s t t o t h e a i r
s p a c e a t t h e t o p o f t h e c e l l and t h e s u r f a c e r e m a i n s a p p r o x i m a t e l y i s o t h e r m a l f o r
t h e d u r a t i o n of t h e e x p e r i m e n t s . Motion o f t h e water i n t h e c e l l i s d e t e c t e d by
t h e i n j e c t i o n o f dye t h r o u g h small h o l e s i n one o f t h e p l a t e s , g e n e r a t i n g s t r e a k l i n e s
which a r e u n f o r t u n a t e l y r a t h e r d i f f u s e due t o t h e g r a d u a l n a t u r e o f t h e f l o w . The
t e m p e r a t u r e t h r o u g h o u t t h e r e g i o n i s v i s u a l i s e d w i t h t h e i n f r a - r e d camera, g i v i n g
a d i r e c t r e p r e s e n t a t i o n o f i s o t h e r m s as i n f i g u r e 3 . 1 .

26
The Results

Two flows are reported, at Rayleigh numbers of approximately 1000 and

1600 for each of the two configurations. The uniformly-heated flows develop from
an intricate "proto-sub1ayer" similar to that described by Elder (1968) as is seen
in figure 3 . 2 , and later exhibit the fluctuating behaviour observed by Caltagirone
et a1.(1971), generating smaller pairs of cells between dominating circulations, as
in figure 3 . 2 and more clearly in figure 3 . 3 .

The flow changes character completely if only half the boundary is heated.
The pattern soon becomes largely unicellular and instead of irregular fluctuations
the behaviour is oscillatory, periodically generating "tongues" of fluid in the
ascending and descending regions of the flow. In figure 3 . 4 ( A to F) these tongues
may best be seen by observing the prominent triangular streakline in the lower left
of ( A ) . In (B) the triangle is depressed at the top as the descending tongue begins
to develop and impressed at the bottom as an ascending thermal forms over the heater.
The triangle is further distorted in (C) as the descending disturbance moves across
and down, while the thermal, now quite prominent, moves leftwards across the heater.
In (D) and ( E ) the upper tongue continues its downward flight while the lower one
reaches the left wall and begins to elongate up it. Finally in (F) the ascending
disturbance has shot rapidly up the left boundary and the descending one has been
completely dissipated. At this time the flow is at a similar stage to figure (A).
The period of this oscillation is approximately 9 0 0 seconds in the experiment, which
corresponds to a non-dimensional time of 0.003. The passage of two ascending
disturbances is seen in the streakline and isotherm plots of figure 3.5 at a Rayleigh
number of 1000 only in this case the descending fluctuations are not apparent.

These results confirm the earlier experimental work of Caltagirone et a1.

(1971) and appear to refute the assertion of Yen ( 1 9 7 4 ) that the flows are steady.
The reason for the contradiction is still not clear, particularly since these
experiments are neither exhaustive nor rigorous. Furthermore there is an additional
point which requires explanation, namely the reason why the solution is more regular
in the half-heated case. In the following chapters an alternative approach is used
to confirm the existence of these effects by representing the flows numerically.

27
Figure 3.1: The Hele Shaw Cell and the AGA Thermovision Camera and
Monitor, and a Typical Set of Thermographs. (This sequence
of isotherms follows that of figure 3.3).
Figure 3.2: Experimental Solution for Rayleigh Number of 1600 - Fully
Heated Boundary. Sequence recorded at 5 minute intervals.
 A B C

D E F

Figure 3.3: Experimental Solution (Streaklines) for Rayleigh Number of

1000 - Fully Heated Boundary.
 A B C

D E F

Figure 3.3 contd.: Experimental Solution (Isotherms) for Rayleigh

Number of 1000 - Fully Heated Boundary. See figure 3.1 for
the continuation of this sequence).
 A B C

D E F

Figure 3.4: Experimental Solution for Rayleigh Number of 1600 - Half

Heated Boundary.
 A B C

D E F

Figure 3.5: Experimental Solution (Streaklines) for Rayleigh Number of

1000 - Half Heated Boundary.
 A B C

D E F

Figure 3.5 contd.: Experimental Solution (Isotherms) for Rayleigh

Number of 1000 - Half Heated Boundary.
Chapter 4 - NUMERICAL METHODS

4.1 PREVIOUS NUMERICAL SOLUTIONS

Caltagirone et al. (1969) discovered the fluctuating flow in convection in

a porous medium experimentally, however all earlier and even later numerical studies
either found steady solutions or did not continue simulations long enough to
determine the true nature of the flow. As was observed in section 1.3, this was
largely due to limitations of the numerical techniques employed. The transient
analysis by Elder (1967b) is of particular interest since the flow region and the
equations of motion are very similar to those proposed here for an enclosed, constant
viscosity model. However the numerical techniques used by Elder (1967b) and also
those of Holst and A z i z (1972b) (which are described in Aziz and Hellums 1967),
namely a successive over-relaxation method for the solution of Poisson's equation
and centred differencing for the first derivatives in the heat flow equation, are
necessarily an iterative and comparatively slow procedure. Therefore until the
application of the more advanced numerical techniques in the present work, numerical
exploration into the type of flow examined by Caltagirone et al. (1971) has not been
possible. It is interesting to note here that in the Newtonian fluid layer problem
the numerical results of Moore and Weiss (1973) superseded those of Fromm (1965) and
Veronis (1966) in the description of an oscillatory regime, largely because of the
extensive and efficient numerical experiments performed. Moore and Weiss (1973)
quote computing times of 5 hours on an IBM 360/44 to perform 4 0 0 0 time steps of their
numerical procedure, but the equations here are simpler and similar runs (on a
Burroughs B6700) would take only 50 minutes.

4.2 FINITE DIFFERENCE METHODS

The numerical solution of convection equations is of interest to numerical

analysts a s well as to fluid dynamicists because of the difficulty in representing
them satisfactorily in difference form. Kreiss and Oliger (1973) have examined a
wide variety of techniques that are applicable to long time numerical integration,
including the finite difference schemes which are considered here.

35
The Advection Terms

Torrance (1968) gives a summary of several finite difference methods, both

explicit and implicit, with special consideration given to the parabolic heat
transport equation (in this case equation 2.1.13). The main difficulty in solving
this equation numerically arises from the representation of the non-linear terms.
In two dimensions these terms appear in Jacobian form

and are generally known as the a d v e c t i o n terms. Arakawa (1966) has explained that
a simple finite difference approximation using central differences causes numerical
instability due to the occurrence of aliasing errors. High order variations in the
fields being calculated cannot be resolved between the grid points of a finite
difference mesh and consequently some lower order variations are represented as
having an unrealistically high kinetic energy that physically would have been
dissipated by higher order effects. This kinetic energy can be "accumulated" by
the numerical procedure and gives rise to the aliasing instability. An example of
such an unstable approximation would be the central difference

where +i,j = $(iAX, jAY, nAT ) ,

and A X , AY and AT are the spatial and time increments respectively.

In this form the approximation of the advection term does not conserve the
kinetic energy of the system, and the numerical solution may produce unrealistically
vigorous flows. The false energy which appears in the advection term is partially
dissipated by the diffusion term in (2.1.13) and thus the effect on the solution is
governed by the relative magnitudes of these two terms. Since the advection is
multiplied by the Rayleigh number, at larger values of R the diffusion terms are
relatively weak. In fact the central differenced solutions are numerically
unstable for Rayleigh numbers above 200 and even below that figure have dubious
accuracy.

The Arakawa Schemes

To avoid aliasing errors Arakawa (1966) developed nine- and thirteen-point

representations of the Jacobian J which conserve the kinetic energy and mean square
temperature and which have a truncation error of similar order to the square and
fourth power of the spatial increment A X (and are accordingly known as the second-
and fourth-order Arakawa schemes - see appendix A ) . Unfortunately the complexity
of the representation on both the nine- and thirteen-point templates makes an
implicit implementation infeasible and it is therefore necessary to use an explicit
method.

36
 The expedience of using Arakawa methods has recently been questioned by
Orszag and Iraeli (1974), largely on the grounds of accuracy and efficiency. In
particular they suggest that the complexity of the Arakawa scheme means that it is
less efficient than a technique using centred differences but with a false dissipa-
tion term added, however clearly there is a danger here than in attempting to
dissipate the aliasing instability the physical instability will also be misrepre-
sented. Therefore the Arakawa technique is preferable in this case, especially if
it is made computationally efficient and the fourth-order scheme is used.

Non-Jacobian Non-Linear Terms

The Arakawa scheme has been derived for products of first derivatives in
Jacobian form, and is not generally extendable to non-Jacobian products, for
example the terms

which appear in the three-dimensional pressure formulation (equation 2 . 4 . 5 ) or the

terms

which appear in the two-dimensional variable viscosity model (equation 2.3.7).

The first of these may be avoided by using the alternative vector potential approach
(see section 2.4) and the variable viscosity equations may also be reformulated to
bring the derivatives into Jacobian form (as was done in section 2.3). However
this formulation is more time consuming since the number of governing equations must
be increased from two to three.

Alternative Spatial Difference Schemes

Historically the first of these is the "upwind" differencing scheme

similar to that of Runchal, Spalding and Wolfshtein (1969) which makes a one-sided
difference approximation to the first derivatives in (4.2.1) or (4.2.4), using
either forward or backward differencing depending on the direction of fluid flow at
the point under consideration. This scheme, with its smaller five-point template,
is readily implemented using an A.D.I. (alternating direction implicit) method
(Mitchell, 1969, p.60) but has neither the conservation properties nor the precision
of either of the Arakawa schemes, being only first-order accurate.

Another method of implementing the upwind differencing scheme is the

strongly implicit procedure (S.I.P.) presented by Curry (1974) who uses it with
central differences. This method is similar to A.D.I. in that it is an approximate
method of solving the complete matrix equation. However, as its name suggests,
S . I . P . is more strongly implicit as it solves the fully updated pentadiagonal
matrix. It does this by factorising the matrix approximately into tridiagonal

37
upper and lower triangular matrices and iterating until the product of the factors
approaches the true matrix. Although the upwind scheme is less accurate than the
second order central differencing methods it is actually less susceptible to
numerical instabilities. Therefore it can be used closer to the range of Rayleigh
numbers at which the flow is unsteady, however the value of doing this is question-
able.

A far more suitable approximation has been suggested by Kreiss and

reported in Orszag and Israeli (1974). This Kreiss differencing is fourth order
accurate and obtains first derivatives

by the solution of the tridiagonal system

where Do 8 ; is the central difference approximation to the derivative. This method

is "highly recommended" by Orszag and Israeli (1974) but unfortunately its stability
has yet to be examined and furthermore it is somewhat slower to compute than a
fourth-order Arakawa implementation. It is useful however in that it can represent
non-Jacobian terms such as (4.2.4) where the Arakawa schemes cannot.

A summary of the different space differencing methods considered is

presented in table I, comparing the computation times f o r calculations of a constant
viscosity model using a stream function formulation on a 3 3 x 33 mesh. The
solutions of a simple closed problem at a Rayleigh number of 2 5 0 using upwind
differencing, the second- and the fourth-order Arakawa methods are compared in
figure 4 . 2 . 1 to the solution which uses the spectral representation described in
section 4 . 3 .

TABLE I Spatial Differencing Techniques

Order Of
Differencing Time/calc.(secs) Advantages Disadvantages
Accuracy

Central Second 1.3 Simple, Unstable

Implicit
Upwind First 1.5 Implicit Unstable
Arakawa Second 1.3 Stable Explicit
Arakawa Fourth 1.4 Stable Explicit
Kreiss Fourth 2.2 Simple Stability
Unknown

38
 (a)

FIGURE 4.2.1 -
Comparison of spatial differencing techniques. Solution generated under the
same conditions using; (a) Second-order Arakawa differencing, ( b ) Fourth-order Arakawa
Isotherms (+ - 4.
differencing, (C) Upwind differencing,(d) Spectral Representation (at earlier time).
and Stream1 i nes (-

The solutions generated using the spectral representation and upwind

differencing methods have been illustrated in figure 4.2.1 before they reached a
steady state and are therefore at an earlier time than the other two solutions.
This is because the upwind scheme is unstable at this Rayleigh number at this early-
stage the scheme has not even preserved the symmetry of the solution and at later
times the results become wildly unrealistic. The spectral scheme is so time
consuming to use that only a relatively short stage of the flow development can ever
be simulated.

39
 The pressure formulation of the equations for the constant viscosity
system in the same simple test case has non-Jacobian non-linear terms and is
therefore best solved by the Kreiss scheme. Figure 4 . 2 . 2 compares the solutions
generated using second-order central differencing and fourth-order Kreiss differen-
cing. The solution by the more accurate method is identical to the Arakawa
solutions in figure 4.2.1 which'is a notable confirmation of the techniques, but
the central differencing method is unsatisfactory since it gives totally unrealistic
solutions even at the early stage at which the diagram is drawn and quickly becomes
unstable at later times.

--
I

t-----------
F i g u r e 4.2.2. - Comparison o f s p a t i a l d i f f e r e n c i n g techniques f o r non-Jacobian n o n - l i n e a r terms.
Left : Second-order c e n t r a l d i f f e r e n c i n g . R i g h t : Fourth- order Kreiss differenclng.
Isotherms (-) and pressure isobars (-----)

Time Differencing

After selection of a difference scheme to represent the spatial derivatives,

it is also necessary to approximate the time derivative in the heat transport
equation. The simplest time differencing scheme is the first-order accurate
forward difference

which has the advantage that it requires only two time levels f o r its calculation,
saving both computer time and memory. The second-order leap-frog method might seem

40
to be more acceptable, however it has been observed by Orszag and Israeli (1974) to
be numerically unstable when used with Arakawa differencing, This instability is
due to the solution at odd and even time steps becoming uncoupled, destroying the
conservation properties of the Arakawa scheme as it develops in time. An alterna-
tive second-order scheme that does not suffer this instability is the Adams-
Bashforth method (Lilly 1966) using three time levels,

which normally requires about twice as many time steps over a given period to
achieve the same accuracy as the leap-frog method (Orszag 1971b).

The fourth-order Runge-Kutta representation described by Carnahan, Luther

and Wilkes (1969, p.363) is also applicable, but requires approximately twice as much
memory space and is four times as time consuming as the forward difference (4.2.5),
since it requires four time levels for each calculation. A more efficient fourth-
order scheme is the Hamming predictor-corrector method, also given by Carnahan et al.
(1969, p.390), which is faster but requires as much memory and more intricate
programming than the Runge-Kutta method.

The time step A T must be very small to retain the semi-conservation

properties of the space differencing schemes. Therefore for a fixed finite time
interval there is little difference between the various time differencing schemes.
In fact, comparison of solutions f o r identical configurations using first-, second-
and fourth-order time differencing reveals that the choice of method is immaterial
over the number of time steps that are required to simulate the appearance of an
oscillatory or fluctuating solution (this is the "worst case" and may take a
maximum of 1000-1200 time steps). The maximum size of the time step in any
particular solution depends largely on the Rayleigh number, varying from at
R=200 to at R=1250. In chapter 8 the solutions are known to be steady so a
variable time step procedure, analogous to a relaxation process, is used, but this
is not suitable for the cases considered SO far in which the flows are likely to be
unsteady.

The Thermal Diffusion Term

The last remaining term in the heat transport equation is the thermal
diffusion term 9'8. This term is less significant than the others in terms of the
overall stability, having only a stabilising influence on the problem. However, if
the accuracy of the solutions is to be consistent, then this term should also
receive due attention. The standard five-point representation of the Laplacian

(4.2.7)

is second-order accurate and has been used almost universally, but where a fourth-
order scheme is used for the advection term, it is also possible to obtain similar
accuracy in the diffusion term. This may be done at little extra cost in computer
time since a large template of points has already been accessed by the program to
evaluate the Arakawa terms. Orszag and Israeli (1974) propose the use of the
d i f f e r e n c e formula

which h a s f o u r t h - o r d e r a c c u r a c y , a l t h o u g h i s o n l y s u i t a b l e f o r s q u a r e meshes
(AX = AY) when i n t h i s form.

The S t r e a m F u n c t i o n E q u a t i o n

Because o f t h e s i m p l e s h a p e of t h e r e g i o n and t h e compact f o r m of t h e

stream f u n c t i o n boundary c o n d i t i o n s , t h e s o l u t i o n of t h e e l l i p t i c e q u a t i o n ( 2 . 3 . 7 )
or ( 2 . 3 . 1 1 ) i s r e l a t i v e l y s i m p l e t o a c h i e v e , by a p p l y i n g t h e Buneman odd- even
r e d u c t i o n method d e s c r i b e d by Busbee, Golub and N i e l s o n (1970) which i s a d i r e c t ,
n o n - i t e r a t i v e scheme for t h e s o l u t i o n of P o i s s o n ’ s e q u a t i o n on a f i n i t e mesh of
points. This algorithm i s considerably f a s t e r than classical successive over-
r e l a x a t i o n t e c h n i q u e s and t h e r e f o r e f a c i l i t a t e s t h e r e p e t i t i v e s o l u t i o n o f t h e stream
f u n c t i o n which i s n e c e s s a r y for s i m u l a t i o n s o f l e n g t h y time d e v e l o p m e n t s .

The f i n i t e d i f f e r e n c e r e p r e s e n t a t i o n o f t h e L a p l a c i a n u s e s t h e f i v e p o i n t
scheme similar t o ( 4 . 2 . 7 ) . The t e m p e r a t u r e g r a d i e n t o n t h e r i g h t hand s i d e of
(2.3.11) i s t h e n e v a l u a t e d u s i n g c e n t r a l d i f f e r e n c e s which a r e o f similar second-
o r d e r accuracy. I t might a p p e a r t h a t t h e f o u r t h - o r d e r a c c u r a t e form o f t h e
L a p l a c i a n ( 4 . 2 . 8 ) c o u l d be used w i t h a c o r r e s p o n d i n g a p p l i c a t i o n of t h e K r e i s s
f o u r t h - o r d e r scheme t o t h e t e m p e r a t u r e g r a d i e n t term. T h i s c a n i n f a c t b e done
a f t e r m o d i f i c a t i o n o f t h e Buneman a l g o r i t h m (which i s s e t o u t i n f u l l i n a p p e n d i x B )
however t h e m o d i f i e d form i s o n l y one q u a r t e r as f a s t and t h e b e n e v o l e n t s i m p l i c i t y
and r a p i d i t y of t h e n u m e r i c a l s o l u t i o n i s d e s t r o y e d . The t y p e o f s o l u t i o n r e q u i r e d
h e r e would be r e n d e r e d i n f e a s i b l e u n l e s s limitless c o m p u t a t i o n a l r e s o u r c e s were
available.

The P r e s s u r e E q u a t i o n

The p r e s s u r e e q u a t i o n ( 2 . 3 . 1 0 ) i s a l s o a form o f P o i s s o n ’ s e q u a t i o n and

may be h a n d l e d by a n a d a p t i o n of t h e Buneman a l g o r i t h m u s e d p r e v i o u s l y . The
m o d i f i e d a l g o r i t h m t a k e s somewhat l o n g e r t o compute s i n c e t h e s o l u t i o n m a t r i x i s
l a r g e r due t o t h e Neumann boundary c o n d i t i o n s (normal d e r i v a t i v e s p e c i f i e d ) on t h e
pressure. F u r t h e r m o r e t h e s o l u t i o n o f P o i s s o n ’ s e q u a t i o n w i t h normal d e r i v a t i v e s
s p e c i f i e d on a l l b o u n d a r i e s i s n o t u n i q u e and it i s n e c e s s a r y t o a d a p t t h e a l g o r i t h m
t o explicitly s p e c i f y a t l e a s t one p r e s s u r e i n t h e r e g i o n ( s e e a p p e n d i x D ) . This
i s a n a l o g o u s t o p r o v i d i n g a gauge p r e s s u r e ,

42
The Governing E q u a t i o n s i n F i n i t e D i f f e r e n c e Form

To summarise, t h e c o m p l e t e f i n i t e d i f f e r e n c e form o f t h e e q u a t i o n s used

i n t h i s work f o r t h e s i m p l e c o n s t a n t v i s c o s i t y , two- dimensional model i s

(4.2.9)

(4.2.10)

where JiYj i s e v a l u a t e d u s i n g t h e f o u r t h - o r d e r Arakawa scheme, (Vz8)2 . is

e v a l u a t e d u s i n g t h e f o u r t h - o r d e r L a p l a c i a n (4.2.8), and e q u a t i o n ( 4 . 2 . 9 )
*!i s second-
o r d e r a c c u r a t e and u s e s c e n t r a l d i f f e r e n c e s t h r o u g h o u t .

Boundary C o n d i t i o n s

The s u b s t i t u t i o n o f boundary c o n d i t i o n s i s o n e f a c e t o f n u m e r i c a l computa-

t i o n s t h a t o f t e n r e q u i r e s c a r e , however f o r t h e s i m p l e problems w i t h no f l u i d s i n k s
o r s o u r c e s t h e r e i s no p a r t i c u l a r problem. The Buneman a l g o r i t h m f o r t h e s o l u t i o n
of t h e stream f u n c t i o n c a n be m o d i f i e d t o a c c e p t e i t h e r t h e c l o s e d , D i r i c h l e t
( J , s p e c i f i e d ) o r t h e recharge, Neumann boundary c o n d i t i o n s and r e p r e s e n t s t h e s e w i t h
t h e same a c c u r a c y as t h e s o l u t i o n i n t h e i n t e r i o r . The e x p l i c i t s o l u t i o n o f t h e
h e a t t r a n s p o r t a l s o a l l o w s d i r e c t s u b s t i t u t i o n of t h e i s o t h e r m a l c o n d i t i o n s ( 2 . 3 . 1 2 )
and ( 2 . 3 . 1 3 ) and t h e N
u
em
n
a i n s u l a t e d c o n d i t i o n s (2.3.14) are a c h i e v e d by i n v o k i n g
symmetry c o n s i d e r a t i o n s and i n c l u d i n g " mirror- image" t e m p e r a t u r e s from beyond t h e
boundary. The l a r g e number o f p o i n t s i n b o t h t h e Arakawa t e m p l a t e s r e s u l t s i n a
boundary r e p r e s e n t a t i o n t h a t i s a t l e a s t as a c c u r a t e as t h e s o l u t i o n a t i n t e r i o r
points.

It s h o u l d be n o t e d h e r e t h a t it is t h e i n c l u s i o n o f t h e s e image t e m p e r a -
t u r e s t h a t a l l o w s t h e u s e o f t h e Arakawa schemes c l o s e t o t h e v e r t i c a l b o u n d a r i e s .
However, one row i n from e i t h e r h o r i z o n t a l boundary t h e t h i r t e e n - p o i n t t e m p l a t e
a t t e m p t s t o r e f e r e n c e p o i n t s which l i e o u t s i d e t h e problem b o u n d a r i e s , and t h e r e f o r e
t h e a p p r o x i m a t i o n i s n o t p e r m i s s i b l e on t h e s e two rows ( u n l i k e t h e v e r t i c a l
b o u n d a r i e s t h e r e a r e no image t e m p e r a t u r e s beyond t h e h o r i z o n t a l b o u n d a r i e s ) .
T h e r e f o r e it is n e c e s s a r y w h e r e v e r f o u r t h - o r d e r Arakawa d i f f e r e n c i n g i s a p p l i e d t o
r e v e r t t o t h e s e c o n d - o r d e r , n i n e - p o i n t t e m p l a t e on t h e second and p e n u l t i m a t e
h o r i z o n t a l rows o f t h e mesh.

G r i d Mesh

The c h o i c e o f mesh s i z e r e q u i r e s some c a u t i o n as it h a s been o b s e r v e d by

Moore and Weiss ( 1 9 7 3 ) t h a t t o o c o a r s e a g r i d may s u p p r e s s u n s t e a d y b e h a v i o u r - should
i t be p r e s e n t . The Buneman a l g o r i t h m o p e r a t e s most e f f i c i e n t l y when t h e r e a r e Zk + 1
s u b - m a t r i c e s i n t h e s o l u t i o n e q u a t i o n (where k i s a n y p o s i t i v e i n t e g e r ) , which means
t h a t o n l y c e r t a i n s i z e s o f mesh a r e p e r m i s s i b l e . If t h e odd- even r e d u c t i o n is
p e r f o r m e d by c o m b i n a t i o n o f row s o l u t i o n s u b - m a t r i c e s t h e n t h e r e must be g k + 1 rows
i n t h e o r i g i n a l mesh ( f o r p r a c t i c a l problems t h i s i m p l i e s 1 7 , 33 o r 65 r o w s ) .

43
The r e d u c t i o n c o u l d e q u a l l y well be performed by combining column s u b - m a t r i c e s ,
a t least f o r t h e s i m p l e two- dimensional s t r e a m f u n c t i o n s o l u t i o n s , i n which case
t h e number of columns must be 2k + 1. The column s o l u t i o n h a s t h e a d v a n t a g e t h a t
it e n a b l e s c o n v e n i e n t i n c l u s i o n o f Neumann r e c h a r g e boundary c o n d i t i o n s o n t h e t o p
h o r i z o n t a l boundary w i t h o u t s i g n i f i c a n t l y r e d u c i n g t h e s o l u t i o n time, a l t h o u g h f o r
computer programs w r i t t e n i n FORTRAN t h e a r r a y s a r e u s u a l l y h a n d l e d more e f f i c i e n t l y
by t h e row s o l u t i o n method. It i s a l s o a d v a n t a g e o u s t o u s e a s q u a r e mesh ( A X = AY)
i n t h e f i n i t e d i f f e r e n c e schemes as t h i s o p t i m i s e s t h e c o m p u t a t i o n time o f t h e
Buneman a l g o r i t h m .

With t h e s e r e q u i r e m e n t s i n mind, it r e m a i n s t o choose a mesh s i z e by

s e l e c t i n g a value of k . G e n e r a l l y a more v i g o r o u s s o l u t i o n r e q u i r e s a f i n e r mesh,
so f o r v a l u e s of t h e R a y l e i g h number below 500 a 1 7 x 17 mesh is c o n s t r u c t e d (for
a s q u a r e , e n c l o s e d r e g i o n ) w h i l e above t h i s v a l u e a 3 3 x 3 3 g r i d i s u s e d . A still
f i n e r 65 x 65 mesh would b e n e c e s s a r y f o r f l o w s more e n e r g e t i c t h a n t h o s e m o d e l l e d
i n t h i s work, b u t i n t h e s e cases t h e s o l u t i o n p r o c e s s becomes l e n g t h y . The
r e c h a r g e and p o r o u s i n s u l a t o r c o n f i g u r a t i o n s r e q u i r e s p e c i a l c o n s i d e r a t i o n s i n t h e
choice of a mesh s i z e f o r t h e i r s o l u t i o n , and t h e s e a r e d e s c r i b e d l a t e r as t h e y a r i s e .

S i n k s and S o u r c e s

The a d d i t i o n o f a s i n k o n a v e r t i c a l boundary r e q u i r e s no s p e c i a l n u m e r i c a l
t r e a t m e n t as t h e r e i s o n l y a minor c h a n g e i n t h e boundary c o n d i t i o n s , which r e t a i n
t h e i r Dirichlet type. However when t h e s i n k i s l o c a t e d i n t h e i n t e r i o r , t h e a l t e r -
n a t i v e a n a l y s i s u s i n g e q u a t i o n s (2.5.8) and ( 2 . 5 . 9 ) must b e a p p l i e d , which r e q u i r e s
t h e e v a l u a t i o n o f t h e terms u , v , g , x
av and &.T h i s p r e s e n t s no g r e a t p r o b l e m
a s t h e s e a r e a l l a n a l y t i c a l l y d e f i n e d f u n c t i o n s and have e a s i l y c a l c u l a b l e v a l u e s a t
a n y p o i n t e x c e p t t h e s i n k l o c a t i o n p o i n t (Xo, Yo). A t t h i s point the considerations
of z e r o n e t v o r t i c i t y and t h e r m a l e n e r g y p r o d u c t i o n a l l o w s t h e t w o c o m b i n a t i o n s o f
t h e s e terms t o be e q u a t e d t o z e r o (see s e c t i o n 2 . 5 ) . Once a g a i n t h e v a l u e s o f t h e
boundary c o n d i t i o n s a r e a l t e r e d b u t t h e i r t y p e is r e t a i n e d and t h e same s o l u t i o n
a l g o r i t h m s may be u s e d .

4.3 VARIATIONAL TECHNIQUES

The G a l e r k i n Method

A n u m e r i c a l method of s o l u t i o n for i n c o m p r e s s i b l e f l o w p r o b l e m s which h a s

r e c e n t l y g a i n e d f a v o u r i s t h e s e m i - a n a l y t i c a l s p e c t r a l r e p r e s e n t a t i o n p r e s e n t e d by
O r s z a g (1971 a , b , c ) and Orszag and I s r a e l i ( 1 9 7 4 ) which i s based on t h e G a l e r k i n
t e c h n i q u e d e s c r i b e d v a r i o u s l y by S c h e c h t e r ( 1 9 6 5 ) , K a n t o r o v i c h and Krylov ( 1 9 6 4 )
and M i k h l i n and S m o l i t s k y (1967). The s p a t i a l dependence of t h e t e m p e r a t u r e and
f l o w f i e l d s a r e expanded i n t o a t r u n c a t e d s e r i e s o f o r t h o g o n a l f u n c t i o n s , f o r example

44
for some finite cutoff value N , where the orthogonal base functions a n
(
:
) must
be consistent with the boundary conditions but are otherwise arbitrary. The
Galerkin equations for the coefficients cn(t) are derived as in Schechter (1965)
by minimising the error caused by substituting the approximation (4.3.1) into the
equations which describe the field y(z,t) . In this way a system of partial
differential equations may be reduced to a coupled, ordinary differential set which
is conceptually easier to solve numerically. The well known finite Fourier
transform method is a special case of the Galerkin method in which the base functions
are sines, cosines or complex exponentials. In this investigation the uniformly
heated problem has periodic boundary conditions which makes it amenable to solution
by Fourier analysis (see appendix E), which is the method used to generate the
solution illustrated previously in figure 4 . 2 . 1 . This form of the method is also
used by Straus (1974) for the analytical study of the stability of this problem.

Speed of Calculation

One major disadvantage of the truncated Fourier series expansion is its

complexity and the correspondingly interminable calculations required to compute the
coefficients. For example the direct calculation of sine and cosine Fourier
coefficients in the solution illustrated in figure 4.2.1 is at least an order of
magnitude slower than a similarly accurate solution using the fourth-order Arakawa
differencing scheme. However the method does have the attractive advantages that
it is accurate, has conservation properties similar to the Arakawa schemes and takes
proper account of the boundary conditions. Orszag (1971a) has invented a very
worthwhile optimisation of the method that uses further complex exponential transforms
to speed up calculation of the coefficients to a level that is at least comparable
with finite difference schemes if the fast Fourier transform algorithm described by
Cooley, Lewis and Welch (1967) is available.

This faster method is capable of adaption to the conditions used here, but
unfortunately only in the simple case of uniform isothermal boundaries. Even then
the complex Fourier transform is not itself directly applicable and requires anomalous
definitions of the early terms in the series to avoid inconsistencies of symmetry
(Orszag 1971b). Orszag (1971c) has demonstrated that a Chebyshev polynomial expan-
sion can handle the boundary conditions satisfactorily without defining a special
transform, but in this case fast Fourier transform techniques are no longer directly
applicable. To derive a similar spectral method for the problem at hand is not
infeasible as is evident from the parallel adaptation by Young (1974) but since the
technique would still only be applicable to uniform boundary conditions, such an
analysis is not attempted. The simple and direct calculation method is considerably
easier to apply and although unsuitable for protracted solutions it provides a
useful basis for comparison with the finite difference methods which are more
generally valid.

Finite Element Methods

Another numerical method which is also derived from variational principles

is the finite element approach which has been applied to fluid flow problems by Cullen
(1974). This technique is a more functional method for the examination of real
Slow Heating

However, if the lower boundary is heated slowly to a final temperature

9=1 ,alargely unicellular motion is obtained which is unsteady at all Rayleigh
numbers larger than approximately 280 -
a figure which is in agreement with the range
240 -
280 predicted by Caltagirone et a l . (1971), and identical to the value observed
by Combarnous and Le Fur (1969). This unsteady unicellular motion also develops if
as well as rapid initial heating an initial unicellular perturbation is introduced
to the flow. These fluctuating solutions exhibit a behaviour very similar to that
observed in the Hele-Shaw results (chapter 3 ) which is comparable to the fluctuating
convective state summarised by Bories and Combarnous (1973). The fluctuating
solution at R=500 is illustrated in figure 5.1.2, and the solutions for this section
are summarised in table 11.

FIGURE 5.1.2 - Uniformly heated fluctuatlng solution at R = 500. An equally spaced sequence of
Isotherms (full lines). Broken lines in (a) are streamlines.

48
Chapter 5 - RESULTS FOR THE TWO-DIMENSIONAL
PROBLEMS

5.1 THE UNIFORMLY HEATED MODEL - f = 1

A set of solutions to the uniformly heated, constant viscosity problem was

generated to obtain a representation of the experimental solutions reported by Combarnous
and Le Fur (1969) and by Caltagirone et al. (1971). It was found that the results of
these authors can only be reproduced when certain initial conditions are applied.

Rapid Heating

Beginning the solution with the initial condition that the flow is everywhere
stationary and the lower boundary raised suddenly to the high temperature 8=1 , it is
found that, in contradiction both to the experimental results of chapter 3 and to those
of Caltagirone et al. (1971), the flows become multicellular and steady (or only mildly
transient) at Rayleigh numbers between 300 and 500. Such a solution is seen in
figure 5.1.1.

FIGURE 5 . 1 . 1 - U n i f o r m l y heated steady solution a t R = 500. Left: Isotherms Right: Streamlines

systems w i t h a r b i t r a r y shapes, b u t f o r t h e r e g u l a r t e s t systems c o n s i d e r e d h e r e t h e
f i n i t e d i f f e r e n c e s o l u t i o n s are s i m p l e r a n d f a s t e r . The f i n i t e e l e m e n t method
r e q u i r e s t h e s o l u t i o n of matrices t h a t a r e of similar s i z e t o t h o s e s o l v e d by t h e
Buneman a l g o r i t h m , b u t t h e y a r e l e s s s p a r s e and must be s o l v e d by i n e f f i c i e n t
elimination techniques. T h e r e f o r e t h e f i n i t e e l e m e n t method i s n o t as u s e f u l for
time d e p e n d e n t s o l u t i o n s and i s u n l i k e l y t o be w o r k a b l e i n t h r e e d i m e n s i o n s .

46
 T a b l e II - Results f o r Uniformly Heated Problem

1 Steady Solution Fluctuating Solution

R I Mesh
Size
No. of
Cells 1 Nu.+ I n t e r v a l" Numean

17x17 1.01 - - -
17x17 4.51 c - -
17x17 6.17 0.0132 5.31 5.81
33x33 7.79 0.0068 6.30 7.11
33x33 - 0,0045 8.58 10.04
1000 33x33 - 0.0022 11.81 12.90
1250 33x33 - 0.0017 14.20 15.69

t Nu t h e N u s s e l t number i s a p a r a m e t e r r e l a t i n g t h e c o n v e c t i v e h e a t t r a n s f e r t o
t h e c o n d u c t i v e h e a t transfer - see a p p e n d i x C.

The i n t e r v a l g i v e n i s t h e closest p e r i o d between a n y two s u c c e s s i v e r e l a t i v e

maxima i n t e m p e r a t u r e a t a p a r t i c u l a r p o i n t i n t h e f i e l d .

Heat T r a n s f e r

When t h e s o l u t i o n t o a p r o b l e m s u c h as t h i s t e n d s t o w a r d s a s t e a d y s t a t e
(as it d o e s i n t h e m u l t i c e l l e d s o l u t i o n s ) it d o e s so i n s u c h a way t h a t t h e e n e r g y
t r a n s f e r r e d by t h e s y s t e m is maximised (see Platzman 1 9 6 5 ) . A t a R a y l e i g h number
of 5 0 0 t h e N u s s e l t number (which i s d i r e c t l y r e l a t e d t o t h e amount o f h e a t c o n v e c t e d
across t h e s y s t e m - see a p p e n d i x C ) f o r t h e s t e a d y t r i c e l l u l a r mode i s 7 . 7 9 , whereas
f o r t h e f l u c t u a t i n g s t a t e it v a r i e s i n time w i t h a mean v a l u e o f 6 . 3 0 a n d a maximum
of 7.11 ( c o n s i d e r e d o v e r a p e r i o d i n which f o u r r e l a t i v e maxima a p p e a r ) . This
i n d i c a t e s t h a t t h e s t e a d y s o l u t i o n p r o v i d e s g r e a t e r energy t r a n s f e r a t t h i s Rayleigh
number and i s t h e r e f o r e t h e p r e f e r r e d mode of f l o w . However it a p p e a r s t h a t a n
e n e r g y maximum o c c u r s d u r i n g t h e f o r m a t i o n o f a s t e a d y p a t t e r n , and t h i s may p r e v e n t
s u c h a s o l u t i o n from d e v e l o p i n g .

S o l u t i o n s f o r s t e a d y c o n v e c t i o n i n a p o r o u s l a y e r w i t h no r e s t r a i n i n g s i d e
walls have been o b t a i n e d by Combarnous ( 1 9 7 0 ) and O ' S u l l i v a n ( 1 9 7 3 ) who d i s c o v e r e d
t h a t R a y l e i g h numbers of a p p r o x i m a t e l y 2 8 0 , 4 0 0 and 700 r e s u l t i n c e l l w i d t h s o f 0.5,
0 . 3 3 and 0.25 r e s p e c t i v e l y f o r t h e p r e f e r r e d f l o w . It is not s u r p r i s i n g then t h a t
t h r e e c e l l s a r e p r e f e r r e d i n t h e s o l u t i o n a t R=500 examined above.

The S o l u t i o n A l t e r n a t i v e s

For r a p i d i n i t i a l h e a t i n g a m u l t i c e l l e d " p r o t o - s u b l a y e r " f o r m s i n a manner

similar t o t h a t d e s c r i b e d by E l d e r ( 1 9 6 8 ) and t h e number o f c e l l s a d j u s t s t o form
t h e p r e f e r r e d s t a b l e mode. However o n c e a small i n i t i a l u n i c e l l u l a r motion h a s been
i n t r o d u c e d , e i t h e r a r t i f i c a l l y o r by commencing h e a t i n g a t a n e f f e c t i v e R a y l e i g h
number a t which t h e u n i c e l l u l a r mode i s p r e f e r r e d , t h e s o l u t i o n r e m a i n s l a r g e l y u n i -

49
cellular and adopts the fluctuating convective state. In the experimental solutions
of Caltagirone et al. (1971) and of chapter 3 , slight physical disturbances or non-
instantaneous experimental heating would explain the similar unsteady behaviour.
In the Hele-Shaw experiments of chapter 3 , the heating was sufficiently rapid to
allow a "proto sublayer" to form initially but slow enough to result in an unsteady
flow at later times.

The Steady Solutions

It seems probable that the steady multicellular alternative has less

practical significance since it requires such artificial conditions for its original
generation. Although stable and more "favourable" at later times, this solution is
susceptible to two-dimensional, unicellular perturbation until quite late in its
development. According to the analysis of Straus (1974) a steady two-dimensional
solution is eventually destined to become unstable to three-dimensional perturbations
at higher Rayleigh numbers. The steadiness of the flow in the solutions generated
here becomes uncertain as the Rayleigh number is increased above 400. In the
solutions at Rayleigh numbers around 500 small fluctuations are present in a tri-
cellular mode.

If a steady mode of convection occurs at Rayleigh numbers above this figure

then the preferred number of cells would be four or more, which is beyond the
resolution of a 3 3 x 3 3 mesh. In any event the results of Straus (1974) indicate that
the preferred mode of flow at Rayleigh numbers above 400 (at which the flow is two-
dimensional and tricellular) is three-dimensional. Further study requires the
representation of three-dimensional flow and is considered later in chapter 6.

The Unsteady Solutions

A possible explanation for the unsteady behaviour is that the fluctuations

which appear in the unicellular solutions are an attempt by the system to resort to
the more favourable steady multicellular pattern which is repressed by the dominant
circulation. However there is a finite energy requirement to transform the solution
from one alternative state to the other - once formed, the fluctuating convective state
is stable to temperature perturbations. An attempt to assist the solution between
states was made by the introduction of thermal energy in the form of a 10% random
variation (or "noise") however this proved unsuccessful. This indicates that both
solutions are physically significant although the reason for the existence of two
solutions is still n o t clear.

In the following section the results indicate that altering the heating
element reinforces or reduces the "unfavourability" of the single-celled mode, and
thus solutions to the non-uniformly heated problem clarify the processes observed
here.

50
The Fluctuation Period

The fluctuation period is defined as the closest interval between any two
successive relative temperature maxima at a reference point in the field. The log-
log plot in figure 5.1.3 indicates that the fluctuation period is proportional to
R-3/2

1250-
1000 -
750 -
R
500 -
375-

The Nusselt Number Variation

The mean Nusselt number for the fluctuating solutions and the steady state
Nusselt number for the steady solutions are plotted in figure 5.1.4. The values lie
close to the lower bound of the envelope results of Elder (1967a), Combarnous and
Le Fur (1969), Buretta and Berman (1974) and Gupta and Joseph (1973) that are summarised
in a single plot by Gupta and Joseph (1973 p.513) which is reproduced in outline here.

Nu Steady
Fluctuating

I I I 1 I I
100 250 375 500 750 1000 1250
R
FIGURE 5.1.4 - P l o t o f Nusselt number Nu vs Raylelgh number R f o r uniformly heated s o l u t l o n s .
F u l l l i n e s a r e t h e d e l i m i t e r s of t h e range of values obtained by e a r l i e r authors (see
text). Log-log scale.

51
5.2 THE NON-UNIFORMLY HEATED MODEL - f < 1

The Range of Results

For simulations of flow in a region heated non-uniformly from below and

with constant viscosity, a range of Rayleigh numbers between 250 and 1250 was employed
and a diversity of input conditions was achieved by using values of f=0.25, 0.5 and
0.75. The results of these simulations are summarised in table III and include
completely regular oscillatory flows. Such flows have not been previously obtained
numerically. The period T~ of these oscillations is defined as the interval between
successive temperature maxima at a point in the rising plume of the flow.

Table III - Results for Non-Uniformly Heated Problem

Mesh
r Steady Solutions

No. of
T Oscillatory Solutions

R Size Cells NU =P Numean NUmax

250 17x17 3 3.36

500 17x17 3 6.02
750 17x17 3 7.45
1000 33x33 3 8.41
1250 33x33 3 9.15

250 17x17 1 3.45

375 17x17 1 4.20
500 17x17 0.0093 4.41 4.42
500 33x33 0.0091 4.42 4.43
625 33x33 0.0064 5.18 5.26
750 33x33 0.0053 6.33 6.51
1000 33x33 0.0031 8.05 8.29
1250 33x33 0.0024 9.53 9.87

250 17x17 1 3.29

375 17x17 1 4.09
500 17x17 0.0194 4.63 4.72
750 17x17 0.0089 4.95 5.01
1000 3 3x33 0.0063 5.59 5.77
1250 33x33 0.0051 7.62 7.97

The Oscillatory Solutions

The oscillatory solutions to the half-heated problem are comparable with

the Hele-Shaw results of chapter 3, however certain features that are not immediately
obvious in the experiments are now apparent. It seems that the rising disturbances
are generated at approximately twice the frequency of the descending ones so that two
ascending thermals pass in every complete cycle, the first largely dissipating before

52
 (a)

FIGURE 5.2.1 - H a l f heated o s c i I l a t o r y s o l u t i o n a t R = 750. A sequence o f isotherms ( f u l l Iines)

e q u a l l y spaced I n time. Broken l i n e s i n ( a ) a r e streamlines.

53
the passage of the second. The progress of a solution through a pair of ascending
oscillations of such a flow is illustrated in figure 5.2.1, in this case the flow is
at a Rayleigh number of 750 and the solution generated on a 33x33 mesh.
The sequence is a set of isotherm plots and the diagrams, for greater clarity, are
constructed of the solution and its mirror image, joined at the centre. A summary
plot of the same oscillation is given in figure 5.2.2, which illustrates the variation
of the Nusselt number and of typical temperatures in the ascending and descending
portions of the f l o w .

FIGURE 5.2.2 - Variation in time of the half heated oscillatory solution at R = 750. Nusselt
number ( - - - ) , Reference temperature at a point in the ascending flow (-1, Reference
temperature at a point in the descending flow (+++++).

These oscillatory solutions appear almost identically on both 17x17 and

33x33 meshes at a Rayleigh number of 500, which indicates that this behaviour is not
merely the result of numerical disturbances, thus the existence of the regular
oscillatory solution previously intimated in chapter 3 is confirmed. The results
demonstrate that significant transiency is possible for non-uniform heat input distri-
butions at Rayleigh numbers greater than approximately 480 for f=0.5 and 450 for 0.25.

The Oscillation Period

The dependence of oscillation period T on the Rayleigh number R is

P
indicated by figure 5.2.3 for f=0.5 and 0.25, and in both cases the lines

approximately fit the points, as was the case for the fully heated solutions in the
previous section (figure 5.1.3).

The two experimental values of T for f=0.5 obtained in chapter 3 do not

P
differ greatly from the values expected for numerical simulations at the same Rayleigh
numbers, however they are not plotted since experimental inaccuracies make their
quantitative validity doubtful.
 '"E
DO2 .W3 .ooL
250 .006 ,008 .01

F
.Ob .02

FIGURE 5.2.3 - Plot of oscillation period -rP vs Raylelgh number R for half and quarter heated
solutlons. Log-log scale.

The Three Quarters Heated Solutions

When the lower boundary is three quarters heated the flow patterns are
quite different to the oscillatory solutions, being multicellular and comparatively
steady. The solution at a Rayleigh number of 2 5 0 is illustrated in figure 5.2.4.

FIGURE 5.2.4 - Three quarters heated steady solution at R = 250. Left: Isotherms, Right :
StreamI 1 nes.

55
 The s t e a d i n e s s of t h e f l o w s i n t h i s s e t o f r e s u l t s i s p a r t i c u l a r s i g n i f i c a n t
s i n c e it r a t i f i e s t h e a s s e r t i o n made e a r l i e r t h a t t h e s y s t e m a c h i e v e s o r a t t e m p t s t o
a c h i e v e a p r e f e r e n t i a l and m i n i m a l l y u n s t e a d y s t a t e whenever p o s s i b l e . An u n h e a t e d
p o r t i o n of t h e boundary t h a t i s s i m i l a r t o t h e c e l l w i d t h f o r t h e p r e f e r r e d s t e a d y
f l o w a t a p a r t i c u l a r R a y l e i g h number f o r c e s a more f a v o u r a b l e and s t e a d i e r p a t t e r n t o
occur. The s h o r t e r h e a t e r l e n g t h s r e i n f o r c e t h e f o r m a t i o n of a s i n g l e c e l l e d mode
and d i s c i p l i n e t h e f l o w i n t o r e g u l a r o s c i l l a t i o n s .

R e c t a n g u l a r Regions

I f t h e r e g i o n i s e l o n g a t e d from t h e s q u a r e s h a p e it h a s been g i v e n so f a r ,
t h e t e m p e r a t u r e g r a d i e n t s a r e r e d u c e d and u n s t e a d y e f f e c t s l e s s e a s i l y e s t a b l i s h e d .
M a i n t a i n i n g t h e h e a t e r l e n g t h t o he h a l f a s l o n g as t h e unchanging v e r t i c a l d i m e n s i o n
and e x t e n d i n g t h e r e g i o n t o X = 1 . 3 a t a R a y l e i g h number o f 7 5 0 r e s u l t s i n t h e r e d u c t i o n
o f t h e d e s c e n d i n g e f f e c t s w h i l e t h e a s c e n d i n g t h e r m a l s c o n t i n u e t o form i n p a i r s w i t h a
period of 0.0049. The mean N u s s e l t number i s t h e n 5 . 5 3 (maximum 5 . 6 8 ) and t h e i s o t h e r m
and s t r e a m l i n e p l o t s a r e i l l u s t r a t e d i n f i g u r e 5 . 2 . 5 .

FIGURE 5.2.5 - Unsteady solution i n rectangular region X = 1.3 a t R = 750. I s o t h e r m (-)

and stream1 i nes (- - -9.

56
Insulated Lower Boundary

Replacing the low temperature condition on the unheated portion of the

lower boundary by a restriction of zero heat withdrawal does not significantly affect
the generation of thermals over the heater (see figure 5.2.7) except that the value
of the Rayleigh number above which oscillatory flows occur is increased to around 900.

.
FIGURE 5.2.7 - Unsteady solution in h a l f heated region with insulated lower boundary at R = 1000.

An insulated unheated portion of the lower boundary is more realistic physically

because it does not produce the singular temperature gradients between the heated and
unheated sections of the boundary that are a feature of the non-uniformly heated
models considered so far. The isothermal condition was used initially to obtain a

57
r e p r e s e n t a t i o n o f t h e c o n d i t i o n s i n t h e Hele-Shaw c e l l , however as it t u r n e d o u t , s u c h
a c o n f i g u r a t i o n i s b e t t e r f o r o b s e r v i n g u n s t e a d y e f f e c t s which o c c u r a t l o w e r R a y l e i g h
numbers t h a n i n t h e more r e a l i s t i c i n s u l a t e d model.

5.3 REVIEW OF CONSTANT VISCOSITY SOLUTIONS

Before c o n t i n u i n g w i t h t h e d i s c u s s i o n o f s i m u l a t i o n s u s i n g t h e more r e f i n e d
models it i s p o s s i b l e t o p r e s e n t a u n i f i e d s y n o p s i s of t h e r e s u l t s a l r e a d y r e c o r d e d
which by t h e m s e l v e s improve t h e c u r r e n t u n d e r s t a n d i n g of n a t u r a l c o n v e c t i o n i n p o r o u s
media.

The Thermal Boundary Layer

(a) Steady Solutions

A s h a s been mentioned a l r e a d y , Combarnous ( 1 9 7 0 ) and O ' S u l l i v a n ( 1 9 7 3 ) h a v e

o b s e r v e d t h a t f o r a n assumed q u a s i - s t e a d y c o n v e c t i v e s t a t e i n a p o r o u s l a y e r , t h e
p r e f e r r e d c e l l w i d t h d e c r e a s e s w i t h i n c r e a s i n g R a y l e i g h number. Although q u a s i -
s t e a d y f l o w s may n o t o c c u r , t h i s d e c r e a s e i n c e l l w i d t h d o e s a t l e a s t i n d i c a t e a
tendency of t h e s o l u t i o n s .

Now, i n t h e f u l l y h e a t e d problem c o n s i d e r e d i n s e c t i o n 5 . 1 , when t h e l o w e r

boundary i s s u d d e n l y h e a t e d a t h e r m a l boundary l a y e r f o r m s by c o n d u c t i o n close t o t h e
heater. A c r o s s t h i s t h i n l a y e r t h e t e m p e r a t u r e v a r i e s f r o m i t s maximum 8.1 ,to a
v a l u e c l o s e t o i t s minimum 820 . Thus i n t h i s r e g i o n t h e effective R a y l e i g h number
increases with t h e thickness of the layer. The l a y e r e x t e n d s by c o n d u c t i o n u n t i l it
r e a c h e s a c e r t a i n t h i c k n e s s a t which t h e e f f e c t i v e R a y l e i g h number a t t a i n s i t s c r i t i c a l
value. It may t h e n be e x p e c t e d t h a t a l a r g e number o f small s q u a r e c e l l s w i l l g e s t a t e
inside the layer. F o l l o w i n g t h i s , two p r o c e s s e s o c c u r . F i r s t l y t h e thermal l a y e r
c o n t i n u e s t o e x t e n d by c o n d u c t i o n and c o n v e c t i o n , i n c r e a s i n g t h e e f f e c t i v e R a y l e i g h
number and e l o n g a t i n g t h e c o n v e c t i o n c e l l s . Although t h e a s p e c t r a t i o o f t h e c e l l s
would n o r m a l l y t e n d t o i n c r e a s e w i t h R a y l e i g h number it d o e s so a t a l e s s e r r a t e t h a n
t h a t c a u s e d by t h e c e l l l e n g t h e n i n g and t h u s t h e c e l l s become t o o t a l l and t h i n f o r
t h e i r t h e r m a l environment. S e c o n d l y , b e c a u s e t h e s e c e l l s a r e now t o o numerous t h e r e
i s i n s u f f i c i e n t e n e r g y i n p u t a t t h e l o w e r boundary t o m a i n t a i n them a l l , and t h o s e
which happen t o be smaller are swept away and d i s s i p a t e d by l a r g e r o n e s . I n t h i s way
t h e s y s t e m a d j u s t s s o t h a t by t h e time h e a t i s t r a n s f e r r e d r i g h t across t h e r e g i o n t h e
number o f c e l l s h a s r e d u c e d t o a p p r o x i m a t e l y t h e p r e f e r r e d f i g u r e . This process i s
t h e development and c o l l a p s e of t h o " p r o t o - s u b l a y e r " o b s e r v e d by E l d e r ( 1 9 6 8 ) .
 (b) Unsteady S o l u t i o n s

If, on t h e o t h e r h a n d , t h e l o w e r boundary i s h e a t e d s l o w l y enough, t h e

t e m p e r a t u r e d i f f e r e n c e a c r o s s t h e boundary l a y e r i s s u c h t h a t t h e e f f e c t i v e R a y l e i g h
number d o e s n o t r e a c h i t s c r i t i c a l v a l u e u n t i l t h e l a y e r h a s e x t e n d e d well a c r o s s t h e
r e g i o n i n which case o n l y a s i n g l e c e l l f o r m s . The same n e t r e s u l t o c c u r s i f a u n i -
c e l l u l a r p e r t u r b a t i o n i s a p p l i e d t o a uniformly s t r a t i f i e d conduction s o l u t i o n . Once
t h i s l o w e r o r d e r f l o w r e g i m e h a s become dominant t h e n h i g h e r o r d e r and more f a v o u r a b l e
flow p a t t e r n s a r e suppressed. The s m a l l e r number o f c e l l s d o e s n o t remove h e a t from
t h e l o w e r boundary as f a s t a s t h e p r e f e r r e d p a t t e r n would, and t h e r e f o r e a t h e r m a l
boundary l a y e r f o r m s by c o n d u c t i o n i n r e g i o n s of stagnant f l o w above t h e h e a t e r .
A small s q u a r e c e l l t h e n b e g i n s t o form w i t h i n t h i s l a y e r , b u t as i n t h e s t e a d y c a s e
t h i s cell h a s o n l y a b r i e f l i f e and never reaches s u f f i c i e n t magnitude t o e n t e r t h e
o v e r a l l flow p a t t e r n . I n t h i s c a s e it i s e n s n a r e d i n t o t h e dominant c i r c u l a t i o n and

. . . .
d i s s i p a t e d., .l e.a v .i n g b e h i n d it a d e p l e t e d t h e r m a l boundary l a y e r .

(c) Experimental S o l u t i o n s

The f l o w s o b s e r v e d i n t h e e x p e r i m e n t s o f c h a p t e r 3 show f e a t u r e s o f b o t h t h e
u n s t e a d y and s t e a d y a l t e r n a t i v e f l o w s , namely t h e i n i t i a l " p r o t o s u b l a y e r " w i t h a
d e c r e a s i n g number o f c e l l s and t h e l a t e r g e n e r a t i o n and d i s s i p a t i o n o f new c e l l s between
dominant p a i r s . I n f i g u r e 3 . 2 t h e r e a r e i n i t i a l l y e i g h t cells a c r o s s t h e base of t h e
a p p a r a t u s ( i n t h e p h o t o g r a p h s a t 5 and 1 0 m i n u t e s a f t e r commencement of h e a t i n g ) , b u t
a f t e r 65 m i n u t e s t h e c e l l number h a s r e d u c e d t o f o u r . By 90 m i n u t e s t h e c e l l number
i n c r e a s e s t o e i g h t a g a i n b u t 4 0 m i n u t e s l a t e r ( a t 120) t h e r e a r e o n l y f o u r . Thus i n
t h e s e f l o w s t h e t h e r m a l boundary l a y e r d o e s n o t form s l o w l y enough t o p r o d u c e j u s t a
s i n g l e c e l l , however i t i s n o t so r a p i d t h a t t h e r e are s u f f i c i e n t c e l l s i n t h e s y s t e m
t o r e.a c.h .t.h. e p r e f e r r e d number a n d t h e r e f o r e t. h e f l o w r e m a i n s u n s t e a d y .

Adjustment o f Cell Number

T h i s d i s c u s s i o n o f t h e h e h a v i o u r o f t h e t h e r m a l boundary l a y e r h a s
i n t r o d u c e d a f u n d a m e n t a l a n d m a j o r p o i n t c o n c e r n i n g t h e t r a n s i e n c e of c o n v e c t i v e
f l o w i n porous media. The m u l t i c e l l u l a r f l o w becomes s t e a d y a f t e r a n i n i t i a l
p e r i o d o f development o n l y b e c a u s e it i s a b l e t o reduce t h e number o f c e l l s i n t h e
f l o w t o a t t a i n a p r e f e r e n t i a l p a t t e r n which r e s u l t s i n maximum e n e r g y t r a n s f e r .
However t h e u n i c e l l u l a r f l o w remains t r a n s i e n t b e c a u s e it c a n n o t i n c r e a s e t h e number
o f c e l l s i n t h e p a t t e r n , as new c e l l s are d e s t r o y e d by t h e e x i s t i n g l a r g e r one b e f o r e
they can absorb s u f f i c i e n t energy. The s y s t e m c o n t i n u a l l y a t t e m p t s t o i n c r e a s e t h e
r a t e o f h e a t t r a n s f e r from t h e l o w e r boundary by i m p r o v i n g t h e f a v o u r a b i l i t y o f t h e
number o f c e l l s .

The R e f o r m a t i o n I n t e r v a l

After t h e f l i g h t o f a n a s c e n t c e l l t h e d e p l e t e d boundary l a y e r r e q u i r e s a
certain time i n t e r v a l , d e p e n d e n t on t h e R a y l e i g h number, t o r e f o r m t o s u c h a n e x t e n t !
t h a t y e t another p r o s p e c t i v e c e l l can evolve. These s t a g e s o f g e s t a t i o n and f l i g h t

59
are also apparent in the fluid layer problem considered by Denton and Wood (1974) who
observed rising thermal disturbances similar to those first described by Howard (1964).
Sparrow, Husar and Goldstein (1970) found that the rate of generation (in the same
fluid layer problem) was proportional to R2'3 , a result which has also been separately
obtained by Krishnamurti (1970b) and by Willis and Deardorff (1970). An example of
the approach of Sparrow, Husar and Goldstein (1970) applied to the present problem is
given in appendix F. Their approach neglects the flow of fluid over the heater and
considers only conductive heat transfer, which is a not unreasonable approximation in
the case of a fluid layer problem in which there is a fluid boundary layer. However
there is no fluid boundary layer in the porous medium problem and the neglection of
hydrodynamic conditions in appendix F leads to the prediction of shorter reformation
intervals T a R-2 than are observed here.

Fluctuation Regularity

The fluctuations observed in the unsteady fully heated flows are less
regular than the oscillations in the half-heated flows. This is due to the lack of
constraint on the location at which the thermal boundary layer can fcrm in the
uniformly heated configuration. When f=0.5 it is inevitable that for a unicellular
flow the fluid must rise over the heater and fall on the opposite side of the region.
Thus the thermal boundary layer is restricted to form within a certain small area of
the heater, and the flight of thermals occurs at regular intervals from the same
location. On the other hand a disturbance can evolve anywhere on a uniformly heated
boundary and when f = 1 the disturbances are released at irregular intervals from
indefinite locations. The descending "antithermals" in the half-heated case occur
more randomly than the corresponding ascending disturbances because they form on a
uniformly isothermal boundary. However these effects experience a certain induced
regularity due to the periodic impingement on the upper boundary of oscillations in
the rising plume, and therefore are not as random as those in the fully heated case.

Other Boundary Conditions

The solutions for rectangular regions confirm that the thermal boundary
layer provides the mechanism for the production of transient effects, and also
indicates the effect of the confining vertical boundaries. With very long, cool
portions of the region, the flow is slower and the isotherms spread further than in a
similarly heated square region. Thus the rising plume is wider which results in a
weaker thermal boundary layer. It should be explained here that a thermal boundary
layer is indicated by an area of maximum of thermal gradient which generally implies
a conductive region (in a convective flow the advective regions of the flow transport
heat more effectively than conductive regions and do not permit accumulations of thermal
energy). Comparing the isothermal plots of figures 5.2.5 and 5.2.6 it is seen that
the conductive region (in which isotherms and streamlines are approximately parallel)
in the wider cell is approximately as thick (remembering the difference in vertical
length scales) but has a much reduced temperature differential across it. Thus the
effective Rayleigh number never reaches its critical value and the solution remains
steady.

60
 A similar effect occurs in the flows in a region with an insulated lower
boundary. The fluid transported over the downstream end of the heater is warmer
than in the isothermal lower boundary case and consequently the temperature differen-
tial and thus the effective Rayleigh number in the thermal boundary layer are both
reduced.

Thus it has been demonstrated that, for the simple enclosed porous region,
the natural convective flow regime is influenced both by the presence of vertical
boundaries (since the steadiness of the pattern depends on the ability of the preferred
number of cells to form within the problem boundaries) and by the method and extent of
heat input employed.

5.4 THE RECHARGE-DISCHARGE SOLUTIONS

"Natural" Conditions

Solutions f o r the half-heated, constant viscosity, surface recharge prcblem

(in the absence of any point fluid sinks o r sources, corresponding to the natural
state of a geothermal field) were generated on 33x33 meshes for selected values of the
Rayleigh number in the range 250 to 1000. The finer mesh is imperative at lower
values of R than those previously requiring this grid because the outflow of fluid
through the surface draws the rising plume right to the upper boundary, resulting in
severe thermal gradients in this area. In fact at higher values R 750 a slight
numerical instability becomes apparent at the top of the rising column, however the
effects of this are advected out of the region with the surface discharge.

When the Rayleigh number exceeds a value in the vicinity of 350 the flow is
regular oscillatory, periodically generating thermals or pairs of thermals in the
ascending portion of the flow only. This behaviour may be seen in the summary plot
of figure 5.4.1 and a thermal is illustrated in mid-flight in figure 5.4.2.

I, .m.

-73

t&

-5.

--__

FIGURE 5 . 4 . 1 - V a r i a t i o n I n time of t h e o s c i l l a t o r y recharge s o l u t i o n a t R = 500. N u s s e l t number

(---I and Reference temperature a t a p o i n t w i t h i n t h e r i s i n g plume (-1.

61
 - Oscillatory solution for recharge problem at R = 500.
FIGURE 5 . 4 . 2
- .
stream1 i nes (- -)
Isotherms 1-( and

The value of the Rayleigh number at which the solution becomes unsteady
(350) is lower than the value 480 observed f o r the half-heated, closed boundary model
of section 5.2. The recharge from the surface transports cold fluid down to the
heater, contracting the rising plume and reinforcing the thermal boundary layer.
The large temperature differential across the layer means that the effective Rayleigh
number is high, and temperature disturbances form more rapidly. The flow of fluid
into the surface also prohibits the formation of a thermal gradient at the upper
boundary above the descending portion of the flow, and therefore prevents the genera-
tion of descending temperature disturbances. The ascending pulses are generated by
the same mechanism as their enclosed system counterparts, but with higher frequency.

Comparison of the temperature profiles on the left vertical boundary in

figure 5.4.3 (which is a pair of plots of a superposition of the profiles at several
successive time intervals) reveals that the recharge disturbances have a larger
amplitude and move both faster and further. A summary of these natural recharge
solutions is presented in table IV.

62
 FIGURE 5.4.3- Plots of temperature profiles along left hand vertical boundary at various times
for R = 750. Left: Closed system of Figure 5.2.1. Right: Recharge system.

Table IV Recharge - Discharge Solutions

R z
NUmin P

250 2.68 2.68 0.057 -

312 2.68 2.68 0.053 -
375 2.42 3.18 0.051 0.0068
500/lst pulse 2.29 3.58 0.051 0.0037
500/2nd pulse 2.19 3.88 0.052 0.0042
750/lst pulse 2.84 4.78 0.053 0.0021
750/2nd pulse 2.34 4.56 0.054 0.0024
1000 2.72 4.82 0.058 0.0017
I I I I

Biperiodic effects

At intermediate values of the Rayleigh number, namely 500 and 750, successive
thermals have different magnitudes and evolve unevenly spaced in pairs. For instance
at R=500 a large pulse precedes a smaller one by an interval of 0.0037, and the
subsequent large pulse, identical to the first, follows 0.0042 after that. This
biperiodicity is apparent in figure 5.4.1 above. As one thermal rises through the
porous matrix it creates a velocity perturbation that interrupts the gestation of the
thermal which has formed on the heater behind it. This correspondingly smaller
thermal is drawn off earlier than its predecessor and makes a premature flight.

63
 When the Rayleigh number is larger, successive disturbances appear so rapidly
that the interaction between them extends beyond adjacent disturbances, and when
R = 1000 the oscillations are regular, On the other hand at a lower Rayleigh number
the thermals are widely spaced in both time and position and do not interact signifi-
cantly at all, s o when R = 375 the oscillations are again regular.

There appears to be a similar biperiodic effect in the experimental results of

Sparrow, Husar and Goldstein (1970, figure 2a, p.797) for a Newtonian fluid layer
heated from below. A l s o Krishnamurti (1970b) and Willis and Deardorff (1970) have
mentioned paired effects with reference to this problem. In the half-heated closed
boundary model of section 5 . 2 the thermals are weaker than here and do not interact
noticeably in this way, however there is another biperiodicity due to the descending
disturbances having approximately half the frequency. This second type of biperiodi-
city, due to the interaction of two different fluctuating sections of the flow, may
be observed in the results of Denton and Wood (1974) for a stratified fluid layer
problem except that there the different disturbances are separated physically by a
fluid interface.

These effects emphasise the importance of the coupling between the velocity and
the temperature fields and stress the necessity of maintaining accuracy when represent-
ing the advection terms.

The Addition of Fluid Sinks

With the existence of oscillatory effects established in a model more realistic

than the enclosed model in which they were discovered, the effect of including fixed
"unnatural" fluid withdrawals may be evaluated. Taking the solution at R = 500 as a
typical case, sinks of strengths between 0.012 and 0.10 are installed on the left
vertical boundary (corresponding to a borehole above a heat island if this boundary
is envisaged as a line of symmetry). The strengths of these sinks range from k to
twice the natural discharge ($max -. 0.05) observed in the preceding results, In its
present state, the drawoff from the Wairakei field is approximately four or five times
the natural discharge of 440 kg/sec although this fluid is withdrawn through many wells.
On the scale of the model used here, several of these wells may be considered to act
together as a single point sink which corresponds to the configuration modelled. The
remainder of the bores are neglected for the meantime since it is the effect of the .
sink that is of interest rather than the faithful representation of the Wairakei system
alone. The results of the flows for this "centreline sink" configuration are
tabulated in table V.
 Table V Sink - Recharge Solutions R = 500

Sink Strength
NUmin NUmax %MX =P

0/lst pulse 2.29 3.58 0,051 0.0037

0/2nd pulse 2.19 3.88 0.052 0.0042
0.012/lst pulse 3.62 4.56 0.041 0.0037
0.012/2nd pulse 3.58 4.72 0.042 0.0042
0.025/lst pulse 5.86 1.39 0.054 0.0037
0.025/2nd pulse 5.41 7.50 0.055 0.0042
0.05 3.91 8.42 0.056 0.0042
0.10/lst pulse 5.13 12.71 0.100 0.0040
0.10/2nd pulse 4.99 13.79 0.100 0.0043

It was found during introductory tests on the model that the instant at
which the sink was "turned on" was not an important factor and the solution eventually
reached the same state regardless of when the sink had been introduced. Therefore
in each case the flow was permitted to develop from time zero with the sink in force.
Typical flows at sink strengths of 0.025 and 0.1 are illustrated in figures 5.4.4 and
5.4.5, which should be compared to the zero withdrawal solution presented previously
in figure 5.4.2.

I
I
I

FIGURE 5.4.4 - O s c i l l a t o r y s o l u t i o n f o r recharge problem a t R = 500 w i t h s i n k of s t r e n g t h 0.025 on

l e f t v e r t i c a l boundary ( c e n t r e l i n e i n t h i s diagram).

65
 FIGURE 5.4.5 - Oscillatory solution for recharge problem a t R = 500 w i t h s i n k of strength 0.1 on
l e f t vertical boundary.

The E f f e c t of S i n k S t r e n g t h

Although t h e r i s i n g plume c o n t r a c t s u n d e r t h e i n f l u e n c e of t h e s i n k and t h e

N u s s e l t number i s c o n s i d e r a b l y i n c r e a s e d , t h e g e n e r a t i o n of t h e t h e r m a l s i s o n l y
slightly affected. When t h e s i n k s t r e n g t h i s 0 . 1 t h e m o t i o n i s dominated by t h e s i n k
f l o w ( i n d i c a t e d by t h e v a l u e o f Jlmax) b u t t h e time p e r i o d l e n g t h e n e d o n l y s l i g h t l y a s
t h e t h e r m a l boundary l a y e r e v o l v e s l e s s e a s i l y . The s o l u t i o n i s u n u s u a l when t h e
sink strength i s c l o s e t o t h a t of the natural discharge (-. 0.05) as t h e maximum stream
f u n c t i o n i s u n c h a r a c t e r i s t i c a l l y l a r g e and t h e f l o w c o m p l e t e l y r e g u l a r . The f o r c e d
f l o w o v e r r i d e s t h e weaker c o u p l i n g between t h e v e l o c i t y and t e m p e r a t u r e f i e l d s and
d i s r u p t s t h e i n t e r a c t i o n between thermal pairs.

Sink Location

A s i m i l a r r e g u l a r i s a t i o n o c c u r s i f t h e s i n k i s p l a c e d o v e r t h e i n t e r i o r end
of t h e h e a t e r , a l t h o u g h i n t h i s case t h e a m p l i t u d e of t h e o s c i l l a t i o n i s r e d u c e d t o
a p p r o x i m a t e l y 2 0 % ( s e e f i g u r e 5 . 4 . 6 ) and t h e p e r i o d becomes 0 . 0 0 4 1 .

This repositioning of the sink a l s o r e s u l t s i n a reduction of t h e heat

t r a n s p o r t e d t h r o u g h t h e s y s t e m as t h e N u s s e l t number v a r i e s o n l y between 2.60 and
3.56. The i n c r e a s e d f l o w t h r o u g h t h e r e g i o n i s choked by t h e flow between t h e
s u r f a c e and t h e s i n k and c o n s e q u e n t l y t h e volume o f f l u i d p a s s i n g t h e h e a t e r i s
reduced. T h i s i s i n d i c a t e d by t h e s h o r t e r g e s t a t i o n p e r i o d o f t h e t h e r m a l boundary
layer. The s i n k i s a l s o drawing f l u i d of much lower t e m p e r a t u r e when i n t h i s
p o s i t i o n and i t s u s e f u l n e s s i s r e d u c e d t o s u c h a n e x t e n t t h a t t h e s y s t e m p r o d u c e s l e s s

66
 FIGURE 5.4.6 -
Oscillatory solution f o r recharge problem a t R = 500 w i t h s i n k of strength 0.05
over interior e n d o f heater (mid-region).

heat than i n i t s “natural” s t a t e . Moving t h e s i n k f u r t h e r o v e r t h e h e a t e r as i n

f i g u r e 5.4.7 i n c r e a s e s t h e N u s s e l t number t o v a r y between 3.79 and 4 . 6 2 and i n t h i s
case t h e solution i s biperiodic again.

I I I
I
I
I

I
I

FIGURE 5.4.7 -
Oscillatory s o l u t i o n for recharge problem a t R = 500 w i t h s i n k o f strength 0.05
over centre of heater.

67
The Upper R e s e r v o i r

Now t h a t a method h a s bean d e v e l o p e d t o r e p r e s e n t s u r f a c e r e c h a r g e and f l u i d

w i t h d r a w a l , it i s p o s s i b l e t o c o n s t r u c t a model o f t h e “ u p p e r r e s e r v o i r ” which was
d e s c r i b e d i n c h a p t e r 1. The u p p e r r e s e r v o i r may be c o n s i d e r e d as a c o n v e c t i v e s y s t e m
i n i t s e l f , f e d w i t h h o t water from below, as s u g g e s t e d by Donaldson (1974). To
r e p r e s e n t t h i s s i t u a t i o n t h e boundary c o n d i t i o n s r e m a i n similar t o t h o s e used i n t h e
d e e p , n a t u r a l r e c h a r g e model e x c e p t t h a t now f l u i d i s i n t r o d u c e d t h r o u g h a s o u r c e i n
t h e l o w e r boundary, a s i n f i g u r e 5 . 4 . 8 .

FIGURE 5.4.8 - Steady solution for recharge problem a t R = 500 w i t h point source of strength
0.05 a t centre of heater.

I n t h i s case t h e p o i n t s o u r c e i s of s t r e n g t h 0 . 0 5 and i s p o s i t i o n e d a t t h e
centre of t h e heater. T h i s c o r r e s p o n d s t o t h e i n t r o d u c t i o n o f h o t water a t t h e
b a s e o f t h e system through a major f i s s u r e . The f l o w f r o m t h e s o u r c e e x p e l s t h e
t h e r m a l boundary l a y e r from t h e l o w e r boundary and p r e v e n t s t r a n s i e n t b e h a v i o u r .
The f l o w p a t t e r n s a r e s i m i l a r l y s t e a d y i f t h e i n f l u x of f l u i d i s d i s t r i b u t e d o v e r t h e
e n t i r e h e a t e r , as i n f i g u r e 5 . 4 . 9 . The r e s u l t s of s i m u l a t i o n s u s i n g t h e s e f o r c e d
i n f l o w models a r e t a b u l a t e d i n t a b l e V I .

T a b l e VI Source - Discharge S o l u t i o n s R = 500

0.050 distributed - 0.046 0.096

68
 FIGURE 5.4.9 -
Steady solution for recharge problem at R = 500 with distributed source of
strength 0.05 over heater.

The addition of a fluid sink of strength 0 . 0 5 corresponding to exploitation

of the upper reservoir, is shown in figure 5.4.10. In this case the Nusselt number
is increased by a factor of three, to approximately 21.

These results demonstrate that unsteady effects cannot develop at higher

levels of a geothermal region if the rising plume of hot fluid is already steady.
However, when disturbances originate at depth as has been observed in the more com-
plete models used earlier in this work, clearly these thermal anomalies will rise
through the upper reservoir. Thus restricting the long term modelling of geothermal
regions to just shallower sub-systems may give limited understanding of the problem.

FIGURE 5.4.10 - Steady solution for recharge problem at R = 500 with source of strength 0.05
over heater and sink of strength 0.05 on left vertical boundary (centreline).

69
Fluid Reinjection

R e c e n t l y i t h a s been s u g g e s t e d t h a t t h e u s e f u l o u t p u t from g e o t h e r m a l a r e a s
may be improved by t h e r e i n j e c t i o n o f water t o r e p l a c e t h a t removed t h r o u g h b o r e h o l e s .
T h i s i s done a t t h e G e y s e r s f i e l d i n t h e U.S.A., l a r g e l y as a means o f d i s p o s i n g o f
e x c e s s w a r m water f r o m t h e power s t a t i o n . In order t o achieve t h e best possible
r e s u l t s f r o m r e i n j e c t i o n it i s n e c e s s a r y t o i n j e c t t h e r e c h a r g e f l u i d i n t o t h e
d e s c e n d i n g r e g i o n of t h e f l o w , a s i n f i g u r e 5 . 4 . 1 1 .

'I
I /
/ /

FlGURE 5.4.11 -
Oscillatory solution for recharge problem a t R=500 w i t h source of strength
0.05 on r i g h t vertical boundary and s i n k of strength 0.05 on left vertical boundary
(centre1 i n e ) .

The i n j e c t e d f l u i d t h e n makes i t s way down t o t h e h e a t s o u r c e w i t h t h e " n a t u r a l "

r e c h a r g e from t h e s u r f a c e . It s h o u l d be n o t e d t h a t it i s o n l y n e c e s s a r y t o r e i n j e c t
a t a d e p t h a t which t h e s u r r o u n d i n g s have t h e same t e m p e r a t u r e a s t h e i n j e c t e d f l u i d .
I n t h i s way no h e a t i s l o s t from t h e f l u i d t o t h e s u r r o u n d i n g s , p r o v i d e d t h e s o u r c e i s
n o t so s t r o n g t h a t i n j e c t e d f l u i d t e n d s t o r i s e t o w a r d s t h e s u r f a c e a g a i n . A source
of s t r e n g t h similar t o t h e maximum stream f u n c t i o n o f t h e n a t u r a l f l o w a t t h e time t h e
s o u r c e is i n t r o d u c e d s h o u l d s a t i s f y t h i s r e q u i r e m e n t . T h e r e i s l i t t l e t o be g a i n e d
by i n t r o d u c i n g t h e f l u i d a t g r e a t e r d e p t h , which would n e c e s s i t a t e d e e p e r wells and
t h e i n s t a l l a t i o n o f h i g h e r p r e s s u r e pumps. The r e i n j e c t i o n s y s t e m i l l u s t r a t e d i n
f i g u r e 5 . 4 . 1 1 u s e s a s o u r c e o f s t r e n g t h 0 . 0 5 a t t h e same d e p t h as a s i n k o f s t r e n g t h
- 0.05 i n t h e a s c e n d i n g p l u m e . T h i s r e s u l t s i n a maximum Nusselt number of 8 . 6 0
compared t o t h e v a l u e 8 . 4 2 for t h e same s y s t e m w i t h o u t t h e r e i n j e c t i o n . It s h o u l d
be noted t h a t t h e behaviour of t h e system i s s t i l l o s c i l l a t o r y .

70
Forced Convection

F i n a l l y , it s h o u l d be n o t e d t h a t t h e models u s e d i n t h i s s e c t i o n h a v e
d e v i a t e d from t h e f r e e c o n v e c t i v e problem f o r which t h e g o v e r n i n g e q u a t i o n s were
derived i n section 2.1. However t h e Reynolds number o f t h e s e s o l u t i o n s i n i n c r e a s e d
a t most by a f a c t o r of two, and n o t by t h e o r d e r o f magnitude t h a t would i n v a l i d a t e
t h e use o f t h e a p p r o x i m a t i o n s t h a t D a r c y ' s Law i s v a l i d , t h a t e n e r g y d i s s i p a t i o n i s
n e g l i g i b l e , t h a t i n e r t i a e f f e c t s may be i g n o r e d and t h a t t h e Boussinesq approxima-
tion is appropriate.

5.5 THE VARIABLE VISCOSITY MODEL

Physical Parameters

From e m p i r i c a l t a b l e s g i v e n by E n g i n e e r i n g S c i e n c e s Data ( 1 9 6 8 ) a l e a s t
s q u a r e s f i t i s o b t a i n e d t o d e t e r m i n e t h e c o e f f i c i e n t s f o r t h e v a r i a t i o n of t h e
v i s c o s i t y and d e n s i t y o f water w i t h t e m p e r a t u r e . T h e s e v a l u e s are:
a = 0.9229 AT I

81 = 2.606 AT ,

8, = 1.335 (AT)Z ,
and 6 3 = -0.4762(AT)' ,
where AT i s e x p r e s s e d i n u n i t s of 100°C, r e s u l t i n g i n a n a p p r o x i m a t i o n t o t h e
v i s c o s i t y dependence i l l u s t r a t e d i n f i g u r e 5.5.1.

L
1 I I I I
25 5; 75 W
l 125 150 175 200 225 250
1 I'CI
FIGURE 5.5, I -
Cubic approxlmatlon t o t h e v a r i a t i o n o f non-dlmensional k i n e m a t i c v i s c o s i t y v i
with temperature T i n pure water.

71
 As has been mentioned previously, when using the variable viscosity equations,
convection that is at least as vigorous as that in the constant viscosity model may
occur at considerably lower a p p a r e n t Rayleigh numbers (which are based on v o ) .
Consequently for a temperature differential AT of l5OoC a range of R between 50 and
160 is considered which corresponds to a range of "cold water" Rayleigh numbers
between 400 and 1250.

The results are collated in table VII for the closed boundary conditions as
in section 5.2 and the recharge conditions as in section 5.4. A 33x33 finite
difference grid is necessary in all cases.

Table VII Variable Viscosity Solutions, f = 0.5, AT = 1.5

IBoundaries
R
NUmin JImax TP
Flow regime

L
Closed 50 3.60 5.10 0.259 0.0248 Regular oscillatory
80 5.59 5.71 0.281 0.0024 Regular oscillatory
160 8.13 8.23 0.312 0.0013 Regular oscillatory

L
Recharge 50 3.29 3.29 0.342 - Steady
80 2.80 3.55 0.265 0.0031 Approximately
regular oscillatory
160 2.75 3.80 0.295 -0.00025 Irregular fluctuation

The Flow Patterns

These solutions show greatest similarity to the constant viscosity

models at the intermediate value 80 in the range of Rayleigh numbers; figures 5.5.2
and 5.5.3 are illustrations of the two configurations at this value.

FIGURE 5.5.2 - Unsteady solution for variable viscosity, closed, half heated problem at R = 80.

72
 FIGURE 5.5.3 - Unsteady solution for variable viscosity, recharge problem at R = 80.

At the high value of 160 the recharge solution exhibits an almost triperiodic
behaviour, generating three evenly spaced thermals at irregular intervals. The time
interval reported in table VII is the spacing between these disturbances, but the sets
of three occur at intervals up to six times this value. As a result of the sharp
decrease in viscosity as a thermal boundary layer forms, the local effective Rayleigh
number close to the heater increases more than it would have done had the viscosity
been constant. The thermal boundary may then be thought of as a reservoir of less
viscous fluid as well as of thermal energy. Thermals then form in a series that is
typical of the effective Rayleigh number, stripping away the boundary layer. After
a certain quantity of heat has been removed from the layer, usually after the flight
of just three thermals, the reservoir of less viscous fluid is expended and the
disturbances lapse until a new thermal boundary layer has gestated. Although this
process is not identical to those in the constant viscosity model, the mechanism is
basically similar.

At the lower values R = 50 and 80 the closed solution is distinctive since

the comparatively small Rayleigh number and large temperature difference emphasise
the viscosity effects. In this case the rising plume follows a less viscous
"channel" that was formed early in the development of the flow when the pattern was
bicellular. In all previous half-heated solutions the leftmost of the two cells was
dissipated, shifting the plume towards the adjacent vertical wall as that boundary
heated up. Here however the smaller reverse cell is maintained (see figure 5.5.4),
varying in size as the plume wavers during the formation of the descending disturbances
in the thermal boundary layer at the surface.

73
 FIGURE 5 . 5 . 4 - Unsteady solution f o r variable viscosity, closed problem at R = 50.

Fluid Speeds

The descending disturbances are much more gradual than the ascending ones
due to the acceleration of the fluid in higher temperature regions. Comparing the
streamlines to those in a constant viscosity model, as in figure 5 . 5 . 5 , the fluid is
less viscous close to the heater and flows more rapidly than it does along the
surface where the streamlines are more widely spaced.

FlGURE 5.5.5 - Comparison of streamlines in vat-table and constant viscosity models. Left:
Variable viscosity recharge solution at R = 80. Right: Constant viscosity recharge
solution at R = 500.
 As a further comparison, a fully heated solution is generated at a Rayleigh
number of 5 0 , and displays the now familiar three-celled pattern. The flow is
unsteady with fluctuations appearing at intervals of the order 0.0045 and a Nusselt
number varying approximately within the range 6-10. As before, the solution shows an
acceleration of the flow close to the heater, as in figure 5 . 5 . 6 .

3
FIGURE 5.5.6 - Unsteady solution for variable v i s c o s i t y , closed, uniformly heated problem at
R = 50. Left: Isotherms, Right: Streamlines.

Applicability of Constant Viscosity Models

Although the simulations performed using the variable viscosity model are by
no means exhaustive, they reveal that the commonly used constant viscosity models are
inexact but not entirely without use. Generally the flows are similar in character
for the two models, however extensive numerical experimentation is more expensive with
the variable viscosity model which takes approximately twice as long to compute.

It must be remembered when using the elementary model that the effective
Rayleigh number may be ten times larger than its apparent "low temperature" value.
. . Consequently the stream function, which is scaled by the Rayleigh number, seems
unrealistically large, however the actual fluid velocities are similar in both models
(except for the differences noted already).

75
Chapter 6 - THREE-DIMENSIONAL TRANSIENT FLOW

6.1 INTRODUCTION

As the solution of the equations governing convective flow through porous

media is so time consuming in two dimensions, it is not surprising that the three-
dimensional situation is considered less frequently. Most of the research reported
has been experimental, for example Bories and Thirriot (1969), Combarnous and Le Fur
(19691, Caltagirone et al. (1971) and Combarnous and Bia (1971). However Holst and
Aziz (1972b) have performed a short transient numerical solution, and Beck (19721,
Busse and Joseph (1972), Gupta and Joseph (1973) and Straus (1974) have viewed the
problem analytically. Mercer (1973) has a different approach to hydrothermal
modelling, sectioning the region horizontally instead of vertically and using a finite
element approach, but this does not take proper account of vertical fluid motion and
therefore is not a three-dimensional analysis. A s was explained earlier, the
numerical methods of Holst and Aziz (1972b) using central differencing and iterative
successive over-relaxation techniques are restrictive in that only low Rayleigh
numbers can be considered, and even then with dubious accuracy. Since the numerical
analysis used here has proved successful so far, it seems expedient to apply it to
this new case also. The behaviour of a three-dimensional system with a non-uniform
heat input is more relevant to actual geophysical systems than the simple flows
modelled in previous sections since the flow patterns are unlikely to be two-
dimensional in a real situation.

6.2 THE RANGE O F SOLUTIONS

The Alternative Formulations

The pressure formulation is the less attractive of the two alternatives

since it has non-linear terms in a non-Jacobian form which would require the use of
unsatisfactory central differencing or very time consuming Kreiss techniques,
Furthermore the Neumann type of pressure boundary conditions ( 2 . 4 . 8 ) produce a
singular solution sub-matrix in the Buneman algorithm which requires special treatment
unless the upper surface boundary is a recharge boundary (see appendix D).

76
 The advantages of the vector potential formulation are twofold. Firstly
the non-linear terms are in Jacobian form and are therefore amenable to Arakawa
differencing and secondly the boundary conditions are of Neumann type on only two of
the six boundaries, a configuration which can be handled by a much more economical
form of the Buneman algorithm. In fact it is possible to solve both vector
potential equations (2.4.11) and (2.4.12) faster than the single pressure equation
(2.4.4), although memory requirements are slightly larger. On the Burroughs B6700
calculation of a single time step of the complete pressure solution on a 17x17x17 mesh
(using central differencing), requires 28 secs of processor time compared to 21 secs
for the vector potential solution (using Arakawa differencing), although the latter
requires 24K words of data storage against 19K for the pressure solution program.

Boundary Configurations

With such large machine requirements even the faster, more accurate vector
potential formulation does not permit extensive numerical experimentation and specific
solutions must be sought. For the purposes of this work it is desirable to investigate
further the recent work by Straus (1974) who shows that for Rayleigh numbers larger
than about 380 no two-dimensional solutions are stable in an infinitely wide region,
and explains the transition below this value observed by Combarnous and Le Fur (1969)
as being from two-dimensional to three-dimensional modes of flow. Now although this
transition had a l r e a d y been observed by Caltagirone et al. (1971) in a layer of finite
width to be a passage into a fluctuating two-dimensional state, it is not possible to
confirm this with certainty from the results of chapter 5 in which the analysis is
two-dimensional. Even though Straus' approach specifically precludes the existence
of oscillatory solutions, his conclusions raise the question as to whether or not any
pseudo-two-dimensional unsteady solution is a correct representation of a three-
dimensional flow.

Having already determined in chapter 5 that in some instances there are

alternative solutions in this type of flow depending on the initial conditions, it is
believed that the proposed three-dimensional flows of Straus (1974) may represent an
alternative to the two-dimensional flows of Caltagirone et al. (1969). This
possibility is not precluded by the analysis of Straus (1974) who in stability
considerations calculated only the eigenvalue with the lowest absolute value.

To resolve the apparent contradiction, flows were simulated at a Rayleigh

number of 5 0 0 in two closed cubic boxes, one heated uniformly and the other heated
over only half the bottom surface.

6.3 THE NUMERICAL RESULTS

Despite attempts to perturb the solutions into three-dimensional modes of

flow by raising in temperature one point adjacent to the heater not on a line of
symmetry, both configurations maintain two-dimensional flows. The uniformly heated
case is strictly two-dimensional and generates smaller unsteady cells exactly as

77
observed in the experiments of Caltagirone et al. (1971) (see figure 6.3.1), while
the half-heated case is almost steady except for a small, non-increasing three-
dimensional secondary flow just above the heater that is a residual of the introduced
perturbation (figure 6.3.2). The left and centre illustrations in figure 6.3.1 and
6.3.2 represent the streamlines (see appendix G ) and isotherms on a vertical section
passing through a diagonal of the horizontal surfaces and the right hand diagram is
and isothermal plot on a section at the horizontal mid-plane.

FIGURE 6.3.1 - Unsteady solution for three-dimensional uniformly heated region at R = 500.
Left: Streamlines (vertical diagonal section). Centre: isotherms (vertical diagonal
section). Right: Isotherms (horizontal mid-plane).

FIGURE 6.3.2 - Steady solution for three-dimensional half heated region at R = 500.

Uniformly Heated Boundary

Thus it appears that fluctuating two-dimensional r o l l s a r e the most likely

mode of flow in the uniformly heated problem and although the three-dimensional
solution of Straus ( 1 9 7 4 ) may be the preferred s t e a d y flow at this Rayleigh number,
it does not necessarily occur. This behaviour corresponds to the alternative steady/
unsteady solutions in section 5.1 where the supposedly preferred steady solution,
which is determined originally by considering an infinitely wide layer (as is the
solution of Straus 1974), is commonly rejected in favour of the unsteady solution.
Once again the solution is altered by the presence of confining boundaries.

78
 Half-Heated Boundary

Although at a Rayleigh number of 500 the solution to the half-heated

problem should just be oscillatory (the cutoff value is ~480) this behaviour is not
apparent for two possible reasons. Either the 17x17x17 mesh used may be too coarse
(a finer mesh is not practical for three-dimensional problems) or more probably the
duration of simulation is too short to reach the oscillatory stage, which takes some
time for marginally unsteady solutions.

Three- Dimensional Non-Uniformity

The flow is truly three-dimensional but still unsteady if only one quadrant
of the lower boundary is heated and in this case regular oscillations occur at intervals
of 0.0073 (see figures 6.3.3 and 6.3.4).

FIGURE 6.3.3 - Oscillatory s o l u t i o n f o r three- dimensional q u a r t e r heated region a t R = 500.

1.7

75-

5-
-
.25-

FIGURE 6.3.4 -
V a r i a t i o n i n t i m e o f r e f e r e n c e temperature a t t h e mid- point o f t h e v e r t i c a l edge
boundary adjacent t o t h e h e a t e r i n t h e q u a r t e r heated three- dimensional r e g i o n .

79
 These solutions were all generated on a 17x17x17 mesh and took of the
order of 350 minutes each. This represents a considerable expenditure on
computing resources, and it is expected that this type of three-dimensional
modelling will have limited use in geothermal applications except for s p e c i f i c
representations. A more suitable technique would be to use an axisymmetric
analysis - this would be more useful f o r modelling geothermal problems but could
not be used to investigate the physics of the flow as has been done here.

80
Chapter 7 - THERMOHALINE CONVECTION I N POROUS
MEDIA

7.1 INTRODUCTION

To this point the working fluid considered has been a single component
liquid, which corresponds to a pure water geothermal system, but this need not always
be the case. Although Wairakei is believed to be a water dominated system (about
12000 ppm dissolved solids), there are hydrothermal areas (for example Reykjaves in
Iceland and Niland, U . S . A . ) where the saturating fluid is an aqueous solution,
containing a significant proportion (Reykjaves 40,000 ppm, Niland 200,000 ppm)
of dissolved mineral salts, which has a concentration dependent density. Although
the unsteady effects observed so far have been in the thermal problem it is also known
that opposing stabilising and destabilising influences in thermohaline convection can
cause oscillatory instability. Thermohaline or double diffusive convection in a
fluid layer has been studied numerically by Elder (1969), experimentally by Foster
(1968 and 1969), Shirtcliffe and Turner (1970), Lambert and Demenkow (1971), Hurle and
Jakeman (1971), Platten and Chavepayer (1973) and Shirtcliffe (1973), and analytically
by Nield (1967), Baines and Gill (1969), Joseph (1970), Legros, Platten and Poty (1972)
and Hart (1973). In several cases "overstable" solutions were predicted and a good
summary plot of the stability of this problem is given by Baines and Gill (1969) in the
Rayleigh number/solutal Rayleigh number plane. These analyses have all been performed
for the fluid layer problem and represent only a small proportion of research in this
field. The thermohaline problem in a porous medium which is of interest here has
been less extensively investigated. Taunton, Lightfoot ana Green (1972) have performed
a similar linear stability analysis to Baines and Gill (1969) in the porous medium case,
and extended the earlier study by Nield (1968). Both investigations noted the
possibility of overstable oscillatory solutions and determined the conditions under
which they can occur, but neither revealed what form these oscillations might take or
what processes cause them. An energy stability analysis was considered by Wankat and
Schowalter (1970) but without specific reference to oscillatory solutions.

It seems desirable then to adapt the numerical approach used in chapters 5

and 6 to the two-dimensional thermohaline problem in a porous medium in order to
explore another instability mechanism which may be present in geothermal systems.

81
7.2 THE E Q U A T I O N S OF MOTION

The d e r i v a t i o n o f t h e g o v e r n i n g e q u a t i o n s c l o s e l y f o l l o w s t h a t o f s e c t i o n 2 . 1 ,
e x c e p t t h a t t h e f l u i d d e n s i t y v a r i a t i o n i s now a f u n c t i o n o f b o t h t h e c h a n g e i n d i s s o l v e d
s a l t c o n c e n t r a t i o n and t h e e x c e s s t e m p e r a t u r e , Thus (2.1.5) i s r e p l a c e d by

where C i s t h e s a l t c o n c e n t r a t i o n and t h e prime d e n o t e s q u a n t i t i e s a s s o c i a t e d w i t h it

( t h i s makes a t t h e c o e f f i c i e n t o f s o l u t a l e x p a n s i o n o f t h e f l u i d ) . Combining w i t h
( 2 . 1 . 1 ) and ( 2 . 1 . 2 ) and i n v o k i n g t h e B o u s s i n e s q a p p r o x i m a t i o n , t h e e q u a t i o n s become

-
aqi
axi
= o , (7.2.2)

I t i s a l s o n e c e s s a r y t o i n c o r p o r a t e an e q u a t i o n t o d e s c r i b e t h e t r a n s p o r t of d i s s o l v e d
s a l t , and t h i s i s t h e w e l l known d i f f u s i o n e q u a t i o n (Holman 1 9 6 8 p . 3 3 6 )

(7.2.5)

where K' is the solutal diffusivity. A s b e f o r e t h e i n e r t i a t e r m s may b e n e g l e c t e d

and t h e e q u a t i o n s r e w r i t t e n i n non- dimensional form. Defining t h e a d d i t i o n a l v a r i a b l e

c - co
c* = - ,
- co

and i n t r o d u c i n g $ as i n (2.3.6), t h e equations (7.2.3) become ( d r o p p i n g t h e *

immediately f o r convenience),

and

which may be s i m p l i f i e d by c r o s s - d i f f e r e n t i a t i o n and s u b t r a c t i o n , t o

v2JI = a0
- a - y
ac
3x , (7.2.6)

where Y i s t h e buoyancy r a t i o %?%.L%%)

a (TI- To)

E q u a t i o n s (7.2.4) and (7.2.5) become r e s p e c t i v e l y

vze = - R 13 E
ay ax - 9
ax 3 0 1 , (7.2.7)

82
and G O ~= C~ - R / A c $ $ ~ - %%I . (7.2.8)

T h e s e l a s t three equations c o n s t i t u t e t h e set governing t h e flow,

Boundary C o n d i t i o n s

The problem b o u n d a r i e s a r e t h e same as t h o s e c o n s i d e r e d i n t h e o r i g i n a l

e n c l o s e d model ( s e c t i o n 2.2), b u t s i n c e it i s e x p e c t e d t h a t t h e a p p e a r a n c e of a
r e g u l a r o s c i l l a t o r y s o l u t i o n i s now d e p e n d e n t on t h e p r e s e n c e o f t h e s a l i n i t y g r a d i e n t
it i s a d v a n t a g e o u s t o c o n s i d e r a u n i f o r m h i g h t e m p e r a t u r e s o u r c e (f = 1) which would
n o t normally produce a r e g u l a r l y f l u c t u a t i n g flow. The d i s s o l v e d s a l t c o n c e n t r a t i o n
v a r i e s from a minimum v a l u e Co a t t h e s u r f a c e t o a maximum o f C1 a t t h e b a s e , or
0 and 1 r e s p e c t i v e l y i n t e r m s o f t h e n o n - d i m e n s i o n a l c o n c e n t r a t i o n . This corresponds
t o a c o n f i g u r a t i o n i n which h o t t e r f l u i d h o l d s more s o l u t e t h a n c o l d e r f l u i d c l o s e r t o
the surface. A u n i f o r m s t r a t i f i c a t i o n between t h e s e two limits i s i n i t i a l l y s t a b l e
i f t h e s a l t s o l u t i o n i s d e n s e r t h a n t h e s o l v e n t (as i s u s u a l l y s o ) , i n which case a'
and t h e buoyancy r a t i o y are b o t h n e g a t i v e .

Physical Parameters

The f l o w i s d e s t a b i l i s e d by t h e thermal g r a d i e n t and s t a b i l i s e d by t h e

s a l i n i t y e f f e c t s , t h e i n t e r a c t i o n between t h e s e o p p o s i n g i n f l u e n c e s b e i n g d e p e n d e n t
on t h e v a l u e s of t h e d i f f u s i v i t y r a t i o , t h e buoyancy r a t i o y and t h e R a y l e i g h
number R . The f l o w i s i n h e r e n t l y more s t e a d y t h a n t h e s i n g l e component c a s e due
to the stabilising solutal effects and t h u s f o r c o n v e c t i v e s o l u t i o n s t h e R a y l e i g h
number i s c o m p a r a t i v e l y l a r g e ( > 2000). For t h e p u r p o s e s o f t h i s s e c t i o n t h e s o l i d
p o r o u s m a t r i x i s c o n s i d e r e d t o b e a p o o r c o n d u c t o r so t h e h e a t c a p a c i t y r a t i o X i s
t a k e n t o be u n i t y , and from E l d e r (1969) and T a u n t o n , L i g h t f o o t and Green (1972) t h e
buoyancy r a t i o i s t y p i c a l l y o f o r d e r -1. and t h e c o n d u c t i v i t y r a t i o o f o r d e r 0 . 0 1 .
S i n c e t h e s o l u t a l R a y l e i g h number S u s e d by t h e s e a u t h o r s is r e l a t e d t o t h e t h e r m a l
R a y l e i g h number R by t h e e x p r e s s i o n

(7.2.9)

t h e n t h e s e v a l u e s r e p r e s e n t a s i t u a t i o n which l i e s w e l l w i t h i n t h e c o m p l e t e l y non-
convecting zone i n t h e s t a b i l i t y l o c u s p l o t o f T a u n t o n , L i g h t f o o t and Green (1972) w i t h
a t e n d e n c y t o w a r d s o v e r s t a b i l i t y r a t h e r t h a n monotonic c o n v e c t i o n . The r e s u l t s of
t h i s s e c t i o n c o n f i r m t h a t f o r any r e a l i s t i c s i t u a t i o n t h a t i s l i k e l y t o o c c u r i n a n
a c t u a l g e o t h e r m a l s y s t e m , a permanently u n s t a b l e s o l u t i o n i s i m p r o b a b l e w i t h t h e s e
boundary c o n d i t i o n s . However t e m p o r a r i l y t r a n s i e n t f l o w s c a n o c c u r f o r s h o r t
d u r a t i o n s and t h e p h y s i c a l p a r a m e t e r s may be e n l a r g e d beyond t h e i r l i k e l y p h y s i c a l
m a g n i t u d e i n o r d e r t o o b s e r v e what form t h e s e o v e r s t a b l e s o l u t i o n s may t a k e . Steady
c o n v e c t i o n p a t t e r n s can a l s o o c c u r f o r v e r y weak s a l i n i t y g r a d i e n t s b u t t h e s e are less
l i k e l y t o o c c u r t h a n t h e o v e r s t a b l e f l o w s , and i n a n y c a s e a r e c l o s e l y a l i g n e d t o t h e
t h e r m a l c o n v e c t i v e s o l u t i o n s which have a l r e a d y b e e n o b t a i n e d i n s e c t i o n 5 . 1 . It
would be p o s s i b l e t o s t u d y t h e e f f e c t o f a d e s t a b i l i s i n g s a l i n i t y g r a d i e n t (as might
be t h e s i t u a t i o n i n t h e c a s e o f h e a v i l y m i n e r a l l e d w a t e r s f l a s h i n g t o s t e a m a t higher
l e v e l s and d e p o s i t i n g t h e i r s o l u t e s ) b u t as t h e e q u a t i o n s a r e so s i m i l a r t h e s o l u t i o n s
would a l s o be n o t u n l i k e t h e t h e r m a l s o l u t i o n s o f t h e p r e v i o u s c h a p t e r s .
7.3 THE NUMERICAL RESULTS

The numerical representation of (7.2.6), (7.2.7) and (7.2.8) may be achieved

in a manner similar to that of chapter 4 , using fourth order Arakawa differencing for
the advection terms and Bunemen odd-even reduction f o r the Poisson equation. The
system is initially presented with a conduction solution and a small unicellular
perturbation. The solutions so generated are summarised in table VIII and fall
into three main groups -
a set at realistic values of the diffusivity and buoyancy
ratios over a range of Rayleigh numbers, a set at similar Rayleigh numbers with a
more positively stabilising salinity gradient, and a third set with an unrealistically
strong salinity gradient. In all solutions the final result is a motionless flow
with uniformly stratified temperature and salinity fields, but in some cases a decay-
ing oscillation with period T occurs beforehand.
P

Table VIII Thermohaline Solutions

K'/C V 1 Process Observed

P
0.01 -1.0 - Conduction
0.01 -1.0 - Conduction
0.01 -1.0 - Conduction
0.01 -1.0 0.0465 Decaying Oscillation
0.01 -1.0 0.0250 Decaying Oscillation
3500 0.01 -1.0 0.0170 Decaying Oscillation

1000 0.05 -1.0 - Conduction

1000 0.05 -0.5 - Conduction
1000 0.05 -0.75 0.024 Decaying Oscillation
500 0.10 -1.0 - Conduction
1000 0.10 -1.0 - Conduction
1500 0.10 -1.0 0.055 Decaying Oscillation
2000 0.10 -1.0 0.025 Decaying Oscillation
2500 0.10 -1.0 0.013 Decaying Oscillation

500 0.5 -1.0 - Conduction

500 1.0 -1.0 - Conduction
1000 1.0 -1.0 - Conduction
2000 1.0 -1.0 - Conduction
3000 1.0 -1.0 - Conduction

This oscillation takes the form of a pair of cells that periodically change their
direction of rotation so that t h e central plume between them alternatively rises and
falls (see figure 7.3.1).
FIGURE 7.3.1 - O s c i l l a t o r y s o l u t i ' o n f o r thermohaline problem a t R = 3500, y = - 1 , d l f f u s i v i t y
r a t l o = 0.01. Evenly spaced sequence of isotherms ( l e f t ) and streamlines ( r i g h t ) .

85
This results in a characteristic "flag-waving" motion of the isotherms which is
reflected in the salinity isoclines only rather more diffusely. Increasing the
stabilising salinity gradient intensifies the effect which then takes place with
greater frequency.

The Oscillation Mechanism

The mechanism of these oscillations depends on the coupling between the

temperature and concentration fields. The two components of the system diffuse at
different rates (the diffusivity of heat is 100 times greater) and thus a parcel of
fluid which is displaced vertically loses excess heat faster than excess salinity.
The parcel of fluid then has a strongly stabilising force acting upon it and it tends
to descend again, producing the reversal of the plume. As it descends the opposite
effect occurs as heat diffuses in faster than salinity diffuses out and the concentra-
tion effects have a decreasingly stabilising influence as the lower and denser levels
of the salinity stratification are reached. The oscillation continues several more
times, decaying due to thermal dissipation. At no time does the flow become
significantly convective with major circulations like those in the thermal solutions
in previous chapters, and the motion is due entirely to the temperature/concentration
interaction (this is indicated by the fact that the oscillations do not appear when
the diffusivity ratio is increased).

Interpreting these results in a physical sense, it is seen that stabilising

salinity gradients could have the effect of preventing motion in a geothermal system.
However, this is not the case in actual systems such as Reykjaves and Niland in
which convective heat transfer is prominent, where more complex processes occur.
Firstly a major characteristic of a brine geothermal system is the importance of the
steam phase. Flashing of water to steam at shallow levels produces a heavy concen-
tration of salts at the surface, which produces a local destabilising gradient.
Secondly a constant heat flux source at the base cf the formation will accumulate
thermal energy in a stabilised, non-convecting system until a stage is reached where
locally the destabilising temperature gradient may overcome the stabi1ising salinity
gradient and the system being to convect. These two phenomena lie beyond the
capabilities of the present model which considers only single phase flow and an
isothermal heat source. This simpler model is useful for demonstrating the mechanics
of the interactions which may occur between conflicting influences on the flow,
without paying too much attention to specific instances of occurrence.

86
Chapter 8 - THE POROUS INSULATOR PROBLEM

8.1 INTRODUCTION

The model regions considered in chapters 4 -

7 have all been heated from below,
and although this is the more commonly studied case, a porous enclosure raised in
temperature on one of its vertical sides as in figure 8.1.1 is of some interest also.

#=O p.0

9-1 9=O

-X

FIGURE 8. I . I - Problem boundaries for porous insulator problem.

Such a situation arises in porous insulators; for example within nuclear reactors,
fibreglass building materials, hot water pipe lagging and refrigerator walls. If
convection occurs in a homogeneous layer of insulating fluid, for instance an air gap
in a house wall, the heat transfer across the layer is greater than if only conduction
takes place. If the gap is of finite thickness but infinite height and width it has
been proven by Gill (1969) that the introduction of a porous material reduces
inertial effects to such a point that convection cannot happen, and improves the
effectiveness of the system as an insulator (provided of course that the permeable
solid is not highly conductive). However for a sealed porous slab of finite propor-
tions Betbeder and Jolas (1972) found experimentally and numerically that convective
motion is possible, which confirmed the earlier numerical results of Chan, Ivey and
Barry (1970). Jannot, Naudin and Viannay (1973) investigated a different case in
which the unheated and cooler vertical boundary is open to the environment with a
forced flow of fluid through it, but they too observed convective solutions.

87
8.2 THE NUMERICAL SOLUTIONS

The s o l u t i o n p r o c e s s f o r t h i s problem is e s s e n t i a l l y similar t o t h a t f o r

t h e fully- heated region i n chapter 2 except t h a t t h e gravity vector i s r o t a t e d
t h r o u g h 90' with r e s p e c t t o t h e region boundaries -
s o t h a t (2.3.10) becomes

Q2J, = - ae , (8.2.1)

and t h e h e a t t r a n s p o r t e q u a t i o n (2.3.9) r e m a i n s unchanged. A s it h a p p e n s t h e flows

o b t a i n e d are a l w a y s s t e a d y and u n i c e l l u l a r which p e r m i t s t h e u s e o f a v a r i a b l e s i z e d
time s t e p i n t h e n u m e r i c a l scheme, i n c r e a s i n g t h e i n c r e m e n t as t h e s o l u t i o n a p p r o a c h e s
a s t e a d y s t a t e , t h e r e b y r e d u c i n g t h e l e n g t h and i n c r e a s i n g t h e a c c u r a c y o f t h e
calculation. A t e a c h s t a g e of t h e c o m p u t a t i o n t h e p r o s p e c t i v e a l t e r a t i o n s t o t h e
t e m p e r a t u r e m a t r i x (E
ae .A T ) a r e scanned and t h e n s c a l e d u n i f o r m l y so t h a t t h e
maximum a l t e r a t i o n i s a l w a y s by t h e same amount a t e v e r y time s t e p . This ensures
b o t h t h a t t h e time d i f f e r e n c i n g d o e s n o t become n u m e r i c a l l y u n s t a b l e due t o t h e
t e m p e r a t u r e c h a n g e s becoming t o o l a r g e and t h a t t h e s o l u t i o n p r o c e e d s as r a p i d l y as
p o s s i b l e w i t h o u t it becoming so. T h i s procedure i s analogous t o s u c c e s s i v e over-
r e l a x a t i o n i n a q u a s i - s t e a d y s o l u t i o n method.

The r e s u l t s p r e s e n t e d below i n t a b l e I X are n o t as r e m a r k a b l e as t h e

u n s t e a d y f l o w s o b s e r v e d i n e a r l i e r c h a p t e r s , however they c o n f i r m and e x t e n d t h e
e x p e r i m e n t a l r e s u l t s o f B e t b e d e r and J o l a s (1972) and v e r i f y o n c e a g a i n t h e v i a b i l i t y
of t h e n u m e r i c a l method.

Table I X H o r i z o n t a l Temperature Gradient S o l u t i o n s

I Square C e l l T Rectangular C e l l (2x1)

R Mesh Nu
Size JImax

25 17x17 1.11 0.065 17x33 1.10 0.055

50 17x17 1.99 0.057 17x33 1.42 0.053
200 17x17 4.89 0.038 17x33 3.03 0,040
500 33x33 8.78 0.028 33x65 4.98 0.029
1000 3 3x3 3 12.40 0.022 3 3x6 5 7.21 0.021
1250 3 3x3 3 13.09 0.020 3 3x6 5 8.11 0.019
1500 3 3x3 3 14.42 0.018 3 3x6 5 8.65 0.018

N u s s e l t Number Dependence

Comparison of t h e r e s u l t s f o r t h e s q u a r e and f o r t h e t a l l r e c t a n g u l a r c e l l
r e v e a l s t h a t i f t h e height of t h e c e l l i s i n c r e a s e d w i t h o u t c h a n g i n g i t s w i d t h or t h e
v a l u e of t h e R a y l e i g h number, t h e n t h e N u s s e l t number d e c r e a s e s , a s i n f i g u r e ( 8 . 2 . 1 ) ,
i n a c c o r d a n c e w i t h t h e a s s e r t i o n s of B e t b e d e r and J o l a s ( 1 9 7 2 ) which a r e a l s o com-
p a t i b l e with t h e conclusions of G i l l (1969) f o r the i n f i n i t e l y t a l l cell. Typical
s o l u t i o n s a t R a y l e i g h numbers of 50 and 1 0 0 0 a r e i l l u s t r a t e d i n f i g u r e s 8 . 2 . 2 ,
8.2.3, 8 . 2 . 4 and 8 . 2 . 5 .

88
 15-

10-

7-

5-
NU

3-

2-

R
FIGURE8.2.1 - Plot of Nusselt number Nu vs Rayleigh number R for porous insulator problem.

H o r i z o n t a l Boundaries

This analysis serves t o h i g h l i g h t t h e a r t i f i c i a l i t y of t h e conditions

c o n s i d e r e d by G i l l ( 1 9 6 9 ) - a l t h o u g h i n t h e i n f i n i t e case t h e r e i s no c o n v e c t i v e h e a t
t r a n s f e r , f o r a l l r e a l r e g i o n s t h e f l o w i s c o n v e c t i v e even a t v e r y small R a y l e i g h
numbers. The i n t r o d u c t i o n of a h o r i z o n t a l boundary i n t o t h e p a t h o f f l u i d r i s i n g
a d j a c e n t t o t h e v e r t i c a l h e a t e r r e d i r e c t s t h e f l o w p e r p e n d i c u l a r l y and t h u s r e s u l t s
i n a transverse advection of heat. Thus i n t h i s c a s e it i s t h e p r e s e n c e o f t h e
enclosing h o r i z o n t a l boundaries t h a t i s instrumental i n producing a convective
solution. Although t h i s e f f e c t i s c l e a r l y u n l i k e t h e i n f l u e n c e t h a t t h e v e r t i c a l
b o u n d a r i e s have on t h e h y p o t h e t i c a l s t e a d y m u l t i c e l l u l a r m o t i o n s o f s e c t i o n 5 . 1 and
t h r e e - d i m e n s i o n a l m o t i o n s o f c h a p t e r 6 , t h i s i s t h e t h i r d t i m e t h a t it h a s been
d e m o n s t r a t e d t h a t s o l u t i o n s d e r i v e d f o r a n i n f i n i t e r e g i o n are i n a p p l i c a b l e t o r e a l
boundaries.

89
 FIGURE 8.2.2 - Porous insulator solution, R = 50. square cell. Isotherms (---I and streamlines
(---- ).

FIGURE 8 . 2 . 3 - Porous insulator problem, R = 1000, square cell

FIGURE 8.2.4 - Porous insulator problem, R = 50, FIGURE 8.2.5 -

Porous Insulator problem, R = 1000
rectangular cell. Isotherms (left) and rectangular ce I I.
streamlines (right).

90
Chapter 9 - SUMMARY AND CONCLUSIONS

9.1 EVALUATION O F THE RESULTS

The character of the flow of a fluid in a porous medium is not only

governed by its equations of motion, but is also shown to be influenced by the
presence of the boundaries of the region. This configuration dependency is evident
in the results both for regions heated from below and those heated from the side.
The dependency of the flow on inherent fluid properties is demonstrated in the thermo-
haline instabilities in chapter 7 and in the generation of thermals observed in
chapters 5 and 6 (most clearly in the recharge solutions of section 5.4). Although
the two different dependencies are contiguous and often cannot be disassociated, it
is useful to review them separately.

9.2 THE THERMAL BOUNDARY LAYER

The generation of thermal disturbances is the most notable characteristic

of the behaviour of fluid flowing in a porous medium. The fundamental mechanism
of this process is the coupling between the temperature and velocity fields within a
thermal boundary layer which is formed over the heat source. Thermal anomalies
evolve in this layer by conduction and are swept away and dissipated by dominant
circulating velocities.

The appearance of the thermal boundary layer depends on the flow conditions
that already exist in the region. In general, a flow pattern with a larger number
of convection cells removes energy from the heater more quickly than a single celled
flow, and so the thermal boundary layer may not necessarily form sufficiently to
produce thermal disturbances when the cell number is increased. The layer may also
be expelled by the injection of fluid close to the heater or by the expansion of the
rising convective plume caused by the lateral extension of a rectangular region.

91
9.3 THE PREFERRED SOLUTION

At any particular Rayleigh number, a region of a certain shape has a

p r e f e r r e d s o l u t i o n that results in maximum heat transfer across the system. For
example in a square, uniformly heated, closed boundary region at R = 4 0 0 the
preferred solution is a steady two-dimensional tricellular flow. The preferred
solution is Rayleigh number dependent and so the alteration of the effective Rayleigh
number during the course of the development of a flow can have quite remarkable
effects on the eventual solution (which need not necessarily be the preferred one).
This is one of the most significant consequences of the results presented in this
work.

If the actual Rayleigh number is l e s s than the Rayleigh number which is

characteristic of the number of cells in the flow at the time, then it is possible
f o r the flow pattern to adjust by dissipating excess cells to form the preferred
solution. However when the actual Rayleigh number exceeds the value implied by the
number of cells, the flow pattern cannot adjust because potential new cells do not
form properly before being absorbed by the dominant flow. The flow then becomes
fluctuating or regularly oscillatory as new cells continue to form and disappear.

9.4 DISTURBANCE INTERACTIONS

The regularity of an unsteady solution depends on the interaction between

the temperature disturbances. In the closed boundary model there is a forward
interaction, as ascending disturbances may influence descending ones which are forming
in front of them, however the effect is only weak. A much stronger backward inter-
action is seen in the recharge solutions in section 5.4 ; this is caused by the
interruption of a gestating disturbance by its predecessor.

These reactions are different to those proposed by Keller (1966) and

Welander (1967) for a simple loop model, in which a disturbance is envisaged as being
preserved throughout a complete circuit of the convection cell (a loop in their case).
The disturbance then receives a boost on rearrival over the heater and makes another
circuit of the cell. This conception may be used to explain periodic effects in the
loop models, and also in the fluid layer model of Moore and Weiss (1973), but cannot
account for periodicity in convection through porous media. The presence o f the
porous matrix dissipates the thermal disturbances before they can complete a circuit
o f the cell. Furthermore the periodicity persists even in the recharge solutions
in which fluid enters and leaves the system through the upper surface and only
passes the heater once in its path. In the porous medium problem the regularity is
inherent in the g e n e r a t i o n of thermal anomalies rather than their behaviour after
formation as in the fluid loop. The intervals between the disturbances is Rayleigh
number dependent and obeys a 3 1 2 power relationship.

92
9.5 TEMPERATURE/SALINITY EFFECTS

In the thermohaline convection problem in a p orous medium there is a

coupling between the temperature and salinity concentration equations (via the
momentum equation), and it is the interaction of these two influences that produces
the motion. The pulsating flow which ensues is simpler to explain than the
unsteady flows in the thermal problem since in this case the effects are explicitly
opposing and there is no overlying convective motion. A thermal disturbance
carries with it an increasingly stabilising concentration anomaly that eventually
halts it and reverses the direction of travel. This behaviour is less prevalent
than in the fluid layer thermohaline problem due to the less significant inertial
effects.

9.6 THE PRESENCE OF BOUNDARIES

One of the most unexpected discoveries in this investigation is the signifi-

cance of physical boundaries in this problem. Elder (1967a) refers to "end effects"
in his solutions at low Rayleigh numbers, however it is well established here that at
higher Rayleigh numbers certain major features of the flow in porous media are
produced by the presence of boundaries. The flow in the uniformly heated regions of
section 5.1 has alternative steady and unsteady modes depending on how much of the
preferred solution has been introduced into the closed region. In an infinite region
the comment made earlier, that the number of convection cells cannot increase, does
not apply. Therefore the unsteady two-dimensional rolls occur only in finite regions.

A similar argument may be used when considering the three-dimensional flows.

Once a two-dimensional mode of flow has been set up in the region it is not possible
to increase the number of cells in the (finite) third dimension. The flow remains
two-dimensional even though the preferred solution in an infinite region would be
three-dimensional.

The flow in the horizontal temperature gradient configurations of chapter 8

also differs significantly from the infinite region solution, since the flows are con-
vective if horizontal boundaries are present.

9.7 THE PHYSICAL IMPLICATIONS

Assuming a depth of 5 km fluctuations such as those observed in chapter 5

would appear in a real geothermal system at intervals of the order of 500-1000 years
and the simulations predict that they would require up to 50,000 years from the
initial evolution of the system before they occurred. The calculated history of the

93
Wairakei system suggest a past life of around lo6 years (Elder 1966a). It has been
suggested by Hochstein (1974) that geological evidence from bore samples and gravi-
metric surveys of the Broadlands geothermal field indicate that mineral deposits are
in excess of those expected from the system in its current form. This suggests that
the region has existed as a steam or brine system at some time in the past, as might
have been the case if a higher temperature anomaly had risen up through the system.
Using a statistical survey of the world's hydrothermal systems Hochstein (1974)
deduced that the interval for which the anomalous system might exist would be of the
order of 500 years. The large variability of the parameters and the simplistic
structures modelled in this work do not permit quantitative conclusions concerning
the effects, however the implication of the transient behaviour observed in the
numerical models is suggestive.

9.8 EXPLOITATION

Although the removal of water from a geothermal system would not prevent
the appearance of low frequency transiency, the reinjection of water might do so.
If cold water is introduced into fissures which feed fluid to the heat source the
thermal boundary layer over the hot rock is interrupted and unsteady effects would
not occur. The more realistic injection system which introduces fluid at shallower
depths, results in an increase in the quantity of heat that may be retrieved from
the system, however in this case transient behaviour is still possible. In order
to achieve maximum heat transfer through the system from the energy source to the
surface, it is necessary to properly locate the collection and reinjection bores.
Such an optimisation has not been performed here, but the possible wastage of heat by
the incorrect positioning of wells is indicated.

9.9 FUTURE MODELLING OF GEOTHERMAL FIELDS

The range of conditions considered in this work falls short of a complete

representation of a real geothermal region. In the Wairakei geothernal system the
additional features of two-phase flow, non-isotropic permeability and mechanical
response of the porous rock still need to be accommodated. Two-phase considerations
are necessary for the simulation of upper "reservoir" although they may be omitted
from the deep models used here. Incorporation of permeability variation requires a
more concise description of the geophysical structure of the system than is
currently available. The mechanical response of the medium to alterations in the
flow (changes from water to steam, closing of fissures, subsidence etc.) is a problem
in solid mechanics and is as important a field for investigation as the investigation
of the convective flow itself. Although three-dimensional flows in a simple model
have been simulated in this work, the full scale numerical representation of an
actual geothermal system in three dimensions is likely to be enormously expensive
(although not impossible). However this difficulty may be evaded by the development
of an axisymmetric model - in terms of applicability to geothermal regions this
would lie somewhere between the two-dimensional and three-dimensional models derived
in this work.

95
APPENDIX A - THE ARAKAWA DIFFERENCING SCHEMES

The Arakawa f i n i t e d i f f e r e n c e r e p r e s e n t a t i o n s o f t h e a d v e c t i o n term

( e q u a t i o n 4 . 2 . 1 ) are as follows (Arakawa 1966):

t h i s b e i n g t h e s e c o n d - o r d e r scheme, and

forms t h e f o u r t h - o r d e r c o r r e c t i o n where t h e f o u r t h o r d e r scheme i s g i v e n by

Equation ( A . 2 ) d i f f e r s from t h a t i n t h e o r i g i n a l p a p e r i n which t h e r e i s a t y p o g r a p h i c a l

error. Also t h e e i , j t e r m s i n ( A . 1 ) a r e r e d u n d a n t and may be c a n c e l l e d o u t , r e s u l t -
ing i n a s l i g h t l y faster calculation.

96
APPENDIX B - AN EXTENSION OF THE BUNEMAN ALGORITHM TO FOURTH-ORDER ACCURACY

The a l g o r i t h m o f Busbee, Golub and N i e l s o n ( 1 9 7 0 ) f o r t h e s o l u t i o n o f

Poisson's equation

v2x Y , (B.1)

w i t h Dirichlet boundary c o n d i t i o n s , uses t h e u s u a l s e c o n d - o r d e r accurate r e p r e s e n t a -

t i o n of t h e L a p l a c i a n

(B.2)

where h i s t h e s p a t i a l i n c r e m e n t i n a s q u a r e MxN mesh (M = N ) , and c o n s t r u c t s a

s o l u t i o n m a t r i x equation i n t h e form

where -xj = xi,j i = 2,3,4 ... M - 1

i = 2,3,4 .. . M - 1 ,
and A i s a t r i d i a g o n a l m a t r i x o f t h e form

and T i s a n (M-2)x(M-2) i d e n t i t y m a t r i x . Now t h e s u b s t i t u t i o n I f o r T a l l o w s t h e

d e r i v a t i o n o f a s i m p l e and p a r t i c u l a r l y u s e f u l c o m p u t a t i o n a l p r o c e d u r e , however i f t h e
fourth- order representation

,jtl ' Xit1,j-I ' Xi-1,jt1 ' Xi-1,j-l )I , (B. 5)

i s used as s u g g e s t e d by Orszag and I s r a e l i (1974) t h e n A i s o f t h e form

97
 -
-10
3
2
5
2 -E 2
A = 3 3 3 , (B.6)
2
5
-x
3

-
-10
3

and T is of the form

T = (B.7)

In this case the simplified derivation of Busbee, Golub and Nielson (1970)
does not follow, and must be generalised to accommodate the case T # I .
Considering the rows ( j - l ) , j and (j+l) of the original (and perfectly
general) equations ( B . 3 ) , dropping the vector signs for simplicity-

The first reduction is achieved by multiplying the first and third equations by T
and the second by A and subtracting the second from the sum of the first and third -

This reduction is performed on every alternate row j = 3,5,7 etc. and therefore
halves the number of rows in the solution matrix. Following the complete reduction
new contracted matrices and vectors may be defined as

or

where A(1) (2T2 - A’)

98
or introducing

and qil)

then (1)
yj

Now after performing similar reductions r+l times

y?+l)

and

Combining these last two equations

and substituting

implies

Now if N = gk+’ + 1 , then k reductions may be performed resulting in

the single vector equation (provided the boundary conditions are homogeneous)

which is readily solved for the vector x2k+l , since A(k) may be reduced to a set
of qk tridiagonal factor matrices, after which the remaining vectors x may be solved
by back solving

o r in rewritten form

The algorithm hinges on the calculation of the sets of vectors ( r ) and

qir) defined by ( B . 8 ) and ( B . 9). It is seen that these equations are greatly
pj
simplified if T (and hence all T ( r ’ ) ) are identity matrices since in this case the

99
s o l u t i o n reduces t o t h e solving of t h e 2r t r i d i a g o n a l f a c t o r s of A") , a process
which i s c o m p u t a t i o n a l l y e f f i c i e n t . It i s t h e r e f o r e a s i n g l e matrix s o l u t i o n f o r
pir) t h e n s i m p l e a d d i t i o n t o o b t a i n q i r ) , and it i s t h i s b e n e v o l e n t s i m p l i c i t y
t h a t makes t h e Buneman a l g o r i t h m so c o m p u t a t i o n a l l y s t r e a m l i n e d . However i f T is
n o t a n i d e n t i t y m a t r i x t h e r e a r e t h r e e a d d i t i o n a l m a t r i x s o l u t i o n s t o be p e r f o r m e d
a t e a c h r e d u c t i o n , w i t h t h e r e s u l t t h a t t h e problem is g r e a t l y c o m p l i c a t e d and t h e
p r o c e d u r e as a whole t a k e s l o n g e r .

100
APPENDIX C - THE NUSSELT NUMBER

The Nusselt number is defined as the ratio of the total heat transfer
through a system to the heat transfer by conduction alone. For example, in the
simple partially heated model proposed in section 2 . 2 the heat lost (per unit
width) through the unheated and uninsulated boundaries will be

where k in this case is the conductivity of the fluid filled medium, and the purely
conductive heat loss would be

where 8' is the temperature field which would exist if fluid motion was disallowed.
Provided the boundary is held at temperature 9 = 0 , these equations are equally viable
for closed or recharge boundaries. Thus the Nusselt number is given by the ratio

ae
Nu = (C. 3)
ae'

which may be calculated approximately from the fields 8 and 9 ' using Simpson's
rule. The purely conductive terms in the denominator are termed the "conduction
coefficient" and need only be calculated once for each configuration that is considered.
A table of conduction coefficients is given in table X below.

Table X - Conduction Coefficients

Dimensionality Shape Heat input Conduction

Coefficient
~~
~~

Two-dimensional Square Uniform f=1.0 1.0

Non-uniform f=0.75 1.43
f=0.5 1.40
f=0.2 5 0.94
Rectangular nxl Uniform n
Rectangular 2x1 Non-uniform f=0.5 1.13
4x3 Non-uniform f=0.5 1.23
Square-insulated
lower boundary Non-uniform f=0.5 0.55

Three-dimensional Cube Uniform 8.0

Two quadrants 9.94
One quadrant 1.37

101
 The presence of a source or sink a l s o affects the heat transfer through
the system. The flow of fluid out of the system (per unit width) wil be R.q K
(taking into account the non-dimensionalisation) which results in an additional
heat transfer of

which is added to the transfer (C.1) and appears on the top line of ( C . 3 ) as a term
R.q.8 since the f a c t o r is unity from the definition of the thermal diffusivity.

102
APPENDIX D -
AN EXTENSION OF THE BUNEMAN ALGORITHM TO NEUMANN-TYPE BOUNDARY
CONDITIONS IN TWO AND THREE DIMENSIONS

The Buneman algorithm as previously described by Busbee, Golub and Nielson

(1970) and in appendix B, can be extended to the three-dimensional solution of
Poisson's equation, in which case the vectors x. and yj in (B.3) are replaced
-1
by corresponding plane matrices

c Xk 1 = xi,j,k i = 2,3 ..., M-1

' Yk yi,j ,k j = 2,3 ..., N-1

and the new A sub-matrix will itself be of the same form as the entire solution
matrix on the left hand side of (B.3) except with a - 6 instead of a -4 down the
leading principal diagonal. Since the second-order representation is used, T is
an identity matrix. The solutions of the reduction equations of the type

are now each achieved by using a complete two-dimensional odd-even reduction sequence
that is similar to the entire solution in the previous case (the only difference
being the value of the diagonal element). In the case of Dirichlet boundary conditions
this is not difficult to perform, but the Neumann conditions required for the pressure
solutions of section 4 . 2 introduce difficulties as outlined below.

When the ACr) matrices are factorised they take the form
2r
A(') = - II (A + 2cOse(r)) , (D.1)
j=1 j

for Dirichlet boundary conditions, and are always non-singular. However when Neumann
conditions are specified the final solution matrix becomes

and in two dimensions one of the factors (when j = 2k is singular. This is to be

expected as the solution of Poisson's equation is not unique in this case, however this
one singular factor matrix may be modified such that effectively a Dirichlet condition
is specified at one end of the row. This is equivalent to setting a reference
pressure at a single point in the region to be zero, thereby restraining the entire
field to be relative to that point.

In three dimensions the factor matrices must themselves be factorised and

in this case the A in (D.2) is of the form
zr
- (A' + 2c0sa(~))
j
3

103
 :
where A' i s t h e o r i g i n a l t r i d i a g o n a l matrix

*i
-6 2
1 -6 1
1 -6 1 ....
i -6

Thus t h e g e n e r a l form o f t h e most b a s i c f a c t o r m a t r i x i s

- d 2
1 d 1
1 d 1 .....

where d = -6 + 2cosa + 2 ~ 0 ~ 9
j j .

T h i s m a t r i x h a s z e r o d e t e r m i n a n t and one z e r o e i g e n v a l u e when

d=0 or d = -
+ 2

which is o n l y p o s s i b l e when

or case.1 = 2 - COS^
j
2;c+ln
Thus t h e m a t r i x i s s i n g u l a r i f b u t o n l y i f a j and e j a r e b o t h -.
Once a g a i n t h i s i s e q u i v a l e n t t o s p e c i f y i n g a r e f e r e n c e p r e s s u r e i n t h e 2k
field.

104
APPENDIX E - SPECTRAL REPRESENTATION OF EQUATIONS OF MOTION

Considering the equations of motion (2.3.8) and (2.3.11) with the uniform
boundary conditions (2.3.12), (2.3.13) and (2.3.14) and f = 1, the transformation into
Fourier space is made by defining

M N

M N
and e = (1-Y) + mgo nZ1 e& cosmnX sin nnY (E.2)

The conduction solution ( 1 - Y ) is separated making the Fourier boundary conditions

homogeneous and identically satisfied by the choice of transform. To simplify the
notation the stars are dropped on the Fourier coefficients e&, and @&, -
confusion
may be avoided by remembering that only the transformed variables carry subscripts.

The transformed equations are determined by substitution of (E.1) and (E.2)

into the two original equations (2.3.8) and (2.3.11), multiplying throughout by
sin knX sin LnY and cosknX sinLnY respectively, integrating over the range 0 2 X 2 1,
0 5 Y 5 1 and substituting the Euler properties of transcendental integrals. This
procedure is well known and is not set out fully here.

The streamfunction equation transforms to

which may be substituted into the transformed temperature equation which is

provided m + k 5 M and n + E L N and n # L in appropriate terms (which are otherwise

zero). Since the Euler property is only valid for non-zero integers the case k = 0
requires special attention and is described by the equation

105
 These equations may be solved simultaneously using the fourth-order Runge-
Kutta method described in Carnahan, Luther and Wilkes (1969 p.361). The solution
process is much lengthier than a similar finite difference procedure since each "point"
in the spectral mesh references every other "point". In a finite difference solution
the temperature at each point in the grid depends only on the values at the points on
the grid which are immediately adjacent to it. The spectral method may be computed
far more efficiently by using further transforms (Orszag 1971a) although it is then
not possible to use the simple transforms defined here in (E.1) and (E.2).

106
APPENDIX F - EVOLUTION OF THE THERMAL BOUNDARY LAYER

As an example o f t h e a n a l y s i s o f Sparrow, Husar and G o l d s t e i n ( 1 9 7 0 ) t o

d e t e r m i n e t h e r a t e o f e v o l u t i o n o f t h e t h e r m a l boundary l a y e r , t h e hydrodynamic
c o n d i t i o n s are i g n o r e d and o n l y c o n d u c t i v e h e a t t r a n s f e r c o n s i d e r e d . For a sudden
t e m p e r a t u r e r i s e a t t h e boundary of a s e m i - i n f i n i t e r e g i o n , t h e t e m p e r a t u r e e in
t h e medium o b e y s t h e r e l a t i o n s h i p

as g i v e n by Holman ( 1 9 6 8 p . 7 9 ) . In t h e range 0 - 0.8 t h e e r r o r f u n c t i o n i s approxi-

m a t e l y l i n e a r , so i f t h e c o n d u c t i o n l a y e r i s c o n s i d e r e d t o l i e between t e m p e r a t u r e s o f
1 . 0 and 0 . 2 t h e n t h e non- dimensional t h i c k n e s s 6 of t h i s l a y e r i s

Now s i n c e t h e e f f e c t i v e l o c a l R a y l e i g h number i s d i r e c t l y d e p e n d e n t on t h e t h i c k n e s s
6 , t h e n it f o l l o w s t h a t Rlocal i n t h e boundary l a y e r i n c r e a s e s as 6 u n t i l it
r e a c h e s i t s c r i t i c a l v a l u e when c o n v e c t i o n b e g i n s t o o c c u r i n t h e l a y e r . The i n t e r v a l
7 r e q u i r e d f o r t h e s o l u t i o n t o r e a c h t h i s s t a g e i s d e p e n d e n t on t h e t e m p e r a t u r e
P
d i f f e r e n t i a l a c r o s s t h e l a y e r , which i s d e r i v e d from t h e t r u e R a y l e i g h number f o r t h e
o v e r a l l system. Thus t h e i n t e r v a l i s r e l a t e d t o R by t h e e x p r e s s i o n

7
P
a R - ~ . (F.3)

T h i s r e l a t i o n s h i p d e s c r i b e s t h e r a t e o f f o r m a t i o n o f t h e boundary l a y e r i n
t h e absence of o t h e r cooling e f f e c t s , p a r t i c u l a r l y t h e flow of f l u i d a c r o s s t h e
heater. It i s n o t s u r p r i s i n g t h e n t h a t (F.3) p r e d i c t s s h o r t e r g e s t a t i o n t i m e s t h a n
the relationship

t h a t i s o b s e r v e d i n t h e f l o w s o f c h a p t e r 5, i n which t h e r e i s no f l u i d boundary l a y e r .

U n f o r t u n a t e l y it i s n o t p o s s i b l e t o p e r f o r m a similar a n a l y s i s f o r t h e c a s e
o f f o r c e d c o o l i n g o f t h e boundary l a y e r s i n c e t h e v e l o c i t i e s a r e n o t c o n s t a n t .
However t h e r e s u l t (F.3) d o e s i n d i c a t e a t e n d e n c y o f t h e f l o w and c o r r e c t l y p r e d i c t s
t h a t t h e boundary l a y e r e v o l v e s more q u i c k l y t h a n i n t h e f l u i d l a y e r (Benard)
p r o b l e m . For a p o r o u s medium problem w i t h o u t major c i r c u l a t i n g v e l o c i t i e s , Wooding
( 1 9 6 9 ) d e t e r m i n e d t h a t t h e c e l l w i d t h is d e p e n d e n t on t h e s q u a r e r o o t o f t h e time
i n t e r v a l , which c o n f i r m s t h e r a t e o f growth o f t h e boundary l a y e r p r e d i c t e d by (F.2).

107
Therefore defining a pseudo-streamfunction

$* = 1
-
JT
($3 - $1) ,

then v 1 -$ ,

and U ' = $ ,

and a representation of the streamlines on the centre diagonal plane may be obtained
by plotting contours of this function.

109
APPENDIX G - STREAMLINE REPRESENTATION IN THREE-DIMENSIONAL FLOWS
In two dimensions streamlines are easily determined as lines of equal
stream function value, but in three dimensions their visualisation is not so simple.
The velocity field in three dimensions is given by

Now the horizontal axes X and Z are rotated through 4 5 O as in figure G.l so that
the new axis X' lies along the diagonal of the region.

FIGURE G.1 - Rotation of Axes

so that the vertical velocity

and, provided X' is an axis of symmetry,

v = - -l a ($3 - $1) . (G.1)

6
The velocity along the X' axis is given by
REFERENCES

Arakawa, A., 1966, J. Comp. Phys. 1, 119

Aravin, V.I., and Numerov, S.N., 1965, "Theory of Fluid Flow in Undeformable Porous
Media". Israel Program for Scientific Translations, Jerusalem.
Aziz, K., and Hellums, J.D., 1967, Phys. of Fluids 10, 314
Baines, P.G., and Gill, A.E., 1969, J. Fluid Mech. 37, 289
Batchelor, G.K., 1967, "An Introduction to Fluid Dynamics", Cambridge University Press
Beck, J.L., 1972, Phys. of Fluids 15, 1377
Betbeder, J., and Jolas, P., 1972, Int. J. Heat Mass Transfer 15, 721
Birikh, R.V., Gershuni, G.Z., Zhukhovitskii, E.M., and Rudakov, R.N., 1972, P.M.M.
36, 745
Bories, S., and Combarnous, M.A., 1973, J. Fluid Mech. 57, 63
Bories, S., Combarnous, M.A., and Jaffrenou, J.Y., 1972 , Comptes Rendus Acad. Sci.
Paris 275A, 857
Bories, S., and Thirriot, C., 1969, La Houille Blanche 24, 237
Brown, W.S., 1973, J . Fluid Mech. 60, 539
Buretta, R., and Berman, A.S., 1974, "Convective heat transfer in a liquid-saturated
porous layer" (To appear)
Busbee, B.L., Golub, C.H., and Nielson, C.W., 1970, S.I.A.M. Journ. on Num. Anal.
7, 627
Busse, F.H., and Joseph, D.D., 1972, J. Fluid Mech. 54, 521
Busse, F.H., and Whitehead, J.A., 1971, J. Fluid Mech. 47, 305
1974, J. Fluid Mech. 66, 67
Caltagirone, J.P., Cloupeau, M., and Combarnous, M.A., 1971, Comptes Rendus Acad.
Sci. Paris 273B, 833
Carnahan, B., Luther, H.A., and Wilkes, J.D., 1969, "Applied Numerical Methods"
John Wiley and Sons, New York
Chandrasekhar, S . , 1961, "Hydrodynamic and Hydromagnetic Stability", Oxford Clarendon
Press, London
Chan, B.K.C., Ivey, C.M., and Barry, J.M., 1970, J. Heat Transfer, Trans. ASME 92
Series C, 21
Combarnous, M.A., 1970, Revue Generale de Thermique 108, 1355
1972, Comptes Rendus Acad. Sci. Paris 275A, 1375
Combarnous, M.A., and Bia, F., 1971, Soc. Petr. Eng. Journ. December, 399
Combarnous, M.A., and Bories, S., 1974, Int. J. Heat Mass Transfer 17, 505
Combarnous, M.A., and LeFur, B., 1969, Comptes Rendus Acad. Sci. Paris 269B, 1009
Cooley, J.W., Lewis, P.A.W., and Welch, P.D., 1967, I.E.E.E. Transactions AU-15, 79
Crowley, W.P., 1967, J. Comp. Phys. 1, 471
Cullen, M.J.P., 1974, J. Inst. Maths. Applics. 13, 233
Curry, D.M., 1974, N.A.S.A. Technical Note TN D-7606
Denton, R.A., and Wood, I.R., 1974, Proc. 5th Austr. Conf. Hydr. and Fluid Mech., 361
Donaldson, I.G., 1962, J. Geoph. Res. 67, 3449
1968, N.Z. Journ. Sci. 11, 3
1970, Geothermics - Special Issue 2, 649
1974, Private communication, D.S.I.R. Wellington
Dybbs, A., and Schweitzer, S . , 1973, Journ. Hydrology 20, 171
Engineering Sciences Data, (1968), Items no. 68009, 68010.
Elder, J.W., 1966a, N . Z . D.S.I.R. Bulletin 169
1966b, J . Fluid Mech. 2 4 , 823
1967a, J. Fluid Mech. 27, 29

110
Elder, J.W., 1967b, J. Fluid Mech. 27, 609
1968, J. Fluid Mech. 32, 69
1969, Phys. of Fluids Suppl. II 12, 194
Fisher, R.G., 1964, N.Z. Jour. Geol. Geophys. 7, 172
Foster, T.D., 1968, J. Geoph. Res. 73, 1933
1969, J. Geoph. Res. 74, 6967
F r o m , J.E., 1965, Phys. of Fluids 8, 1757
1969a, Phys. of Fluids Suppl. II 12, 3
1969b, J. Comp. Phys. 3, 176
Gebhart, B., 1973, Ann. Rev. Fluid Mech. 5, 213
Gill, A . E . , 1969, J. Fluid Mech. 35, 545
Gill, A . E . , and Davey, A., 1969, J. Fluid Mech. 35
Gill, A . E . , and Kirkham, C.C., 1970, J. Fluid Mech. 4 2 , 125
Gupta, V.P., and Joseph, D.D., 1973, J. Fluid Mech. 57, 491
Hart, J.E., 1973, J. Fluid Mech. 59, 47
Healy, J., and Hochstein, M.P., 1973, Journ. Hydrology (N.Z.) 12, 71
Hochstein, M.P., 1974, Private communication, University of Auckland
Holman, J.P., 1968, "Heat Transfer", McGraw-Hill
Holst, P.H., and Aziz, K., 1972a, Canadian J. Chem, Eng. 50, 232
1972b, Int. J. Heat Mass Transfer 15, 73
Homsy, G.M., 1973, J. Fluid Mech. 60, 129
1974, J. Fluid Mech. 62, 387
Horne, R.N., and O'Sullivan, M.J., 1974a, J. Fluid Mech. 66, 339
1974b, Proc. 5th Austr. Conf. Hydr. and Fluid
Mech., 231
Horton, C.W., and Rogers, F.T.Jr., 1945, J. Appl. Phys. 16, 367
Howard, L.N., 1964, Proc. 11th Int. Congr. Appl. Mech, Munich, ed. H. Gortler,
Berlin : Springer-Verlag.
Hurle, D.T.J., and Jakeman, E., 1971, J. Fluid Mech. 47, 667
Jannot, M., Naudin, P., and Viannay, S., 1973, Int. J. Heat Mass Transfer 16, 395
Joseph, D.D., 1971, J. Fluid Mech. 47, 257
Kaneko, T., Mohtadi, M.F., and Aziz, K., 1974, Int. J. Heat Mass Transfer 17, 485
Kantorovich, L.V., and Krylov, V.I., 1964, "Approximate Methods of Higher Analysis",
Groningen : P. Noordhof Ltd.
Katto, Y . , and Masuoka, T., 1967, Int. J. Heat Mass Transfer 10, 297
Keller, J.B., 1966, J. Fluid Mech. 26, 599
Kreiss, H., and Oliger, J., 1973, "Methods for the Approximate Solution of Time
Dependent Problems", G.A.R.P. Publ. Ser. 10, World Meteor. Org.
Krishnamurti, R., 1970a, J. Fluid Mech. 4 2 , 295
1970b, J. Fluid Mech. 42, 309
Kruger, P. (ed.), 1973, "Geothermal Energy Resources, Production and Stimulation",
Stanford University Press.
Lambert, R.B., and Demenkow, J.W., 1971, J. Fluid Mech. 54, 627
Lapwood, E.R., 1948, Proc. Camb. Phil. Soc. 44, 508
Legros, J.C., Platten, J.K., and Poty, P.G., 1972, Phys. of Fluids 15, 1383
Lilly, D.K., 1965, Mon. Weath. Rev. 93, 11
McNabb, A., 1965, Proc. 2nd Austr. Conf. Hydr. and Fluid Mech., C 161
Maksimov, E.A., and Stradomskii, M.V., 1971, Inzhenerno-Fizicheskii Zhurnal 2 0 , 588
Mercer, J.W., 1973, "Finite Element Approach to the Modelling of Hydrothermal
Systems", Ph.D. Thesis, Univ. of Illinois
1974, Private communication
Mikhlin, S.G., and Smolitsky, K.L., 1967, "Approximate Methods f o r the Solution of
Differential and Integral Equations", Elsevier, N . Y .
M i t c h e l l , A.R., 1 9 6 9 , " C o m p u t a t i o n a l Methods i n P a r t i a l D i f f e r e n t i a l E q u a t i o n s " ,
J o h n Wiley
Moore, D . R . , a n d Weiss, N . O . , 1 9 7 3 , J. F l u i d Mech. 58, 289
N i e l d , D.A., 1 9 6 7 , J . F l u i d Mech. 2 9 , 545
1 9 6 8 , Water R e s o u r c e s R e s e a r c h 4, 553
Orszag, S . A . , 1 9 7 1 a , M.I.T. S t u d . Appl. Math. 50, 293
1 9 7 1 b , J. F l u i d Mech. 4 9 , 75
1 9 7 1 c , J. F l u i d Mech. 50, 689
Orszag, S . A . , and I s r a e l i , M . , 1 9 7 4 , Ann. Rev. F l u i d Mech. 6, 281
O ' S u l l i v a n , M.J., 1 9 7 3 , " S t e a d y Convection i n a P o r o u s Medium", P r i v a t e communication
Palm, E . , Weber, J . E . , a n d K v e r n v o l d , O., 1 9 7 2 , J . F l u i d Mech. 54, 1 5 3
Platten, J.K., a n d C h a v e p a y e r , G., 1 9 7 3 , J. F l u i d Mech. 60, 305
P l a t z m a n , G . W . , 1 9 6 5 , J. F l u i d Mech. 2 3 , 4 8 1
Plows, W.H., 1 9 6 8 , Phys. of F l u i d s 11, 1 5 9 3
P r a t s , M . , 1 9 6 6 , J. Geoph. Res. 7 1 , 4835
R u b i n , H., 1 9 7 4 , J o u r n . Hydrology 21, 1 7 3
Runchal, A . K . , S p a l d i n g , P.B., and Wolfshtein, M . , 1 9 5 9 , Phys. of F l u i d s S u p p l . 11,
12, 2 1
S c h e c t e r , R.S., 1 9 6 7 , "The V a r i a t i o n a l Method i n E n g i n e e r i n g " , M c G r a w - H i l l
S h i r t c l i f f e , T.G.L., 1 9 7 3 , J . F l u i d Mech. 57, 27 ,
S h i r t c l i f f e , T.G.L., a n d T u r n e r , J . S . , 1 9 7 0 , J. F l u i d Mech. 41, 707
Sparrow, E.M., Husar, R.B., and G o l d s t e i n , R.J., 1 9 7 0 , J. F l u i d Mech. 41, 793
S t r a u s , J.M., 1 9 7 4 , J. F l u i d Mech. 64, 5 1
T a u n t o n , J.W., a n d L i g h t f o o t , E.N., 1 9 7 0 , Chem. Eng. S e i . 25, 1939
T a u n t o n , J.W., L i g h t f o o t , E . N . , a n d Green, T . , 1 9 7 2 , Phys. o f F l u i d s 1 5 , 748
T o r r a n c e , K.E., 1 9 6 8 , J. Res. N.B.S. 72B, 281
T o r r a n c e , K.E. a n d T u r c o t t e , D . L . , 1971, J . F l u i d Mech. 47, 1 1 3
V e r o n i s , G . , 1 9 6 6 , J. F l u i d Mech. 2 6 , 49
Wankat, P.C., and S c h o w a l t e r , W.R., 1 9 7 0 , Phys. of F l u i d s 1 3 , 2418
Weber, J . E . , 1 9 7 3 , I n t . J. Heat Mass T r a n s f e r 16, 961
Welander, P . , 1 9 6 7 , J . F l u i d Mech. 29, 1 7
1 9 7 1 , J . F l u i d Mech. 4 7 , 51
Westbrook, D . R . , 1 9 6 9 , Phys. of F l u i d s 1 2 , 1547
Wilkes, J . O . , and C h u r c h i l l , S.W., 1966, Am. I n s t . Chem. Eng. J . 1 2 , 161
Willis, G.E., a n d D e a r d o r f f , J.W., 1 9 6 7 , Phys. of F l u i d s 10, 931
1 9 7 0 , J. F l u i d Mech. 44, 6 6 1
W i r t z , R.A., B r i g g s , D.G., and Chen, C.F., 1 9 7 2 , Geoph. F l u i d Dyn. 3 , 265
Wooding R . A . , 1 9 5 7 , J. F l u i d Mech. 2 , 273
1 9 6 0 , J. F l u i d Mech. 7 , 5 0 1
1 9 6 9 , J . F l u i d Mech. 39, 477
Yen, Y.C., 1 9 7 4 , I n t . J. Heat Mass T r a n s f e r 1 7 , 1349
Yih, C.S., 1 9 6 9 , " F l u i d Mechanics" , M c G r a w - H i l l
Young, R . E . , 1 9 7 4 , J . F l u i d Mech. 6 3 , 695

112