Mixed Convection in Fluid Superposed

Porous Layers 1st Edition John M.

Dixon
Visit to download the full and correct content document:
https://textbookfull.com/product/mixed-convection-in-fluid-superposed-porous-layers-
1st-edition-john-m-dixon/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Convection in Porous Media 5th Edition Donald A. Nield

https://textbookfull.com/product/convection-in-porous-media-5th-
edition-donald-a-nield/

Multiphase Fluid Flow in Porous and Fractured

Reservoirs 1st Edition Wu

https://textbookfull.com/product/multiphase-fluid-flow-in-porous-
and-fractured-reservoirs-1st-edition-wu/

Fluid Dynamics: Part 3 Boundary Layers Anatoly I. Ruban

https://textbookfull.com/product/fluid-dynamics-part-3-boundary-
layers-anatoly-i-ruban/

Applications of Heat Mass and Fluid Boundary Layers 1st

Edition R. O. Fagbenle

https://textbookfull.com/product/applications-of-heat-mass-and-
fluid-boundary-layers-1st-edition-r-o-fagbenle/
Physicochemical Fluid Dynamics in Porous Media
Applications in Petroleum Geosciences and Petroleum
Engineering Mikhail Panfilov

https://textbookfull.com/product/physicochemical-fluid-dynamics-
in-porous-media-applications-in-petroleum-geosciences-and-
petroleum-engineering-mikhail-panfilov/

Diffusion Nmr of Confined Systems Fluid Transport in

Porous Solids and Heterogeneous Materials Rustem
Valiullin

https://textbookfull.com/product/diffusion-nmr-of-confined-
systems-fluid-transport-in-porous-solids-and-heterogeneous-
materials-rustem-valiullin/

How to Publish in Biomedicine 500 Tips for Success

Third Edition John Dixon

https://textbookfull.com/product/how-to-publish-in-
biomedicine-500-tips-for-success-third-edition-john-dixon/

Are Some Languages Better than Others? Robert M. W.

Dixon

https://textbookfull.com/product/are-some-languages-better-than-
others-robert-m-w-dixon/

Australia s Original Languages An introduction R. M. W.

Dixon

https://textbookfull.com/product/australia-s-original-languages-
an-introduction-r-m-w-dixon/
SPRINGER BRIEFS IN
APPLIED SCIENCES AND TECHNOLOGY

John M. Dixon
Francis A. Kulacki

Mixed Convection
in Fluid
Superposed
Porous Layers

123
SpringerBriefs in Applied Sciences
and Technology

More information about this series at http://www.springer.com/series/8884

John M. Dixon • Francis A. Kulacki

Mixed Convection in Fluid

Superposed Porous Layers
John M. Dixon Francis A. Kulacki
Department of Mechanical Engineering Department of Mechanical Engineering
University of South Florida University of Minnesota
Tampa, FL, USA Minneapolis, MN, USA

ISSN 2191-530X ISSN 2191-5318 (electronic)

SpringerBriefs in Applied Sciences and Technology
ISBN 978-3-319-50786-6 ISBN 978-3-319-50787-3 (eBook)
DOI 10.1007/978-3-319-50787-3

Library of Congress Control Number: 2016961237

© The Author(s) 2017

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or
dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt
from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained
herein or for any errors or omissions that may have been made.

Printed on acid-free paper

This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface

Extensive research has been done in the field of natural, mixed, and forced
convection in a porous layer. Several studies have investigated natural and forced
convection in a system that includes a porous and a superposed fluid layer, but
mixed convection has not been addressed. The present monograph is motivated to
fill the gap in the literature regarding mixed convection.
We investigate mixed convective heat transfer in a long channel that is partially
filled with a porous layer and has a fluid layer above it. The channel is heated on the
bottom over a finite length with cross flow along the length of the channel. The two
sublayers are treated as a single domain numerically, and the porosity is used as a
switching parameter, causing the governing conservation equations to transition
from an extended form of the Darcy-Brinkman-Forchheimer equation in the porous
sublayer to the Navier-Stokes equations in the fluid sublayer. This methodology
avoids the need for boundary conditions at the interface between the two domains.
Dimensionless groups are varied and include the Péclet number, Rayleigh number,
the porous sublayer height, Darcy number, Prandtl number, and the conductivity
ratio between the solid and fluid phases. The impact of the various additional terms
in the extended form of Darcy’s law is also investigated.
The conductivity ratio, Darcy number, porous sublayer height, Rayleigh num-
ber, and Péclet number all have a strong effect on the overall Nusselt number, while
the Prandtl number, Brinkman effect, Forchheimer effect, and convective terms
have negligible effects on Nusselt numbers. A Péclet number is observed at which
the Nusselt number is a minimum and is shown to be proportional to the Rayleigh-
Darcy number, the product of the Rayleigh and Darcy numbers, and inversely
proportional to the porous sublayer height. This Péclet number is termed the
“critical Péclet number.” A critical porous sublayer height ratio is also observed
at which the Nusselt number is a minimum and is proportional to the Rayleigh-
Darcy number and inversely proportional to the Péclet number. Streamlines capture
the transition from the natural convection regime to the forced convection regime.
In the transition region, flow patterns have characteristics of both. Isotherms

v
vi Preface

capture the thermal plume above the heated wall and show the influence of cross
flow on the shape and character of the plume.
An experimental apparatus is designed in order to collect data over a similar
range of parameters explored numerically. The numerical results show good agree-
ment with the experimental data within the bounds of uncertainty. The experiments
confirm the presence of the critical Péclet number. However, they do not show the
same trends at intermediate porous layer heights. The effect of the dimensionless
porous sublayer height on the Nusselt number is shown to be small from 0.5 to 1.

Tampa, FL John M. Dixon

Minneapolis, MN Francis A. Kulacki
Contents

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 General Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Effective Conductivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3 Interfacial Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.4 One-Domain Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.5 Mixed Convection in Saturated Porous Media . . . . . . . . . . . . . . . . . 8
1.6 Natural Convection in Porous Media . . . . . . . . . . . . . . . . . . . . . . . . 12
1.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2 Mathematical Formulation and Numerical Methods . . . . . . . . . . . . . . 15
2.1 Solution Domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.2 Governing Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3 One-Domain Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.4 Numerical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3 Numerical Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.1 Verification of Solution Method . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.1.1 Rayleigh-Bénard Convection . . . . . . . . . . . . . . . . . . . . . . . . 21
3.1.2 The Horton-Rogers Lapwood Problem . . . . . . . . . . . . . . . . . 21
3.1.3 Natural Convection in Fluid-Superposed Porous Layers . . . . . 24
3.1.4 Mixed Convection in a Porous Layer . . . . . . . . . . . . . . . . . . 26
3.2 Mixed Convection in a Fluid-Superposed Porous Layer . . . . . . . . . . 27
3.2.1 Variation of Thermal Conductivity . . . . . . . . . . . . . . . . . . . . 27
3.2.2 Inclusion of Brinkman and Forchheimer Terms . . . . . . . . . . . 30
3.2.3 Effect of Prandtl and Darcy Numbers . . . . . . . . . . . . . . . . . . 32
3.2.4 Sublayer Critical Rayleigh Number . . . . . . . . . . . . . . . . . . . 34
3.2.5 Critical Péclet Number . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.2.6 Transition from Natural to Forced Convection . . . . . . . . . . . 37
3.2.7 Critical Sublayer Height . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

vii
viii Contents

4 Measurement of the Heat Transfer Coefficient . . . . . . . . . . . . . . . . . . . 47

4.1 Apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.2 Procedure and Data Reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.3 Experimental Uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.4.1 Validation of Numerical Results . . . . . . . . . . . . . . . . . . . . . 51
4.5 Correlation Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
5 Summary of Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
5.1 Critical Péclet Number . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
5.2 Variation of Physical Quantifies . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
5.3 Streamlines and Isotherms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
5.4 Heat Transfer Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

Appendix: Heat Transfer Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
Nomenclature

a Area (m2)
A Heated wall area (m2)
CF Forchheimer coefficient, Eqn. (2.8)
cp Specific heat at constant pressure (J/kg K)
d Particle diameter (m)
Da Darcy number, K/H2
DH Channel hydraulic diameter
g Gravitational acceleration, (0,
g) (m/s2)
h Heat transfer coefficient (W/m2K)
H Height (m)
k Thermal conductivity (W/mK)
K Permeability (m2)
Lh Length of heated wall (m)
Nu Nusselt number, hH/km
p Pressure (Pa)
p
Dimensionless pressure, pH2/ρ0α2f
Pe Péclet number, U0H/αf
Pr Prandtl number, νf/αf
q Heat transfer (W)
q000 Energy generation (W)
Ra Rayleigh number, gβH3(Tw
T0)/αfνf
t Time (s)
T Temperature (K)
U0 Volumetric flow/area (m/s)
v Fluid velocity, (u,v) (m/s)
v
Dimensionless velocity, (u/U0,v/U0)
w Weight factor, Eqn. (1.22)
x Horizontal coordinate (m)
y Vertical coordinate (m)

ix
x Nomenclature

Greek Symbols

α Thermal diffusivity (m2/s)

β Isobaric coefficient of thermal expansion (K
1)
γ Beavers-Joseph constant, Eqn. (1.19)
δ Dimensionless heated wall length, Lh/H
ϕ Porosity (
)
η Dimensionless porous sublayer height, Hp/H
θ Dimensionless temperature, (T–T0)/(Tw–T0)
κ Conductivity ratio, ks/kf
λ Diffusivity ratio, αe/αf
μ Viscosity (Pa s)
μ
e Brinkman viscosity (Pa s)
μ
^ Dimensionless Brinkman viscosity μ e=μ
ν Kinematic viscosity (m2/s)
τ Dimensionless time, tαf/H2
ρ Density (kg/m3)
σ Heat capacity ratio, (ρcp)m/(ρcp)f
τ Dimensionless time, tαf/H2
τp Dimensionless oscillatory period, Eqn. (1.21)
ω Vorticity

Subscripts

c Critical value
e Effective
f Fluid
fs Fluid-solid interface
h Heater
H Total height
m Average value
p Porous
s Solid
w Wall
0 Reference
Chapter 1
Introduction

1.1 General Considerations

Research on energy and momentum transport in bottom-heated saturated porous

layers has established an extensive literature, but mixed convection in horizontal
fluid-superposed porous layers remains an open topic. The research reported in this
monograph is aimed at closing this gap in the knowledge base. A continuing issue
for the fluid-superposed layer is the treatment of the interface at the porous
sublayer. Much work has been done on this problem over the years [1–7], but
nothing appears to have been produced that is directly applicable to buoyancy-
driven flows. Recent work [1] on natural convection in horizontal fluid-superposed
porous layers heated from below has shown disagreement between numerical pre-
dictions and experimental results. This could be due to the limitations of the current
methods being used and further highlights the need for an appropriate method to
handle the interface between the pure fluid layer and fluid-saturated porous layer.
From the earliest stages of research on energy transport in porous media, the
mathematical approach has been to treat the porous medium as a continuum with a
representative elementary volume that averages out pore-size flow characteristics,
replacing them with a larger-scale average and thus enabling governing equations
to treat the porous medium as a continuum [2, 3]. However, some of the physics of
transport is lost in the process of averaging. Herein rises the challenge of defining
the boundary conditions between a porous layer and its superposed porous layer.
The research presented in this monograph (1) obtains a numerical solution to the
problem of mixed convection in horizontal fluid-superposed porous layers heated
from below with a variable porous sublayer height; (2) determines a suitable
method for handling the boundary between a fluid and porous layer for
buoyancy-driven flows and uses it to numerically solve the governing partial
differential equation for mass, momentum, and energy conservation; and (3) deter-
mines the heat transfer laws pertaining to a range of the governing hydrodynamic

© The Author(s) 2017 1

J.M. Dixon, F.A. Kulacki, Mixed Convection in Fluid Superposed Porous Layers,
SpringerBriefs in Applied Sciences and Technology,
DOI 10.1007/978-3-319-50787-3_1
2 1 Introduction

Fig. 1.1 Cross flow in a porous layer with a superposed fluid

and geometrical parameters defining the system. The problem space is limited to
heating along a finite length of the lower boundary of the porous sublayer and a cold
sink temperature on the upper surface of the fluid sublayer. Figure 1.1 shows the
general features of the problem domain.
Heat transfer and fluid flow in porous media has been studied for over 150 years.
Textbooks are available that provide clear derivations of the governing equations
and give a condensed review of the relevant literature. Nield and Bejan [2] provide
a critical review of literature through 2006, and Whitaker [3] derives various forms
of Darcy’s law through an exposition of volume-averaging methods. Approxi-
mately 200 papers in recent years have been published annually on the topic
[2]. In the following paragraphs, we establish the general background to of the
problem and point to several relevant prior investigations of the heat transfer
problem.
Darcy’s law was first established in 1856 and is given as:

K
v ¼  ∇p: ð1:1Þ
μf

This equation directly relates the bulk, or drift, velocity to the pressure gradient,
∇p; the fluid viscosity, μf; and the bulk permeability, K. This equation is simple and
easy to use, but it can clearly be seen that it does not allow for any variation in the
velocity profile. As a consequence, for a fixed pressure gradient, the velocity profile
is flat, thus ignoring any edge effects due to solid or free boundaries.
The Forchheimer term was later added to account for inertial effects in the flow
[4]. The proper definition of the constant on the Forchheimer term, CF, has been
debated [2, 5], but the general form is:

μf CF
∇p ¼  v  1=2 ρf jvjv: ð1:2Þ
K K
1.1 General Considerations 3

The Forchheimer term is also known as the quadratic drag term and becomes
significant at higher flow velocities. The pressure drop for flow through a restric-
tion, e.g., orifice, pipe fitting, or nozzle, is proportional to the square of the average
fluid velocity and including the Forchheimer term provides a more physically
realistic governing equation. However, it can safely be neglected at the typical
low flow velocities found in many porous media applications.
Brinkman [6] derived a new equation treating the problem as flow over a sphere,
resulting in what is commonly known as the Brinkman equation:
μf
∇p ¼  vþμ
e∇2 v, ð1:3Þ
K

where μ e is denoted as the Brinkman viscosity, or porous viscosity, which is often

taken to be the fluid viscosity but not necessarily equal to it. This equation is very
similar to Darcy’s law, but the Brinkman term is not universally accepted, has a
limited range of validity, and is typically negligible outside of boundary regions
[2]. Including the Brinkman term allows for the variation in the velocity profile as
occurs at solid boundaries.
To account for time-varying flow in porous media, many authors have used the
work of Wooding [7] and have added similar acceleration terms (Eqn. (1.4)), but the
appropriateness of these additions has been questioned [2, 5]:
 
∂v μ
ρf ϕ1 þ ϕ2 ðv  ∇Þv ¼ ∇p  f v: ð1:4Þ
∂t K

The convective term has been considered an inappropriate expression of the

nonlinear drag forces, and the Forchheimer term is preferred. Other expressions
have been suggested to account for transient effects but have been shown to be
negligible for many applications. For example, for liquid metal flowing through an
exceptionally high permeability medium, the characteristic time is 1 s [2]. Thus
transient terms can safely be neglected in most applications. It should also be noted
that the inclusion of the porosity in the transient and convective terms is ad hoc and
no rigorous development or justification is available [5].
When considering the effects of buoyancy, an additional term is added to
Darcy’s law:

K
v¼ ð∇p þ ρf gÞ: ð1:5Þ
μf

The Boussinesq approximation for the equation of state allows the density in the
buoyant term to vary linearly with temperature and is widely used when the
temperature variation is not too large. Carr and Straughan [8] consider water near
the freezing point and suggest using a quadratic function that allows for the
minimum density of water to occur at 4  C.
4 1 Introduction

Combining the Forchheimer, Brinkman, Wooding, and buoyancy terms with

Darcy’s law, and using the Boussinesq approximation, the Darcy-Brinkman-
Forchheimer (DBF) equation governs the flow:
 
∂v
ρf ϕ1 þ ϕ2 ðv  ∇Þv
∂t
ð1:6Þ
μ CF ρf jvjv
¼ ∇P  f v þ μ e ∇2 v  1  ρf , 0 gð1  βðT  T 0 ÞÞ:
K K2

The l.h.s. represents the transient and convective terms. The first term on the
r.h.s. is the pressure gradient, which is the driving force of any bulk flow. The
second term on the r.h.s. represents the linear drag forces, and the fourth term
represents the quadratic drag forces. The third term accounts for the edge effects in
the flow, and without it, the basic velocity profile would be uniform. The fifth term
accounts for the buoyancy effects.
The energy equation for a porous medium is much less debated than the
momentum equation. However, there are several aspects to consider owing to the
two-phase nature of a porous medium. The basic energy equations are:

∂T s
ð1  ϕÞρs cp, s ¼ ð1  ϕÞ∇  ðks ∇T s Þ þ ð1  ϕÞqs 000 ð1:7Þ
∂t
∂T f
ϕρf cp, f þ ρf cp, f v  ∇T f ¼ ϕ∇  ðkf ∇T f Þ þ ð1  ϕÞqf 000 , ð1:8Þ
∂t

where Eqn. (1.7) represents the temperature of the solid phase and Eqn. (1.8) the
temperature of the fluid phase.
Heat transfer between the solid and fluid phases occurs at the interface and has
been handled in different ways [2, 3], but the results are most simply explained by
defining an effective heat transfer coefficient between them and altering Eqns. (1.7)
and (1.8) to account for the interphase transfer:

∂T s
ð1  ϕÞρs cp, s ¼ ð1  ϕÞ∇  ðks ∇T s Þ þ ð1  ϕÞqs 000 þ hfs ðT f  T s Þ ð1:9Þ
∂t
∂T f
ϕρf cp, f þ ρf cp, f v  ∇T f ¼ ϕ∇  ðkf ∇T f Þ þ ð1  ϕÞqf 000
∂t
þ hfs ðT s  T f Þ, ð1:10Þ

where volumetric energy sources in the solid and fluid phases are formally
included. Determining the values of the interphase heat transfer coefficient, hfs,
remains a challenge, but they were experimentally determined by Polyaev [9] and
shown to be in general agreement with the theoretical work of Dixon and Cresswell
[10]. For a bed of particles, the following formulas define the heat transfer
1.2 Effective Conductivity 5

coefficient [2], where afs is the surface area per unit volume, d is the particle
diameter, and B ¼ 10 for spherical particles:

hfs ¼ afs h* ð1:11Þ

afs ¼ 6ð1  ϕÞ=d ð1:12Þ
1 d d
*
¼ þ ð1:13Þ
h Nufs kf Bks
 
0:255 1= 2=
Nufs ¼ Pr 3 Red3 , ð1:14Þ
ϕ

where the particle Reynolds number, Red, is defined in terms of the particle
diameter, d, drift velocity in the porous medium, |v|, and fluid viscosity, μf. Other
models [11] have been shown to produce good agreement with measurement but are
omitted from the present review.
Thorough analysis of interface heat transfer has been conducted, and volume
averaging techniques have been employed, resulting in criteria for local thermal
equilibrium [3, 7]. In many applications it is often appropriate to make the assump-
tion of local thermal equilibrium. The criteria are always met in steady state and
reduce the equations to a single governing equation:

∂T
ρm c p , m þ ρf cp, f v  ∇T ¼ ∇  ðkm ∇T Þ þ qm 000 ð1:15Þ
∂t

where the subscript “m” represents the volume-averaged thermophysical properties

of the two-phase system:

ρm cp, m ¼ ð1  ϕÞρs cp, s þ ϕρf cp, f ð1:16Þ

000 000 000
qm ¼ ð1  ϕÞqs þ ϕqf ð1:17Þ

km ¼ ð1  ϕÞks þ ϕkf ð1:18Þ

1.2 Effective Conductivity

Characterizing the heat transfer and fluid flow characteristics in a porous medium
requires large-scale averaging of the flow. In the averaging process, pore level
physics is lost and is added back in a variety of ways. For example, consider a
porous medium of glass spheres filled with stagnant water and heated at the bottom.
Heat is transferred into the water and into the glass spheres, and the overall
conductivity is an average of the conductivities of the water and the glass spheres.
6 1 Introduction

The resulting value depends on the amount of each of the phases (porosity), on the
arrangement of the solid phase, and on the conductivities of each of the two phases.
When there is fluid motion, Darcy’s law provides the bulk flow velocity, which may
appear very laminar and orderly, but at the pore level, the flow behaves much
differently. Every particle that is encountered splits the flow effectively dispersing
it. The flow path is therefore a tortuous path. In laminar flow in fluids, heat transfer
perpendicular to the flow occurs strictly through conduction in the fluid, but flow in
porous media much more effectively transfers heat in the direction perpendicular to
the flow due to tortuosity and dispersion [12]. Dispersion and tortuosity not only
affect heat transfer and effective thermal conductivity but also momentum transfer
or effective viscosity.
For the most general case of a non-isotropic porous medium with varying flow
direction, a full dispersion tensor is used [3], defining unique values for each
direction. For simpler cases, a longitudinal dispersion and a transverse dispersion
are separately defined [2, 13], with longitudinal dispersion in the direction of
streamlines and transverse dispersion perpendicular to streamlines. However, for
easy implementation into standard numerical and analytical schemes, a single,
effective value is preferred. Many authors have approached this problem and
have developed correlations and theoretical predictions for a large range of values.
Various modeling approaches and measurement techniques have been devel-
oped for estimating effective thermal conductivity [5, 14], and further review will
not be considered at this point. Generally, the effective thermal conductivity is a
function of the solid-fluid conductivity ratio; the porosity; the structure of the solid
phase, e.g., foam metal matrix, monodisperse spherical particles, polydisperse
miscellaneous shapes, etc.; and the geometry of the solid-fluid interface. For
κ ~ 1, most of the methods produce similar results, and there is a negligible
difference between predictions within the uncertainty of most measurement
methods.
It is important to properly account for the effects of dispersion due to bulk
motion of the fluid. Combining the existing experimental data, correlations have
been developed [15] that provide the longitudinal and transverse dispersion coef-
ficients, thus increasing the effective diffusivity in both directions. Depending on
the geometry of the porous medium and the local Péclet number, transverse
dispersion can either approximately equal longitudinal dispersion or be up to
100 times greater [16]. At high flow rates, most of the longitudinal heat transfer
is due to advection, and transverse dispersion can be neglected.

1.3 Interfacial Boundary Conditions

Investigation of the boundary condition between a porous layer and a fluid layer
was conducted by Beavers and Joseph [17]. At the time of their research, the most
basic form of Darcy’s law was being used, providing a flat velocity profile in the
porous region, and a no-slip velocity profile was used at the interface between the
1.3 Interfacial Boundary Conditions 7

porous region and the fluid, fixing the edge velocity in the fluid to be equal to that of
the porous region. However, experimental data for Poiseuille flow over a porous
block revealed higher than expected flow rates in the porous region due to a
penetration of fluid flow. The penetrating flow creates a slip velocity at the
interface. The analytical expressions is:

∂u γ
¼ 1=2 ðu  um Þ, ð1:19Þ
∂y K

where um is the mean fluid velocity above the porous layer and γ is an empirical
constant. This expression provides a simple and elegant way of calculating the
velocity profile but relies on the empirical constant to assure evaluation of the slip
velocity. However, the Brinkman term in Darcy’s law can be used instead of the
Beavers-Joseph boundary condition to achieve the same result [18].
Jones [19] built on the work of Beavers and Joseph and considers flow over a thin
spherical shell. He assumes the slip velocity was based on the shear stress, resulting
in an equation that considers both horizontal and vertical velocity components:

∂u ∂v γ
þ ¼ 1=2 ðu  um Þ: ð1:20Þ
∂y ∂x K

However this relation has not yet been verified [20].

Ochoa-Tapia and Whitaker [21–23] use volume-averaging methods to develop
boundary conditions that have a jump in the stress and the heat flux but not the
velocity or temperature, and Sahraoui and Kaviani [24] propose a slip in the
temperature and continuity in the heat flux, velocity, and shear stress. Experiments
are required however to determine an unknown parameter in each of their
formulations.
Alamzi and Vafai [25] show that all of these boundary conditions produce
similar results for Poiseuille flow, but no other conditions were considered. They
find small discrepancies in the fluid flow at the interface, smaller discrepancies in
the temperature profiles, and even smaller discrepancies in the Nusselt number
predictions. However, all of these boundary conditions have been developed for
simple Poiseuille flow and have neglected the effects of buoyancy and flow normal
to the interface.
Recent investigations have used the lattice Boltzmann method (LBM) to inves-
tigate interfacial boundary conditions. Nabovati [26] considers parallel flow and
confirmed the slip parameter of Beavers-Joseph, finding it to depend on the
porosity. Bai and Yu [27] also implement the LBM for flows other than Poiseuille
flow using Ochoa-Tapia and Whitaker’s stress-jump boundary conditions. Both
Nabovati and Bai and Yu neglect buoyancy effects. It is sufficient to note here that
the LMB confirms prior findings, but it has not yet provided new fundamental
insight on the boundary conditions between porous and fluid layers.
8 1 Introduction

1.4 One-Domain Formulation

When considering convection in a porous medium with an overlying fluid layer,

there are two distinct domains or sublayers. In the porous sublayer, some form of
Darcy’s law governs the flow, and in the fluid sublayer, the Navier-Stokes equation
governs the flow. The approach thus far has kept each of the two sublayers separate
and has joined them with a variety of interfacial boundary conditions. This is
termed the two-domain approach. Some authors use a form of Darcy’s law that
reduces to the Navier-Stokes equations as the porosity goes to unity, allowing for
the problem to be treated in a single domain and thus avoid the need to explicitly
define the interfacial boundary conditions. This method allows for a far simpler
numerical solution in many situations and is termed the one-domain approach.
Although historically the two-domain approach has been the standard model, more
recently the one-domain approach has been widely used [1].
Hirata and Goyeau [28] analyze natural convection in a porous layer with a
superposed fluid sublayer. They consider the one- and two-domain approaches and
show that the one-domain approach can be made to match the results of the
two-domain approach provided that the sharp change in porosity at the interface
is properly accounted for. Hirata and Goyeau accomplish this by defining the
porosity, Darcy number, and thermal diffusivity with Heaviside functions and
include the derivatives of these functions in the mathematical formulation. This
effectively adds variable porosity terms to the source term at the interface but
nowhere else in the domain.

1.5 Mixed Convection in Saturated Porous Media

Mixed convection in porous media has been investigated by many authors over the
years. This research has provided a great deal of insight into the present work, even
though the specific problem of mixed convection in a horizontal porous layer with a
superposed fluid layer has not been addressed in the literature. In addition, studies
of natural convection in porous layers, particularly in porous layers with an
overlying fluid layer, provide valuable insight.
Wooding [29] first studied mixed convection in porous media, analyzing the
effect of large Reynolds number on the convection characteristics of the system.
Prats [30] continued the theoretical study, considering the effect of cross flow on the
convection currents. Sutton [31] revisited the classical onset of convection problem,
but included a net cross flow, and was the first to discover that the critical Rayleigh
number increases with the increasing cross flow. As the flow rate approaches zero,
the critical Rayleigh number approaches the theoretical limit, 4π 2. This result was
confirmed by Homsy and Sherwood [32].
Rubin [13] investigated the effect of heat dispersion due to bulk fluid motion,
allowing for different thermal diffusivity in the transverse and longitudinal
1.5 Mixed Convection in Saturated Porous Media 9

directions. His theoretical results further confirm the finding that the critical Ray-
leigh number increases with an increasing Péclet number.
Combarnous and Bia [33] experimentally studied mixed convection in horizon-
tal layers at low Péclet number. Their data do not show any change in the onset of
convection due to cross flow. Reda [34] conducted mixed convection experiments
in a vertical chamber around a cylindrical heat source with the flow opposing the
buoyancy induced upward motion. The parameter Ra/Pe characterizes the heat
transfer of the system, and the transition from mixed to forced convection occurs
at Ra/Pe ~ 0.5, independently of the size of the heater or the amount of total power.
Lai and Kulacki [35] investigated mixed convection in porous media analyti-
cally, numerically, and experimentally in horizontal porous layers with localized
heating from below over a range of Péclet and Rayleigh numbers. Nusselt numbers
show steady behavior in the buoyancy-dominated regime and steady behavior in the
forced flow regime but oscillatory behavior in the transition region between natural
and forced convection (Fig. 1.2). This oscillatory behavior is initially shown to
stabilize, but later findings disagreed, showing ongoing oscillations with period:

τp ¼ 2:41Pe1:096 : ð1:21Þ

Vertical flows in porous media with a finite length wall heat source have also
been considered [36–38]. Numerical solutions were obtained using the basic form
of Darcy’s law with the buoyancy term included. In the opposed flow case, the
results show a Péclet number (denoted as the critical Péclet number) that results in a
minimum heat transfer rate, but no minimum was shown to occur in aided flow. The
same problem was later considered but includes the Forchheimer term and the

15 15
Pe Pe= 8
Nu 10 Nu 10
1
0.1 5
5
5 0 0.5 1.0 1.5 2.0 0 0.5 1.0 1.5 2.0
τ τ
15 15
Pe = 2 Pe = 10
Nu 10 Nu 10

5 5
0 0.5 1.0 1.5 2.0 0 0.5 1.0 1.5 2.0
τ τ
15 15
Pe = 4 Pe = 12
Nu 10 Nu 10

5 5
0 0.5 1.0 1.5 2.0 0 0.5 1.0 1.5 2.0
τ τ
15 15
Pe = 6 Pe = 18
Nu 10 Nu 10

5 5
0 0.5 1.0 1.5 2.0 0 0.5 1.0 1.5 2.0
τ τ

Fig. 1.2 Nusselt number oscillations in mixed convection in a porous layer [35]
10 1 Introduction

I Entrance Section
II Test Section
III Exit Section
IV Cooling Chamber

1 Tube Bundle ( Flow

Straightener )
2 Strip Heater
3 Screen
4 Copper Plate
5 Air Vent

Fig. 1.3 Apparatus for mixed convection in a saturated porous heated from below [39]

effect of thermal dispersion in their governing equations. Similarity solutions show

that including the Forchheimer term decreases the Nusselt number and including
the effects of dispersion greatly increases the Nusselt number. The general form of
the heat transfer correlations remained the same. Horizontal flow over a slender
body of revolution was then considered, and similarity solutions were obtained.
Nusselt numbers show asymptotes for the natural and forced convection limiting
flows.
Mixed convection in horizontal porous layers heated locally from below was
then experimentally studied [39]. A long test chamber (Fig. 1.3) was constructed
with heater lengths tested over a range of Rayleigh and Péclet numbers (1 < δ < 5,
1 < Ra < 1000, and 0.1 < Pe < 120). Spherical glass beads filled the test section,
and fluid was forced horizontally through the chamber. The top surface was
maintained at a constant temperature by circulating cooling water over a copper
plate. Using a more sophisticated (though not justified) effective thermal conduc-
tivity, the authors were able to correct the experimental results, leading agreement
between numerical and experimental data.
The effective thermal conductivity was calculated with the following equations:

ke ¼ ð1  wÞkm þ wkf : ð1:22Þ

Nuconduction
w¼1 : ð1:23Þ
Num
1.5 Mixed Convection in Saturated Porous Media 11

This method of correcting the Nusselt number leads to:

  0:375
Nu Ra
¼ 1:895 þ 0:20 ð1:24Þ
Pe0:5 Pe1:5

for the experimental data, and

  0:372
Nu Ra
¼ 1:917 þ 0:21 ð1:25Þ
Pe0:5 Pe1:5

for the numerical results. In Eqns. (1.22)–(1.25), the length scale is the height of the
porous layer. A critical Péclet number was numerically confirmed, and the mea-
surements showed it becomes more prominent as the length of the heated section
exceeds the total layer height. Similar experiments were then reported [40] with a
sudden expansion at the beginning of the flow channel just prior to the heated wall
section. A critical Péclet number is also observed, but inclusion of non-Darcy terms
had a negligible effect even beyond Pe ~ 100.
A series of papers by various authors investigate mixed convection about a line
source in an infinite saturated porous domain. Pop et al. [41] develop a similarity
solution using the Darcy’s law with the buoyancy terms included. Jang and Shiang
[42] use an implicit finite difference method to numerically solve the DBF equation.
Nakayama [43] divides the flow regime into five distinct regions and presents
similarity solutions for each and shows where the addition of the Forchheimer
and buoyancy terms is necessary. Zhou and Lai [44] numerically solve the vertical
flow problem for Darcy flow over a heated cylinder for Reynolds number from 10 to
100 and Grashof number from 0 to 400. Oscillatory behavior is again observed for
opposed flow, and streamlines and isotherms showing the transient behavior were
presented.
Kumar et al. [45] considered mixed convection with both a fluid layer and a
fluid-saturated porous layer in vertical flow. They investigated isothermal and
isoflux boundary conditions and used the DBF equation with the buoyancy term,
algebraically solving for the steady-state velocity and temperature profiles in a
semi-infinite domain. Their solutions cover a range of viscosity ratio, conductivity
ratio, Darcy number, Grashof number, and width ratios, employing continuity and
no-slip boundary conditions. Although they were the first to consider the
two-domain system in mixed convection, the results are of limited value in the
present context owing to exceedingly high Darcy numbers, 0.125–0.5.
Bhargavi [46] studied forced convection in a porous layer with a superposed
fluid layer. He considers two-dimensional flow and solves the governing equations
for fully developed conditions in an infinite channel. Entrance effects are also
considered. Various boundary conditions are considered for the heated wall and
top surfaces, and Ochoa-Tapia-Whitaker’s stress-jump condition is applied at the
interface between the porous and fluid regions. Results from the Darcy equation
are compared to those from the Darcy-Brinkman equation and to those from
the Forchheimer equation, showing significant differences at some conditions.
12 1 Introduction

In addition, the effects of viscous dissipation were considered, including five forms
of the dissipation function. Within a certain range of Darcy number (0.1–0.001), a
peak Nusselt number occurs at a dimensionless porous sublayer height of 0.6–0.8.

1.6 Natural Convection in Porous Media

Elder [47] first studied natural convection from a localized heat source, numerically
solving the problem with a finite difference method. The domain considered is
similar to that for mixed convection studies, being much longer than tall, with an
aspect ratio of 10. Streamlines and isotherms are presented for a range of heater
sizes and Rayleigh numbers. With a dimensionless heater length of δ ¼ 1, the fluid
plume rises in a single narrow column. With δ ¼ 2, two recirculating regions form,
rotating in opposite directions. With δ ¼ 3, a pair of closed cells are formed and at
δ ¼ 6, three pairs of cells. This same problem of localized heating has been
investigated by Prasad and Kulacki [48, 49]. They solve the problem numerically
over a range of aspect ratios and dimensionless heater lengths. At a fixed heater
size, it is shown that changing the aspect ratio has minimal effect on the overall heat
transfer coefficient. At small dimensionless heater lengths (δ < 2), recirculation due
to natural convection is always unicellular, but for δ > 2 recirculation becomes
multicellular.
Bagchi [1] extensively reviews the literature on natural convection in a porous
medium with an overlying fluid layer as a background for his study with a heated
segment on the bottom (Fig. 1.4). He used porosity as a switching parameter in the
DBF equation to produce the Navier-Stokes equation in the fluid sublayer. However

Fig. 1.4 Problem domain for natural convection study [1]

1.7 Conclusion 13

derivatives of the porosity are not considered to account for the rapid change at the
interface. Various heater sizes, dimensionless porous sublayer height, Rayleigh
number, Prandtl number, Darcy number, conductivity ratio, and overall aspect
ratio are investigated. Numerical results showed excellent agreement with the
classical solutions for the Rayleigh-Bénard and Horton-Rogers-Lapwood problems.
As the porous sublayer height is decreased (η < 1), the strength of the thermal
plume greatly increases in the fluid region, and circulating flow was primarily
contained in the fluid region as well. It is shown that increasing the Darcy number
increases the heat transfer coefficient due to a decreased restriction of bulk fluid
motion, that the Prandtl number had little effect on the overall heat transfer, and that
the conductivity ratio has a large impact due to an increase in convective motion.
Disagreement between numerical and experimental results pointed to either the
need for refinement of the numerical solution or a source of error in the
experiments.

1.7 Conclusion

The development of the governing equations has been discussed, historical work
considering the interfacial boundary conditions has been reviewed, various
approaches to the effective thermal conductivity have been explored, and applica-
ble research on natural and mixed convection in porous media have been reviewed.
The current state of research clearly favors the one-domain approach using porosity
as a switching parameter to avoid explicit definition of the boundary condition at
the interface. This method leaves open the question of porosity variation near the
interface. However, there is no established method to properly transition from the
porous layer to the fluid layer, and appropriate methods need to be determined.
Determining the effective thermal conductivity of a porous medium has been a
challenge, and there is an abundance of research on the subject. The effective
thermal conductivity becomes somewhat of a catchall to include various other
effects as well, such as dispersion and tortuosity. The research on stagnant thermal
conductivity provides many correlation equations that have a great deal of variation
at high conductivity ratio but general convergence at low conductivity ratios. The
effects of dispersion, tortuosity, etc., can easily be included by using readily
available correlation equations that correct for these effects.
Natural convection in porous media, particularly in a porous layer with an
overlying fluid layer, has been thoroughly addressed in the literature. Mixed
convection, on the other hand, has not. There are many investigations of mixed
convection that consider a full porous domain, but the literature is lacking on mixed
convection in a porous layer with a superposed fluid layer.
Chapter 2
Mathematical Formulation
and Numerical Methods

2.1 Solution Domain

The general features of the solution domain and inlet velocity profile are shown in
Fig. 1.1. The top of the flow channel is held at a constant temperature, T0, and has a
no-slip velocity boundary condition. The bottom also has a no-slip boundary and
has a constant temperature, Tw, on a finite length, Lh. The bottom is otherwise
insulated. The inlet temperature is also at T0. Far upstream and downstream, the
inlet and outlet are also taken as adiabatic (∂T/∂x ! 0) and present two options for
the velocity distribution. For the natural convection limit, no-slip velocity boundary
conditions are applied at the left and right boundaries, and for mixed convection, a
fixed velocity profile based on steady-state conditions is applied at the inlet and
outlet.

2.2 Governing Equations

Equations governing the flow and the temperature for each sublayer reduce to a
single set of equations as porosity transitions from ϕ ¼ ϕ0 to ϕ ¼ 1 when the
solution shifts from the porous to the overlying fluid sublayer. For incompressible
flow, the governing equations for the porous sublayer are the continuity equation,
the DBF equations with the addition of the buoyancy term and transient and
convective terms, and the volume-averaged energy equation.

© The Author(s) 2017 15

J.M. Dixon, F.A. Kulacki, Mixed Convection in Fluid Superposed Porous Layers,
SpringerBriefs in Applied Sciences and Technology,
DOI 10.1007/978-3-319-50787-3_2
16 2 Mathematical Formulation and Numerical Methods

The continuity equation for incompressible flow is:

∂u ∂v
þ ¼ 0: ð2:1Þ
∂x ∂y

The DBF equations with the Boussinesq equation of state are:



1 ∂u 2 ∂u ∂u
ρf ϕ þϕ u þv
∂t ∂x ∂y
! ð2:2Þ
2 2
∂P μf ∂ u ∂ u CF
¼  uþμ
e þ  pffiffiffiffiρf jvju
∂x K ∂x2 ∂y2 K


1 ∂v 2 ∂v ∂v
ρf ϕ þϕ u þv
∂t ∂x ∂y
!
2 2
∂P μf ∂ v ∂ v CF ð2:3Þ
¼  vμ
e þ  pffiffiffiffiρf jvjv
∂y K ∂x2 ∂y2 K
 ρf , 0 gð1  βðT  T 0 ÞÞ:

Applying the conservation of energy principle yields the combined energy

equation for the porous domain:



∂T ∂T ∂T ∂ ∂T ∂ ∂T
σ þu þv ¼ αe þ αe , ð2:4Þ
∂t ∂x ∂y ∂x ∂x ∂y ∂y

where σ is the ratio of volumetric heat capacities, (ρcp)m/(ρcp)f. This equation

assumes local thermodynamic equilibrium and allows for property variation.
Equations (2.1)–(2.4) are made dimensionless with length scale H, velocity scale
U0, and temperature Tw – T0. The momentum equations are multiplied by H3 =α2f ρ0 .
The resulting dimensionless equations are:


∂
u ∂
u ∂
u
ϕ1 Pe þ ϕ2 Pe2 u þ v
∂τ ∂
x ∂
y
!
2 2
∂
p PrPe ∂ u ∂ u
¼  u þ μ
bPrPe þ 2 ð2:5Þ
∂
x Da ∂
x2 ∂y
CF Pe2
 pffiffiffiffiffiffi jvj
u
Da
2.2 Governing Equations 17


∂
v
1 2 ∂
v ∂
v
ϕ Pe þ ϕ Pe u þ v
2
∂τ ∂
x ∂
y
!
2 2
∂
p PrPe ∂ v ∂ v
¼  v þ μ
bPrPe þ 2 ð2:6Þ
∂
y Da ∂
x2 ∂y
CF Pe2 gH 3
 pffiffiffiffiffiffi v þ 2 þ PrRaθ:
Da αf

The r.h.s. of Eqns. (2.5) and (2.6) contain four to six terms, the first being the
pressure gradient which is the driving force for forced flow. The second term is the
Darcy term and is generally the dominant term in most natural and mixed convec-
tion applications due to the low values of the Darcy number. The third term is the
Brinkman term which captures the viscous effects due to boundaries within the
porous domain. The fourth term is the Forchheimer term which can become the
dominant term at large flow velocities. The transition from the Darcy regime to the
Forchheimer regime will vary depending on the Darcy number and the flow
velocity. The sixth term in Eqn. (2.6) captures the effects of buoyancy, causing
flow to accelerate upward when the temperature is elevated above that of the free
stream.
The energy equation is multiplied by H2/αf, and the resulting dimensionless is:



∂θ ∂θ ∂θ ∂ ∂θ ∂ ∂θ
σ þ Pe
u þ Pe
v ¼ λ þ λ : ð2:7Þ
∂τ ∂
x ∂
y ∂ x ∂
x ∂
y ∂
y

This equation neglects viscous dissipation and does not include volumetric
energy generation.
The momentum equations are next reduced to the stream function—vorticity
form to eliminate the pressure gradient. The partial derivatives of Eqn. (2.5) with
respect to y and Eqn. (2.6) with respect to x are taken to derive the variable porosity
form of the equations. The Darcy number is a function of the porosity and therefore
varies in the y-direction.
Although the Forchheimer constant is written as a function of porosity by some
authors, the more widely accepted form is not a function of porosity but rather the
bead diameter and the hydraulic diameter, DH, of the channel [2]:


d
CF ¼ 0:55 1  5:5 : ð2:8Þ
DH
18 2 Mathematical Formulation and Numerical Methods

With the over bars dropped, the dimensionless momentum equations are written as:

∂uy  
ϕ1 Pe þ ϕ2 Pe2 uuxy þ uy ux þ vuyy þ vy uy
∂τ
Dauy  uDay  
¼ Pxy  PrPe 2
þμbPrPe uxxy þ uyyy ð2:9Þ
Da
pffiffiffiffiffiffi 1 1
Da jvjuy þ jvjy u  jvju Da2 Day
 CF Pe2 2
Da
∂vx  
ϕ1 Pe þ ϕ2 Pe2 uvxx þ ux vx þ vvxy þ vx vy
∂τ
PrPe  
¼ Pxy  vx þ μ
bPrPe vxxx þ vxyy ð2:10Þ
Da
CF Pe2  
 pffiffiffiffiffiffi jvjvx þ jvjx v þ PrRaθx :
Da

Subtracting Eqn. (2.9) from Eqn. (2.10), the governing equation for the porous
sublayer is:

∂ω h i
ϕ1 Pe þ ϕ2 Pe2 ðuωÞx þ ðvωÞy
∂τ
 
¼μbPrPe ωxx þ ωyy


PrPe CF Pe2
þ   pffiffiffiffiffiffi jvj ω
Da Da ð2:11Þ


jvjy u  jvjx v
þ PrRaθx þ CF Pe2 pffiffiffiffiffiffi
Da


pffiffiffiffiffiffi 
Day CF Pe2 Dajvj
þ u 2 PrPe þ :
Da 2

The dimensionless vorticity form of the Navier-Stokes and energy equations for
the fluid sublayer is, respectively:

∂ω h i  
Pe þ Pe2 ðuωÞx þ ðvωÞy ¼ PrPe ωxx þ ωyy þ PrRaθx ð2:12Þ
∂τ



∂θ ∂θ ∂θ ∂ ∂θ ∂ ∂θ
þ Peu þ Pev ¼ þ : ð2:13Þ
∂τ ∂x ∂y ∂x ∂x ∂y ∂y
2.4 Numerical Methods 19

2.3 One-Domain Formulation

To allow for implementation of well-established numerical schemes, a single set of

governing equations is preferred. If the porosity is allowed to vary, ranging from
ϕ ¼ ϕ0 to ϕ ¼ 1 where ϕ ¼ 1 in fluid region, it follows that Da ! 1, λ ! 1, μ b ! 1,
and Day ! 0. Using these values, the governing equations in the porous sublayer
reduce to those in the fluid sublayer. This is a simple concept, but the sophistication
of the method lies in the variation of the porosity and the additional terms in the
governing equations that result.
Several authors assume a porosity variation of the following form [2, 50]:
h i
ϕ ¼ ϕ0 1 þ ð1:4ÞeðNdÞ ,
y
ð2:14Þ

where 5 < N < 6 for ϕ0 ~ 0.4. This equation can be used to allow for either the
channeling effect near a solid boundary or as is done in the present case, porosity
variation at the interface between the porous and fluid sublayers. For the results
reported in this monograph, N ¼ 5.
To summarize, the porous region is governed by the DBF equations with the
addition of the buoyancy, transient, and convection terms, while the fluid region is
governed by the Navier-Stokes equations. However, the single set of equations
developed is suitable for both regions, provided the porosity is varied, resulting in a
single problem domain. The resulting equation contains additional terms, wherein
the effects of porosity variation are realized in the Darcy number for flow and the
conductivity variation in the energy equation.

2.4 Numerical Methods

Finite volume methods are chosen to implement the numerical solution of the
governing equations. Finite volume methods provide better conservation of energy
and mass than traditional finite difference methods across the finite difference grid.
In many of the derivative terms, central differencing is employed, and for the
convective terms, upwind differencing is employed. Although quadratic upwind
differencing has been used [1], standard upwind differencing is less problematic
owing to inherent instabilities in regions of sharp gradients. A variable grid is
chosen over a constant grid to allow coarser overall grid except in desired regions to
achieve a grid-independent solution with fewer total grid points [51, 52]. Second
order difference equations are used for most equations with the exception of
boundary conditions where the grid points were clustered, thus, achieving second
order accuracy with first order equations. Complete details of the numerical solu-
tions are given by Dixon [53].
Chapter 3
Numerical Results

3.1 Verification of Solution Method

3.1.1 Rayleigh-Bénard Convection

Rayleigh-Bénard convection considers an infinite two-dimensional fluid layer with a

fully heated bottom. No-slip boundary conditions are applied at the top and bottom,
constant temperatures are applied at the top and bottom, and adiabatic conditions are
applied at the side walls. Hollands [54] gathered experimental data from a wide range
of studies and developed a correlation equation for the Nusselt number. Numerical
solution results using the present numerical solution method are compared to this
correlation equation for no-slip and no-shear conditions (Fig. 3.1). The no-shear
boundary condition accurately captures the critical Rayleigh number at the onset of
convection but overpredicts the Nusselt number at higher Rayleigh numbers. The
no-shear boundary condition more accurately tracks the experimental results at
higher Rayleigh number but overpredicts the critical Rayleigh number (Fig. 3.2).

3.1.2 The Horton-Rogers Lapwood Problem

The Horton-Rogers-Lapwood problem investigates natural convection in a

two-dimensional domain containing a saturated porous layer and a fully heated
bottom. No-slip boundary conditions are applied at all surfaces, and constant
temperatures are applied at the top and bottom. The side walls are adiabatic. Bagchi
[1] compared his numerical results to several existing numerical and experimental
results with good agreement (Figs. 3.3 and 3.4). The numerical results of the present
investigation are compared to Bagchi’s results in Fig. 3.5, showing excellent
agreement.

© The Author(s) 2017 21

J.M. Dixon, F.A. Kulacki, Mixed Convection in Fluid Superposed Porous Layers,
SpringerBriefs in Applied Sciences and Technology,
DOI 10.1007/978-3-319-50787-3_3
Another random document with
no related content on Scribd:
 CHAPTER XIV
FURTHER RESEARCHES

Last week, we had a few days of such cool, bright, windless

weather, that it seemed as if a St. Luke’s summer had set in. Now,
however, the icy gales from the east are once more blowing round the
boma of Newala, and we had rain on Michaelmas Day, which was
somewhat early. This must have been a signal universally understood
by young and old; for I am no longer besieged by the hitherto
inevitable boys, and my old men, too, have ceased their visits.
Fortunately, I have been able to pump the old gentlemen so
effectually in the course of the last few weeks that I could leave at
once, quite happy in the possession of an enormous stock of notes,
were I not detained by the linguistic inquiries which I am now set on
making. It is quite impossible to give here even the merest indication
of the knowledge so far gained as to all these more or less strange
customs and usages. The details will have a place in official and other
documents to the preparation of which the leisure of many coming
terms will have to be sacrificed; here I can only indicate such
prominent points as are calculated to interest every civilized person.
Personal names among the natives offer an unlimited field for
research. Where Islam has already gained a footing, Arab names are
prevalent. The Makonde askari Saidi bin Musa keeps step with his
comrade Ali bin Pinga from Nyasa, and Hasani from Mkhutu
marches behind the Yao porter Hamisi. Among the interior tribes the
division into clans predominates as a principle of social
classification, and therefore, even in the case of converts to
Christianity, the baptismal name is followed by the clan name. Daudi
(David) Machina is the name of the native pastor at Chingulungulu,
and the presumptive successor to Matola I and Matola II calls
himself Claudio Matola. We shall have something more to say about
these clan names later on.
The meaning of the names is often equally interesting. My carriers
alone have already provided me with a good deal of amusement in
this respect, the appellations they go by being in most cases
exceedingly absurd. Pesa mbili (“Mr. Twopence”) is as familiar to us
as his friend Kofia tule, the tall man with the little flat cap, Kazi
Ulaya, the man who works for the European and Mambo sasa
—“Affairs of to-day.” Besides these, the following gentlemen are
running about among the two dozen who compose my faithful
retinue:—Mr. Blanket (Kinyamwezi bulangeti, corrupted from the
English word), Mr. Cigarette (no commentary needed), Kamba Ulaya
(European rope, i.e. hemp rope as distinguished from native cordage
of cocoa-nut fibre or palm-leaf twist), Mr. Mountain (Kilima) and
Messrs. Kompania and Kapella (Company and Band—from the
German Kapelle). The names Mashua (boat) and Meli (steamer,
from the English “mail”) have a nautical suggestion and Sita (Six) an
arithmetical one—and, to wind up with, we have Mpenda kula—(“He
who loves eating”).
The names used by the interior tribes are free from the noticeable
European touch found in these designations of the carriers, but here,
too, we come across amusing specimens. I notice at the same time
that these names are certainly not the first to grace their bearers. As
is so often the case with primitive peoples, and with the Japanese at
the present day, we find that every individual on being formally
admitted to the duties and responsibilities of adult life assumes a
new name. The natives hereabouts do not know or have forgotten the
original significance of this change, but we are not likely to be wrong
in supposing that the new name also means a new person, who
stands in quite a different relation to his kinsmen and his tribe from
his former one. Officially, every adult Yao, Makua, Makonde or
Matambwe has the right to offer himself as godfather, but I have the
impression that the majority of names one hears are really
nicknames, casually given by acquaintances.[53] It is well known that
the native has a very acute sense of the weak points and absurdities
of others.
Che Likoswe (“Mr. Rat”) will be remembered by his war-songs at
Chingulungulu, and with him may be classed Che Chipembere (“Mr.
Rhinoceros”). The latter is liable to fits of sudden rage, like the
pachyderm, hence his name. The name of the old beer-drinker,
Akundonde, is a reminiscence of his original kinship with the
Wandonde tribe. Che Kamenya is he who is victorious in fight; there
was joy at the birth of Machina; Makwenya gathers everything to
himself, but Che Mduulaga, on the other hand, thinks nothing of
himself,—he is modesty in person. In the same way, Mkotima is a
quiet man, Siliwindi is named after a song-bird so-called; and,
finally, Mkokora is he who carries away dirt in his hands.
These are some Yao men’s names. I will only mention the
following women’s names for this tribe:—Che Malaga means “She is
left alone”—all her relations have died. Che Chelayero, “She who has
a hard time.” Che Tulaye, “She who fares poorly,” and Che Waope,
“She is yours.”
The personal names of the other tribes have on the whole the same
character—Kunanyupu is an old Makua, who, according to his own
statement, has killed many gnus (nyupu) in his youth. Nantiaka is
the Don Juan who flits from one attraction to another. A similar train
of thought has suggested the name of Ntindinganya, the joker, who
contrives to saddle others with the blame of his own tricks.
Linyongonyo is the weakling; Nyopa the ambitious man who strives
to make himself feared by others; Madriga is the sad, melancholy
man; Dambwala the lazy one.
Among the women Alwenenge is “the one who knows her own
worth,”—her lord and master has, it is true, taken another wife, but
he will not remain with her, but return penitently to Alwenenge, as
she very well knows. Much less fortunate is Nantupuli; she wanders
about the world and finds nothing at all, neither a husband nor
anything else. Other unfortunates are Atupimiri and Achinaga—the
former has a husband who is always on his travels and only comes
home from time to time to “measure” (pima) his wife, i.e., to see how
she is behaving. Achinaga’s husband, on the other hand, is ill and
cannot work, so that she has to do everything by herself. There is also
a Pesa mbili among the Makonde women. The name implies that she
formerly stood high in the estimation of men, but now she has grown
old and is only worth two pice. Beauty has its market value even with
the negro.
A field of inquiry, extremely difficult to work, but which will
everywhere well repay cultivation, is that of the customs
accompanying the life of the individual from his cradle to his grave.
The native infant—which is not black, but at first as pink as our
own new-born babies—has come into the world in its mother’s hut.
The father is far from the spot, the women having sent him out of the
way in good time. The baby is carefully washed, and wrapped in a
piece of new bark-cloth. At the same time its ears are anointed with
oil, that it may hear well, and the ligature under the tongue loosened
with a razor, to ensure its learning to speak. Boys are everywhere
welcomed; but with regard to girls, the feeling varies in different
tribes, and, just as is the case among ourselves, in different families.
It is often stated in ethnographical works that primitive peoples
rejoice on purely interested grounds at the birth of girls, on account
of the price they will bring when married. Up to a certain point such
considerations may have weight here, too, but in general people are
glad of daughters if only because they can soon begin to help their
mother in her numerous outdoor and indoor tasks. Their marriage,
moreover, brings an additional faithful and unpaid worker into the
household. For this is the land of exogamy, where the young wife
does not go to her husband’s home, or enter his family, but, on the
contrary, the man leaves his father and mother and either moves
directly into the house of his wife’s parents, or builds his own close
beside it. In any case for some years, until his own family
circumstances necessitate a different arrangement, he devotes all his
powers to keeping up his mother-in-law’s establishment. He sees to
the planting of the crops and their ingathering, he breaks up new
ground, in short he renders every possible service, and anticipates
her every wish. I have often been ashamed when the conversation
turned on this and other features of native life, to remember the
tenor of those venerable jests of which our comic papers never
weary. Of course, a mere passing traveller like myself is no judge of
the more intimate side of family life, but Knudsen, who has lived in
the country long enough to become thoroughly familiar with the
people’s ways of thinking and acting, confirms the impression I had
arrived at, that, not only is the relation between mother and son-in-
law nothing short of ideal, but that the behaviour of young people to
their elders in general deserves to be called exemplary. We who
belong to the highest stage of culture, or, according to the view held
by most of us, the stage of culture, spend half our lives in educational
establishments of various kinds and grades, and the final result is
shown by statistics in the diminishing percentage of illiterates in our
population. But let all who have eyes to see and ears to hear observe
how little ethical sense and how much downright brutality make up
the daily life of these very representatives of culture. I am far from
wishing to say anything against our system of education and our
schools—I am a kind of schoolmaster myself—but it gives food for
serious reflection to see how worm-eaten, in spite of all the care
bestowed on it, is much of the fruit they produce, and how ethically
sound is the life we meet with among these barbarians. And this is
the outcome of a training extending over three or four months and
received from teachers who have passed through no school or
college.
The treatment of twins is different among
the various tribes in this part of the country.
The Wayao welcome them with unmixed joy,
while the Makonde look on their birth as a
terrible event, to be averted if possible by all
sorts of charms. But even here the parents are
not so cruel as to kill them if they do come
into the world; they are allowed to live and
treated in the same way as by the Wayao, i.e.,
their clothing (such as it is) is always alike. If
this were not done, it is believed that one of
them would certainly die.
For the first year the African infant remains
in close contact with its mother. When it is WOMAN CARRYING A
only a few days old, she takes it out for the BABY ON HER BACK.
FROM A DRAWING BY
first time, to be shown to the admiring PESA MBILI
neighbours. Like a little lump of misery it
squats in the large coloured cloth enfolding
the upper part of its mother’s person. It usually hangs on the
mother’s back, but she very often swings it round to one hip. When
the time comes for feeding the baby, it and the bag containing it are
brought round to the front. Nothing so impresses me with the idea of
poverty and squalor as this treatment of infants: no change of
clothing for mother or child—for there is no supply of extra garments
—no drying, no powdering, no napkins, no regular bath after the first
few days, no care of the mouth. On the contrary, every child has sore
places where the skin has been chafed, especially at the joints, and in
folds and depressions of the body; half-healed scabs, where nature is
getting the upper hand in spite of neglect; eyes nearly always bleared
and running in consequence of the perpetual attacks of flies, and,
finally, individual cases, here and there, of thrush-ulcers on such a
scale that fungoid growths actually protrude from mouth and
nostrils. It would be well if the Government and the Missions could
unite to put an end to this frightful state of things, not so much by
medical work, which is naturally limited to certain localities, as by
training the mothers, as extensively as possible, in the simplest rules
of hygiene and cleanliness.
I have been half-an-hour in a native village.
The men and boys were all assembled within
two minutes of my arrival; the women are
gathering more slowly; the little girls,
curiously enough, are altogether absent. Just
as with us, the women have at once gathered
into a closely-packed group. A shy silence
reigns at first, but no sooner have they had
time to get used to the sight of the white man,
THREE MAKUA than there is an outburst of talk in every key,
VEGETARIANS in spite of the hugest of peleles. At least half
these women are carrying babies, but this
term is tolerably comprehensive. Great boys
and girls of two, or even three years old, are sprawling on the slight
backs of delicate-looking mothers, or making violent attacks on the
maternal fount of nourishment. My camera apparently affords the
pretext for this last manœuvre; for, as if at a given signal, the whole
little black band is propelled forward into position at the very
moment when I press the bulb.
The later stages of childhood among the natives are passed in a
way not materially differing from our own youthful recollections. The
little boys band themselves together in troops and carry on their
games in the village and the bush; while the little girls begin at an
early age to help their mothers indoors and out. Wherever I have
been able to carry on my activity as a collector, I have been
particularly assiduous in getting together all toys and games in use in
the country. There is one point deserving of special notice in
connection with children’s games, and this is that almost from the
first day of its existence the child is present wherever anything is
going on. When the mother joins in the dance, the baby on her back
goes through every movement with her, and thus learns dancing, so
to speak, instinctively. By the time it can stand on its own little feet,
it joins in with the same certainty as that with which the partridge
chick just out of the egg goes to pick up its food. Whether native
children have outside these dances anything that can be called
concerted games, I cannot say, but so far I have seen nothing of the
sort,[54] unless we might count the great skill shown in clapping the
hands in unison, in which, with its pleasing rhythm and (one might
almost say) variety of tune, they are as much at home as their elders.
Otherwise every child seems to be dependent on itself, at least as far
as toys are concerned. For boys, bow and arrows are the sine qua non
in the first place. If I had been willing to buy all the toy bows offered
me, I should have had enough to load a small ship. Here in Africa the
weapon is as much of a survival as in most other countries. The fact
of its being confined to children shows that, as in Europe, it is no
longer seriously used in war, but only in play, or at most, in the
chase. We find, as might be expected, that the grown men are no
better archers than the boys, and vice versa. Where firearms have
once been introduced, more primitive weapons are no longer valued.

USE OF THE THROWING STICK

 It is not easy to form an ethnographical collection in this country.
It is only in consequence of my very resolute attitude—which is far
more effectual than my bags of copper coin—that the people make up
their minds to bring anything at all, and then it is chiefly rubbish. In
order to obtain the more valuable class of articles, such as the more
important household implements, or the carved masks and other
works of art, I am frequently compelled to resort to a mild display of
force, by making the headman of the village morally responsible for
the production of the specimens. And yet every article is liberally
paid for. How peculiarly difficult it is to obtain toys, of all things,
people at home have no notion. I would suggest the following
explanation for this fact. If a Japanese ethnographer, for instance,
were to visit Germany in the autumn he would find it easy enough to
make a large collection of kites, but tops—to take one of our most
typical children’s toys—he would only be enabled to see and procure
if he definitely inquired for them. It is just the same here; everything
has its season, and toys above all. Having once grasped this truth, I
always made short work of the business, by delivering to the
assembled villagers a lecture on all the playthings of mankind,
winding up with, “If you have so and so—or so and so—be quick and
bring it here.” In many cases neither my own linguistic acquirements
nor the interpreter are sufficient, and gestures have to supply the
lack of words. I was quite startled at my success, one day at
Chingulungulu, when, on having gone through the vigorous
movement of slinging a stone, I saw Salim Matola, the all-
accomplished, return in a short time with two remarkable objects,
which, on his demonstration of the way in which they were used,
proved to be a veritable throwing-stick and a sling—an amentum. I
have rarely had such a feeling of complete success as at this moment.
Who would have thought to find the throwing-stick and the sling in
Eastern Africa, a region hitherto considered so barren as regards
ethnography? The former is an implement intended to serve no other
purpose than the lengthening of the lower arm in order to throw a
spear or a stone; it represents, therefore, in terms of physics, the
lengthened arm of a lever. Its principal region of distribution is
Australia; it also exists in some parts of the Western Pacific, among
the Hyperboreans, and in isolated parts of America. Hitherto it has
been assumed that the African had not arrived at this invention. The
sling in the same way serves to lengthen a lever, but the spear or
stone is in this case not thrown by means of a catch on the stick or
board, but by a string fastened to the root of the forefinger, while the
free end is coiled round the missile. If the warrior throws his arm
forward, the weapon leaves the hand by centrifugal force, uncoils
itself from the string and flies away with great initial velocity.

THROWING WITH THE SLING

Where such antiquities as these occur—I

reflected at the time—surely there are more
discoveries to be made. This expectation was
in fact fulfilled, though I had first to fight my
way through a superfluity of another species
of toy. One day, in the course of the lecture
already referred to, I happened to make the
gesture of whipping something over the
ground, and it was at once correctly
understood, for from that time forward the
SPINNING A TOP young people simply overwhelmed me with
tops. No less than four kinds are in use here.
One exactly corresponds to our European peg-top,[55] and is, like it,
driven with a whip, a second has a round or square piece of gourd
fixed on a short, stout wooden peg as axis of rotation;[56] a third is
similar to the last, but has a second disc under the first (which is
about the size of a five-shilling piece), in order to place the centre of
gravity higher up. Finally, we have a very complicated mechanism
whose action resembles that of our humming-top. The second and
third require no whip, but are spun with the thumb and middle
finger. The fourth, on the other hand, needs a “frame” to spin it. This
is represented by a piece of maize-cob perforated lengthwise,
through which the string wound round the top is quickly pulled back.
Like many other things, the art of spinning tops is not made easy for
native boys, the soft, sandy ground being ill-suited to this game; yet
the little fellows show great skill in it.

IKOMA DANCE AT THE GIRLS’ UNYAGO, ACHIKOMU

 XYLOPHONE (MGOROMONDO)

With one exception, children have no musical instruments peculiar

to themselves. Whether they fiddle on the sese, the one-stringed
violin, or maltreat the ulimba, that instrument on which all Africans
strum—the box with wooden or iron keys fixed to its surface, and
struck with the finger-tips—or strike the mgoromondo, that
antediluvian xylophone in which the keys rest on a layer of straw, or
play on the lugombo, the musical bow with calabash resonator,
which is so widely distributed over East and South Africa—in every
case the instruments are only clumsy imitations of those used by
grown-up people. The only one whose use is confined to the young is
the natura—a friction-drum, made from a bottle-gourd or the fruit of
the baobab, cut across and covered in, like a drum, with the skin of
some small animal. A blade of grass passes through the middle of the
diaphragm, and thence down through the bottom of the shell. By
rubbing a wetted thumb and forefinger down the stalk, as the little
wretches are perpetually doing, a noise is produced so excruciating
that even my carriers—who are not precisely sufferers from nerves—
take to flight when they hear it. But young people are not only
capable of preserving ancient survivals in culture through thousands
of years, but also have the advantage of a greater receptivity for
novelties. I have in my collection two charming specimens of an
African telephone, consisting of two miniature drums, beautifully
carved and covered with the delicate skin of some small animal,
perforated in the middle to allow the passage of a thin string, which
is kept from slipping through by a knot on the inside of the skin. I
never thought, at first, of taking this thing seriously, but one day,
having a spare quarter of an hour, I put one of the drums into
Knudsen’s hand, and told him to walk away till the string—about a
hundred yards in length—was stretched tight. I held the membrane
to my ear, and heard quite clearly, “Good-day, Professor. Can you
hear me?” So the thing really acts, and all that remains for us to do is
to develop it and boldly link ourselves up with the coast and that
centre of civilization, Lindi![57] There can be no question of
independent invention in this case; the telephone is undoubtedly
borrowed—but the fact of the borrowing, and the way it is applied by
children are not without interest.
Such an important epoch in
native life as that represented
by the unyago, with all its joys
and woes, its games and
dances, cannot be without
influence on the habits of the
young people, even before it
arrives. Thus I have some ipivi
flutes obtained from little
fellows far too young to be
admitted to the mysteries.
Anyone who wishes to excel in
an art must begin his training
early, and the flute players of
the ndagala practise their
instrument for years
beforehand. Moreover, boys,
who had evidently not yet
passed through the unyago,
have more than once brought
me specimens of the kakale, PLAYING THE
the long sticks, painted black NATURA
and white in alternate rings
with a little trophy at the top,
NATURA
(FRICTION-
consisting of the shell of some fruit with a plume of
DRUM) feathers stuck in it. These two insignia of maturity,
therefore, are also found in the capacity of toys.
There is nothing surprising in this so far as the boys
are concerned, for the native has no secrets from them. At the
ceremonies I witnessed at Achikomu, as well as at Niuchi and
Mangupa, there was always a whole troop of little fellows, covered
with dirt and ashes, running about. Strangely enough, there were
never any half-grown girls to be seen on these occasions; everything
relating to the mysteries seems to be carefully kept secret from them.
It was only during my long residence at Newala, with its possibilities
of free intercourse between me and the different tribes, as well as
among natives of different ages, that I could see and photograph any
of these young things. They seem to be brought up much more within
the walls of the hut and its compound than we are accustomed to
suppose; and even in the hundreds of visits I have paid to native
homes, I have seldom been able to see the young daughters of the
house face to face. As a rule, I only caught sight of a slender little
figure retreating swiftly through the back door of the hut.
Under these circumstances, of
course, I cannot say how the little
native girl actually grows up, and
whether she enjoys anything even
faintly resembling the happy
childhood of our own loved ones
—but nothing leads us to suppose
that she does; though there is no
question that the native shares in
the universal instinct which
inspires all parents with affection
for their offspring; he feeds his
UNASIKIA?: “DO YOU HEAR?” children and protects them when
they need protection; he rejoices
when they thrive and mourns
over their illness and death. I can still see Matola, as he came to me
one day—his usual expression of gentle melancholy heightened to
one of deep grief and anxiety—carrying a little girl of some five or six
years. She was not even his own child, but a relative, for whom he
entreated my help. To my sincere regret, it was impossible for me to
do anything—the poor little thing was suffering from a malignant
gangrene, which had eaten away the whole front of one thigh, so that
the tendons were laid bare and the bones were beginning to bend. I
spoke very seriously to Matola, asking whether he were as much of a
mshenzi as his people, who were
perishing through their own
stupidity and apathy. He, the
headman, and a clever man at
that, knew very well, so I told
him, that there were German
doctors at Lindi, who could cure
even such cases as this, if the
patients were brought to them.
He ought therefore, to send the
child down at once, unless he
wished her to die, as all her elder NDIO: “YES”
brothers and sisters had done.
Matola gazed at me for some time, evidently
wavering between hope and doubt; but in the
end he followed my advice; and I have since
heard that the child is well on the way to
recovery. But it is astonishing and perplexing
that such an enlightened man as the chief of
Chingulungulu should have allowed the
disease to go on so long before taking any
serious steps to obtain assistance. What then
could be expected of a man from the bush,
who consulted me immediately after my
arrival, asking me for medicine for his sick
child?
“What is the matter with your child?”
“A wound on her foot.”
“But, my good man,” I said, “I can’t give you
medicine to take home,—you would not know
how to put it on. You must bring your child
here. Where do you live?”
“Mbali—a long way off—Bwana,” he
answers, lengthening the vowel to signify
inexpressible distance.
NATIVE TELEPHONE
“How far?”
“Well—about two hours.”
 “Oh! you call that far, do you? you mshenzi! if you were going to a
beer-drink, twenty hours would be karibu sana. Off with you now,
and come back at eight to-morrow morning.” But neither at eight nor
at any later hour was there any sign of the noble father from the
Makonde bush. It was not till the fifteenth day after the preliminary
consultation that he appeared, bringing with him a little girl of five or
six. I did not at first remember him, but at once recalled his previous
visit when the child, overcoming her natural shyness, held out her
foot. Nothing was to be seen but a horrible mass of dirt and sand
cemented together with blood. I started at once on the cleansing
process, with the help of Stamburi, my trusty hospital orderly; and
when at last the foot was laid bare, we found that the whole ball was
eaten away to the bone—whether owing to jiggers, or through the
cumulative effect of various other circumstances, my medical
knowledge is insufficient to decide. When at last I glanced at the
father, I saw him staring like one hypnotized at a leg of antelope
intended for the next day’s dinner, which Knudsen had hung up just
over my table. Having recalled him to reality, I bade Moritz give him
the softest part from the skin of a recently killed wild pig, and told
him to make a shoe, or at least a sandal such as are certainly not
unknown in this country, as he must see for himself that the child
could not walk through the dirty sand with her freshly-bandaged
foot. He had his knife with him—let him get to work without delay!
We two practitioners devoted ourselves once more to the treatment
of the wound, which was in truth a terrible one; and in a little while
the bandage was put on as correctly as we knew how. A second look
at the father showed that he was still staring at the raw joint, as
intently as if he had really eaten his way into it. It is a good thing,
after all, in such cases, to have the kiboko within reach. In another
quarter of an hour the well-wrapped foot was protected by a very
serviceable pigskin slipper. But that is the last I ever saw or heard of
the gentleman, and he never so much as thanked me either for the
treatment or for the thrashing.
 MAKONDE CHILDREN

Boys and girls, as a rule, reach the age of eight or nine, perhaps
ten, before any event of importance interrupts the even tenor of their
lives. Then the assembly of the men, which when the harvest is over,
meets daily in the baraza, decides where the unyago is to be
celebrated in the current year. Since all the adjacent districts have
now taken their turn in bearing the expense of the ceremony, it is a
point of honour that our village should invite them this time. The
resolution is soon carried into effect; the moon is already on the
wane, and the celebration must take place before the new moon. The
unyago presents exactly the same features in all the tribes of this
region. The men erect a circle—larger or smaller as circumstances
may require—of simple grass huts in an open space near the village.
In this space the opening and closing ceremonies are performed; the
huts are intended for the candidates to live and sleep in. Such an
arena, with all its appurtenances in excellent preservation, was the
circle of something over fifty yards’ diameter which I was enabled to
photograph when visiting the echiputu at Akuchikomu. The charred
remains of a similar lisakasa, as the system of huts is called in Yao,
were to be seen near the road on this side of Akundonde’s—the relics
of a former festival.

MASEWE DANCE OF THE MAKUAS IN THE BOMA AT

NEWALA

It is inherent in the nature of the whole institution that both boys

and girls should be passive throughout. They sit silent, inactive and
motionless in their huts while, on the first night of the festival, the
grown-ups feast and drink and enjoy themselves in the wild masewe
dance. Next day the boys, each one in charge of his instructor, are
conducted to the bush by the chief director. There they sleep one
night without any shelter whatever. For a short time, on the
following day, they may do as they please, but during the remainder
they have to set to work with their anamungwi (teachers) and build
the ndagala. As soon as this airy construction is finished, one after
another of the boys is laid on a very primitive couch of millet-straw,
and the jua michila performs the operation. For weeks the little
patients lie there in a row, unable to do anything to accelerate the
slow process of healing. Not till this is complete and the subsequent
moral and other instruction has begun do the wari, as the boys are
now called, acquire the right to take part in public life. In the high
spirits engendered by the pride of their new position, they indulge in
many a mad freak. Woe to the unhappy woman or girl who, ignorant
of the situation of the ndagala, strays into this region of the bush.
Like a troop of mischievous imps, the boys rush on her, tease her,
perhaps even tie her up and ill-use her. According to tribal custom,
they are quite within their rights in so doing, for their abode in the
bush is supposed to be utterly unknown to women. When he goes out
into the pori the boy is dead to his mother,—when he returns, he will
be a different person with a new name, and nothing to connect him
with his former relationship.
I have already tried to describe the course taken by the instruction
imparted in the ndagala. Old Akundonde and his councillor, in the
candour induced by their libations, were certainly trustworthy
informants in this respect. It is an irreparable misfortune that the
liquor supply coming to an end when it did (in such a surprisingly
short time) deprived me of the conclusion of the address to the wari,
but the fragment already given is quite sufficient to indicate the
character of the teaching.
The lupanda reaches its culminating point only with the closing
ceremony. The preparations on both sides are extensive: in the bush
the wari are being restored by their mentors by means of head-
shaving, baths, anointing with oil, and a supply of new cloth, to a
condition worthy of human beings. In the village, meanwhile, the
mothers, long before the time fixed, have been brewing large
quantities of beer and preparing still larger piles of food for the
festivity. When the great day at last arrives, the boys come back in
procession, in their clean new garments, with their faces, necks, and
freshly-shorn scalps all shining with oil, and carrying in their right
hands the kakale, the sticks headed with rattles which have already
been described. Men and women line the road in joyful expectation.
Ever louder and more piercingly, the “lu-lu-lu” of the women
vibrates across the arena, and yonder the drums strike up with their
inspiriting rhythm, while the hoarse throats of the men utter the first
notes of a ng’oma song. In short, everything is going on in the most
satisfactory and genuinely African way.
 KAKALE PROCESSION ON THE LAST DAY OF THE UNYAGO

The Africans, being human, like ourselves, it is only to be expected

that all their works and ways are subject to as many changes and
inconsistencies as our own. I have devoted a disproportionate part of
the time (over a month) spent at Newala to the task of fixing the
typical course of all these ceremonies. This has been a most severe
labour, for if, in my wish to obtain unimpeachably accurate results, I
arranged to let my informants of each tribe come separately, I might
be sure that the two or three old men who made their appearance
would say little or nothing. The native intellect seems not to become
active till awakened and stimulated by sharp retort and rejoinder in a
numerous circle of men. I have thus been compelled to go back again
and again to my original method of assembling the whole senate of
“those who know,” some fifteen aged Yaos, Makua and Makonde in a
heterogeneous crowd round my feet. This was so far successful as to
produce a lively discussion every time, but it becomes very difficult
to distinguish between the elements belonging to different tribes. Yet
I venture to think that, with a great deal of luck, and some little skill,
I have succeeded so far as to get a general outline of these matters. I
feel quite easy in my mind at leaving to my successors the task of
filling up the gaps and correcting the inaccuracies which no doubt
exist.
Further, it must be remembered that my notes on the initiation
ceremonies of these three tribes would, if given in full, take up far too
much space to allow of their reproduction here. Two other points
must be borne in mind. What I saw with my own eyes of the unyago,
I have here related in full, with that local colouring of which actual
experience alone enables a writer to render the effect. But those
scenes at Achikomu, Niuchi and Mangupa are only tiny fractions of
the very extensive fasti represented by the girls’ unyago in reality;
while, as to the remainder, I can only repeat what I have heard from
my informants. Quotations, however, always produce an impression
of dryness and tedium, which is what I would seek to avoid at any
price. I therefore think it better to refer those interested in the details
of such things to the larger work in which it will be my duty,
according to agreement, to report to the Colonial Office on my doings
in Africa and their scientific results.
The last point belongs to another department. The negro is not in
the least sophisticated as regards the relation between the sexes.
Everything pertaining to it seems to him something quite natural,
about which his people are accustomed to speak quite freely among
themselves,—only in extreme cases showing a certain reticence
before members of the alien white race. Now the part played by sex
in the life of the African is very great, incomparably greater than with
us. It would be too much to say that all his thoughts and desires
revolve round this point, but a very large proportion thereof is
undoubtedly concerned with it. This is shown in the clearest way, not
only in the unyago itself, but in the representation which I
subsequently witnessed. In the present state of opinion resulting
from the popular system of education, such delicate matters can only
be treated in the most strictly scientific publications, being debarred
from reproduction in a book of any other character. This is necessary
—I must once more emphasize the fact,—not on account of the
subject itself, but out of consideration for the misguided feelings of
the public. It may be regrettable, but it is true.
Of all the tribes in the South of German East Africa, the Yaos seem
to be, not only the most progressive, but the most prosaic and
unimaginative, and in fact their initiation ceremonies are very
simple, compared with those of the Makonde and Makua. Those of
the latter have to a certain extent a dramatic character. The Makua
choose a branch of a particular shape, and forked several times,
which they plant in the midst of the open space where the festival is