Dynamics of Liquid Solidification - Thermal Resistance of Contact Layer

Mathematical Engineering
ZygmuntLipnicki
Dynamics
of Liquid
Solidification
Thermal Resistance of Contact Layer
Series editors
Jrg Schrder, Essen, Germany
Bernhard Weigand, Stuttgart, Germany
Today, the development of high-tech systems is unthinkable without mathematical
modeling and analysis of system behavior. As such, many elds in the modern
engineering sciences (e.g. control engineering, communications engineering,
mechanical engineering, and robotics) call for sophisticated mathematical methods
in order to solve the tasks at hand.
The series Mathematical Engineering presents new or heretofore little-known
methods to support engineers in nding suitable answers to their questions,
presenting those methods in such manner as to make them ideally comprehensible
and applicable in practice.
Therefore, the primary focus iswithout neglecting mathematical accuracyon
comprehensibility and real-world applicability.
To submit a proposal or request further information, please use the PDF Proposal
Form or contact directly: Dr. Jan-Philip Schmidt, Publishing Editor (jan-philip.
schmidt@springer.com).
More information about this series at http://www.springer.com/series/8445

Zygmunt Lipnicki
Dynamics of Liquid
Solidication
Thermal Resistance of Contact Layer
123
Zygmunt Lipnicki
University of Zielona Gra
Zielona Gra, Lubuskie
Poland
ISSN 2192-4732 ISSN 2192-4740 (electronic)

ISBN 978-3-319-53431-2 ISBN 978-3-319-53432-9 (eBook)
DOI 10.1007/978-3-319-53432-9
Library of Congress Control Number: 2017931050
Springer International Publishing AG 2017

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microlms 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 specic 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. The publisher remains neutral with regard to
jurisdictional claims in published maps and institutional afliations.
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
Acknowledgements
I would like to thank Prof. Bernhard Weigand from Stuttgart University for
long-term scientic cooperation. Thanks to it came into being a number of joint
publications, which are the basis of this book.
I also thank to Prof. Waldemar Woczyski from the Polish Academy of
Sciences in Krakw and Prof. Jerzy Godziszewski from University of Zielona Gra
for their valuable advice and assistance in editing this book.
I would also like to dedicate my book to my wife Krystyna and my daughter
Monika.
v
Contents
1 Solidication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 The General Solidication Problem . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Equations Describing the Solidication Problem . . . . . . . . . . . . . . 2
2 Solidication of PCMs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.1 Heat Storage Based on the Solidication Process . . . . . . . . . . . . . . 7
2.1.1 Multidimensional Solidication Heat Transfer. . . . . . . . . . . 8
2.1.2 Solutions of the Practical Solidication Problem . . . . . . . . . 9
2.2 The Geometry of Outer PCMs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3 The Criterion for the Selection of Accumulator Design . . . . . . . . . 12
3 The Model of Solidication of a Liquid with the Contact Layer . . . . 15
3.1 Equations Describing the Problem with the Contact Layer . . . . . . . 15
3.2 Strict and Approximate Solution of the One-Dimensional
Solidication Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.2.1 Strict Solution of the One-Dimensional Solidication
Problem. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.2.2 Approximate Solution of the One-Dimensional
Solidication Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
4 Solidication on a Rectangular Geometrics . . . . . . . . . . . . . . . . . . . . . 25
4.1 The Immovable Layer of Liquid Solidication . . . . . . . . . . . . . . . . 25
4.2 Forced Convection Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.2.1 The Effect of the Boundary Layer on the Solidication
Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
4.2.2 Solidication on a Plate with Variable Temperature . . . . . . 38
4.3 Free Convection Effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.3.1 Quasi-Stationary Convection in a Vertical Channel . . . . . . . 46
4.3.2 Quasi-Stationary Convection in a Horizontal Channel. . . . . 50
4.3.3 The Simplied Model of Convection
with Solidication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
vii
viii Contents
5 Solidication in an Annular Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

5.1 The Solidication of a Non-heated Liquid . . . . . . . . . . . . . . . . . . . 63
5.2 Solidication of a Superheated Liquid . . . . . . . . . . . . . . . . . . . . . . 67
5.2.1 Theoretical Solution of the Solidication Problem . . . . . . . 68
5.2.2 The Beginning of a Solidication ProcessAsymptotic
Solution. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
5.3 The Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
5.3.1 Research Equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
5.3.2 The Course of the Experiment . . . . . . . . . . . . . . . . . . . . . . 74
5.3.3 The Results of the Research and Discussion . . . . . . . . . . . . 78
5.3.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
6 Solidication of a Liquid Flowing into the Channel . . . . . ......... 83
6.1 The General Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . ......... 83
6.2 The Solidication of a Liquid Flowing into a Channel
with Weakly Conducting Walls . . . . . . . . . . . . . . . . . . ......... 87
with Conducting Walls . . . . . . . . . . . . . . . . . . . . . . . . . ......... 92
7 Role of the Contact Layer in a Solidication Process . . . . . . . . . .... 97
7.1 Studies of the Thermal Contact Resistance . . . . . . . . . . . . . . . .... 97
7.2 A Theoretical Analysis of Thermal Contact Resistance
on the Basis of Experimental Investigations . . . . . . . . . . . . . . .... 101
7.2.1 Thermal Contact Resistance for Solidication
on a Plate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 101
on a Cylinder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 103
7.3 Development of the Contact Layer During the Solidication
Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 105
7.3.1 Development of Contact Layer and Its Role
in Phase-Change Process on a Rectangular Plate . . . . .... 108
7.3.2 Development of a Contact Layer in Solidication
on an Annular Surface . . . . . . . . . . . . . . . . . . . . . . . . .... 111
8 Phase Heat Accumulator. . . . . . . . . . . . . . . . . . . . . . . . . . . ......... 119
8.1 Selected Review of the Units Used for Heat Storage . . ......... 119
8.2 Cooperation of the Heat Accumulator Storage
with the Heat Pump . . . . . . . . . . . . . . . . . . . . . . . . . . . ......... 120
8.3 Example of Calculations . . . . . . . . . . . . . . . . . . . . . . . . ......... 123
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
Annexe 1 (Zalba i in., 2003; Domaski, Moszyski, 1983;
and Others) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
About the Author
Zygmunt Lipnicki is the professor of the Institute of

Environmental Engineering at Zielona Gra University
in Poland. He received his doctoral degree in 1985 at
Pozna Technical University and his habilitation
degree from Pozna Technical University in 2000.
Dr. Lipnickis teaching focuses primarily on fluid
dynamics, thermodynamics and heat transfer. His cur-
rent research programmes include analytical and
experimental studies of thermal resistance of contact
layer in solidication process.
ix
Introduction
The resistance of heat flowing through a transition layer which occurs on the border
of different heat-exchanging phases is a phenomenon common both in technology
and in nature. It has been a subject of many scientic papers discussed in detail in
the works of Madhusudana (1996) and Furmaski and Winiewski (2002). In the
process of solidication occurring, for example, in casting moulds, a contact layer is
formed between the wall of the container and the solidied liquid determining the
intensity of heat exchange between the wall and the liquid. This particular matter
was the research subject of many authors (Beck 1969; Longa 1985; Wang and
Matthys 1994; Sahai 1998; Loulou et al. 1999a, b).
The transition layer is a spatial element bounded by two not very precisely
dened surfaces contacting with other bodies. The literature uses a less precise
name of this phenomenon: contact layer. This name will be also applied in this
book.
To a great degree, this book is a summary of its authors works published in the
last years concerning the research on thermal resistance of a transition layer in the
solidication process of the flowing liquid phase (Lipnicki 1999, 2003; Lipnicki
et al. 2005; Lipnicki and Bydaek 2008; Lipnicki and Kuczma 2009; Lipnicki and
Weigand 2011, 2012; Lipnick et al. 2014a, b; Lipnicki and Panto 2014; Weigand
and Lipnicki 2016). It is the summing up of the book published in Polish (Lipnicki
2012) and in a great part can be considered its second edition. However, this one
includes new elements. Some slight mistakes were spotted and removed, and more
chapters and new approach were included.
Storage and release of heat during a phase transition of melting or solidication
are interesting and important phenomena that can be practically applied in the
energy sector (Domaski and Moszyski 1983; Alexiades and Solomon 1993;
Jaworski 2001; Wnuk 2001; Suguwara et al. 2008; Mehling and Cabeza 2008;
Krasowski and Lipnicki 2009; Lipnicki and Krasowski 2010). The main feature of
those phenomena is a relatively small volume of the heat-accumulating units and
stable temperature of transition processes. Moreover, by applying PCM materials of
high transformation energy accompanying heat storage or release during a
xi
xii Introduction
solidication or melting process in heat-storing units, it is possible to reduce the

capital cost of the heat-accumulating equipment.
PCM materials are usually poor thermal conductors; therefore, the amount of
heat flowing from the battery to the environment is small. Considerable heat
resistance of the layer solidied during the phase change makes it necessary to
increase the surface of heat transfer in relation to the undergoing transition of outer
material. Therefore, the impact of the outer geometry of PCM materials on the heat
flow is discussed in this book. The second important heat flow-inhibiting factor is
the transition layer itself formed between the solidifying liquid and the heat-
consuming cold wall. An analysis of the impact of heat resistance of that layer on
the flow of heat between the wall and the liquid is the main subject of this book.
This book consists of eight chapters and the summary. In Chap. 1, a general
model of liquid solidication is discussed. Conservation equations concerning the
flow of both mass and heat and boundary conditions are formulated. Conservation
equations are subjected to appropriate physical interpretation. Possible solutions
of these equations are supplied.
In Chap. 2, processes of solidication of liquid PCM materials are analysed with
special attention to both their geometric arrangement and location of channels with
PCM materials in relation to gravitational force.
Chapter 3 concentrates mainly on formulating the fluid solidication with taking
into consideration the transition layernamed also the contact layer that is formed
between a cold wall and a solidied layer. Solutions for a one-dimension theoretical
model of fluid solidication, quasi-stationary, strict, and approximate ones, have
been compared.
In Chap. 4, the following processes are discussed: solidication of immovable
liquid within a rectangular space, solidication with a contact layer, and solidi-
cation of flowing liquid in contact with a cold plate, and solidication of liquid
under natural convection. Examples of solidication both of overheated and of not
overheated liquids are provided.
In Chap. 5, solidication of a liquid in closed, annular, cylindrical units is
discussed.
Dynamic solidication of a liquid in a cooling channel is discussed in Chap. 6.
Chapter 7 treats both the development of a contact layer in the solidication
process and the effect of its resistance on the process.
In Chap. 8, a project of an exemplary phase accumulator is presented.
This work is addressed to a scientist, a power engineer, or a mechanic engineer
interested in the phenomenon of heat flow accompanying the phase changes and in
dening the dynamics of liquid solidication that is the process of development
of the solidied layer, as well as in the relation between temperature distribution
and generated heat flow and time. Apart from the amount of heat absorbed or
released during the transition phase, an important parameter is the thermal power
that in a simple way depends on the size of the outer surface of the PCM material
releasing or absorbing the heat. Some applicable, in practice, basic equations
dening the energy flow in the solidication process of a liquid are presented in this
book.
Chapter 1
Solidication
1.1 The General Solidication Problem
Solidication, i.e. transition of the liquid phase into the solid one, is a phenomenon
that often occurs in both nature and engineering. In nature, to the most common
solidication processes belong to: water freezing and ice forming, ground frost and
solidication of lava flowing out of a volcano crater. In engineering, the techno-
logical processes are used in the production. Solidication is applied in metals and
alloys founding, in welding and soldering and also in food freezing.
The solidication process is accompanied by an energy effect, i.e. emission of
heat affecting the process intensity. In both the specialist and the scientic literature,
a great number of papers have been devoted to this problem; these works were
reviewed amongst others by: Viskanta (1983), Fukusako and Yamada (1994), and
Hu and Argyropulus (1996).
Figure 1.1 presents the phenomenon of solidication of a liquid initially lling
up the whole at random selected space X that during the process is divided by a
surface (called the solidication front) into two separable parts, the liquid XL and
the solid one XS X XL XS . The shape of the space X depends on the geo-
metric arrangement of containers or channels in which flows the liquid subjected to
the process of solidication. In general, the interface division F can be a curved one
and dened by a perpendicular vector ~ nF that moves towards the liquid with
!
solidication velocity U F and depends on the location on the surface in which the
phase transition heat is generated. The phenomenon of phase transition occurs
because of the total heat removal through surface @X which bounds the considered
space
Z
_Q1 Q_ 2 k grad T x; y; z; t ~
nd @X [ 0; 1:1
@X
Springer International Publishing AG 2017 1

Z. Lipnicki, Dynamics of Liquid Solidication, Mathematical Engineering,
DOI 10.1007/978-3-319-53432-9_1
2 1 Solidication
where x; y; z; t are space coordinates in relation to Cartesian system and time, with T
temperature and kthermal conductivity coefcient of the substance solidifying
on the border of the space.
The above is an acute inequality and it satises the necessary condition of
solidication. The heat removed through the outer surface of space @X diminishes
the internal energy of substance accumulated in space X.
The heat flow Q_ 1 is removed through the cold surface XS absorbing the heat
released in the process of phase transition, and the heat flow Q_ 2 can be transported
to the liquid solidifying through surface XL . With the advancement of time, the
volume of solidied layer XS is increased and the volume of liquid XL is reduced.
Because of the interface movement, equations dening the considered process of
solidication are nonlinear and that is one of the main difculties in obtaining an
analytic solution for the problem of solidication. Dening the shape and the
velocity of movement of the interface is one of the numerous important tasks
concerning the solidication process.
1.2 Equations Describing the Solidication Problem
The considered solidifying liquid is affected by body force ~ f which is usually equal
~
to gravitational force f ~
g, if the solidication process occurs in gravitational eld.
The impact of body forces can be the cause for free liquid convection which effect
on the solidication process requires to be taken into consideration in a theoretical
analysis. The location of the liquid and solid elements is related to Cartesian
coordinates system x, y, and z.
Theoretical problems of this kind are considered to be those with a movable
boundary condition known in the literature as Stefan problem (Rubinstein 1967). In
the scientic literature, there are many works concerning the solution of liquid
solidication problem (Rubinstein 1967; Viskanta 1983; Crank 1984; Mochnacki
and Suchy 1993).
The phenomenon mentioned above is very complex. It is mathematically dened
by means of a system of coupled equations of mass conservation and of motion and
of energy equations, either within a liquid or a solid body.
In the considered case of a liquid, the equation of mass conservation is as
follows:
!
div U x; y; z; t 0; 1:2
!
with velocity vector U . In the above equation, it was also accepted that the liquid is
incompressible, i.e. it is of constant density, what is justied at not too high
pressures.
The equation of incompressible liquid motion is also dened by NavierStokes
equation of momentum conservation
1.2 Equations Describing the Solidication Problem 3
!
@ U x; y; z; t ! !
U x; y; z; t grad U x; y; z; t
@t 1:3
1 !
~ f grad px; y; z; t mr2 U x; y; z; t:
qL
In the above equation, p means pressure inside the liquid and m means kinematic
viscosity. The expression on the left side of Eq. (1.3) denominates the relation of
material velocity to time and space variables x, y and z as the summing-up of
! !
derivatives of localisation @ U =@t and convection U gradU . ~
f on the right side
of the equation denominates the body strength unit vector; grad
~i@=@x ~j@=@y ~ k@=@z and r2 @ 2 =@x2 @ 2 =@y2 @ 2 =@z2 are differential
operators of rectangular coordinate system.
The equation of energy conservation for a liquid flowing through low velocity
without considering the energy dissipation because of meagre liquid viscosity and
lack of internal heat sources is as follows:
@TL x; y; z; t ! kL
U grad TL x; y; z; t jL r2 TL x; y; z; t; jL ; 1:4
@t qL cL
In the above equation, an additional constant coefcient of liquid heat conduc-

tion kL and constant specic heat cL were accepted. Parameter jL is the coefcient
of the liquid heat diffusion. The expression on the left side of Eq. (1.4), similarly as
in the equation of momentum conservation, denominates the material temperature
derivative TL , and r2 on the right side is the Laplaces operator. This equation has
been known in the literature as FourierKirchhoff equation.
The equation of heat flow in the immovable solidied part is described by
Fourier equation
@Ts x; y; z; t ks
js r2 Ts x; y; z; t; js ; 1:5
@t qs cs
where the values of the solidied layer are presented by: Ts temperature, js coef-
cient of solid heat diffusion, ks coefcient of solid heat conduction, qs solid
density and cs solid specic heat.
Solution of the system of Eqs. (1.2), (1.3), (1.4) and (1.5) can be obtained from
the following boundary conditions on the front of the phase divisions:
The temperature of the liquid and the solidied layer is equal to the temperature
of liquid solidication TF on the solidication front F
TL x; y; z; tjF Ts x; y; z; tjF TF ; 1:6

4 1 Solidication
n detail of interface
solid
f
iL
Q1 solid
S nF UF Q2
L iLF liquid
L iS
UF n
iSF
.
S q
z L
y
x liquid front of solidification
front of solidification
Fig. 1.1 Solidication of a liquid in a space
The difference in heat flow conducted through front F of phase divisions is equal
to the velocity of heat release on the solidication front in the direction per-
pendicular to the front (left side of the equation)
qS LUFn ks~
nF grad Ts x; y; z; tjF kLs~
nF grad TL x; y; z; tjF 1:7
where L is the heat of solidication of the liquid, also called the latent heat of
phase transition, UFn is the velocity of front solidication in the direction
perpendicular to the front and kS and kL are coefcients of conduction of both
solid and liquid bodies.
The solidication heat is equal to the difference in the heat contents of liquid and
solid bodies in the solidication temperature (see Fig. 1.1)
L iLF iSF 1:8
On the outer front of the solidied layer, the boundary condition of solidication
temperature is as follows:
Ts x; y; z; t T@XS ; 1:9
where T@XS means the temperature on the border of the solidied layer and the wall.
The liquid velocity on the walls of the channel surrounding the solidifying liquid is
equal to 0.

!
U 0: 1:10
@XL
It is presumed that on the border of the wall and the cooling liquid, the streams
of heat transported to the cold wall and of that absorbed from it are equal
1.2 Equations Describing the Solidication Problem 5
@TW
kSC hTW T0 ; 1:11
@n
where T0 is the temperature of the environment, h means coefcient of heat

transmission on the border of the wall and the cooling liquid and n means direction
perpendicular to the surface of the wall.
A general analysis of the immovable liquid solidication problem in the same X
area was conducted by Shamsunder and Sparrow (1975). They suggested the
application of an enthalpy model to the research on immovable liquid solidication.
The change in time of enthalpy of the solidifying medium contained in space X, as
it results from the principle of energy conservation, is described by the equation
Z Z
d
q ix; y; z; tdX k grad T x; y; z; t ~
nd @X; 1:12
dt
X @X
where i means the specic enthalpy of the body contained in the considered space,
depending on location and time.
The right side of the above equation presents the total heat flowing through the
outer surface @X. The mathematical analysis of the solidication process inside the
space X shows that the left side of Eq. (1.12) expressing the change of enthalpy
may be reduced to the following (Shamsunder and Sparrow 1975)
Z Z Z
d
qidX k grad T ~
nd @X kS~
nF grad TS jF kL~
nF grad TL jF qLUFn dF:
dt
X @X F
1:13
As it results from the principle of the energy conservation, a change in the

enthalpy inside the space X following the solidication change undergoing inside
the space cannot cause a change of the total enthalpy if the net amount of heat
transported to the considered area does not change.
Comparisons of Eqs. (1.12) and (1.13) allow to obtain an energetic condition on
the solidication surface
kS~
nF grad TS jF kL~
nF grad TL jF qLUFn 0 )
1:14
kS~
nF grad TS jF kL~
nF grad TL jF qLUFn
compatible with the condition (1.7).

Because the above equations are nonlinear, their full analysis is very difcult
especially in multidimensional geometric systems. They are generally solved by
applying approximate methods, i.e. analytical, semi-analytical and numerical ones.
An analytical method allows to describe the phenomenon in a reasonable simple
way to supply sufciently precise and practically useful solutions. Worthy of
attention are the works of Viskanta (1983), Chung and Epstein (1984), Alexiades
6 1 Solidication
and Salomon (1993), Fukusako and Yamada (1994) and Weigand et al. (1997,
1999), and also other authors research and their own solutions of solidication and
melting of a flowing liquid as well as methods of solving the problem were
reviewed. Research on a possible simple solution of the above system of equations
is also a subject of this work.
Chapter 2
Solidication of PCMs
2.1 Heat Storage Based on the Solidication Process
Heat storage is a subject of considerable importance in power engineering. The

most common methods of heat storage exploit the thermal capacity, the latent heat,
the heat of reaction and of photo chemical reaction and production of fuels. This
work is concentrated on heat storage by exploiting the phenomenon of solidication
as a special case of numerous phase changes occurring in reality.
The discussed heat storage and release system include a solidifying PCM
(phase-change material), a reservoir and a heat-transporting medium. An example
of heat storage and release is presented (Fig. 2.1). The PCM contained in a
heat-isolated reservoir is successively subjected to melting when heat is stored and
to solidication when it is released.
The agent acting as an intermediary in the storage or release of heat Q_ flows
inside the channel or channels submerged in the material undergoing the phase
change. The solidication process starts when PCM reaches the solidication
temperature. In this phase very important for the process are thermophysical fea-
tures of the material, its geometric shape, thermophysical features of the reservoir
wall and the conditions of heat absorption. Equilibrium processes of solidication
of different PCMs are presented (Fig. 2.2).
The process of solidication as it was discussed in Chap. 1 is the change of a
liquid into a solid body with heat release at the same time. The phenomenon of
equilibrium solidication of homogeneous bodies and compounds is presented
(Fig. 2.2), where 1 and 2 appropriately mark the starting and the nal points of the
process. An interface between the solid and liquid parts of geometry depending on
the heat transfer conditions separates a sharp solidication front of temperature TF
in case of homogeneous materials and a biphasic surface of temperature within
liquidus TL and solidus TS interval in case of inhomogeneous ones. Within the not
yet solidied liquid part, free or forced convection can still advance. Full

DOI 10.1007/978-3-319-53432-9_2
8 2 Solidication of PCMs
reservoir PCM - liquid
.
Q
intermedium
Fig. 2.1 Heat storage and release
T T
1 2 TL 1
TF 2
TS
Q1 2 Q 1 2
S S
Fig. 2.2 PCM solidication processes
solidication is accompanied by the generation of heat Q12 and flow of entropy

S to the heat receiver. A large eld of research on the storage of heat or cold
resulting from the phenomenon of solidication or melting was presented in the
review works of Viskanta (1983), Zalb et al. (2003), Prashant et al. (2008) and
Mehling and Cabez (2008). Examples of mathematic modelling of solidication
processes are described among others in works: Mochnacki and Suchy (1993), and
Alexides and Solomon (1993).
2.1.1 Multidimensional Solidication Heat Transfer
In practice, geometric shapes of applied PCMs are complex and multidimensional

and depend on the shape of the reservoir (channel) in which they are contained.
2.1 Heat Storage Based on the Solidication Process 9
Release of the accumulated heat from liquid PCMs during solidication is, as it has
been already mentioned, a very complex phenomenon because of a movable surface
called the solidication front (the case of movable boundary condition) of unknown
shape separating the solid phase from the liquid one. Thus, a full description and a
complete analytic solution are very difcult or even impossible.
In the solving of the two- or three-dimensional problems of heat transfer during a
phase change, most helpful method is the numerical ones. With some chosen
numerical methods and their review one, can get acquainted in works, among other
of Shamsunder et al. (1975), Talmon et al. (1981), Viskanta (1983) and Mochnacki
and Suchy (1993). Upon an analysis of presented work, the following main
numerical methods applied in heat exchange during phase changes can be
distinguished:
Methods of nite differences (MRS) in regard to temperature as a denite
parameter (method of nite explicit difference, method of innite inexplicit
difference, method of movable boundary immobilization, method of isotherm
migration);
Methods of nite differences (MRS) in regard to enthalpy, i.e. the enthalpy
method;
Methods of nite elements (MES).
Numerical methods are especially helpful in situations where analytic ones are
not applicable, particularly in case of multidimensional phenomena. As particular
errors of numerical methods can be listed those of process modelling (the applied
mathematical model does not accurately reflect the reality, errors committed when
tracing the original area by a computational one, errors of discretisation and those of
rounding. The numerical methods are also disadvantageous because of the necessity
to control the numerical error, the latter depending on net density and on temporal
step. The analytic methods are free of numerical errors, and the solutions obtained
are clearer in the form and easy to interpret.
2.1.2 Solutions of the Practical Solidication Problem
In practice, the attainable PCMs due to their temperature and the phase-change heat
are characterized by a relatively small coefcient of heat transfer (Zalba 2003). To
ensure effective storage of heat of high-thermal stream intensity, the volume of the
heat accumulating materials should be small in relation to the boundary surface.
However, to ensure storage of appropriately big amount of heat, the total volume of
the material subjected to phase change should be also appropriately considerable.
To make that discrepancy compatible, the total PCM volume should be divided into
components of relatively small volume. Then, the time of charging and discharging
of heat accumulators based on phase transition can be relatively short and regulated
through appropriate construction of heat exchangers (heat accumulators).
Fig. 2.3 Element of heat

phase exchanger
H g
n2
solid layer
n1 q2
l
q1
PCM
2
cooling liquid
The flow of heat in single elements can be modelled as one-dimensional heat

flow. Then, a theoretical solution of such a fragmentary task is not theoretically
difcult since there are simple analytical methods of solving this problem.
Figure 2.3 shows an example of such element lled with PCM where H is the width
of the channel and l its length. The width of the channel is assumed as small in
relation to its length.
H
1: 2:1
l
PCM is on both sides bounded by outer surfaces @X1 and @X2 being at the same
time the channel walls marked by perpendicular ~
n1 and ~
n2 vectors, with intensities
of the heat streams described by equations

@T @T
q_ 1 kS and q_ 2 kS ; 2:2
@n1 @X1 @n2 @X2
with Tthe temperature and kS the coefcient of heat conductivity through a

PCM.
When analysing the phenomenon occurring in simple geometric elements, some
effects of free convection occurring within the liquid PCM in the result of the
gravitation ~
g as well as heat resistance of the contact layer formed between the
solidied layer and the cool wall of the channel can be easily taken into consid-
eration. The resistance of the contact layer is important for the construction aims.
Constructors of heat exchangers based on phase transition (solidifying and
melting) and on PCM exploitation should aim at providing such construction
solutions in which heat exchange is effective.
2.2 The Geometry of Outer PCMs 11
2.2 The Geometry of Outer PCMs
Generally, the geometry of PCMs is very complex. It is the main difculty when
theoretically describing the flow of heat which accompanies the transition phase
occurring within the material considered. The radius of the wall curvature is of
signicant importance. A similar problem can be found in casting of alloys where in
case of casts of different outer geometries, the problem of solidication is attempted
to be considered irrespective of their shape by introducing a representative
parameter, i.e. supplementary wall thickness (Braszczyski 1989), introduced by
Chvorinov, expressing the relation of the cast volume to its outer surface. The cast
solidication time is proportional to the square of the above-mentioned parameter.
Such a simple description of solidication is difcult to be accepted for PCMs
since the latter are usually poor heat conductors. An attempt to overcome this
inconvenience is the research on the phase change of PCM substances with added
metallic supplements, thereby creating a unit of porous structure. Such research was
conducted in the work of Weaver and Viskanta (1986). To improve the effectiveness
of heat conductivity of the material undergoing the phase change, the above authors
were sinking metal globules, aluminium or copper ones in its volume. Thus, created
system was characterized by highly effective heat exchange. Theoretical research on
similar porous medium was conducted in the work (Lipnicki and Weigand 2008).
Another way for the improvement of the heat exchange effectiveness is the
choice of outer geometry of PCM. Figure 2.4 shows some examples of various
shapes of materials undergoing phase changes. Each case requires an individual
theoretical approach; however, theoretical difculties rise proportionally to the
outer geometry complexity.
The presented above particular cases of outer geometry of materials, i.e. a thin
plate (a), a cylinder (b), a hollow cylinder (c) and a globe (d), in the case of heat
accumulators can be of different degree of usability. An accurate identication of the
most advantageous heat accumulator is determined by material, the geometry and
outer conditions. Its construction is very complex because of a great number of
(a) (b) (c) R1
H R
(d)
L L
L R
R2
Fig. 2.4 Examples of outer geometry of PCMs

interrelation combinations between those factors. The outer conditions are dened
among others by heat resistance of the layer near the contact surface. An accurate
choice of the most preferable case requires lots of theoretical analyses. As it has been
already mentioned, heat conductivity of the PCMs is very poor. This makes the time
of heat release longer; i.e., the power of heat flow is relatively low. To increase the
strength of the heat flow, one should attempt to enlarge the outer surface of the
element, and in what follows, there is a need to divide the construction into several
relatively thin components. The outer geometry of the element determines the
process course. Usually, it is advised to obtain a relatively high power of the
accumulator. To full this task, the investigation of phenomena may be limited to
some chosen shapes of elements of PCMs (Mehling and Cabeza 2008).
2.3 The Criterion for the Selection of Accumulator Design
The constructions of heat accumulators based on phase change of PCMs may differ in
relation to their intended use, size and shape. The intended use of a heat accumulator,
the method of heat delivery or removal directly influence the choice of geometry of
PCMs and their outer sizes. The shape of elements of an accumulator and the method
of heat absorption determine the phenomena indirectly realized during the process.
Figure 2.5 shows various examples of proles of phase changes in PCMs together
(a) (b) (c) (d)

g g
solid solid
layer layer
q q
solid q q q q
PCM
layer
PCM PCM insulation
PCM solid insulation
layer
insulation
(e) (f) (g) g
PCM
solid
layer solid
insulation layer
q q
PCM PCM
q q
q solid
(h) g
layer
solid
layer
PCM
insulation
Fig. 2.5 Exemplary models of solidication of a flat liquid layer

2.3 The Criterion for Selection of Accumulator Design 13
with the direction of heat removal q. These are the models of solidication of a flat
layer of a liquid in the case of two-sided cooling (a) and one-sided cooling (b);
models of solidication of a vertical layer of liquid under the impact of free con-
vection during two-sided cooling (c) and one-sided cooling (d); models of solidi-
cation of a liquid layer during forced flow at two-sided cooling (e) and one-sided
cooling (f); and models of solidication of horizontal immovable layer of a solidi-
fying liquid cooled from the bottom (g) and under free convection (h) cooled from
above. During free convection, an important role is played by the orientation of the
element of the heat accumulator to the gravitation vector ~
g. In a similar way, one can
consider the models of solidication of elements constructed of PCMs of curved
outer surfaces in which more complex phenomena of heat flow and exchange occur
and their theoretical description is still more complex. Every one of the mentioned
models should follow basic equations of mass conservation and those of momentum
and energy if both the initial and boundary conditions have been fullled.
It remains to determine which of the above-presented theoretical models reflex
the real phenomenon in the best way? An attempt to establish the choice criterion
regarding the construction of a heat accumulator in view of its intended use is one
of the aims of this work. To answer the question of the choice criterion for a heat
accumulator, its main exploitation parameters, rst of all capacity and thermal
power, should be dened. The listed above features are derivatives of shape and of
volume of the PCM, of geometric sizes and outer surface within which PCM is
contained, of difference between the solidication and of cooling temperatures,
solidication heat, specic heat, density, heat conductivity and viscosity.
Introductory estimation of the amount of accumulated heat and the heat stream size
in regard to the applied PCM and the type of construction requires an analysis and
theoretical evaluation based on the accepted physical model. In other words, con-
struction of a heat accumulator requires a complex theoretical analysis of the whole
system of the heat accumulator and its elements.
To establish a general choice criterion for a heat accumulator considering its
outer needs and as well as to choose the best solution in regard to the posed
problem is a complex task requiring answers to a lot of elementary questions and
joining them together.
The features of an accumulator are dened by heat power and capacity,
geometry and thermal conductivity of material. The phenomena appearing in the
process are the flow of heat in immobile material and free convection or the forced
one. Outer conditions will dene the boundary ones and the method of heat transfer.
These factors should decide on the choice of the accumulator. The collection of the
elements of a heat accumulator is a collection of various types, and some elements
are interdependent what hinders their equal treatment.
The factors determining the features of a heat accumulator can be divided into
the following groups:
The factors deciding on free convection:
type of PCM,
spatial arrangement of the reservoir in which PCM is contained,
condition of heat absorption,

geometry of outer surface of the reservoir in which PCM is contained and
thermal resistance of the contact layer between PCM and the boundary wall.
The factors deciding on the type of forced convection:
type of PCM,
geometry of the flow channel,
conditions of heat absorption from the flowing liquid and
The factors deciding on mixed convection and forced and free one:
type of PCM,
participation of factors influencing this kind of convection (pressure gradient,
temperature gradient),
geometry of the flow channel,
conditions of the heat absorption from the flowing liquid and
When considering the presented above models of the phenomena accompanying
liquid solidication and the groups of factors deciding on the capacity and power of
a heat accumulator, it can be seen that solidication of PCMs to release their earlier
accumulated heat is a very complex problem. The author offers an individual
approach to each particular case. Examples of such solutions are presented in the
next parts of this work.
Chapter 3
The Model of Solidication of a Liquid
with the Contact Layer
3.1 Equations Describing the Problem with

the Contact Layer
A liquid PCM subjected to solidication in a reservoir is treated as a Newton liquid

with its solidied volume on one side bounded by the cold wall of the reservoir and
on the other side by the solidication front surface.
To simplify further reflections and to conrm the appearance of free or forced
convection of the liquid before the solidication front, the thermal conductivity
within the liquid on the interface boundary in Eq. (1.6) is replaced by the following
one

@TL
kL hTF Tf ; 3:1
@n xd
where the macroscopic parameter of heat transmission h describes the amount of

heat transfer between the liquid and the front of the layer solidied in temperature
TF and where temperature Tf is that of liquid remote from the boundary layer.
Direction n means perpendicular direction to the solidication surface.
Equation of heat flow in the solidied layer of a movable edge, i.e. in an area of
variable volume, at constant thermal conductivity coefcient in case of isotropic
bodies is dened by quasi-stationary Fourier equation
r2 Ts x; y; z 0; 3:2
with temperature Ts of the solidied layer.

In relation to the accepted space coordinate system, Laplace differential operator
is as follows:

DOI 10.1007/978-3-319-53432-9_3
16 3 The Model of Solidication of a Liquid with the Contact Layer
for Cartesian x, y and z system
@2 @2 @2
r2 ;
@x2 @y2 @y2
for cylindrical r; u; z system

1@ @ 1 @2 @2
r2 r 2 2 2;
r @r @r r @u @z
for spherical r; u; / system

1 @ 2 @ 1 @ @ 1 @2
r2 r sin / 2 2 :
r @r
2 @r r sin / @/
2 @/ r sin / @u2
In the process of liquid solidication, a gap, further called the contact layer, is
created at the cold wall (Fig. 3.1). On one side, the gap is bounded by the surface of
the layer solidied at temperature T and on the opposite side by the outer cold wall
of surface temperature Tw . The formed contact layer causes additional heat transfer
resistance q_ from the solidifying liquid to the surrounding wall, and the temperature
within the contact layer decreases
DT T Tw : 3:3
The shape of the solidication front on which condition (3.1) is satised depends
on the geometry of solidifying body and the direction of phase-change heat transfer.
For thin-layered solidifying PCMs (Fig. 3.1), the temperature prole across a
solidied layer of d thickness in majority of cases will be considered as a linear one
within 0 d0 d interval
Fig. 3.1 Model of H

solidication of a liquid
g solidification
front
.
q

Tf
liquid TF
T cold wall
TW
U
. T
z
y
x solid layer
contact layer
3.1 Equations Describing the Problem with the Contact Layer 17
0
d
T T T TF : 3:4
d
Thus, the average temperature in the solidied layer will be equal to

Rd
Tdd0 T TF
Tsr 0
; 3:5
d 2
and the change of heat accumulated in the solidied layer will be
Zd
d dd
qs cs Tdd0 qs cs Tsr ; 3:6
dt dt
0
and is small in comparison with the phase-change heat stream, i.e. qS cs Tsr qS L.
In majority of tasks discussed in this work, the above condition is taken into
consideration.The flow of heat through the contact layer is convectional and dened
by Newtonian equation
q_ hCON T Tw ; 3:7
where hCON parameter represents the convective heat transfer coefcient of the
contact layer.
The structure of the contact layer and its impact on the heat flow require special
attention. A theoretical elaboration of the effect of the transition layer on the
conditions of heat exchange is a very important as well as difcult and not quite
satisfactorily solved task. Any knowledge concerning the phenomenon of creation
of the mentioned layer in a solidication process and its heat resistance is based on
the rst place in already obtained experimental assessments. Experimental works
contrary to theoretical research allow to reach a solution of this problem, and such
papers can be found in the literature (Wang and Matthyes 1994; Loulou et al. 1998;
1999a, b). They mainly depend on measuring hCON coefcient of heat transfer
from the contact layer by means of various methods. This problem is given a special
attention in Chapter 4 of this work.
On the border of the wall and the cooling liquid, depending on the absence or the
presence of the contact layer, a condition of heat flow equality is posed
@Ts
ks h0 TW T0 ; 3:8
@n
where T0 is the temperature of the environment, h0 is the coefcient of heat transfer

on the surface of the outer wall and the cooling liquid and n indicates the direction
perpendicular to the surface of the wall.
3.2 Strict and Approximate Solution

of the One-Dimensional Solidication Problem
In order to conduct a comparative analysis of solutions of the solidication prob-

lem, a description of one-dimensional solidication of a flat layer of immovable and
not heated liquid (Fig. 3.2) with temperature equal to the solidication temperature
TF and thickness H is recommended. Between the surface of the solidied layer of
T temperature and a cold wall of TW temperature, during the solidication process,
a contact layer is formed and its impact on the solidication process is being
investigated. The generating force of the process is the cooling of the layer by
means of heat stream q. _ The surface of the solidication layer of thickness dt
moves at dd=dt velocity (solidication front velocity).
3.2.1 Strict Solution of the One-Dimensional

Solidication Problem
To solve the problem of solidication of a liquid layer, two methods have been used.
The rst applies a classical solution known in the literature as Stefan problem solution
(Rubinstein 1967, Viskanta 1983) The relation of temperature TS within the solidied
layer to time t and to coordinate of position x is presented in differential equation
@TS @ 2 TS
jS ; 3:9
@t @x2
After introducing denotations
TS T x ~d d ; jS t
hS ; ~x ; s FO Ste; FO ;
TF T H H H2
c T
T

Ste
S F
;
L
Fig. 3.2 Solidication of a TF

liquid rectangular layer
liquid
d
H solidification front
dt
solid
layer TS
0
contact layer q T TW
3.2 Strict and Approximate Solution of the One-Dimensional 19
Equation (3.9) takes the following form
@hS 1 @ 2 hS
: 3:10
@s Ste @~x2
The limiting conditions (boundary and initial ones) for the solidied layer are as
follows:
1 hS j~x0 0; s [ 0; 2 hS j~x~d 1; s [ 0; 3 hS js0 1; ~x [ 0;

@hS d~
d 3:11
4 ; s [ 0; 5 ~ds0 0:
@~x ~x~d ds
They are in turn as follows: 1initial temperature of the layer solidied on the
contact surface TS T ; 2temperature of the solidication front surface
TS TF ; 3temperature of the liquid within the whole space at the beginning
TS TF ; 4heat stream on the solidication front; and 5initial condition
dening the thickness of the solidied layer in the beginning of the solidication
process. Conditions 2 and 4 directly result from the equations dening the con-
ditions on the surface of solidication (1.7) and (1.9).
Solution of Eq. (3.10) is as follows
r!
~x Ste
hS C1 C2 erf ; 3:12
2 s
where C1 and C2 are integration constants.

Application of boundary condition 2 leads to the following
r!
~d Ste
1 C1 C2 erf ;
2 s
An analysis of the above equation shows that the argument of error function
(erf) should be also independent from time and equal to the constant K further
called the solidication constant
r
~d Ste
K : 3:13
2 s
Finally, the integration constants determined by the boundary conditions 1 and

2 are equal to
1
C1 0 and C2 :
erf K
When substituting the integration constants into Eq. (3.12), one obtains an
equation dening the eld of temperature within the solidied layer, depending on
the position and time
r!
1 ~x
Ste
hS erf : 3:14
erf K 2 s
Next, by means of boundary condition 4, following equation is obtained
Ste
p K exp K 2 erf K; 3:15
p
which denes solidication constant K in Stefan number Ste function.

By transforming Eq. (3.13), one obtains equation dening the thickness of the
solidied layer
~d p2K p
s: 3:16
Ste
The heat stream flowing through the contact layer is equal to that of transported
through the solidied layer

@TS
hCON T TW kS ; 3:17
@x x0
with the coefcient hCON of the heat transfer within the contact layer. The above
equation takes the form of

1 @hS
hW ; 3:18
BiCON @~x ~x0
where new values in Eq. (3.18) are in turn dened by hW the cooling wall tem-
perature lower than the contact temperature and by BiCON Biot number describing
the thermal conductivity of the contact layer:
TW T hCON H
hW ; TW \ T; BiCON :
TF T kS
As a result of differential Eq. (3.14), the thermal gradient on the surface of the
solidied layer is represented by following relation
2 r Ste
r
Ste
@hS 1 ~x Ste 1
exp :
@~x ~x0 2 erf K 4s ~x0 s 2 erf K s
When substituting the above into Eq. (3.18), a formula dening the temperature of
a cold wall in time function will be obtained
r
1=BiCON Ste
hW : 3:19
2 erf K s
For a perfect contact (absence of heat flow resistance), BiCON ) 1, hw ) 0, i.e.

the wall temperature is equal to the temperature of the surface of the contact layer
TW T.
Based on the formulae (3.14) and (3.19), the distribution of the temperature in
the solidied layer and in the cold wall is graphically presented in time function
(Figs. 3.3 and 3.4).
S
= 0.00446
1.0
= 0. 0 0 1 = 0.005
0.8
= 0.01 = 0 .0 2
F
0.6
0.4
0.2
0
0.005 0.01 0.015 0.02 x
- 0.2
( = 0.02)
( = 0.01)
- 0.4 S te = 0.0289; 1 BiCON = 0.05; T = 50 C ;
( = 0.005)
- 0.6 = 0.001 W = 0.996 TW = 10 0 C ;
= 0.005 W = 0.445 TW = 7.230 C ;
- 0.8
= 0.01 W = 0.315 TW = 6.580 C ;
- 1.0 ( = 0.001) = 0.02 W = 0.223 TW = 6.130 C ;
Fig. 3.3 Distribution of temperature in the solidied layer and the cold wall
0 0.05 0.1 0.15 0.2

0
1 BiCON = 0.05
- 0.2
- 0.4
1 BiCON = 0.1
- 0.6
1 BiCON = 0 .2
- 0.8
- 1.0
W

Fig. 3.4 Relation of the cooling wall temperature to solidication time h0
One-dimensional solidication of the water layer with initial solidication

temperature at constant temperature of the contact layer T 5 C equal to Stefan
number Ste 0:0289 is graphically presented (Fig. 3.3). Heat resistance of the
contact layer is 1=BiCON 0:05.
The graphs of distributions for different times show that the distribution
temperature of the solidied layer is in linear dependence to position at every time.
However, the temperature TW of the cold wall at xed temperature T on the surface
of the solidied layer increases with time (the points on axis hW ), when
approaching surface temperature of the solidied layer.
For long solidication times (see Fig. 3.4), the temperature difference between
the wall and the solidifying layer decreases and for big values of time tends to reach
the constant value. The jump of temperature through the contact layer decreases
with an increase in Biot number.
3.2.2 Approximate Solution of the One-Dimensional

Solidication Problem
The equation of balance of heat flowing through the solidied layer and the contact
layer is dened by the equations for a quasi-stationary state
dd TF T
qS L kS hCON T TF : 3:20
dt d
The rst expression of the above system of equations presents the heat stream
generated in the result of a liquid solidication, the second shows the heat stream
flowing through the solidied layer, and the third shows the stream of heat flowing
through the contact layer. Presentation of the temperature gradient in the second
expression as a relation of temperature difference and thickness of the solidied

layer is justied by linear distribution of temperature, as shown in Fig. 3.3.
By comparing the second expression with the third one in Eq. (3.20), the
following equation dening the contact temperature is obtained
TF kS TW hCON d
T ; 3:21
kS hCON d
showing that the contact temperature is not constant and depends on the solidied
layer thickness.
After introducing the following values:
~ d jS t cS TF TW hCON H
d ; s FO Ste; FO ; Ste ; BiCON ;
H H2 L kS
the system of Eq. (3.20) becomes a differential non-dimensional one
d~d 1
. : 3:22
ds 1 BiCON ~d
After satisfying the initial condition, the solution of the above equation is a formula
dening the solidied layer thickness in time function (Fig. 3.5)
s
~d 1 1
2
2s: 3:23
BiCON BiCON
.
Relation of the solidication velocity d~d ds and the solidied layer thicknesses ~d
to time s for different values of thermal resistance of the contact layer 1=BiCON
d d
0.5 10
0.4 1 B iC O N = 0 . 0 5 8
d d
0.3 6
0.2 4
1 B i C O N = 0.2
0.1 2
0
0 0.04 0.08 0.12 0.16 0.2
Fig. 3.5 Relation of the solidication velocity and the solidied layer thickness to time
is graphically presented by basing in the formulae (3.22) and (3.23). The speed of
the development of the solidied layer thickness rises along with its thermal
resistance diminishing. With escalation of time, solidication velocity decreases as
a result of an increase in the solidied layer thermal resistance.
As an example, two methods of solution of the problem of water layer solidi-
cation were compared in this chapter: the strict one described in Eq. (3.8) and the
approximate quasi-stationary one described in Eq. (3.23). If a perfect thermal
contact between the cold wall and the solidied layer BiCON ! 1 is assumed, then
Stefan numbers are Ste ! Ste. In example, for Stefan number Ste 0:0289, in case
of a strict solution, the solidication constant is K 0:12, and in what follows,
p
equation dening the solidied layer thickness in relation to time is ~d 1:412 s,
whereas in the case of an approximate, quasi-stationary method, the presented
p
above development of the solidied layer is dened by the relation: ~d 1:414 s.
As it can be seen, the compared formulae are highly compatible.
The above comparison of the methods justies the use of the quasi-stationary
method. This method is comfortable in practical applications to solidication
problems because of its useful mathematic formulae.
An analytic solution of one-dimensional liquid solidication in consideration of
changes in the solidied layer heat capacity, near to a cold wall of constant tem-
perature and perfect thermal contact with the solidifying liquid, were presented by
Saviano and Singel (1969) who applied iterative calculation technique. Their results
showing the temperature distribution in the solidied layer and the development of
the solidied layer thickness are compatible with those presented above.
The above comparison of different methods of solidication time assessment
justies the application of the approximate quasi-stationary method presented in the
following chapters. This method, thanks to its simple mathematic description, is
very comfortable in practically solving one-dimensional problems of solidication.
Chapter 4
Solidication on a Rectangular Geometrics
4.1 The Immovable Layer of Liquid Solidication
Examples of theoretical solutions of the immovable layer of liquid solidication are

included among others in the work of Viskanta (1983). The discussed area may be a
component of some more complex solidication systems.
As it has been mentioned, thermal conductivity of PCM is very poor. That is
why the research on optimal constructional solutions of heat accumulators based on
phase transition should aim at increasing their thermal output. To compensate the
mentioned disadvantage, a PCM of both proper geometry and relatively large
external surface of heat release should be selected.
Search for optimal shapes of heat accumulators should be one of the main tasks
of design engineers. One of particular solutions is the accumulator of relatively
strong thermal power (an example of this accumulator is presented in Fig. 4.1),
composed of very thin rectangular prisms of width H and large lateral surface l b.
Every element of PCM is placed in a housing, mostly plastic or metal one to ensure
invariability of its shape after the change in a solid into a liquid in the process of
melting. In a gap of width s between those elements flows a cooling liquid which
absorbs the PCM latent heat. The conception project as well as the application of
the rectangular heat accumulator was presented in the work (Lipnicki and
Rubaszewski 2000). They proposed to use the discussed accumulator in a car with a
combustion engine to enable fast heating of the power supply in the moment of cold
start. In this work, a similar model with the contact layer thermal resistance
included will be discussed.
A theoretical model of the phenomenon occurring in an element of the accu-
mulator constructed of rectangular parts is presented in Fig. 4.2.
A liquid PCM placed inside a rectangular element undergoes solidication at
constant temperature TF thanks to the heat absorbing, cooling liquid of constant
temperature T0 and of constant coefcient of heat absorption h flowing on the
outside of the element. A contact layer of heat transfer coefcient hCON is formed

DOI 10.1007/978-3-319-53432-9_4
26 4 Solidication on a Rectangular Geometrics
Fig. 4.1 Example of a s H

rectangular heat accumulator
construction
cooling liquid
l
insulation
PCM element xb
Fig. 4.2 A theoretical model

of the phenomenon
between the walls of temperature TW and the solidied layer. Opposite to wall, the
surface of the solidied layer of temperature T indicates the border transition layer.
Position of the solidication front moving from the cold wall towards the inside of
the liquid material denes coordinate d in relation to time t at the same time being
the solidied layer thickness. What is more, the solidifying PCM is characterized by
4.1 The Immovable Layer of Liquid Solidication 27
constant values such as solidication heat L, density qs , specic heat c, thermal

conductivity ks and thermal diffusion js .
The heat stream flowing from the solidifying PCM to flowing outside cooling
liquid is dened by the following equation
dd ks
q_ qs L TF T hCON T TW hTW To : 4:1
dt d
After transformation, the formula denes the surface temperature of the layer
solidied in the contact layer
q L dd
T TW S : 4:2
aCON dt
Further transformation of the above equations allows to obtain the differential

equation dening the heat exchange in non-dimensional form
~
1 d~d2 1 1 dd
1; 4:3
2 ds Bi BiCON ds
where the following values were used: thickness, Stefan number, Fourier number,
thermal resistance of the heat transfer from a PCM element to cooling liquid and
thermal resistance of the contact layer, adequately expressed as
~ d c TF T o js t 1 ks 1 ks
d ; s Ste Fo; Ste ; Fo 2 ; ; :
H L H Bi hH BiCON hCON H
The cases of thermal resistance mentioned above are converses of appropriate

Biot numbers, Bi and BiCON .
Solution of Eq. (4.3) is function

1 ~2 1 1 ~d s 0;
d 4:4
2 Bi BiCON
which satises the initial condition
~d 0; s 0: 4:5
Hence, the equations properly dene the solidied layer thickness
s

2
~ 1 1 1 1
d 2s 4:6
Bi BiCON Bi BiCON
(a) (b)
Fig. 4.3 Development of the solidied layer thickness a layer thickness ~ d and solidication
velocity d~d=ds; b layer thickness d and heat flux q,
_ for C28H58; H = 0.02 m; T0 = 20 C
and the solidication velocity
d~d 1

r 2 : 4:7
ds 1
Bi 1
BiCON 2s
The above analytic solution for a single PCM element is graphically presented in
Fig. 4.3a, but Fig. 4.3b shows some examples of calculations for n-oktasan PCM. It
has been accepted that the thermal resistance of the heat transfer between the liquid
and the guard of the PCM housing is 1=Bi 0:2. The graphs were done for
different thermal resistance of the contact layer 1=BiCON .
As it results from the above solution, the thickness of the solidied layer ~d
increases with time, but solidication velocity d~d=ds decreases. A decrease in time
of solidication velocity and simultaneously of the stream of generated heat q_ is
caused by the increase of thermal resistance of the solidied layer formed during the
solidication process. The impact of thermal resistance of the contact layer 1=BiCON
is also considerable, and as a parameter, it plays a signicant role in the solidi-
cation process. Together with the increase of the thermal resistance of the contact
layer, the velocity of the liquid solidication decreases.
4.2 Forced Convection Effects
In the proper conditions of heat exchange on the plate surface, a flowing liquid
solidication may occur. Then, the outer surface of the solidied layer becomes a
new contact surface between the flowing liquid and the plate. Near the solidied
surface at the same time being the solidication front, two border layers also occur
in the flowing liquid: the hydrodynamic and the thermal one. The research on the
impact on the solidication process of those two layers which are formed on the
border of the solidied liquid is conducted from both the theoretical and practical
4.2 Forced Convection Effects 29
points of view. In practice, this problem occurs in the production of printed elec-
tronic circuits where a wave of liquid metal, called solder for a very short period of
time, contacts with a cold plate usually made of other alloy. This problem also
occurs in practical application of the alloy casting technology based on the solid-
ication of a liquid alloy flowing in channels. In some types of cooling installations,
the phenomenon of a flowing liquid solidication can also occur, which is disad-
vantageous in this case.
The development of border layers on a cold plate was investigated in the past
(oc 1962; Cebeci and Bradshaw 1984; Lienhard and Lienhard 2004;
Kays et al. 2005). The rst problem of a flow in the proximity of a wall was solved
by Prandtl in 1904. Prosnak (1970) divided the space of flow into two subsections.
The subsection directly contacting the plate, in which the forces of inner friction are
of considerable importance and of great impact on the flow of the liquid he named
the boundary layer. But the flow through the second subsection was considered to
be inviscid. The theory of the boundary layer based on the Prandtl model had a
very important place in the development of the fluid mechanics and in engineering.
There are many scientic works (Lapadula and Muller 1966; Saviano and Siegel
1969; Epstein 1976; Viskanta 1983) dealing with the problem of a flowing liquid
solidication on the cold plate. They in majority concern the solidication on a cold
isothermal plate and on cold surfaces of small heat capacity. The solidication of a
liquid flowing on the cold surface was theoretically researched by Epstein (1976).
In his paper, he investigated the problem of non-stationary development of the
solidied layer of a liquid flowing on cold surfaces and his research was that of the
analytical kind. Saviano (1969) solved the problem of solidication of a liquid on a
cold isothermal plate analytically. But the influence of the hydraulic and the thermal
boundary layers in the solidication process has been less analysed.
The thermal resistance of the contact layer in the process of a liquid solidication
on a cold plate was by some authors theoretically investigated, among others by
Ignaszak et al. (2002), Lipnicki (2003), Lipnicki et al. (2005), Lipnicki and Bydaek
(2005, 2008). Some interesting experiments connected with the measurement of the
heat flowing through a contact layer were conducted by Wang and Matthyes (1994)
and Loulou et al. (1998, 1999).
In this chapter, two important problems mentioned above are theoretically
analysed, i.e. the effect of the boundary layer on the solidication process (Sect. 4.
2.1) and the solidication of a liquid in contact with a plate of variable temperature
(Sect. 4.2.2).
4.2.1 The Effect of the Boundary Layer on the Solidication

Process
The effect of the boundary layer on the solidication process was analysed by
Lipnicki and Weigand (2011). Presently, a model of the solidication on a
rectangular plate with a perfect thermal contact between the plate and the solidi-
fying liquid is investigated.
A liquid of density q, kinematic viscosity m, heat diffusion coefcient j and
specic heat cp flow above the cold plate with velocity u1 and temperature T1
(Fig. 4.4). In the result of the solidication process, a solidied layer of density d is
formed above the plate of length l and width b. The coordinate system x, y with
level coordinate x is placed on the plate surface, and its beginning coincides with
the beginning of the plate. The temperature of the plate surface TW is lower than the
solidication temperature TF of the liquid flowing above. The flowing liquid and the
cold plate were separated from each other by the solidied layer. In the result of the
solidication, the interface d moves at velocity @d=@t, depending on both the time
t and location on the plate.
Between the solid body and the stream of the liquid a hydraulic boundary dl and
a thermal boundary dt are formed. The beginnings of the boundary layers coincide
with the beginning of the plate. Depending on Prandtl number Pr, the hydrody-
namic boundary layer can be thicker or thinner than the thermal boundary layer.
Following the rst law of thermodynamics for quasi-stationary conditions and
for the two-dimensional closed control surface presented in Fig. 4.4, the heat bal-
ance equation is
q_ 1 xdt xb q_ 2 x dxdt x dxb q_ 3 dxb qdxb

_ 4:8
There are no heat sources inside the space considered. A vertical heat stream
perpendicular to the investigated surface does not flow, q_ 3 0, because this surface
is an adiabatic one (without the temperature vertical gradient, T T1 ). Still the
horizontal heat streams q_ 1 ; q_ 2 are different than 0. The heat stream q_ flowing from
the investigated surface depends on the heat of the liquid transition into the solid
L and flows on to the cold plate through the solidied layer of heat conductivity ks :

@d @T TF TW
q_ qs L ks ks ;
@t @y d
4:9
TF TW @d
q_ ks qs L :
d @t
Fig. 4.4 Solidication of the y t

flowing liquid
liquid on the cold plate u q&3
(Lipnicki and Weigand 2011) T u
l
q&1 q&2
d 0 u x
TF
q& Tw cold plate
x dx
solid layer
l
A. The heat transfer within the flow space

The equation of energy conservation which denes the quasi-stationary condi-
tions of the heat flow within the flow space was reduced to the following equation
(Lienhrd and Lienhard 2004)
Zdt Zdt " #

@T @T @T @T
ux dy uy dy jL : 4:10
@x @y @y ydt @y y0
0 0
By applying the equation of mass conservation, the above can be reduced to the
following
Zdt
@ @T
uT T1 dy jL : 4:11
@x @y y0
0
Within the boundary layers, hydrodynamic dl and thermal dt , the following

assumptions are fullled
@u @u ddl @T @T ddt
ii ; hh 1 and ii ; hh 1:
@y @x dx @y @x dx
Thus, formulated assumptions express the change in parameter value of the

boundary layer and allow to apply a simplied analysis. The eld of velocity u and
temperature T can be approximated by fourth grade polynomials according to the
model accepted by Loitiaski (1962). In that model, the following assumptions of
both the velocity and
temperature were accepted: @u=@y 0 and u u1 for
y dl ; u 0 and @ 2 u @y2 0 for y 0; T T1 and @T=@y 0 for y dt ; and

T TF and @ 2 T @y2 0 for y 0. By using the above equation, an approximate

prole of temperature
" 3 4 #
y y y
T TF T1 TF 2 2 TF T1 TF h; 4:12
dt dt dt
and that of velocity were obtained

" 3 4 #
y y y
u u1 2 2 u1 ~u; 4:13
dl dl dl
where h and ~
u are presented by expressions in square brackets.
In the following investigations, the above equations were used to dene the
relation of the thickness of boundary layers, the hydrodynamic and thermal ones
according to Loitiaski approximation (1962)
~dt
r Pr 1=3 ; 4:14
~dl
where the dimensionless thicknesses of boundary layers are denominated by

~
dt dt =l, ~
dl dl =l.
By substituting Eqs. (4.12) and (4.13) into Eq. (4.11), the equation dening the
thickness of the thermal layer is obtained
~
~dt @ dt 2
; 4:15
@~x Re Pr f r
In the above equation, Prandtl number, Reynolds number and the dimensionless
coordinate were dened as follows
m u1 l x
Pr ; Re ; ~x : 4:16
jL m l
If the thermal boundary layer is thicker than the hydraulic one dt [ dl ; r i1,
function f(r) is equal to (Loitiaski 1962)
1 3 2 3 3
f r ; 4:17
180r 5 140r 4 15r 2 10r 10
For the opposite case dt \dl ; r h1, function f(r) takes the following values.
1 4 3 3 2
f r r r r: 4:18
180 140 15
After solving the above simple differential Eq. (4.15) with boundary condition
dt 0 for ~x 0, the following equation dening the thickness of the thermal
boundary layer is obtained
p
2 ~x
~dt p : 4:19
Re Pr f r
The above equation clearly shows that the thickness of the thermal boundary
layer depends on parameter r determining function f(r). By using relation (2.12), the
heat transfer coefcient can be calculated out of the equation

q_ kL @T 2kL
h x : 4:20
TF T1 TF T1 @y y0 dt
whereas the local and medium Nusselt number are dened by the following
equations
p Z1
h xl 2 Re Pr f r p
Nux p and Nu Nux d~x 2 Re Pr f r :
kL ~dt ~x
0
4:21
The relation of parameter r and Nusselt number Nu to Prandtl number Pr was

shown in Fig. 4.5.
B. The analysis of the solidication space
An average thickness of the solidied layer can be estimated by applying the
average value of Nusselt number from the equation of heat balance (4.21) on the

border of phases. The average thickness of the solidied layer ~d is a function of
time. It allows to dene the relation of the solidied layer thickness to other
thermodynamic parameters and can be an important parameter for the designers of
systems where the solidication phenomenon occurs. Interesting is also the relation
of the solidied layer thickness to its location on a cold plate.
After transforming Eq. (4.9), the next one is obtained

# @ ~d
Nu ; 4:22
~d @s
Fig. 4.5 Relation of Nusselt

p Nu/ Re ; r

number Nu Re and 3
parameter r to Prandtl number
(Lipnicki and Weigand 2011)
2
N u / Re
1
Pr
0
0 1 2 3 4 5

where s is the dimensionless time, ~d the dimensionless average thickness of the
solidied layer, # the temperature distribution and Ste the Stefan number. The listed
dimensionless values are dened as follows:
d TF TW c T1 T F
s Fo Ste; ~d ; # ; Ste : 4:23
l T1 TF L
Nusselt number Nu is the parameter connecting two spaces: that of flow and that
of solidication. By dividing the variables in Eq. (4.22), one gets equation
~ ~
d dd
ds ; 4:24
# Nu ~d
and its solution

1 # ~
s # ln Nu d ; 4:25
Nu
2
# Nu ~d
satisfying the initial condition
~
d0 for s 0: 4:26
The application of Eq. (4.25) allows to dene the development velocity of the
solidied layer thickness by calculating the thickness derivative of the solidied
layer

@ ~d # Nu ~ds
~ ; 4:27
@s d s
When analysing the above equation, it can be seen that for time equal to 0,
velocity of the solidied layer approaches innity

@ ~d
! 1: 4:28
@s ~
d!0
The relations of both the thickness of the solidied layer and solidication
velocity on a cold plate to time are presented in Figs. 4.6 and 4.7 for various external
conditions. The thickness of the solidied layer depends on time s, parameter #,
Prandtl number Pr, and Reynolds number Re. Based on the obtained solution, it can
be seen that in the initial phase the thickness of the solidied layer increases very
rapidly to nally reach the assumed one. It occurs because of some deterioration of
the heat flow through the solidied layer caused by the increase of its thickness, i.e.
of its thermal resistance. The analysis of the above gure also shows that with
(a) (b)
~
~

0.2 0.2
0.16 0.16
0.12 0.12
0.08 0.08
0.04 0.04
0 0
0 0.02 0.04 0.06 0.08 0.1 0.02 0.04 0.06 0.08 0.1
Fig. 4.6 Thickness of the solidied layer in time function for a Pr = 1, # 0:5, and b Re = 100,
# 0:5 (Lipnicki and Weigand 2011)
(a) (b)
~ ~
d ~ d d
0.08 400 d
50
0.06 300 40
30
0.04 200
~
20
~
d d
0.02 100 10
0
0 0 0.00004 0.00008
0 0.005 0.01 0.015 0.02
Fig. 4.7 The development of solidied layer for a Pr = 0.1, # 0:5, and b Re = 2000, # 0:5
an increase of Reynolds number the nal thickness of the solidied layer decreases,
whereas the thickness of the solidied layer decreases at small Prandtl number
(Fig. 4.6b). The solidication velocity (Fig. 4.7b) increases at small Prandtl number.
Small Prandtl number dene the liquids considered as good conductors, liquid
metals among them.
C. The boundary case of solidication
Since the surface of the solidied layer is assumed adiabetic, there is not a
thermal contact layer q_ 0; r 0; ~dt 0 in the flowing liquid. The liquid
flowing above an adiabatic surface is not overheated, and its temperature equals that
of solidication TF. A model of this phenomenon was presented (Fig. 2.8). In the
discussed case, the heat balance Eq. (4.22) is a differential one
# @ d~
0; 4:29
~d @s
(b)
(a) ~
t 0
u 1.0
T = TF
flowing liquid 0.8
frozen layer adiabatic surface
TF
0.6
0.4
Re = 100; Pr = 1
cold plate 0.2
0

0 0.05 0.1 0.05 0.2
Fig. 4.8 The flow of not overheated liquid a a model of solidication and b two cases of
solidication for # 3 (Lipnicki and Weigand 2011)
Solution is the following relation satisfying the initial condition (4.26)

p p
~d 2# s: 4:30
The obtained solution in comparison with another case was shown in Fig. 4.8.
The increase of the solidied layer thickness in time is higher than that in the
chosen case (Re 100; Pr 1) with outer heat transfer. The convectional heat
transfer between the cold plate and the flowing liquid brings about some decrease in
solidication velocity.
D. Thickness of the layer solidied in steady state and depending on time
conditions
In the last chapter, the medium thickness of a solidied layer was presented in
relation to parameters Re, Pr and #; here, the results of investigations of local
thickness of the solidied layer are presented as the function of time and location.
Hence, the medium value of Nusselt number Nu was replaced in Eq. (4.25) by local
value of Nusselt number Nux . Thus, the analysis resulting from the approximate
solution of the thickness of a solidied layer remains the same. Hence, the equation
for the local thickness of the solidied layer is ~d~x; s.

1 # ~d :
s # ln Nu 4:31
# Nux ~d
x
Nu2x
Equation (4.31) is formally the same, and in Eq. (4.25), only Nu is replaced by
Nux . This equation satises initial condition ~d~x; 0 0 as well as boundary con-
dition ~
d0; s 0, at the beginning of the solidied layer ~x 0, because at ~x ! 0,
local Nusselt number is Nux ! 1, which results from Eq. (4.21). Distribution of
thickness of the layer solidied in steady-state conditions can be obtained from
Eq. (4.31) by substituting @ ~d=@s ! 0. Thus, the local thickness of the layer
solidied in steady-state conditions is described by equation
p
# ~x
~dS p : 4:32
Re Pr f r
The above equation indirectly describes the heat flow through a control volume
of width d~x.
Figure 4.9a, b show the thicknesses of the layer solidied on a plate. Figure 4.10
shows the development of thickness ~d of the solidied layer as the function of
location on the plate ~x for different times from the beginning of the solidication
until the steady-state conditions were reached.
E. Conclusions
Solidication of a flowing liquid on a cold plate depends on both the type of
cooling and thermal resistance occurring between the flowing liquid and the cold
(a) (b)
S S
0.2 0.2
0.16 0.16
0.12 0.12
=1 = 0.8
0.08 0.08
=0. 5
0.04 0.04
0 x 0
= 0.1
x
0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0
Fig. 4.9 Thickness of a layer solidied in the steady-state conditions a Pr = 0.5 and b Re = 2000;
Pr = 0.5 (Lipnicki and Weigand 2011)
Fig. 4.10 Thicknesses of a

layer solidied on a cold plate 0.1
for Pr = 0.5 and Re = 2000
0.08
0.06
steady state =0.01
0.04
=0.013
0.02 =0.007 =0.004
0
=0.001 x
0 0.2 0.4 0.6 0.8 1.0
plate. This resistance is related to actual thickness of the solidied layer in the
analysed time. Convectional heat transfer from the liquid to the solidied layer
depends on the development of two boundary contact layers, hydrodynamic
thickness of the boundary contact layer dl and thermal thickness of the boundary
layer dt , and on their relation to one another. Parameter r denes the relation of the
layer thickness to Prandtl number. Reynolds number is of a considerable impact on
the solidication velocity. In the case of big Reynolds number, the solidication
process is slower. It results from the fact that heat transfer from the flowing liquid to
the solidied layer is slower for big Reynolds number. At a small Prandtl number,
there is a rise of increase of the solidied layer. A considerable increase of the
solidied layer thickness in time can be noticed. Temperature coefcient # influ-
ences the thickness increase of the solidied layer which is proportional to that
coefcient.
A similar analysis should be continued for turbulent flows of liquid solidifying
above a cold plate.
4.2.2 Solidication on a Plate with Variable Temperature
The results of analytical research on solidication of heat conducting liquid metal

flowing near a cold plate with variable temperature are presented (Lipnicki 2003).
Solidication of liquid metal flowing on a cold plate is important for practical
designs. Thickness of a solidied metal layer is especially important in the pro-
duction technology of electronic circuits based on printed circuit boards. In general,
the production of such elements consists in coating cold copper plates with a layer
of tin. A contact layer created between the solidied layer and the cold plate and its
role in the heat flow from the solidifying metal to the cold plate is the subject of
present considerations. The research was concentrated on setting the conditions and
theoretical relations concerning the creation of a layer solidied in this process.
It was assumed that the temperature within the cold plate depends on both
location and time.
A. Solidication on a cold plate
Figure 4.11 presents a picture of the contact layer of a cold copper plate and of
solidied tin which were observed by means of X-rays. The discussed contact layer
was formed in the result of flowing liquid solidication Sn (liquid tin) above a cold
plate Cu (copper plate). On the tin/copper contact occurs tin/copper solution (white
spots). Copper atoms are represented by red spots. The structure of this layer is an
example of connecting one element with another as a solution. On the discussed
area, there also occurs additional resistance of heat transfer from the warmer to the
colder place.
The knowledge of thermal resistance in the contact layer is important from practical
point of view. It is applicable in metals welding and soldering. Therefore there is need
for new theoretical and experimental investigation of this phenomenon. In the
Fig. 4.11 Picture of the 2,5 m

contact layer in X-rays
(Lipnicki 2003)
Sn
Cu
pure tin solid solution of tin in copper
following part of this work, a theoretical analysis of solidication of a liquid flowing

above a cold plate with variable temperature is presented.
The flowing liquid (Fig. 4.12) is assumed to be overheated, and its temperature
is equal to that of solidication TF. The theoretical analysis aims to dene the
thickness of the solidied layer d formed on the plate and the velocity of its
development dd=dt, as well as the eld of temperature T in a time function,
depending on the cooling conditions dened by the features of both the cold plate
and liquid metal. The temperature TW of the cold surface is related to time
t. Between the solidied layer and the cold plate surface, a contact layer is formed
causing additional thermal resistance of flow. In the transition layer, a difference in
temperatures appears DT T TW , where T is the temperature of the solidied
layer opposite to the cold plate surface. The heat flow through the contact layer is
dened by the coefcient of heat transfer hCON and the heat flow between the
solidifying liquid and the surface of the solidied layer by coefcient of heat
transfer h. The outer surface of the cold plate is an adiabatic one.
Fig. 4.12 A theoretical contact layer

model (Lipnicki 2003)
d H
TF cold
plate
frozen T
layer
T
TW adiabatic
liquid . surface
q
0 x
Rys. 2.12
In the result of solidication, a solidication front (interface) moves from the

boundary surface of the cold plate into the liquid at the velocity of d~d=ds. On the
liquid/solid interface, solidication heat is emitted. The stream of heat q released
from the solidication front is conducted through the solidied layer to the cold
plate. The cold plate accumulates the emitted heat which causes an increase of its
temperature.
The solidication front is sharp and rectangular. In addition, it is assumed that
the heat accumulation by the solidied layer is very small in comparison with a cold
plate. All thermodynamic parameters are considered constant.
The energy balance describing the flow of heat from the flowing and solidifying
liquid to the cold plate is dened by the following equations (Lipnicki 2003)
dd TF T TF T
qs L t h T 1 T F k s ; ks hCON T TW : 4:33
d d d
The temperature of a solidied layer is obtained from the above equation
L dd h
T TW qs t T1 TF :
hCON d hCON
To facilitate the analysis of the problem, a set of dimensionless was introduced:

time, thickness of the solidied layer, coordinate axis, temperature and the wall
temperature, appropriately dened by
d x T T0 TW T0
s Fo Ste; ~d ; ~x ; h ; hW ;
H H T F T0 TF T0
Dimensionless Fourier, Stefan and Biot numbers
t js c s TF T 0 H hCON
Fo ; Ste ; BiCON ;
H2 L ks
and two dimensionless parameters were introduced
jC ~ kC
~
j ;k :
jS kS
Since the flowing liquid is not overheated, there is not free convection heat flow.
Considering the above, Eq. (4.33) can be written down in the following dimen-
sionless form:
d dd 1 dd
1 hW : 4:34
ds BiCON ds
To solve Eq. (4.34), some former knowledge of temperature hw s is needed.

The mentioned temperature is also necessary to dene boundary conditions
between the plate surface and the solidied layer. These will be dened in the
following analysis of the problem.
Differential equation of thermal conductivity in a cold plate allowing to dene
the eld of temperature h~x; s takes the following form
@h ~ @2h
j
: 4:35
@s Ste @~x2
In the considered case, one edge of the cold plate ~x 0 (Fig. 4.12) is of tem-
perature hw , and the other ~x 1 is thermally insulated, i.e. @h=@~x 0, and the
temperature within the whole volume of the plate in the beginning of the process is
equal to initial temperature h~x; 0 0. Both discussed boundary conditions and
the initial one as well can be written down in a compact form
@h
h~x; 0 0; 1; s 0; h0; s hw : 4:36
@~x
Equations (4.34), (4.35) and (4.36) create conjugated system of differential

equations to be solved together.
B. Solution of solidication on a plate
By exploiting experiences concerning cooling conditions, the relation of a cold
plate surface temperature to parameter s was introduced as
hW 1 expbs; 4:37
where unknown constant b will be dened based on the analysis mentioned below.
By substituting the above temperature function to Eq. (4.34), the following is
obtained
~d d~d 1 d~d
expb s : 4:38
ds BiCON ds
Integration of the above equation from s 0 to s 1 allows to dene

parameter b
Zs ~d2 ~d
expbs ds
2 BiCON
0
Z1 ~d2 ~dmax
) expbs ds max
4:39
2 BiCON
0
1
) b ~2 ~dmax
:
dmax
2 BiCON
In this case, the thickness of the solidied layer varies from ~d 0 to ~d ~dmax ,
and constant b is dened by ~dmax .
By using the equation of total energy balance for a cold plate, maximum
thickness of the cold plate is obtained
qs L H ~dmax cs qs TF T0 H
~k Ste 4:40
) ~dmax :
~
j
Thus, constant b nally takes the form
1 1
b ~2 ~dmax
2 : 4:41
dmax ~kSte ~kSte
1
1
2 BiCON 2 ~
j BiCON ~
j
The thickness of the solidied layers in two boundary cases is calculated as

follows:
In case of absence of thermal resistance of the transition layer 1=BiCON 0 the
formula for Eq. (4.34), is dened by equation
s
d2 1 2
1 expb s ) d 1 expb s; 4:42
2 b b
in case of thermal resistance of the transition layer 1=BiCON [ 0 (see Eq. 4.34),
the thickness of the solidied layer is dened by equation
s
1 1 2
d 1 expb s: 4:43
BiCON BiCON b
2
The solution of equation of thermal conductivity (4.35) in a nite area 0h~xh1, at

initial and boundary conditions (4.36), in time si0 is as follows:
To solve Eq. (4.35), the thermal eld is presented in a form including a new
function w~x; s (Smirnow 1962)
h~x; s 1 ebs w~x; s; 4:44

based on the solutions of similar equations which can be found in the scientic
literature (Carslaw and Jaeger 1959; Smirnow 1962). After supplementing the
above expression into the main equation, a new differential equation is obtained
@w ~ @2w
j
b ebs ; 4:45
@s Ste @~x2
with new, more simpler boundary conditions
@w
w0; s 0 and w~x; 0 0; 1; s 0: 4:46
@~x
The following form of a solution of Eq. (4.45) is postulated
X
1 j~ 2
w~x; s An eStekn s ebs sinkn ~x: 4:47
n1
It is subsequently checked whether the postulated function w~x; s satises the

boundary and initial conditions. As it can be easily noticed, the above equation
satises the initial condition; the second, boundary one will be satised for the set
of parameters
p
kn p n 1:
2
Constant An is dened by equation
X
1
~
j
An b kn2 sinkn ~x b;
n1
Ste
which was obtained from Eqs. (4.45) and (4.47). Integrating the n-term of the above
series within the limits from 0 to 1, according to formula
Z1 Z1
~
j
A n b kn
2
sin kn ~x d~x b sinkn~x d~x;
2
Ste
0 0
A relation dening constant An was obtained, equal to
2b
An ~
j
:
kn kn2 Ste b
Finally, the exact analytic solution of Eq. (4.35) takes the form of innite series
(Lipnicki 2003)
tin copper
F 1
= 0.3 0.2
= 0.3
= 0.25 = 0.25
= 0.2 = 0.15 0.15
= 0.2 = 0.1
= 0.15 0.1
= 0.1 0.05
solidification front 0 0.2 0.4 0.6 0.8 1 x

Fig. 4.13 Effect of time-s on the thickness of the solidied layer-d and on the temperature
distribution h for a perfect contact between the cold plate and the solidied layer 1=BiCON 0
(Lipnicki 2003)
X1
1 1 h j~ 2 i
h~x; s 1 ebs 2 b e Stekn s
e bs
sinkn ~x: 4:48
k j~ 2
n1 n Ste kn b
Solutions of Eqs. (4.41), (4.42), (4.43) and (4.48) are graphically presented in
gures from (4.13) to (4.16).
Figure 4.13 presents the position of the solidication front and the temperature
distribution within the plate for 1=BiCON 0. The distribution of temperature is
monotonic for different times.
The impact of the thermal resistance of the transition layer on the development
of the solidied layer is evident. At a smaller thermal resistance of the contact layer,
the thickness of the solidied layer increases faster with time (see Fig. 4.14).
Figures 4.15 and 4.16 present a comparison of anticipated proles of temperature
distribution for two different conditions.
C. Conclusions
The presented theoretical model analytically describes non-stationary solidi-
cation of a liquid on the cold plate. Solution regarding an increase of thickness of
the layer solidied on the cold plate and non-stationary eld of temperature inside
the plate was obtained (Figs.
. 4.14, 4.15 and 4.16). The theoretical model depends
~
on three parameters: kSte j ~, j
~=Ste and 1=BiCON . The rst is equal to maximum
thickness of the solidied layer dmax , the second represents thermodynamic features
of the two contacting bodies, liquid and plate, whereas the third parameter denes
thermal resistance of the contact layer related to the contact features. There is also
the fourth parameter b dening the conditions of the plate cooling. The results
reached based on the proposed theoretical model comply with those obtained by the
method of computer simulation, Calcosoft-2 (Ignaszak et al. 2002). Thermal
resistance of the contact layer between the cold plate and the solidied layer during
Fig. 4.14 Layer thickness for

different thermal resistances
of a tin/copper contact layer 1
(Lipnicki 2003)
0.8
1/BiCON = 0 1/BiCON = 0.2
0.6
0.4
1/BiCON = 1 1/BiCON = 2
0.2
0
0 0.1 0.2 0.3 0.4 0.5
Fig. 4.15 Comparison of

temperature distribution in a
copper plate for different 0.5
thermal resistances (Lipnicki 1/BiCON = 0.2 = 0.8
0.4 1/BiCON = 0
2003)
0.3
0.2 1/BiCON = 0
1/BiCON = 0.2 = 0.6
= 0.2
0.1 1/BiCON = 0
1/BiCON = 0.2 x
0
0 0.2 0.4 0.6 0.8 1
the solidication process affects distribution of temperature in both the cooling plate
and the solidied layer (see Figs. 4.15 and 4.16) and is very important in the
solidication process.
4.3 Free Convection Effect
Natural convection of a liquid in a closed space occurs when it is affected by mass

forces, and the temperature of walls of the reservoir differs from that of the liquid.
These conditions are satised if the liquid is either heated or cooled. An omni-
present mass force is that of gravitation. Natural convection also called free con-
vection is a phenomenon of a liquid flow from spaces of higher thickness to those of
lower generated by the difference in thicknesses caused by the difference in tem-
peratures. Thus, created circulation process usually increases the heat transfer in
comparison with that through an immovable liquid. The above phenomenon is
often accompanied by liquid solidication. An analysis of solidication of a liquid
flowing in the result of free convection within rectangular space will be conducted
in this chapter.
tin copper
F 1
contact layer
0.2
= 0.2, 1/Bi= 0
= 0.2, 1/Bi= 0 0.15
= 0.2, 1/Bi=0.2 = 0.1, 1/Bi= 0
= 0.2, 1/Bi=0.2 0.1
= 0.1, 1/Bi= 0.2
= 0.1, 1/Bi= 0
0.05
= 0.1, 1/Bi= 0.2
solidification front 0 0.2 0.4 0.6 0.8 1 x
Fig. 4.16 Temperature distribution and development of the solidied layer for different thermal
resistances of the contact layer (Lipnicki 2003)
4.3.1 Quasi-Stationary Convection in a Vertical Channel
The liquid is contained in a rectangular reservoir bounded by two vertical walls of

thickness H, with orientated coordinate system x, y and vertical axis y coinciding
with that of the cold wall (Fig. 4.17). The cold wall of the channel is being kept at
temperature Tw lower than that of liquid solidication. On this wall, a solidied
layer of identical thickness d and vertical surface is formed. Temperature T1 of the
second, opposite wall, is higher than that of solidication. Rectangular solidication
front of solidication temperature TF moves inwards the channel. Between the
walls, a liquid of temperature T, velocity u2 , thickness q, coefcient of heat dif-
fusion j, pressure p, kinematic viscosity m and coefcient of thermal expansion b
flow in conditions of free convection decreasing during the process. The flow width
of the channel is H d. Quantitative development of solidication layer thickness
was not dened in this chapter.
Equations of conservation (1.3 to 1.5) in the considered space (Fig. 4.17) were
veried according to Boussinesq approximation (1903) for quasi-stationary con-
ditions to the following system of equations:
equation of momentum conservation
2
@u1 @u1 1 @p @ u1 @ 2 u1
u1 u2 m ;
@x @y q @x @x2 @y2
2 4:49
@u2 @u2 1 @p @ u2 @ 2 u2
u1 u2 m gbT TF ;
@x @y q @y @x2 @y2
4.3 Free Convection Effect 47
Fig. 4.17 Solidication in a solidification

rectangular channel with free y front g
convection
q
PCM
frozen
layer Tf
T1
TF
warm
T wall
TW
contact u2
layer liquid
u1
cold x
wall
H
equation of energy conservation
2
@T @T @ T @2T
u1 u2 j 2 ; 4:50
@x @y @x2 @y
equation of mass conservation
@u1 @u2
0: 4:51
@x @y
In the accepted flow model, it was assumed that u1 0, what allows to reduce
the above system of equations to a simpler one, i.e.
1 @p
0 ;
q @x
1 @p @ 2 u2
0 m 2 gbT TF ; 4:52
q @y @x
@ T
2
0j 2:
@x
Based on the above equations, it can be stated that pressure does not depend on
coordinate x, and the pressure gradient is @p=@y const.
By introducing dimensionless values
x u2 H T TF H 3 1 @p gbT1 TF H 3
~x ;~
u ;h iP ; Ra Gr Pr ;
H j T1 TF jm q @y mj
system of Eqs. (4.52) can be reduced to a dimensionless form
d 2 ~u
Ra h P;
d~x2 4:53
d2h
0:
d~x2
compressed by temperature h and having three unknowns: ~u; h; P, where function

P is constant, independent of position within the liquid space. To complete this
system of equations, the condition of global mass conservation in a liquid flowing
across the reservoir is also applied
Z1
~ud~x 0: 4:54
~d
Additionally, the system of equations should satisfy the boundary conditions:
~x ~d; ~u 0; h 0 i ~x 1; ~u 0; h 1: 4:55
The system of equations was reduced to that of ordinary differential equations of

quadratic order. The solution in the rst place bases on solving the equation of heat
conduction and supplementing it to that of momentum conservation. The solution
of the equation of heat conduction is presented in linear function:
h C1~x C2 ; 4:56
with integration constants C1 and C2.

After satisfying the boundary conditions and supplementing integration con-
stants to Eq. (4.56), the function dening the eld of temperature is obtained
1 ~d
h ~x : 4:57
1 ~d 1 ~d
By supplementing the received equation to that of momentum conservation and

after double integration, the eld of velocity is obtained:
!
Ra 1 Ra ~d
u
~ ~x 3
P ~x2 C3~x C4 : 4:58
~
6 1d 2 1d ~
By using the boundary conditions of velocity fading on both the wall and the
solidication front (4.55), integration constants C3 and C4 were calculated. Finally,
the eld of velocity in the form of cubic polynomial is obtained.
Ra h i
~
u 2~x3 3 ~d 1 ~x2 1 4~d ~d2 ~x ~d 1 ~d : 4:59
12 1 ~
d
In the above equation also, the relation between Rayleigh Ra number and
constant P
1
P Ra; 4:60
2
calculated out of the condition of global law of mass conservation (4.54) was
applied.
Figure 4.18 shows the velocity eld and liquid temperature in a flat, thin
channel. Theoretically obtained velocity eld ~u in its form is similar to wave
function. Near the solidication front, the liquid flows down, but near the warm
wall, it flows upwards. The velocity eld is directly proportional to the value of
Rayleigh Ra number, and the temperature eld h is linear, running from 0 to 1. As it
can be seen, together with the increase of the solidied layer thickness ~d, the space
of the liquid flow decreases 1 ~d. It can also be noticed that maximum flow
velocity decreases ~u. Within ~d ! 1, the solidication velocity decreases down to
zero.
In the considered model, the liquid flow caused by free convection does not
additionally contribute to heat transfer as vertical gradient of temperature
u Ra
0.01 1
0.006 0.8

0.002 0.6
x
0
- 0.002 0.4
- 0.006 0.2
x
- 0.01 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Fig. 4.18 Field of velocity and temperature in the channel, in the function of solidication
@T=@y 0 is assumed absent. The heat stream flows crosswise from the warm
wall to the interface surface and then, together with the interface stream, follows to
the cold wall. Thus, the free convection liquid flow is considered the deciding one.
That is why in slim reservoirs (much higher than wide), free convection does not
considerably affect the solidication process.
Solidication in a vertical, rectangular channel lled with wet, porous medium
was solved in the work of Lipnicki and Weigand (2008).
4.3.2 Quasi-Stationary Convection in a Horizontal Channel
In the history, one of the rst experimental research on the stability of a flat layer of
liquid heated from the bottom and cooled from above was that of Bernard (1900).
Bernard noticed hexagonal cells periodically reappearing in a liquid in which free
convection flows occurred. The theoretical base to describe this phenomenon was
then set by Rayleigh (1916). He showed that free convection occurs at a specic
vertical boundary gradient of temperature and also referred to Bernards experi-
ments. His research was continued by Jeffreys (1927) and Pellew and Southwell
(1940). This theoretically and practically important problem was followed by many
other authors investigating that eld further (Bisshopp 1966; Edler 1965; Davis
et al. 1984). Also, the influence of a liquid surface tension on the mentioned
convection process known as Benarda-Marangoniegoconvection (Chandracekhar
1961), dominating in a thin layer and hampering the free convection, was dis-
covered. The effect of surface tension decreases in thicker layers of the liquid. Apart
from hexagonal cells, a great amount of other forms of flow was also discovered.
From a practical point of view, the most interesting are investigations concerning
the heat transfer in the liquid layer between two horizontal and stiff rectangular
walls, which in general are reduced to dening Nusselt number.
In this part of the chapter, two cases of solidication of a horizontal and rect-
angular layer of PCM subjected to solidication (Fig. 4.19) are considered. In the
rst case, solidication proceeds from above (a) and in the second from the bottom
(b). The two cases diametrically differ. In the rst one, the free convection plays the
(a)
Tot contact layer (b) g liquid
g cooling liquid frozen layer T1
TW 0
TF z

H
Tf H
z
T0 PCM TF
PCM
0 TW cooling liquid
T1 liquid
frozen layer T frozen layer
ot
Fig. 4.19 Variants of solidication of a rectangular, horizontal liquid layer

domineering role, and in the second, only pure heat conduction, without free
convection, occurs.
The rst case is more difcult from theoretical point of view and was given most
attention. Here, a horizontal layer of a liquid (PCM) cooled from above is con-
sidered. Solidication will cause forming of a rectangular layer solidied in the
upper part that will move downstairs. In the remaining part of the liquid, free
convection may or may not occur. In the result of heat transfer, the liquid density in
lower part is smaller than in the upper one. Heterogeneous distribution of liquid
density in the layer contributes to the forming of convection movement countered
by liquid viscosity. If the vertical gradient of temperature is high enough and the
liquid becomes unstable, free convection occurs. The task of this work is also to
dene the conditions of the loss of liquid stability.
In the rst case, the investigated horizontal liquid layer is of initial width H. The
upper surface of the liquid is cooled by a liquid flowing outside. Because of that, the
wall temperature TW is lower than that of liquid solidication TF , and consequently,
a solidied layer of width d is formed. In the remaining volume of the rectangular
liquid layer of width variable H 0 H d, free convection may appear in the layer
of stable temperatures of the boundary walls: upper TF and lower T1 , and the
temperature of the lower wall is higher than that of the upper one T1 [ TF . The task
thus posed is an attempt to adapt Rayleigh model to the conditions accepted. The
considered model will be treating the set of quasi-stationary Rayleigh models for
layers of diminishing height.
Instability of a liquid layer is caused by the vertical gradient of temperature
f dT=dz exceeding the critical value. It is more convenient to use the dimen-
sionless Rayleigh similarity number
gfbH 04
Ra0 Rad ; 4:61
jm
where g, b, j and m represent acceleration, coefcient of liquid thermal expansion,

coefcient of liquid thermal diffusion and liquid kinematic viscosity.
Rayleigh number where the variable thickness of a layer is a characteristic
feature diminishes along with solidication development Ra0 Rad\Ra. That
shows that the most favourable moment for the start of free convention is the
beginning of solidication.
Equation of mass, momentum and energy conservation in Boussinesq approxi-
mation takes the following forms:
u 0;
div~ 4:62

d~u p
t grad gz bgT T1 ~
k mr2~
u; 4:63
d q1
dT
t ar2 T: 4:64
d
The equilibrium is distributed by introducing slight disturbances of velocity u0 ,

temperature h0 and pressure p0 to Boussinesq approximation (http://home.iitk.ac.in/
*sghorai/NOTES/benard/benard.html)
~ u0 x; y; z; t; T T0 z h0 x; y; z; t and p p0 z p0 :
u ~
For recording convenience signs were omitted in the above system of equations
recorded as dimensionless coordinates which after elimination of small parts of
higher order take the form:
r ~
~u 0; 4:65
@~
~u
r~p Ra Prh~
k PrD~
~u; 4:66
@s
@h
~u3 Dh; 4:67
@s

where ~~
u ~u1 ; ~
u2 ; ~u3 , Ra fgbH 4 jm and Pr m=j.
To mark appearing in Eqs. (4.65), (4.66) and (4.67)
dimensionless coordinates,
the following unites were used: length H, time H j, velocity j=H, temperature
2

difference T1 TF fH and pressure j2 q H 2 .

The above system of equations should satisfy the following boundary conditions
on the surfaces of the layer:
~u1 ~u2 ~u3 0 and h 0 for ~z 0; 1: 4:68
Rayleigh solution to the above presented problem of liquid layer stability was
shown by Chandracekhara (1961). Rayleigh proposed a solution in the following
form
u3 ~x; ~y; ~z; s W ~zf ~x; ~yers ; h~x; ~y; ~z; s H~zf ~x; ~yers ;
~ 4:69
with W ~z and H~z the amplitudes of velocity and temperature, f the periodic
function
2
2 in the
surface
of horizontal wave number a satisfying differential equation
@ @~x @ 2 @~y2 f a2 f 0 and r the proper value.
In the result of the theoretical analysis conducted by Rayleigh (1916), Jeffreys
(1926) for a horizontal layer of liquid bounded by rigid horizontal edges, the lowest
Rayleigh critical number and corresponding wave number were obtained for an
even solution
Ra 1707; 672 1708; a 3:117: 4:70
The above already classical analytical solution is consistent with theoretical

research based on calculus of variations conducted by Pellew and Southwell (940).
2.0
Nusselt number, Nu
1.5
1.0
0.8
1000 1500 2000 2500 3000 3500
Rayleigh number, R
Fig. 4.20 Experiments conducted by Silveston for different liquids: silicone oil AK 350;
silicone oil AK 3; ethanediol; heptane; water (Chandrasekhar 1960)
Obtained theoretical results are in a great part conrmed by experiments conducted

by Silveston (Fig. 4.20).
Figure 4.20 shows experimental research on Nusselt number conducted by
Silveston. The obtained data show that Rayleigh critical number is approximately
1700 (considerable consistence with the theoretical value of 1708 was reached).
Summarizing, when the value of Rayleigh number is over 1700, in the horizontal
layer, free convection occurs. The liquid starts to move from the bottom upwards,
and in the considered models, the phenomenon laminar free convection should be
assumed.
Examples of different convective flows in a layer of liquid heated from the
bottom are presented below (Bisshopp 1960):
cylinder cells
DW ~z 2p 2p
~u1 sin ~x;
a2 ~l ~l
2p
~u2 0; a ;
~l
2p
~u3 W ~z cos ~x;
~l
rectangular cells
DW ~z 2p 2p 2p
~
u1 sin ~x cos ~y;
a 2 ~
lx ~
lx ~ly
s
DW ~z 2p 2p 2p 1 1
~
u2 cos ~x sin ~y; a 2p ;
a 2 ~
ly ~
lx ~ly ~l2 ~l2
x y
2p 2p
~
u3 W ~z cos ~x cos ~y;
~lx ~ly
hexagonal cells
DW ~z 4p 2p 2p
~
u1 p sin p ~x cos ~y;
3a2 ~
3l 3l~ 3~l

DW ~z 4p 2p 2p 2p 4p
~
u2 cos p
~
x 2 cos ~
y sin ~y; a :
3a2 3~l 3l~ 3~l 3~l 3~l

1 2p 2p 4p
u3 W ~z 2 cos p ~x cos ~y cos ~y ;
~
3 3~l 3~l 3~l
Interpretation of the above results allows to dene the conditions of free con-
vection which occur in a rectangular liquid layer during solidication advancing
from above (case I). As it can be seen, solidication is accompanied by free
convection when Rayleigh critical number is exceeded. Together with the devel-
opment of the solidication process, conditions unfriendly to free convection occur.
But when the layer of liquid solidies from the bottom, there is no free convection,
but pure heat conduction takes place (case II).
In particular, interesting are also conclusions of the analysis conducted in regard
to heat transfer on the solid/liquid interface. Thus, it is necessary to calculate
Nusselt number to dene the solidication process.
When the cylindrical receivers of the interface heat are positioned regularly
within the liquid PCM, the results of research on free convection in such geo-
metrical arrangements may be also helpful. The problem of free convection in the
mentioned arrangement has been considerably well discussed on the literature.
A theoretical analysis of free convection within a vertical arrangement of regularly
positioned thin cylinders together with a review of other investigations was pre-
sented in the work of Kowalski and Lipnicki (1989). Research on liquid solidi-
cation in an arrangement of rectangular cold cylinders hexagonally positioned in a
horizontal layer with consideration of free convection was done in the work of
Lipnicki et al. (2014).
4.3.3 The Simplied Model of Convection

with Solidication
Orientation in relation to force of gravitation of a liquid closed within rectangular

space and subjected to solidication is decided on the type or absence of free
convection. The second important factor is the method of both heating and cooling
the liquid layer.
In a rectangular layer with boundary walls cooled or heated, the forming of free
convection is determined by temperature distribution. Within a horizontal layer,
free convection is connected with exceeding critical Rayleigh number. Apart from
the already discussed particular cases of a layer positioned within the gravitation
eld (rectangular and horizontal layer), there is a series of positions depending on
the angle of layer inclination to vertical direction. All such cases discussed in the
work (Cengel 2003) are presented in Table 4.1. The nal effect was the presentation
of medium values of Nusselt number occurring in different geometrical arrange-
ments necessary to solve solidication processes.
The position of channels (Table 4.1) lled with a solidifying liquid affects the
values of Nusselt number and thus indirectly the value of convective heat transfer
coefcient on the surface of solidication. The knowledge of convective heat
transfer coefcient on the surface of solidication allows for a simple description of
the liquid solidication process. As a macroscopic parameter, convective heat
transfer coefcient depends on the position of the channel in relation to vertical
direction (gravitational acceleration ~
g).
4.3.3.1 Solidication in the Channel with Constant Temperature

of the Outer Wall
For the solidication model presented in chapter one, the type of liquid convection
(forced or free) affects the value of convective heat transfer coefcient and thus also
the Nusselt number. Thus, the calculation of Nusselt number is one of the main
tasks when analysing and solving the solidication problem. In the presented
simplied model, the approach to solve this problem is similar.
The heat flow on the border of solid and liquid phases is dened by the equation
of energy conservation for quasi-stationary solidication (Fig. 4.17). To obtain an
approximate solution of the solidied layer thickness d, equation of heat transfer
equilibrium on interface is applied

@TL @Ts @d
kL ks qs L ; 4:71
@x xd @x xd @t
Table 4.1 Free convection in closed rectangular spaces

Medium Nusselt number Nu for free convection in closed spaces
Mac Gregor

and Emery (Kays et al. 2005)
10\H L\40; 1\Pr\2000; 104 \RaH \107
L=H 0:3
1=4
NuH

0:42 RaH Pr
0:012
1\L H\40; 1\Pr\20; 106 \RaH \109

1=3
NuL 0:46 RaL
Berkovski

and Polevikov (after Kays et al. 2005)
2\L H\10; Pr\; 105 ; RaH \1010
NuH 0:22Pr RaH =0:2 Pr 0:28 L=H 1:4
RaL gbT1 jm
T2 H 3
RaH \1708free convection does not occur

NuH 1
RaH \108 Hollands (after Cengel 2003)
h i
NuH 1 1:44 1 1708
RaH Ra18 1
H
If expression is negative, should be equal zero

RaH gbT1 jm
T2 H 3
RaH \105 ; 0\/\700 Hollands (after Cengel 2003)

!
1708 1708sin 181:6
NuH 1 1:44 1 1
RaH cos / RaH cos /
" #
RaH cos /1=3
1
18
If expression is negative, should be equal zero
RaH gbT1 jm
T2 H 3
G. de Vahl Davies R.W. Thomas (after Kumar and Kalam

1991)
1\L H\33; 0:5\Pr\104 ; RaH \2 103

NuH 0:286 Ra0:258
H L=H 0:239 R1 =R2 0:442
RaL gbT1 jm
T2 H 3
with latent heat. Free convection occurs in the gap, and the heat transfer from the
solidied layer to the liquid can be dened by the boundary condition of third type

@TL
kL hTF Tf ; 4:72
@x xd
with h the coefcient of heat transfer between the solidied layer and flowing
liquid.
By taking into consideration the above equation and thermal resistance of the
transition layer, a simplied equation of heat transfer is obtained
dd TF T
qs L hTf TF ks hCON T TW : 4:73
dt d
After transformation, the following was obtained from Eq. (2.62)
q L dd h
T TW s Tf TF 4:74
hCON dt hCON
By introducing the following dimensionless values to Eq. (4.73)
~ d ks TF TW hH hCON H
d ;B ; Nu ; BiCON ; s Ste Fo;
H kL Tf TW kL ks
cp Tf TF jt ks ~ ~ kL
Ste ; Fo 2 ; j k; k ;
L H qs cp ks
differential equation of heat transfer with solidication in dimensionless form was

obtained

1 d~d 1 1 1
1 Nu B Nu : 4:75
~ ds
d BiCON ~d BiCON
In this equation number 1=BiCON represents the thermal resistance of the tran-
sition layer, and number Nu denes the thermal impact of the flowing liquid. By
satisfying the initial condition
~d 1 for s 0; 4:76
solution of differential Eq. (2.64) was obtained in the form of dening both the time
and thickness of the solidied layer

1 1 B Nu BiCON
1
s 1 ln ; ~d
Nu BiCON B Nu BiCON
1
Nu~d
" !#
1 Nu Nu
B 1 exp s : 4:77
Nu BiCON 1 BiCON
1
Analytical solution (4.77) is also presented in a graphic form for different

parameters (Figs. 4.21, 4.22 and 4.23).
The above diagrams show that the thickness of a solidied layer increases with
time, at rst very quickly to reach the established thickness in a long enough period.
It can be explained by the solidied layer thermal resistance more and more
increasing in time. The effect of parameters of the investigated model, i.e. Nusselt
Fig. 4.21 Solidied layer

thickness for Nu 5 and 1
B5 1 / BiCON = 0 .1
0.8
1 /BiCON = 0 .2
0.6
1 / BiC ON = 0 . 3
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1

thickness for 1=BiCON 0:1 1
and B 5
Nu = 5
0.8
Nu = 7
0.6
Nu = 10
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1

thickness for 1=BiCON 0:1
and Nu 5 0.6
B = 10 B=7
0.4
B =5
0.2
0
0 0.02 0.04 0.06 0.08 0.1
number Nu, relation of temperatures B and contact layer thermal resistance 1=BiCON
on development of the solidied layer thickness, is considerable. The increase of
both the thermal resistance of the solidied layer 1=BiCON and of the Nusselt
number Nu causes decrease of the solidication velocity and of the solidied layer
thickness and increased parameter B cause increase of the solidied layer maximum
thickness.
Fig. 4.24 Model of

solidication of PCM cooled
with cooling water
4.3.3.2 Solidication in the Channel with External Convective Cooling
A model of solidication of a PCM cooled with outside flowing liquid of tem-

perature T0 and convective heat transfer coefcient h0 is presented in Fig. 4.24. The
heat flow on the border of the solid and liquid phases as in the earlier case is
described by the equation of energy conservation for quasi-stationary solidication
process. Approximate equation dening the solidied layer thickness d is obtained
by applying equations of heat transfer in quasi-stationary conditions. Simplied
heat balance for the model with a cooling system can be recorded by equation.
dd ks
hTf TF qs L TF T hCON T TW h0 TW T0 4:78
dt d
After transformation of the above equation, the following formulae for tem-
peratures both in the gap and on the wall surface were obtained
h q L dd h q L dd
T TW Tf TF s ; TW T0 Tf TF s :
hCON hCON dt h0 h0 dt
Equation (4.78) after eliminating the calculated above temperatures can be

written in a dimensionless form
" #
1 d~d 1 1 ~ 1 1 1 1 d~d
h 1 hNuk ; 4:79
k ds Nu~k ~d
Nu~ BiCON Bi0 BiCON Bi0 ds
where
Tf TF ~ d c TF T 0 jt hH
h ; d ; s Ste Fo; Ste ; Fo 2 ; Nu ; BiCON
TF T0 H L H kL
hCON H h0 H
; Bi0 :
ks ks
By dividing the variables in the differential Eq. (2.68), the following was
obtained

~dd~d 1 1 d~d
BiCON Bi0
ds
1 hNu~k BiCON Bi0 hNu~k~d
1 1
~ d~
dd Ad~d
; 4:80
1 hNu~kA hNu~k~d 1 hNu~kA hNu~k~d
where to simplify the recording, an additional parameter representing the contact

layer thermal resistance and heat transfer from the PCM-bounded surface to the
cooling liquid was introduced. An analytical solution of Eq. (4.80) is as follows

1 ~d 1 ~ ~
s ln 1 hNuk A d C; 4:81
hNu~k hNu~k
where C is integration constant dened upon the initial condition (4.76). Finally, the
time of solidication is dened by equation
!
~d 1 hNu~k
s 2 ln 1 ~d : 4:82
hNu~k hNu~k 1 hNu~kA
The relation of the solidied layer thickness to time in water-cooling conditions

was presented in a graphic form (Fig. 4.25). When analysing the graph, it can be
Fig. 4.25 Relation between

the solidied layer thickness 0.8
and time
Nu =1, A=0.2
0.6
0.4
Nu =1, A=0.5
Nu =2, A=0.5
0.2

0
0 1 2 3
seen that at the beginning of solidication, the solidied layer thickness consid-
erably increases in time and then decreases to nally reach the established value.
Development of the layer thickness brings about an increase of its thermal resis-
tance up to reaching the value at which the established conditions of heat flow
occur. Relation curves were drawn for different parameters hNu and A. The rst
parameter is the product of overheating and of Nusselt number; the second presents
the sum of the contact layer thermal resistances and convective heat transfer
resistance from the surface of PCM to the cooling liquid. As it can be seen, together
with the increase of parameter hNu when the remaining parameters are stable, the
solidied layer thickness decreases. The effect of overheating parameter h itself or
that of Nusselt Nu number are justied, and their increase reduces the solidied
layer thickness. Wheres the increase of parameter A, i.e. the total increase of both
the thermal resistance of the contact layer and the heat transfer resistance during the
cooling process brings about some decrease of the solidied layer thickness.
As the solidication advances, space containing liquid decreases and the con-
ditions of liquid flow change as well. That causes a change in Nusselt number
depending on actual size of the space in which free convection occurs. For a
variable Nusselt number, Eq. (4.80) cannot be easily analytically integrated and
numerical methods must be applied. By transforming formula (4.80), equation for
the solidied layer time is obtained.

Z~d ~d A d~d
s : 4:83
0
1 hNu ~d ~k A ~d
Numerical calculation of the above integral will not be especially difcult under
the condition where
relation between Nusselt number and the solidied layer
~
thickness Nu d is known.
In the literature, there are numerous papers providing Nusselt number in vertical
rectangular channels depending on their geometric sizes. A sample formula for
Nusselt number according to Berkovski and Polevikow for solidication of vertical,
rectangular layer of width H d and length L (after Cengel 2003) is presented
below
0:28 1=4
Pr L
Nu 0:22 RaHd :dla 2\L=H d\10 i RaHd \1010 ;
0:2 Pr Hd
4:84
where Rayleigh number is dened by equation
g b Tf TF H d3
RaHd :
mj
In the discussed case, Nusselt number depends on the development of solidi-

cation (solidied layer thickness) and that is why it is of limited importance.
Very useful formulae for Nusselt number for free convection both laminar and
turbulent in rectangular channels were presented (Kaysa et al. 2005)
1=3 g b Tf TF H d 3
NuHd 0; 046 RaHd ; RaHd GrHd Pr Pr
m2
4:85
In the case of free turbulent convection as it can be seen when comparing the
formulae below, heat transfer coefcient a does not depend on the advancing
solidication but indirectly on the channels width:
r
NuHd 3 gbTf TF
a kL 0:046kL Pr : 4:86
Hd m2
In consequence, it is natural that Nusselt number in this case referring to a

constant but different characteristic size is also constant. Then, Eq. (4.83) can be
analytically solved.
Chapter 5
Solidication in an Annular Space
5.1 The Solidication of a Non-heated Liquid
The outer geometry of PCMs built of concentric annuli can be also used for heat
storage because of their relatively large outer surface.
A cylindrical heat accumulator consists of a series of non-dimensional con-
centric layers with inner radius Rn1 and outer radius Rn2 increasing proportionally to
their distance from the axis of the accumulator (Fig. 5.1). The layers mentioned
above consist of subjected to phase change PCM material of solidication tem-
perature TF , density qs , latent heat L, thermal conductivity coefcient ks and heat
diffusion coefcient j. Two heat streams, inner q0 and outer q00 , are generated from
each layer. In the annular channels between cylindrical surfaces flows, parallel to
the axis of the channel, a cooling heat absorbing liquid of temperature To , with
inside heat transfer coefcient h0o and outside heat transfer coefcient one h00o .
Between the solidifying material and the inner and outer walls are contact layers of
thermal resistance dened by the following: the inner layer heat transfer coefcient
n0CON and that of the outside layer h00CON . In the beginning, the PCM remains in
liquid state of temperature lower and nearer to that of solidication. Because of
cooling with a liquid flowing between layers of material contained in concentric
cylindrical channels, that material undergoes solidication releasing mainly the
phase-change heat. Solidication fronts from both inside and outside are marked d0
and d00 . The temperature of the cooling liquid is assumed constant. Distribution of
temperature in axial intersection (Fig. 5.1) and in a non-dimensional layer (Fig. 5.2)
was presented.

DOI 10.1007/978-3-319-53432-9_5
64 5 Solidication in an Annular Space
Fig. 5.1 Longitudinal R n2

section of the cylindrical heat R n1
accumulator
PCM
PCM
PCM
PCM
PCM
PCM
TF
TO
TW
cooling frozen
liquid layer
Fig. 5.2 Solidication model R n2

of PCM on non-dimensional
layer R n1
TF
TW TW
TO T T TO
. .
q q
PCM
cooling
liquid
contact frozen
layer layer
The phase-change-generated heat streams flowing from both the inner and the
outer layers are in turn described by equations
dd0 ks
q_ 0 qs Ld0 d0 TF T 0 Rn1 h0CON T 0 TW0 Rn1 h00 TW0 To ;
dt ln Rn1
dd00 ks
q_ 00 qs Ld00 Rn2 TF T 00 Rn2 h00CON T 00 TW00 Rn2 h000 TW00 To :
dt ln d00
5:1
5.1 The Solidication of a Non-heated Liquid 65
After transformation, the following formulae for temperatures on outer surfaces

of the contact layer were received

1 dd0 1 1 1 dd0
TW0 T0 0 qs Ld0 0
T T0 0 0 qs Ld0 ;
Rn1 h0 dt Rn1 hCON hO dt
00 00
1 00 dd 1 1 1 00 dd
TW00 T0 q Ld T 00 T0 q Ld :
Rn2 h000 s
dt Rn2 h00CON h00O s
dt
After supplementing the above calculated temperatures, Equation (5.1) were

reduced to two differential equations in dimensionless form
2
1 dd~0 1 ~0 2
ln d A0 1;
2 ds 2
2 5:2
1 dd~00 1 ~00 2 00
ln d A 1;
2 ds 2

where parameters A0 1 Bi0CON 1 Bi0O ;A00 1 Bi00CON 1 Bi00O express the sum
of thermal resistances of contact layers and heat storage on both sides of the
investigated PCM layer.
In Eq. (5.2), the following variables were used
~ d0 ~0 d00 0 cTF T0 jt jt
d0 ;d ; s Ste Fo0 ; s00 Ste Fo00 ; Ste ; Fo0 2 ; Fo00 2 ;
Rn1 Rn2 L Rn1 Rn2
h0 Rn1 h00 Rn2 h0 Rn1 h00 Rn2
Bi0CON CON ; Bi00CON CON ; Bi00 0 ; Bi000 0 :
ks ks ks ks
Solutions of Eq. (5.2) with initial conditions satised
s0 0 ) d~0 1;
5:3
s00 0 ) d~00 1;
are equations expressing solidication time together with given areas of solutions
validity
1 ~02 ~0 ~02 1
s0 d ln d d 1 A0 ~d02 1 dla 1 d0 ~ d0k Rn1 d0 d0k d00k ;
4 2
1 ~002 1
s00 d00 ~d002 1 A00 ~d002 1 dla 1 ~
d ln ~ d00 ~ d00k Rn2 d00 d00k d0k ;
4 2
5:4
where d0k d00k is the nal interface position of both layers.

Real times of the solidication process are the same for both layers and they are
adequately equal

R2n1 0 R2n1 1 ~0 2 ~0 ~0 2 1
0 ~0 2
t s d ln d d 1 A d 1 ;
aSte aSte 4 2

5:5
2
Rn2 00 2
Rn2 1 ~00 2 ~00 ~00 2 1 00 ~00 2
t s d ln d d 1 A d 1 ;
aSte aSte 4 2
After calculating the solidication velocity by applying Eq. (5.4), equations of

heat streams coming from both the inner and the outer layers are obtained
dd0 dd~0 1
q_ 0 2p qs L d0 2p qs L a Ste d~0 0 2p qs L a Ste ;
dt ds ~0
ln d A0
dd00 dd~00 1
q_ 00 2p qs L d00 2p qs L a Ste d~00 00 2p qs L a Ste :
dt ds ~00
ln d A00
5:6
The total stream of solidication heat generated in the layer is equal to the sum
of heat streams released inward q_ 0 and outside layer q_ 00
q_ q_ 0 q_ 00 : 5:7
Based on a model of solidication, examples of thermodynamic calculations for

PCM non-dimensional layer (Fig. 5.2) were conducted. As the PCM, octacosane
(C28H58) of constant thermophysical parameters was applied:
L 253:3 kJ=kg; TF 62 C; k 0:267 W=mK;

q 900 kg m3 ; c 1:92 kJ=kgK; j 1:55 107 m2 s
In addition, it was accepted that the temperatures of walls on both sides are the
same, the cooling liquid temperature is T0 5 C and the parameter of the total of
contact resistances and heat accumulation is assumed as A0 A00 0:2.
Relations of the position of solidication fronts of the solidied layers and of the
emitted heat streams to time, for non-dimensional element of a cylindrical heat
accumulator, were presented (Fig. 5.3). In the nal moment of solidication, the
solidied surfaces meet and the whole channel is lled up with a solid substance.
Emitted heat streams as it is shown in Fig. 5.3 decrease with time. The reason is the
increase of the solidied layers thermal resistance along the solidication process.
Especially interesting is the beginning of work of a heat accumulator characterized
by a large heat stream. This characteristic of an accumulator can be exploited in
installations where a strong initial stream of heat is necessary. A project of a
5.1 The Solidication of a Non-heated Liquid 67
Fig. 5.3 Thickness of the

,
..
solidied layer and heat q,q
streams within kW/m
non-dimensional layer 1.40
Rn1 0:1 m; Rn2 0:12 m .
q
0.12 1.12
0.84
.
q
0.11 0.56
0.28
t, s
0.10
0 1000 2000 3000 4000
cylindrical accumulator and an analysis of its functioning are presented in the work
(Lipnicki et al. 2001).
The proposed theoretical model can be useful when designing heat accumulators
which consist of annular cylindrical PCM forms.
5.2 Solidication of a Superheated Liquid
A liquid heated above the solidication temperature when flowing in a channel can
be subjected to free convection together with solidication if the temperature of the
channel wall is lower than that of solidication. The solidication of a liquid in a
vertical, annular channel in conditions of free convection is of considerable theo-
retical and practical importance and can be a subject of research on heat
accumulation.
Laminar and turbulent free convections which, depending on outer conditions,
occur in vertical, annular channel were investigated separately from solidication
by many authors. As the main ones in this eld can be considered the works of
Edler (1965) who investigated free convection of liquids in annular channels for
both the laminar (19651) and the turbulent cases (19652) as well as those of
Azouni (1987) who investigated turbulent convection. Those works treated free
convection without phase change.
A great amount of theoretical and experimental research was conducted on free
convection together with solidication but in other geometrical arrangements. Some
review of the literature devoted to this eld can be found in the works (Viskanta
1983, 1988; Davis et al. 1984; Betzel and Beer 1988). Especially interesting may be
considered the paper of Davis et al. (1984) who investigated Bernard convection
with solidication between two horizontal plates. However, that literature is short of
experimental investigation on solidication of a liquid flowing in a vertical, annular

channel where one of the walls is cooled and the other heated.
Research on both the PCM solidication and the effect of free convection on the
process undergoing in vertical, annular channelsseen as particularly interesting
arrangementwas presented in detail in the work of Lipnicki and Weigand (2012).
In this research, considerable attention was also given to the effect of transition
layer thermal resistance on solidication. That is why the results included in the
work of Lipnicki and Weigand (2012) were presented in detail. The presented
research on the problem of solidication in a vertical, annular channel also con-
siders thermal resistance of the contact layer between the inner cold wall and the
solidied layer. An analysis of the importance of transition layers in the solidi-
cation process in general was recently investigated, both experimentally and the-
oretically by Loulou et al. (1998, 1999), Lipnicki (2002) and Lipnicki et al. (2005).
The research shows that the effect of a transition layer on the solidication process
is signicant. For that reason, the impact of the transition layer between the cold
wall and the layer solidied in the solidication process is analyzed in detail.
5.2.1 Theoretical Solution of the Solidication Problem
A diagram of the geometry of a channel with a liquid subjected to solidication and

free convection was presented (Fig. 5.4). The investigations are concentrated
mainly on thickness development of the layer solidied inside the channel. The
liquid flowing inside the channel is of density q, kinematic viscosity m, specic
heat at constant pressure cp and medium temperature Tf . The solidied layer of
thickness d develops on the surface of the inner cylinder of radius R and height H.
The axis z of coordinate system r; z coincides with the symmetry axis of the
cylinders and radius r begins on that symmetry axis. The constant temperature of
the inner cylinder is equal to TW and it is lower than the liquid solidication
Fig. 5.4 Geometry of the

channel (Lipnicki and 1 liquid
g
R
Weigand 2012)
r
Tf
TF
H
TW
T.
q
cold w
wall
contact layer
solid solidification front
5.2 Solidication of a Superheated Liquid 69
temperature TF . There also formed a transition layer between the cylinder surface
and the inner surface of the solidied layer. Across the contact layer, the temper-
ature decreases T TW in the result of solidication where T is the temperature
on the inner surface of the solidied layer. Then, the thickness of the solidied layer
formed on the cold wall increases at velocity @d=@t. Equation of energy balance of
the solidication front d takes the form

@TL @Ts @d
kL ks qs L ; 5:8
@r rd @r rd @t
where L is the liquid solidication heat. Free convection occurs in the liquid lling
up the annular space of w width. Free convection results from the heat transferred
through the interface by assuming heat transfer coefcient h between the solidied
layer and the liquid flowing around, what is described by boundary condition III.

@TL
kL h TF Tf ; 5:9
@r rd
Heat transfer coefcient can be set by means of Nusselt number dened by

empirical equations, for the laminar free convection (see Kumar et al. 1991)
0:238 0:442
H Ro
Nuw 0:286 Ra0:258
w Pr 0:006
; 5:10a
w Ri
and for turbulent free convection (see Kays et al. 2005)
Nuw 0:046Raw1=3 ; 5:10b
where Rayleigh number is the product of Grashof and Prandtl numbers

gb Tf TF w3 gb Tf TF w3
Raw Grw Pr Pr :
m2 mj
By applying Eqs. (5.10a, b), heat transfer coefcient can be dened. In the case
of laminar free convection, the value of heat transfer coefcient varies and depends
on solidication phase, but in the case of turbulent free convection, heat transfer
coefcient is independent from the advancing solidication as expressed in
equation
s

Nuw 3 gb T f T F
h kL 0:046 Pr : 5:11
w m2
In the considered model, the main task was to dene the medium thicknesses of
the solidied layer, both on the perimeter and on the axis. The created model can be
of use for designers of heat accumulators based on the phase-change heat because
the amount of heat released in a simple way depends on the change of solidied
layer thickness in time.
For the solidied layer, the temperature distribution can be calculated for
quasi-established conditions in one-dimensional space out of thermal conductivity
equation

1@ @Ts
r 0; 5:12
r @r @r

r R : T T; r d:T TF :
The assumption of quasi-established conditions is fully justied, since the

considered problem concerns small Stefan numbers. By solving the above
Eq. (5.12), one obtains the distribution of temperatures in the solidied layer and
radial gradient of temperature on the solidication front.

ln ~r @h 1
h ; ; 5:13
ln ~d @~r ~r~d ~d ln ~d
where ~r Rr ; ~
d Rd ; h TT T
F T
.
Then, it is very interesting to calculate the perimeter and axial medium value of
the solidied layer thickness as time function. Special attention was given to the
impact of the contact layer that forms between the inner cold cylinder and the
solidied layer (see Fig. 5.4). The stream of heat flowing through the contact layer
can be approximately dened by equation q_ hCON T TW , where hCON is the
contact layer heat transfer coefcient. The formed contact layer causes additional
thermal resistance in the process. The value of the contact layer heat transfer coef-
cient can be dened only experimentally. This is also discussed in the next chapters
of this work. By using Eqs. (5.12) and (5.13), heat balance equation is obtained
dd TF T 1
qs L h Tf TF ks hCON T TW : 5:14
dt d ln Rd
Out of Eq. (5.14), the following was obtained
q L dd h
T TW s T f TF : 5:15
hCON dt hCON
By introducing non-dimensional variables
~d d ; ks TF TF
B ; Nu
hR
;
R kL Tf TF kL
5:16
hCON R
BiCON ; s SteFo;
ks
into Eq. (5.14), the following differential equation was obtained

d~
d 1 1 1 Nu
1 Nu B : 5:17
ds BiCON ~d ln ~d ~d ln ~d BiCON
In this equation, number BiCON represents the effect of contact layer and number
Nu and that of free convection of a liquid on the heat flow. After satisfying the
initial condition
~d 1 for s 0; 5:18
the time of solidication in relation to thickness of the solidied layer in an integral

form is expressed by equation
Z~d ~ Z ~d
d ln ~d 1=BiCON ~ 1 ~d ln ~d 1=BiCON
s dd d~d:
B Nu d~ ln d~ 1=BiCON Nu B=Nu 1=BiCON ~d ln ~d
1 1
5:19
From the analysis of the above equation, it results that the necessary condition of
solidication is to satisfy the inequality B=Nu [ 1=BiCON .
The method of approximate calculation of the expression (5.19) was presented below.
In the attempt to nd an analytical form of solution, integral (5.19) was transformed to
Z~d Z ~d
~
d ln ~d 1=BiCON 1 B d~d
s d~d 1 ~d ;
B Nu ~
d ln ~d 1=BiCON Nu Nu 2
A ~d ln ~d 5:20
1 1
B 1
A :
Nu BiCON
By supplementing ~d ex , the right side of Eq. (5.20) was reduced to
Z~d Z~d
B dd B dx
2 : 5:21
Nu2 A d ln d Nu Aexx
1 1
By approximating function expx with quadratic polynomial, the following

was obtained
ex 1 0:9312 x 0:2992 x2 : 5:22
This equation approximates solutions within 0\x\1 with accuracy exceeding

1%. By supplementing Eq. (5.22) to Eq. (5.20), an expression was obtained that
can be analytically integrated, and nal formula expressing the solidication time in
relation to the solidied layer thickness takes the form
" p! p#
1 ~
B 1 2a ln ~d b c b c
s 1d p ln p ln b pc ;
Nu Nu2 c 2a ln ~d b c 5:23
c 1 0:3297A 1:8624A; 2
a 0:2992A; b 1 0:9312A
Accuracy of calculating by means of the formula (5.23) is considerably high, and

the estimation of calculating error does not exceed 2%. This can be satisfactorily
used in practical calculation of predicted development of the solidied layer
thickness in time. To obtain some more accurate values that could be compared
with the experimental ones, the following integral formula should be used (5.19).
Development of the solidied layer thickens for chosen parameters that is pre-
sented in Fig. 5.5, and the temperature prole in the solidied layer for two times is
presented in Fig. 5.6.
An analysis of presented graphs (Fig. 5.5) shows that along with an increase of
parameter B, solidication velocity rises but an increase of transition layer thermal
resistance 1/BiCON reduces the solidication velocity.
(a) (b)

1.3 1.3
1.2 1.2
1/BiCON = 0.2 1/BiCON = 0.3 B=7
1.1 1.1
B=4
1/BiCON = 0.4
1.0 1.0
B=3
0.9 0.9

0.8 0.8
0 0.002 0.004 0.006 0.008 0.01 0 0.002 0.004 0.006 0.008 0.01
Fig. 5.5 Development of the solidied layer for Nu = 10, a B = 7 and b 1/BiCON = 0.2 (Lipnicki
and Weigand 2012)
Fig. 5.6 Temperature prole (0.05)

in the solidied layer for
B = 8.18; Nu = 4.29; (0.02)
1/BiCON = 0.22 (Lipnicki and
1.2
Weigand 2012)
1.0
b
0.8
= 0.02
0.6
= 0.05
0.4
0.2
0
r
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
5.2.2 The Beginning of a Solidication Process

Asymptotic Solution
At the beginning of a solidication process, the solidied layer is very thin

(D 1; ~
d ! 1. By neglecting logarithmic terms as small of higher order in
Eq. (5.17), the following was reached
1
ds d~d: 5:24
BBiCON Nu
When accepting BiCON number as constant, the equation can be analytically

solved. In the result, after satisfying the initial condition (5.18), a simple relation of
the layer thickness was obtained
~d 1 BBiCON Nus: 5:25
This equation can be applied when calculating the solidied layer thickness for
very small values of time.
(a)
(b)
d
3 d 1.8
25 1.6
Nu=4.29 Nu=8
2 asymptotic solution 20
1.4
15 Nu=10
d 1.2
1 d 10
5 1.0

0
0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
Fig. 5.7 Development of the solidied layer for a B = 8.18; 1/BiCON = 0.22; Nu = 4.29 and for
b B = 8.18; 1/BiCON = 0.22; Nu = 4.2910 (Lipnicki and Weigand 2012)
Figure 5.7a shows that the difference between accurate and asymptotic solutions
considerably rises with time. Figure 5.7a, b shows an increase of the solidied layer
thickness in time. The velocity of the solidication layer thickness decreases, and
for very big values of time approaches 0. The increase of the solidication velocity
is caused by the increase of thermal resistance of the layer.
5.3 The Experiment
5.3.1 Research Equipment
The equipment used in solidication of the investigated liquid on a cold cylindrical

surface is a copper cylinder.
Distilled water of physical features representative for other materials is the
investigated PCM. Obtained results can be used for the elaboration of general
physical regularities controlling the phenomenon of solidication.
A research stand consists mainly of the central unit, two refrigerating units and
two cycles: for both cooling water and a coolant. A draft of a research stand is
presented in a photograph (Fig. 5.8a, b). Figure 5.9 shows the structure of the
central unit. Photographs of the main component of the central unit are presented in
Fig. 5.10.
The central unit (Figs. 5.8a, b and 5.9) is placed in vertical position. The
investigated space in the shape of vertical annulus is bounded by two coaxial
copper cylinders. The sizes of annular space are as follows: outside diameter of the
channel 2RA 200 mm, inside diameter 2R 134 mm and height H 170 mm.
To make the observation of the process and measurement of the thickness of the ice
layer possible, two transparent, rectangular plates made of plexiglass are xed on
the upper and lower parts of the channel. During the experiment in both cylinders,
in their lower, middle and upper parts, thermocouples for temperature measurement
are installed. The copper cylinder, the outer as well as the inner one, is presented in
Fig. 5.10 in the process of being assembled.
5.3.2 The Course of the Experiment
The water in the annular channel is cooled from inside through the middle cylinder
in which cold glycol is flowing and from outside through the outside cylinder of
constant temperature secured by flowing water. Before the experiment, both water
and glycol were cooled to the required values of temperatures. To enable precise
observation of interface between water and ice, water was tinged with methylene
blue. Additionally, the observation stand was illuminated from below by neon light.
5.3 The Experiment 75
(a)
(b) measuring
cylinder
observation
overflow h
channel
water
g
test section piezometer
Venturi nozzle (central unit)
water ethylene
pump glycol
pump
Venturi nozzle
illumination
water inlet ethylene glycol
water
refrigeration
unit
refrigeration
unit
Fig. 5.8 a A research stand. b A draft of an experimental stand (Lipnicki and Weigand 2012)
pyn chodzcy (glikol etylowy)

observation
bolt with flow holes

thermocouples plexiglass
water flow
. .
water g
test
section inner copper
ice layer cylinder
water flow
ilumination
coolant flow (ethylene glycol)
Fig. 5.9 Central unit (Lipnicki and Weigand 2012)
Fig. 5.10 Components of the annular channel, a two sleeves; b top view (Lipnicki and Weigand
2012)
Photographs of different periods of the experiment were taken from the top of the
stand.
The temperature of the glycol, of both inlet and outlet water as well as of walls of
the cylinder, during the experiment was monitored at some chosen points. Water
temperature was measured with the accuracy of 0.1 C, but glycol temperature with
that of 0.2 C. Since the picture was enlarged, ice thickness was measured with the
accuracy of 0.1 mm. The accuracy of temperature calibration was 0.1 C. To
estimate medium thickness of the ice layer formed on the inside cylinder, a mea-
suring cylinder was applied in which the growth of water volume DV caused by
changes of ice density in the relation to water density was measured. The measuring
cylinder was placed in the upper part of the central unit (Figs. 5.8a, b and 5.9).
Volume changes in the measuring cylinder (Fig. 5.11) were used to calculate
medium thickness of the ice layer according to the equation
s
R R2 qL DV :
D 5:29
qL qS pH
This way of measuring the ice layer thickness is an original method (Lipnicki
and Weigand 2012).
Inside the medium cylinder (inner copper cylinder), the required temperature
was maintained. To measure the flows of both water and glycol, venturi tube was
applied.
Fig. 5.11 Method of medium

ice thickness measurement measuring
cylinder
ice layer
subsection 1
g
water
H subsection 2
subsection 3
R
5.3.3 The Results of the Research and Discussion
Cumulative parameters of particular measurements were listed in Table 5.1. The

research area was divided into three subsections (lower, middle and upper) as it is
shown in Fig. 5.10, and each section covers approximately 1/3 of total height of the
channel. The liquid and wall temperatures of each subsection are listed in Table 5.1.
When observing the solidied layer, it is shown that its thickness is almost
uniform along the whole height of the inner cylinder (white layer, Fig. 5.12).
However, small changes towards the periphery may appear (the dark blue colour is
only reflection of light and of no signicance). The development of the solidied
layer depends on the wall temperature and that of the liquid (information regarding
the mentioned temperatures is included in parameter B) and on the temperature of
the thermal resistance of transition layer 1=BiCON . Based on the conducted mea-
surements, it is shown that temperature changes along the height (Figs. 5.13 and
5.14). For that reason, the vertical, annular cylindrical channel was divided into
three subsections: upper subsection1, middle subsection2 and lower subsection
3, of approximately constant temperature. Changes of the channel walls tem-
perature are height function and that affects the changes of contact layer thermal
resistance along the height of the channel. The propelling force deciding on heat
transfer and
free convection from the solidied layer is the temperature difference
Tf TF . It does not affect the heat transfer changes because the ice layer surface
is solid and its temperature is equal to 0 C. The above-mentioned parameter
1=BiCON in Eq. (5.17) can be set by matching the experimental measurements of the
solidied layer thickness to those estimated theoretically. Values thus obtained are
included in Table 5.1 for particular cases of experimental research. Heat transfer
coefcient h, parameter B, Nusselt number Nuw and Yarleigh number Raw are
calculated on the basis of Eqs. (5.9)(5.11). Thermal resistance parameter 1=BiCON
was calculated by comparing theoretical curves with experimental results
Table 5.1 Main cumulative parameters of experiments (Lipnicki and Weigand 2012)
h i
Exp. no. Subsection TW Tf Ste B h; mW Raw Nuw 1=BiCON
[oC] [oC]
2
K
I 1 3.0 2.0 0.0123 5.84 36.50 812,000 4.291 0.16

2 4.2 2.0 0.0123 8.18 36.50 812,000 4.291 0.22
3 8.7 1.5 0.0092 22.64 36.21 797,000 4.266 0.48
II 1 1.0 1.25 0.0077 3.13 34.80 710,000 4.104 0.055
2 1.5 1.25 0.0077 4.69 34.80 710,000 4.104 0.12
3 4.3 1.25 0.0077 13.44 34.80 710,000 4.104 0.39
III 1 9.4 2.20 0.0135 16.63 36.32 798,000 4.266 0.42
2 9.8 2.20 0.0135 17.34 36.32 798,000 4.266 0.44
3 10.1 2.20 0.0135 17.87 36.32 798,000 4.266 0.46
IV 7.5 2.0 0.0123 14.61 36.50 2,700,000 6.405 0.19
Frozen time 200 s Frozen time 990 s
Fig. 5.12 Pictures of the layer frozen in different periods of time (Lipnicki and Weigand 2012)
experiment I
(see Figs. 5.15, 5.16, 5.17 and 5.18). In addition, it can be noticed that solidication
fronts are similar for every subsection. Stefan number is usually lower than 0:014,
which justies the assumption of a quasi-xed development of the ice layer.
(a) (b)
T, 0 C
mm
7 6 top middle
4
6 down
2
5 outside cylinder
0
4 -2 top
3 -4
internal cylinder middle
2 -6
down
1 -8
t, s t, s
0 -10
0 1000 2000 3000 4000 0 1000 2000 3000 4000
Fig. 5.13 Relation of the solidied layer thickness (a) and temperature (b) to time (exp. I)
(a) (b)
T, 0C
mm
6 4
top, middle, down (out cylinder)
0 top
4
middle
-4
2 down
internal cylinder
-8 t, s
t, s
0
0 1500 3000 4500 6000 0 1500 3000 4500 6000
Fig. 5.14 Relation of the solidied layer thickness (a) and temperature (b) to time (exp. II)
The results of experimental research in comparison with theoretical assumption are

shown (Figs. 5.15, 5.16, 5.17 and 5.18).
Figures 5.13 and 5.14 present the relation of an ice layer medium thickness and
of the temperature in three subsections to time in regard to experiments I and II.
Figures 5.13b and 5.14b show time relation of the temperature on the surface of the
cold internal cylinder and the warm outer one. As it is shown in Fig. 5.13b, the
temperature on the cold internal cylinder (middle and upper) changes for the rst
2000s, and then it is approximately stable.
(a) (b)

1.06 1.06
experiment
1.04 1/BiCON = 0.15 1.04

1/BiCON = 0.22
1/BiCON = 0.16 1/BiCON = 0.25
1.02 1.02
1/BiCON = 0.2 experiment
1 1

0 0.0005 0.001 0.0015 0.002 0 0.0005 0.001 0.0015 0.002
Fig. 5.15 Experimental comparison of data with theoretical assessments: a exp. I subsection 1;
b exp. I subsection 2 (Lipnicki and Weigand 2012)
(a)
(b)

1.08 1.1
1.06 1.08
1/BiCON = 0.05
1.06
1.04 1/BiCON = 0.055
1/BiCON = 0.48 1.04
1.02 experiment
1.02
experiment
1 1

0 0.001 0.002 0 0.001 0.002 0.003
Fig. 5.16 Comparison of experimental and theoretical assessments: a exp. I subsection 3;

b exp. II subsection 1 (Lipnicki and Weigand 2012)
(a) (b)

1.1
1.1
1.08
1.08
1.06
1/BiCON = 0.12 1.06
1.04
1/BiCON = 0.39
1.04
1.02 1.02
experiment experiment
1 1

0 0.001 0.002 0.003 0 0.001 0.002 0.003
Fig. 5.17 Comparison of experimental data with theoretical calculations for: a exp. II subsection
2; b exp. II subsection 3 (Lipnicki and Weigand 2012)
Fig. 5.18 Relation of contact

1/BiCON
thermal resistance to wall
temperature (Lipnicki and 0.5
Wiegand 2012) 0.4
0.3
0.2
0.1
TW , 0C
0
-12 -10 -8 -6 -4 -2 0
Comparisons of theoretic calculations and experimental data for experiments I

and II of dimensionless quantity ~d R D=R were presented (Figs. 5.15, 5.16
and 5.17) as time function. The above gures show the influence of different
1/BiCON on the shape of the solidied layer. As an example, it was shown
(Fig. 5.15b) that the research results can be adjusted to the value of the contact
thermal resistance approximately equal1/BiCON * 0.22. As it is shown in
Fig. 5.16 in subsection, 3 values of 1/BiCON are different, since in this subsection,
the cold cylinder temperature varies. That shows that during the solidication
process the contact thermal resistance is of strong impact on the solidication. It is
connected with the layer of ice formed around the cylinder because a density
difference between ice and water causes the ice layer expansion that can be com-
pensated by the frozen layer movement (as in the case of solidication on a rect-
angular plate). Figure 5.17 presents a comparison of experimental data for
experiment II in subsections 2 and 3. It can be seen that the model very precisely
anticipates an expansion of the ice layer if 1/BiCON is adequately selected. In the
case of liquids other than water where there is much less of density change between
the solid and the liquid, the influence of the contact thermal resistance can be much
smaller. However, the thermal resistance of the contact layer should always be
experimentally veried.
As it was mentioned above, the thermal resistance of a contact layer is very important
because it signicantly affects the process of a liquid solidication. Because of a great
amount of influencing factors and their complexity, it is very difcult to dene this
resistance theoretically. The experimentally obtained values of the contact thermal
resistance for respective experiments are presented in the last column of Table 5.1.
Upon these measurements, the relation of the contact layer thermal resistance to
the surface temperature of the cooled copper cylinder can be shown. The contact
thermal resistance as the function of the cylinder surface temperature was graphi-
cally presented (Fig. 5.18). When the temperature decreases, the contact thermal
resistance rises. That can be explained by a decrease of the temperature of ice what
nally leads to the creation of a larger gap (apart from other causes) between the
cold cylinder and the layer of ice. As it is shown in the graph, the measurement data
are of some dispersion. It may so be because thermal resistance is a very complex
parameter also depending on purity of the surface, the degree of its moisture and so
on. The individual influence of particular factors is difcult to establish.
5.3.4 Conclusions
The thickness development of a layer solidied from a liquid undergoing free

convection inside an annular space depends on the method of both heating and
cooling B; Nu as well as on thermal resistance of the contact layer between
flowing liquid and the cold wall 1=BiCON . Experimental results were presented for
medium thickness of an ice or water layer undergoing free convection in the annular
layer. These values are compared with theoretical assumptions and allow to indi-
rectly dene the thermal resistance of the contact layer. The results of measurement
of the contact layer thermal resistance influencing the development of the solidi-
cation front in an annular space with a cold internal cylinder and in the conditions
of free convection are presented in this work.
Chapter 6
Solidication of a Liquid Flowing
into the Channel
6.1 The General Analysis
The literature on the solidication of a liquid flowing into the channel was reviewed
by Cheung and Epstein (1984), Fukusako and Yamado (1994) as well as Weigand
et al. (1997, 1999). Experimental research veried by the former theoretical results
was conducted by Lipnicki and Weigand et al. (Lipnicki and Weigand 1996;
Lipnicki et al. 2009).
Apart from the solidication front, the flowing liquid is characterized by the
front of stream. The front of stream makes the theoretical description of the process
of lling the channel with a liquid additionally difcult. Both the theoretical and
practical signicance of the discussed problem brought about, especially in foundry
engineering, a great number of scientic papers (Ragone et al. 1956; Flemings
1974; Madejski 1976; Mutwil and Bydaek 1986; Bydaek and Lipnicki 1990). The
investigations conducted on the lling of a cold channel with a solidifying liquid in
various external conditions are presented in the works of Lipnicki and Sobich
(Lipnicki 1992, 1999; Lipnicki and Sobich 1994, 1995, 1996).
Because of the high intensity of heat exchange, the discussed phenomenon of
lling the channel with a solidifying liquid can be also exploited in designing heat
accumulators (Fig. 6.1). The streams of the solidifying liquid flowing into the
cooling channel release the heat of the cooling liquid.
The physical and theoretical description of the phenomenon of a liquid solidi-
cation in the cooling channel is a complex task. That is because of both the
variable amount of the flowing liquid (continual supply of the liquid) and its
simultaneous solidication. The forces influencing the fluid movement are also
variable in time. This phenomenon is of undetermined character and depends on
both the space variables and time.
To solve this problem, the application of a simplied model based on
Poiseuilles set of flows in channels of diminishing diameters is proposed.

DOI 10.1007/978-3-319-53432-9_6
84 6 Solidication of a Liquid Flowing into the Channel
cooling medium
solidified liquid(PCM) front of stream
PCM
cylindrical channel
Fig. 6.1 Dynamic heat exchanger
solidification front in time t+dt

solidification front in time t
Q
liquid r(t) r(t+dt)
Q
solid layer
channel
position of the head stream (t)
Fig. 6.2 Model of solidication of the liquid flowing into the channel (Lipnicki 1999)
A simplied model of lling a liquid into the channel with cold walls is pre-
sented in Fig. 6.2. Two successive phases of lling, in two times t and t dt, are
shown. As the time goes, the channel narrows and the liquid flows in a cylindrical
channel of diminishing diameter and increasing length. A detailed analysis of this
problem is included in the work (Lipnicki 1999, 2000).
A homogeneous liquid of density q and constant pressure p at the beginning of
the channel lls up the horizontal cylindrical channel of radius R and axis x. The
physical model of this phenomenon is presented in Fig. 6.3. The liquid of mass m1
6.1 The General Analysis 85
Fig. 6.3 Physical model of the phenomenon and the channel geometry (Lipnicki 1999)
flows through the vertical channel and the loss of mass m2 caused by solidication
occurs also in perpendicular direction; thus, the change in main stream momentum
is caused only by the mass changes of the liquid. Coordinates xt and r t indicate
the position of the stream head and of the interface.
The momentum conservation law can be described by equation
2 3
ZZZ X
d4 ~ dX5 ~
qU Pi 6:1
dt i
Xt
~ is the liquid
where Xt is the fluid (variable) volume of the liquid in the channel, U
velocity and ~
Pi is the set of forces generating the liquid flow.
The expression on the left side in Eq. (6.1) was transformed by assuming the
density constant as follows:
2 3 2 3
ZZZ ZZ
d4 d
~ dX5 4q xt
qU ~ dF 5 d p q xt V t r 2 t
U
dt dt dt
Xt F t
6:2
with F tthe channel flow surface variable in time, perpendicular to the flow
directionand V tthe medium velocity of the liquid flow in the channel cross
section.
The force bringing about the flow Pw t results from the pressures difference,
whereas the resisting force Po from both the friction sx on the phase border and the
surface tension r on the head of stream. By including the above into Eq. (6.1), the
following is obtained:
d
p q xt V t r 2 t Pw Po : 6:3
dt
Fig. 6.4 Model of the liquid

flow in the channel at a
particular time (Lipnicki
1999)
The practical application of the equation discussed above may concern different
liquids and different conditions of outside cooling. The shape and expansion of the
interface dened by relation r t, knowledge of which is vital for solving the
problem of solidication of a liquid flowing in the channel, depends on both the
kind of liquid and cooling conditions.
The liquid flow in the channel (Fig. 6.3) in accordance with the accepted model
is treated as HagenaPoiseuille set of flows in the following space:
0 r 0 r t; 0 xt xmax t; 6:4
with r 0 the radius coordinate and xmax , the maximum length of the liquid in the
channel (Fig. 6.4).
The forces determining the solidifying liquid flow in the channel are as follows:
1. The one forcing the flow is as follows:
Pw p p r 2 t; 6:5
where p is the liquid pressure.

2. The resistance force Po with its value obtained as the result of the following
analysis.
The liquid flows through the channel (Fig. 6.3) of variable radius r t. The liquid
flow velocity in the channel denes Poiseuille equation:
Ur 0 ; t Ct r 2 r 02 ; 6:6
where C t is related to time.

The tangential stress on the interface and resulting flow inhibiting force is
dened by equation:

@U V
sx g 0 4g Ps 8 p g V x: 6:7
@r r0 r r
Another reason for the liquid flow resistance is surface tension affecting the head
stream. The force component caused by the surface tension and inhibiting the flow
horizontally is dened by equation (see Fig. 6.3)
6.1 The General Analysis 87
Pr 2p R r cos h 6:8
with r the surface tensions and h the surface tension angle.

By supplementing relations (6.56.8) to equation of momentum conservation
(6.3), the following non-dimensional differential equation of fluid movement was
obtained
d 2
~r ~x w P ~r 2 2 C 8 ~x w 6:9
d~t
where non-dimensional radius, non-dimensional axial coordinate, non-dimensional

time, non-dimensional pressure parameter and surface tension parameter were
adequately dened
r x R m p R2 Rr
~r ; ~x ; wV ; ~t t ; P ; C cos h
R R m R2 q m2 q m2
In real conditions, both the liquid stream head and interface are the most curved
ones. They are characterized by a curvature, and the heat removal is multidirec-
tional. Beyond that area, the solidied layer thickness at a particular moment can be
considered constant. The assumption of radial heat removal along the whole length
is to simplify the real phenomenon.
The obtained Eq. (6.9) together with that of thermal conductivity with border
conditions is the subject of the following analysis in relation to accepted model of
heat absorption.

with Weakly Conducting Walls
The subject of the analysis is the solidication of a liquid at the following condition
dening the relation of the channel wall heat accumulation to that of the liquid heat
p
ksc csc qsc
b p 1; 6:10
kcq
which in case of a not overheated liquid allows for temporary acceptation of the
temperature eld presented in Fig. 6.5.
The outer space of the channel is dened by coordinate y directed towards
outside from the outer surface of the channel. To simplify the matter, the effect of
the channel curvature on the heat transfer phenomenon was omitted. Then the eld
of temperature is described by one-dimensional equation of thermal conductivity
TO
y solid
TF layer
0
r(t) R x
liquid
wall
Fig. 6.5 Temperature distribution in the liquid and in the wall
@T @2T
jsc 2 6:11
@t @y
with jsc the coefcient of heat diffusion for the channel wall material.
The solution of Eq. (6.11), at border conditions:
T0; t TF ;
Ty; 0 TSC ;
6:12
@T
1; t 0
@y
The eld of temperature is dened by error function

y
Ty; t TF Tsc TF erf p ; 6:13
2 jsc t
with erf the operator of Gauss error function.

The heat conducted through the wall material in a unit length of the channel in
time t is
Zt
@T
Qsc 2 p R ksc 1 dt
@y y0
0
Zt r
ksc csc qsc dt
2pR TF Tsc p 6:14
p t
0
6.2 The Solidication of a Liquid Flowing into a 89
By comparing this heat with that of phase change
Zr
QF 2 p r dr q L 1; 6:15
R
interface equation in non-dimensional form was obtained

p
~r 2 t 1 a ~t; 6:16
where
4 Ste b
a p 6:17
p Pr
is the solidication parameter and
c TF Tsc gc
Ste ; Pr ;
L k
are Stefan and Prandtl similarity numbers. The obtained interface equation remains
in accordance with Chvorinov principle (Chvorinov 1958).
A solution of liquid movement Eq. (6.9) with consideration of both the interface
Eq. (6.16) and the initial condition is as follows:
~x 0 for ~t 0 6:18
are the following solutions (Lipnicki 1999): for the flow path
v
2 ~t 3
u
u Z~t Z
u 2C
~x~t t2 expu 4 P 2 expu d~t5 d~t 6:19
~r
0 0
and the liquid flow velocity in the channel

hR i
~t
expu 0 P 2C ~r2 expu dt dt
~ ~
w~t r
hR i 6:20
R ~t ~t
2 0 expu 0 P ~r2 expu dt d~t2C ~
where
16 2 16 a2
u ~r ln ~r 2
a2 a2
Integrals in the obtained solutions were estimated by applying Gauss-Legendres

quadrature. The result of numerical calculations of velocity and path is presented in
Figs. 6.6 and 6.7 at the absence of the surface tension (C 0) and in Figs. 6.8 and
6.9 at the presence of surface tension (C 6 0).
As it results from the presented theoretical analysis, the solidication process for
a given parameter a determines the liquid flow in the channel and what follows the
length of inundation of the channel. The velocity slightly increases in the beginning
of the process in the effect of decreasing amount of the flowing liquid, then
resistance forces prevail and flow velocity decreases. The surface tension signi-
cantly affects the liquid flow conditions. The maximum path of the liquid flow can
be expressed by analyzing the presented solution by means of a very simple for-
mula (Lipnicki 1999).
Fig. 6.6 Field of the liquid

velocity (C = 0) (Lipnicki
1999)
Fig. 6.7 Path of the liquid

flow (C = 0) (Lipnicki 1999)
6.2 The Solidication of a Liquid Flowing into a 91
Fig. 6.8 Field of the liquid velocity (Lipnicki 1999)
Fig. 6.9 Path of the liquid flow (Lipnicki 1999)
p
P
~xmax 0:2888. . . C 0; 6:21
a
which shows that maximum path is inversely proportional to the value of the
solidication parameter. The above formula is of signicant importance in the
practical application.

with Conducting Walls
The not overheated liquid of the phase-change temperature TF flows into the
channel of radius R (Fig. 6.10). Outside the channel flows the cooling liquid of
temperature TO . Between the surface of the channel and that of the solidied layer is
the contact layer of convective heat transfer coefcient hCON . Heat transfer from the
outer surface of the channel wall to the environment is dened by heat transfer
coefcient hch . Radial distribution of the temperatures is presented in Fig. 6.10. The
constant temperature in the cross section is adopted based on the assumption that
both the solidied layer and the flowing liquid are very good thermal conductors.
The heat balance for a unit of length of the channel shows that the phase-change
heat
Zr
QF 2 p r dr q L 1 6:22
R
passes to the wall and is transferred to the outer cooling liquid
Zt Zt
Q hCH 2 p R T T0 dt 1 hCON 2 p R TF T dt 1:
0 0
6:23
By comparing the above expression, equation dening the contact temperature is

obtained
hCH T0 hCON TF
T 6:24
hCH hCON
Fig. 6.10 Temperature TO

distribution in the liquid and
solid
in the wall cooling liquid TF layer
r(t) R x
liquid
contact layer
cooling liquid
6.3 The Solidication of a Liquid Flowing 93
By exploiting Eqs. (6.22), (6.23) and (6.24), relation dening the interface
equation is obtained
~r 2 ~t 1 a1 ~t; 6:25
where the solidication parameter a1 in the above formula is equal to a1

2 1 1 1
Ste
Pr and,
Bi BiCON
c TF TO hCH R hCON R mcq

Ste ; Bi ; BiCON Pr ;
L k k k
and Stefan number, Biot number, contact layer Biot number and Prandtl number are
adequately dened.
It results from formula (6.25) that the solidication parameter a1 depends mainly
on heat transfer conditions. In this case, the differential equation of fluid motion
(6.9) was analytically solved with the initial condition satised (6.18). For the
velocity eld, the following was obtained
p
2 P ~r 2 ln ~r 2
w~t p for a1 4; 6:26a
~r 4 2 ln ~r 2 1 1
q
8a1

1Pa1
a1 4
2 ~
r
2 a1
~ r 2
w~t q
4
for a1 6 4; 6:26b
1
~
r 4 a1 ~ r 4 a1 a1 4
2 8 8
and the flow path

r
P 4
~x~t ~r 2 ln ~r 2 1 1; for a1 4; 6:27a
32
s

P 1 4 a1 4 a 4
~x~t
1
~r ~r 4 a1 ; for a1 6 4: 6:27b
a1 a1 4 2 8 8
A graphic picture of the solutions (6.26a, b) and (6.27a, b) is presented in

Figs. 6.11 and 6.12.
The liquid flow velocity in the channel with a well heat conducting wall increases
for parameters a1 [ 8. It is in accordance with the conditions of liquid flow in the
channel when the mass loss of the flowing liquid in the result of solidication
complies with exceeding flow resistance resulting from friction. For parameters
a1 8, the velocity of the liquid flow is constantequilibrium of forces generating
the flow is retained. However, for a1 \8, the flow resistances are domineering and
cause a reduction in liquid flow velocity. In all the analyzed cases, the flow of liquid
stops in the moment when the cross section of the channel is totally plugged. The
Fig. 6.11 Liquid flow

velocity (Lipnicki 1999)
Fig. 6.12 Liquid flow path

(Lipnicki 1999)
cases when the velocity differs from zero in the moment of the flow blockade were
shown in the graph by marking the curves ends with little circles.
Based on the conducted analysis of the solutions of Eqs. (6.9) and (6.24), a
simple relation dening maximum length of the liquid flow up to the moment of the
blockade caused by total plugging of the channel was obtained
p
P
~xmax p : 6:28
8 a1
The practical application of the above formula can be signicant when designing
such processes, and parameters a1 allow for comprehensive elaboration of heat
transfer conditions as well as of a liquid thermophysical and of the channel cooling.
6.3 The Solidication of a Liquid Flowing 95
The structural model applied to solve the problem of solidication and that of heat
transfer as well as allowed to dene the functions of both flow velocity and path:
w w~t; a; P; C and ~x ~x~t; a; P; C; 6:29
which can be useful in the theoretical analysis of the phenomenon and in practice.
As a comparison to the dynamic solidication discussed above (non-stationary
process), it is worth to mention the works (Lipnicki et al. 2014; Lipnicki and Panto
2014), although the concern is the stationary flow with liquid metal solidication in
a cooled channel (continuous casting).
Chapter 7
Role of the Contact Layer
in a Solidication Process
In the theoretical research, the thermal resistance of the contact layer was until now
considered a constant parameter in equations dening a liquid solidication of
value assumed a priori, and its effect on the phenomenon was investigated poste-
riorly. In reality, the thermal resistance of a contact layer (Artyukin and
Nenarokomov 1984; Nishida et al. 1986; Wang and Matthys 1994; Loulou et al.
1999a, b; Furmaski and Winiewski 2002; Lipnicki et al. 2005; Lipnicki and
Weigand 2012; Weigand and Lipnicki 2016) depends on the time of the solidi-
cation, the geometric arrangement of cold walls surrounding the solidifying liquid,
and on the intensity of heat transfer in the process. For the above reason, a theo-
retical analysis of the development of a contact layer with special attention given to
the width of the contact layer and its coefcient of thermal resistance 1=BiCON s is
presented.
7.1 Studies of the Thermal Contact Resistance
In a process of solidication, a contact layer of both structure and size not yet
dened but playing an important role in the process is formed along the contact of
the solidied liquid and the wall. It affects the conditions of heat transfer from the
solidifying liquid. Then, the conditions of solidication affect the structure and
other features of the solidied layer which is being formed. Until now, the presence
and the role of a contact layer (the gap) were replaced by the macroscopic parameter
representing the contact thermal resistance 1=BiCON . It comprehensively expresses
numerous single phenomena which occur in the process: gas pressure, gas tem-
perature, and gas movement. This chapter is devoted to the theoretical analysis of
this parameter.
A contact layer and problems connected with it are common in metals and
alloys casting where in the process of solidication, a gap, i.e. discontinuity space,
is formed between the cast and the surface of a foundry mould. Yet the contacting
DOI 10.1007/978-3-319-53432-9_7
98 7 Role of the Contact Layer in a Solidication Process
Fig. 7.1 Example of a

contact layer (Lipnicki 2012) Sn
air bubble
Cu
bodies do not create a solution. During the soldering and welding, another contact
layer is formed. It is a special type of a contact layer. Here, the contact is a diffusive
compound of two metals. The mentioned contact, for example, occurs during the
production of electronic plates. Printed copper plates are coated with a tin layer in a
special generator producing upward bulging waves of liquid tin above which shifts
a cold copper plate with surfaces to be soldered.
Figure 7.1 presents a picture of the contact of solidied tin with a copper plate.
On the contact, a discontinuity appears in the form of a solid solution of one metal
in another. The structure of the contact layer is complex and without distinct
boundaries. Also air bubbles appear in the contact layer. In such place, additional
thermal resistance occurs. It is another type of the contact layer in comparison with
the cases already discussed that also causes additional thermal resistance.
The problem of heat flow through the contact layer formed in the process of
solidication was the subject of a lot of studies. Among the rst works in this eld
were those of Wang and Matthys (1994), Furmaski and Winiewski (2002).
Wang and Matthys (1994) conducted experiments connected with very fast
solidication of tin poured onto a cold aluminium base. The temperature was
measured with thermocouples xed at two different points above the plate. Then by
applying the inversion method for the heat transferred, the values of the contact
thermal resistance were established. It was noticed that the thermal resistance of the
contact layer varies in time of solidication, and its values were dened. At the
beginning of the solidication process, a fast increase in the thermal conductivity
was noticed, and then after reaching the maximum value it starts to decrease equally
violently to reach, in an enough long period of time, the constant value (Fig. 7.2).
Superheating of the solidifying liquid very strongly affects the thermal resistance of
the contact layer.
Research on thermal resistance of a transition layer was conducted by Loulou
et al. (1999). Their experiment concerned the measurement of the contact thermal
resistance in the case of a tin drop solidifying on a cold nickel base. Their concept is
presented in Fig. 7.3. They measured the temperature with thermocouples placed at
chosen points of the nickel base and of the tin drop. Based on the above, they
7.1 Studies of the Thermal Contact Resistance 99
Fig. 7.2 Thermal resistance kW

of a contact layer, Wang and m2 K
Matthys (1994)
50
40
thermal conductivity
30
20
10
0
0 0.2 0.4 0.6 0.8 1 s
time
thermocouple
Fig. 7.3 Concept of a
research stand according to 15 mm
Loulou et al. (1999b) x
drop of tin
x= b
x= a
x=0
nickel base
dened the heat stream and hence resulting in the thermal resistance of the tran-
sition layer. Based on the conducted measurements, the phenomenon of a tin drop
solidication and the formed contact layer thermal resistance along the process was
described.
In their research, Loulou et al. divided the process of solidication into four
successive phases (Fig. 7.4). During the rst solidication phase (A), a very fast
decrease in thermal resistance caused by the distribution of the liquid on the plate
surface was noticed. During this time, the contact thermal resistance depends on the
surface roughness, the surface tension of the liquid metal, the degree of wettability
and the contact angle to the surface, and the type of conned gas and the liquid
metal pressure. The contact thermal resistance at the end of this phase reaches the
minimal value. Both the liquid and the plate contact each other through apexes of
the peaks of irregularity of the rough surface, and nucleation begins at the lowest
temperature points, i.e. where the liquid metal, solid plate, and conned gas contact.
In the next phase (B), the process of nucleation and solidication proceeds. The
contact thermal resistance becomes established. Then, the interfacial gap increases
Fig. 7.4 Evolution of the m2K

kW
thermal resistance of a contact
0.4
layer (Loulou et al. 1999b)
thermal resistance
0.3
0.2
0.1
A B C D
0
s
0 0.2 0.4 0.6 0.8 1.0
time
Fig. 7.5 Comparison of the kW

thermal resistance of contact m2 K
layers
50
40
thermal conductivity
tin/copper [Wang, Matthys,1994]
30
20
tin/nickel [Loulou i in.,1999]
10
0
0 0.2 0.4 0.6 0.8 1 s
time
because of the solidied layer contraction and the plate expansion. That is why the
contact thermal resistance begins to increase and the third phase follows (C). In the
next phase (D), the contact thermal resistance becomes stable.
The studies of Wang and Matthys (1994) as well as those of Loulou et al.
(1999) are of the same quality, what is conrmed by the comparison of their
research in Fig. 7.5. However, it is difcult here to refer to the quantitative com-
parisons since those concern other solidifying metals.
7.2 A Theoretical Analysis of Thermal Contact Resistance 101
7.2 A Theoretical Analysis of Thermal Contact Resistance

on the Basis of Experimental Investigations

on a Plate
A process of a non-heated liquid solidication on a cold plate of variable tem-

perature is described by Eq. (4.40), introduced in Chap. 4. By assuming a variable
thermal contact resistance, it takes the form
~d d~d 1 d ~d
s expb s; 7:1
ds BiCON ds
where constant b is a parameter dening the velocity of heating a plate cooled with
the solidifying liquid. The changeable thermal contact resistance 1=BiCON , as it
results from experimental research, very strongly depends on the time, and for that
reason the nonlinear differential Eq. (7.1) is difcult to be solved analytically. This
problem was analysed (Lipnicki et al. 2005; Lipnicki and Weigand 2012) in dif-
ferent theoretical descriptions of solidication in different periods. In the beginning
of solidication, the thickness of the solidied layer is very small ~d 1; thus,
Eq. (7.1) after rejection of small terms of higher grade is reduced to a simple
differential equation
1 d~d
s expb s; 7:2
BiCON ds
which can be analytically solved without difculty. After separating the variables
and satisfying the initial condition: ~d 0 for s 0, a solution in integral form is
obtained.
Zs
~d 1
expbsds: 7:3
1=BiCON s
0
As it can be seen, to solve the above equation it is necessary to know the

function dening the thermal contact resistance 1=BiCON s. Based on the earlier
discussed experimental research conducted by Wang and Mattyes, Loulou et als
functional relations of the thermal contact resistance to time are introduced. It seems
that the results of research conducted by Wang and Mattyes are very well described
by the function presented in Fig. 7.6. It consists of two basic parts: the rst is a
variable function in the eld 0 s=sst 1, the second a constant one in the eld
1 s=sst 1. The time sste , after which the thermal resistance of a contact layer
Fig. 7.6 Approximated

BiCON
relation of Biot number to
time (Lipnicki et al. 2005;
Lipnicki and Kuczma 2009) 30
20
BiO
10
Biste
/ste
0
0 1 2 4 6 8 10
stabilizes divides the solidication process into two characteristic periods: when the
thermal resistance is strongly dependent on time and when it is relatively stable.
One of many possibilities of obtaining analytical solutions is the attempt to
approximate experimental data by theoretical relations BiCON s during the initial
period of solidication. In the works of (Lipnicki et al. 2005; Lipnicki and Kuczma
2009), a series of random approximations were conducted. When approximating the
data by a parabolic function (see Fig. 7.6), Biot number is presented by the fol-
lowing analytical relation
BiCON s 22Bi0 Biste s=sste 2 4Bi0 Biste s=sste ; 0 s=sste 1;

BiCON s Biste const; 1 s=sste 1;
7:4
where Bi0 is the maximum Biot number within the rst interval (0 s=sst 1) and
Biste the constant Biot number in the second interval (1 s=sst 1). After inte-
grating Eq. (7.3) and taking into consideration the initial condition ~d 0 for s 0,
the following analytical solution was obtained
Zs Zs
~ 22Bi0 Biste 4Bi0 Biste
d s expbsds
2
s expbsds
s2ste sste
0 0
42Bi0 Biste 2 2 7:5
b s 2 bs 1 expbs 1
b sste
3 2
4Bi0 Biste
bs 1 expbs 1:
b2 sste
Figure 7.7b illustrates in a graphical form the above analytical solutions

obtained by various methods of calculations of the solidied layer thickness
development. The proles drawn in Fig. 7.7a, b clearly show that the beginning of
solidication is presented in different ways, depending on the calculation method
(a) (b)
1.2 0.1 analytical solution
analytical solution for small
for small 0.08
0.8
Bi=
CON const. 0.06 Bi=
CON const.
Runge -Kuttas method 0.04
0.4
0.02
/ste /ste
0 0
0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0
Fig. 7.7 Relation of the solidied layer thickness to time-comparison of different methods of
calculation, for: a sste 1; b 1, b sste 0:01; b 1 (Lipnicki et al. 2005)
applied. Analytical solutions for very small times and variable Biot number should
be considered the most reliable and true. The initial solidication velocity presented
by the curve is relatively small; thus, the thickness of the solidied layer increases
very slowly. This is justied by a relatively strong thermal resistance of the contact
layer formed in the process (see Fig. 7.4). However, for bigger times values the
analytical solution is not very precise because of the accepted approximation. In this
case, credible results are given by calculations conducted for a constant Biot
number. This was also conrmed by the calculations conducted by RungeKutta
numerical method.

on a Cylinder
Upon the theoretical analysis conducted in Chap. 5 for a liquid flowing in an

annular space between two coaxial cylinders of different temperatures, it can be
seen that a liquid solidication on the internal surface of a cold cylinder of tem-
perature lower than that of liquid solidication is dened by Eq. (5.19)
Z~d ~d ln ~d 1=BiCON
s d~d: 7:6
1
B Nu d~ ln ~d 1=BiCON
The above parameter 1=BiCON denes the contact thermal resistance of variable
and undened value. Based on the conducted theoretical analysis and by referring it
to the conducted experiment, it was possible to dene the medium value of the
contact thermal resistance. The results were presented in Table 5.1 and in Fig. 5.18.
As it is known, the parameter 1=BiCON denes the features of the contact layer
which are variable in time especially in the beginning of the solidication process.
It should be assumed that in the initial period of solidication, this parameter
differently affects the solidication process on a cylindrical surface. The latter was
conrmed by (Lipnicki and Weigand 2012). In the initial period of solidication,
BiCON number can be dened in similar way as for a rectangular plate and in
accordance to the experiments conducted by Wang and Matthyes (1994) as well as
by Loulou et al. (1998, 1999) by using the equation

ps
Bis Bi0 sin 0:85 for s sst ; 7:7
sst
where sst is the time necessary to obtain the constant thermal resistance. In the
above equation, another approximating function than that in case of a rectangular
plate was applied. However, the choice of approximating function should not
signicantly affect the results obtained.
By supplementing Eqs. (5.7)(3.25), a differential equation was obtained.

ps
B Bio sin 0:85 Nu ds d~d 7:8
sst
The asymptotic solution of the above equation, after the integration for small
values of time and for variable Biot number and after satisfying the initial condition
(5.18) takes the form of

~ B Bio sst s
d 1 Nus 1 cos 0:85 dla s sst : 7:9
0:85p sst
The solidication process in its initial period seen from two angles is presented
in Fig. 7.8a, b. At the beginning of the solidication process when thermal resis-
tance 1=BiCON is relatively high (see Fig. 7.4), the solidied layer increases slowly.
Then, the solidied layer considerably increases.
(a) (b)

1.3 1.025
1.25 1.02
1.2
1/BiCON=const 1.015
1.15 1/BiCON=const
1.01
1.1
1/BiCON()
1.05 1.005
1/BiCON()
1 1
0 0.005 0.01 0.015 0.02 0 0.0004 0.0008
Fig. 7.8 Increase in a solidied layer thickness obtained by different methods for Bi0 7:2,
B 7:8; Nu 6:8, 1=BiCON 0:3; sste 0:001, a long solidication time; b beginning of
solidication (Lipnicki and Weigand 2012)
In the quoted work (Lipnicki and Weigand 2012), a comparison between the
theoretical calculations and the experimental data of the solidied layer thickness as a
time function was conducted. By comparing the investigations, it can be seen that the
value of the contact thermal resistance considerably affects the development of the
solidied layer. It was noticed that by comparing the theoretical and experimental
results of investigations, the contact thermal resistance can be estimated. By means of
the created theoretical model (5.19), it was possible to distinguish from numerous
describing this phenomenon parameters the one of thermal resistance. Phenomenon
parameters that of the thermal resistance of the contact layer and to estimate its value
and effect on the solidication process.
An increase in the ice layer formed around the cylinder and the change of ice
density brought about by the change in temperature causes some deformation of
that layer which is compensated by the formation of a gap. Based on the mea-
surements, the relation of thermal resistance of the contact layer to the surface
temperature of the cooled copper cylinder was presented (Fig. 5.18). With a
decrease in the temperature, the thermal resistance of the contact layer increases.
7.3 Development of the Contact Layer During

the Solidication Process
A contact layer is formed in the solidication process on the boundary of solidied

layer, and a cold wall changes with time as a discontinuity place. The change
concerns its geometric sizes and thermodynamic parameters such as temperature
and thermal conductivity. For that reason, a theoretical description of the phe-
nomenon occurring in the contact layer done by applying modern calculation
techniques (numerical methods) requires time. The small size of the contact layer
requires application of special theoretical models based on the physics of gases of
the gas remaining in narrow gaps as well as returning to simple analytical
descriptions used in such cases. A simplicity of the model is not its drawback when
it allows for solving the problem. Quite the reverse, it is an advantage if the problem
is accurately and properly solved. In the chapter, solving the problem of contact
layer development and its thermal resistance by applying the physics of gases in the
gap is continued with the use of analytical methods.
The formation of a contact layer and the role played by the latter in the solid-
ication process was investigated and presented by Weigand and Lipnicki 2016.
A model of the contact layer in a graphical form and the heat flow through this layer
was presented in Fig. 3.1 in Chap. 3 of this work. The equation of stream balance
through the solidied layer and of the heat absorbed by the liquid on the solidi-
cation front is described by the equation

@TS dd
kS qS L hTL TF ; 7:10
@n nd dt
Table 7.1 Density and thermal conductivity coefcient of ice (Ranjevi 1916)
Temperature (K) Density (kg/m3) Thermal conductivity [W/(mK)]
273.15 917 2.210
253.15 920 2.442
223.15 924 2.780
173.15 928 3.489
with
TS contact layer temperature;
TL liquid temperature;
TF solidication temperature;
kS coefcient of the contact layer thermal conductivity; and
h coefcient of heat transfer on the liquid/solidication layer border.
The main cause of the gap formation between the solidied layer and the cold
wall is a change in the density of the solidied layer near to the cold wall surface.
The relation of the density of an ice layer (solidied) to temperature is presented in
Table 7.1 (Ranjevi 1916). The density change in the solidied layer is the reason
of the change in its volume.
In Fig. 7.9, density changes in the function of solidied layer thickness
d 0 d0 d are presented. Densities of the solidied layer in various places were
0
suitably marked: qW density at the wall surface temperature TW , q density at

temperature T of the surface of the solidied layer surrounding the gap what
coincides with coordinate dC . In Fig. 7.9, the elementary mass of the solidied
layer is marked as dm.
The gap between the cold wall surface and the solidied layer surface is lled
with air under atmospheric pressure. The model is applied to calculate the width of
the gap for solidication on rectangular and cylinder surfaces of small curvature
without taking into consideration the tensions accompanying the solidication
process in the solidied layer. The formed contact layer of width dC causes
Fig. 7.9 Density distribution S dm=S d

of the solidied layer near the

cold wall (Weigand and W
Lipnicki 2016)
F
0 C d
7.3 Development of the Contact Layer During the Solidication 107
Fig. 7.10 Temperature thin gap solid layer

prole in the contact layer
(gap) and in the solidied
layer (Weigand and Lipnicki q
2016) TF
T solidification
02 front
01 Tw
0
C
additional thermal resistance. The temperature distribution in the gap and in the
solidied layer is presented in (Fig. 7.10).
Heat stream q_ flowing through the gap can be dened by assuming a stable
quasi-one-dimensional heat transfer in the gap
kCef kC
q_ T T W T TW ; 7:11
dC dC d0
where kC the coefcient of heat conduction in the gap, kCef the substitute (effective)
coefcient of heat conduction in the gap, which according to the theory of gas
kinetics can be estimated out of equation (Furmaski and Winiewski 2002)
kC
kCef : 7:12
1 2a1
a1 2a1
a1
2ja Kn
ja 1 Pr
Additional computational thickness d0 for the gap is considered to be an addition

as a result of the gas accommodation by the boundary walls. The computational
thickness d0 d01 d02 as in formula (7.10) can be calculated using the formula
(Collie 1989; Furmaski and Winiewski 2002)

2 a1 2 a2 2ja Km
d0 : 7:13
a1 a2 ja 1 Pr
In the formulae (7.12) and (7.13) the following parameters were marked:
a1 ; a2 coefcients of accommodation;
ja isentropic exponent for gas;
Kn Km =dC Kundsen number;
Km medium free path of a gas molecule; and
Pr gas Prandtl number.
It is assumed that air molecules reach the thermal equilibrium with the walls, so
it is accepted that coefcients of accumulation are of equal value.
7.3.1 Development of Contact Layer and Its Role

in Phase-Change Process on a Rectangular Plate
The concept of calculating the width of the gap, i.e. a contact layer between the cold
wall and the surface of a solidied layer (see Fig. 7.9.), relies on comparing the
volumes of the layer solidied at solidifying temperature TF and that solidied near
the cold wall
Zd
qF qW 0
m qF d qS dd0 ; qS qW d; 7:14
d
dC
where qF the density of the layer solidied in temperature TF and qW in temperature

of the cold wall surface TW . Equation (7.14) shows that the relation of sof the
width of the contact layer dC to the width of the solidied layer d is equal
r
dC qF qW q
s 2 W 1 : 7:15
d qW qF qF qF
If the temperature of the cold wall is 20 C, the relation s is
dC
s 1:63044 103 : 7:16
d
By solving Eq. (7.11), the contact layer temperature is obtained
TF kS dC d0 TW kC d
T : 7:17
k S dC d0 k C d
By using Eqs. (7.10), (7.11), (7.15), and (7.17), the following equation in
non-dimensional form is obtained
kC B d~d
Nu; 7:18
~d ~kC s ~d0 ds
where
~ d ~ dC ~d0 d0 ; dC sd; s SteFo;

d dC ;
H H H
cp TL TF jt kL
Ste ; Fo 2 ; j ;
L H qs cp
k S T F TW ~kC kC ; Nu hH
B ;
kL TL TF kS kL
Fig. 7.11 Development of C

the solidied layer and of the
contact layer on a rectangular 0.8 0.008
wall, for B = 42, Nu = 5,
~kC 0:0087 0.6 0.006

0.4 0.004
C
0.2 0.002
0
0 0.002 0.004 0.006 0.008 0.01
The analytical solution of an ordinary differential Eq. (7.18) at initial condition
s 0; ~d 0 7:19
is the equation
2 3
~ ~ ~kC B Nu s ~kC ~d ~d0
d kC B
s ln4 5; 7:20
Nu Nu2 s ~kC ~kC B Nu~d0
which denes the solidication time s in relation to the thickness of the solidied
layer d.
Following the formula (7.20), examples of calculation were conducted for the
solidied layer thickness and that of the contact layer which are presented in a
graphical form in Fig. 7.11.
Examples of calculations of parameters of an air-lled contact layer of width
dC 0:0004 m for temperature T = 253.15 K and pressure p 105 Pa are
presented below:
medium free path of an air molecule in the gap
T 253:15
Km 2:27 105 2:27 105 5:75 108 m;
p 105
Knudsen number
Km 5:75 108
Kn 0:000144;
dC 0:0004
Table 7.2 Results of calculations of a contact layer on the rectangular cold plate, for ice
(H = 0.1 m; TW = 20 C)
s ~dC Kn ~d0 ~dC ~d0 hCONW/m2 K Relative
error (%)
0
5:75 106 3:26 105 0.0176 1:87 106 3:45 105 6614 5.75
2:26 105 6:52 105 0.0088 1:87 106 6:71 105 3399 2.88
4 4 6 4
2:04 10 1:96 10 0.0029 1:87 10 1:98 10 1152 0.95
4 4 6 4
4:62 10 2:93 10 0.0002 1:87 10 2:95 10 773 0.65
8:25 104 3:91 104 0.0001 1:87 106 3:93 104 580 0.52
1:88 103 5:87 104 9:8 104 1:87 106 5:89 104 387 0.26
3 4 4 6 4
4:29 10 8:80 10 6:5 10 1:87 10 8:82 10 260 0.23
2 3 4 6 3
1:36 10 1:53 10 3:8 10 1:87 10 1:53 10 149 0.13
correction
2 a1 2 a2 2ja Km 2 1 2 1:4 5:75 108

d0 2
a1 a2 ja 1 Pr 1 1:4 1 0:716
1:87 107 m
Calculations of the increasing in time thickness of the contact layer and the
heat-transfer coefcients are listed in Table 7.2.
As it can be seen in Table 3.2, at the beginning of the solidication process
(s\2:26 105 ), Kundsen number values are relatively high, Kn [ 0:01; thus,
when calculating the thermal resistance of a contact layer, a leap of temperature, i.e.
an additional extrapolated width ~d0 (Fig. 7.10.), should be taken into consideration.
For longer solidication times, the values of correction ~d0 in comparison with the
width ~ dC of the gap are relatively small, ~d0 ~dC , what means that thermal resis-
tance of the contact layer signicantly depends only on the thickness of the gap
itself. In the last two columns of Table 3.2 were presented the values of the
coefcient hCON and the values of relative errors appearing when calculating this
coefcient with omitting ~d0 . Calculation errors apart from the initial stadium of
solidication are rather slight. The convective heat-transfer coefcient values and
the calculation error decrease with an increase in the solidication time (the
thickness of the contact layer increases). The presented analysis shows that the
effect of gap walls (correction effect) on the thermal resistance is considerable in the
initial phase of solidication, and it is necessary to take the effect of the gap walls
into consideration.
Development of solidied layer thickness on a cold rectangular plate is presented
in Fig. 7.12. The value of the conductive heat-transfer coefcient of the contact
layer hCON is very high at the beginning of the solidication process and then fastly
decreases with time s. For long time, the changes in the conductive heat-transfer
hCON
C
W
1 0.002
m2K
1600 0.8 0.0016

1200 0.6 0.0012
C
800 0.4 0.0008
hCON
400 0.2 0.0004
0
0 0.004 0.008 0.012
Fig. 7.12 Development of both the solidied layer ~ d and contact layer ~
dC thickness and of the
conductive heat-transfer coefcient hCON in the contact layer for TW 253:15 K
coefcient are smaller and smaller, and the value of the conductive heat-transfer
coefcient is getting stable. The presented relation is of the same quality as the data
presented in the literature. On the contrary, the thicknesses of both the solidied
layer ~
d and gap ~
dC increase with the solidication time. For long time, the solidied
layer develops more slowly.
7.3.2 Development of a Contact Layer in Solidication

on an Annular Surface
The role played by thermal stress caused by temperature distribution in the solid-
ied layer seems to be insignicant in solidication of a liquid on cold rectangular
walls. For that reason in the analysis of the contact layer formation in solidication
on rectangular walls, only the contraction of the layer solidied near a cold wall
caused by a change in the solidied layer density was considered. For cold
cylindrical walls, the thermal stresses that occur in the solidied layer and their
effect on the layer deformation depending on temperature are important.
To dene the thickness of a contact layer formed in solidication on a cold
cylindrical surface (Fig. 7.13), a theoretical analysis based on two models was
applied: (1) the model based on concentration (thickening) of the solidied layer
near a cold wall and (2) a model in which deformations of the annular solidied
layer caused by thermal stresses occurring in the layer are considered.
Fig. 7.13 System geometry ice

for a cylindrical surface contact
(Weigand and Lipnicki 2016) layer

R
C
cylinder
TL
TF
TW
T
q
Model-1Contraction of the Solidied Layer

The balance of mass near the cylindrical wall gives the equation
Zd

m 2p qS d0 dd0 p d2 R2 qF ; 7:21
dC
where the substance density in the solidied layer qS is a radius function
lnd0 =R
qS qW qF qW ; R d0 d; 7:22
lnd=R
and dC the radius of the contact layer (Fig. 7.13) is the lower limit of the integral in
Eq. (7.21).
By substituting Eq. (7.22) into Eq. (7.21), the relation between the radii of the
contact and the solidied layers is obtained (Weigand and Lipnicki 2016)
!
q F qW ~
~d2 d~2 qW ~2 ~2 ~2
d2 ln ~d ~d2C ln ~dC C d dC d 1 0 7:23
q ln ~
F d 2 2 qF
The numerical solution of the above equation allows to dene the relation
between the radiusof the contact
layer ~dC and that of the solidication layer ~d. The
difference dC R ~dC 1 is the width of the contact layer (Fig. 7.13).
Model-2Deformation of the Solidied Layer

During a solidication process, a contact layer is formed between the cold wall and
the solidied layer. The solidied layer is subjected to thermal stresses because of
the temperature changes, what causes its deformation. The inner border of the
solidied surface bordering with an air gap undergoes deformation. The second
border of the solidied layer is the solidication front that is created in contact with
the liquid. The thickness of the solidied layer affects the size of the air gap. As it
can be anticipated in the case of a rectangular solidied layer, the thermal stress
does not signicantly affect the size of the gap. In model 2, some deformation of the
solidied layer is considered. The annular shape of the solidied layer that forms on
a cylindrical wall is presented in Fig. 7.13. The process of contact layer formation
in the form of a cylindrical casting mould was presented by aid of approximated
calculations and measurements (Nishida et al. 1986).
The width of the contact layer between the solidied layer and the cylinder
surface is caused by the deformation of the annular solidied layer because of
thermal stresses in the layer. The deformation occurring in the cylindrical solidied
layer at the axially symmetrical heat flow T r , according to the classical theory of
thermal stresses presented by Timoszenko et al. 1961, in relation to both the radius
and temperature, denes the differential equation

d 1 drDdr 1 m dT
b ; 7:24
dr r dr 1 m dr
where m is the Poisson number, b thermal expansion coefcient of the solidied

layer material.
The following equation is the solution of the above differential equation dening
deformation Ddr , at the absence of stress on the unbounded inner and external
borders of the solidied cylindrical layer.
0 r 1
Z Zd
1m b@ 1 2mr 2 R2
Ddr T r rdr T r rdr A: 7:25
1mr d2 R 2
R R
The displacement of the radius of the solidied layer internal surface, by sup-
plementing r dC in Eq. (7.25) is dened by the equation
Zd
2bdC 1 m
Dd dC R T r rdr: 7:26
d2 d2C
dC
It should be noticed that together with the increase in the radius of the external
layer surface described by radius d, radius dC of the internal layer also increases
from its initial position dC R.
By assuming one-dimensional (radial) heat flow in the solidied layer, the

temperature eld is dened by the equation
ln ~d0 $
T r T d 0 TF TW ; 1 ~d0 d: 7:27
ln ~d
By substituting the above equation into Eq. (7.26), the equation dening the
contact layer radius dC in the solidied layer radius function d is obtained (Weigand
and Lipnicki 2016)

~ 2bdC 1 m 1 ~2 ~ ~2 ~ 1 ~2 ~2

dC 1 d ln d dC ln dC d dC : 7:28
~
d2 ~d2C ln ~d 2 4
By applying the equations dening the radii of the contact layer ~dC in the
solidied layer radii function ~d (model1; Eq. 7.23, model2; Eq. 7.28),
numerical calculations were conducted (Table 3.3 and Fig. 7.14). The values of the
contact layer radius ~dC according to model-1 (Eq. 7.23) monotonically decrease to
unity, but according to model-2 (Eq. 7.28) they just approximate to it, ~d ! 1,
displaying a peculiarity of a violent leap ~dC to unity. This peculiarity was shown in
Table 3.3 and in Fig. 7.14 (Table 7.3).
Heat streams flowing through the contact layer and the solidied layer are equal
(see Fig. 7.13) to
kC R kS d
q_ T TW TF T 7:29
dC R d0 d lnd=R
From the above equation, the temperature of the contact layer is obtained
TW kC ln d TF kS dC R d0
T : 7:30
kC R ln d kS dC R d0
Fig. 7.14 Width of the C

contact layer in the solidied thermal deformation
layer width function 1.0008
(Weigand and Lipnicki 2016)
1.0006
1.0004
solid layer contraction
1.0002
1.0
1.0 1.2 1.4
Table 7.3 The width and the thermal conductivity of the contact layer on a cylindrical surface
(R = 0.1 m)
~d Solidied layer contraction Solidied layer thermal deformation
Model-1 Model-2

~dC 1 hCON W m2 K ~dC 1 hCON W m2 K
1 0 0
1.1 0.000165 1382 0.000692 329
1.2 0.000336 679 0.000709 322
1.3 0.000513 444 0.000725 314
1.4 0.000691 329 0.000741 308
1.5 0.000882 259 0.000756 302
By substituting Eqs. (7.29) and (7.30) into Eq. (7.10), differential equations of
heat balance in non-dimensional form are obtained
~k B d~d
C Nu; 7:31
~kC ~d ln ~d s ~d ~d ~d 1 ~d~d0 ds
where
~ d ~dC dC ;~d0 d0 ; dC 1 sdd 1d;

d
R R R
cp TL TF jt kL
s SteFo; Ste ; Fo 2 ; j ;
L R qs cp
k S T F TW kC hR
B ; ~kC ; Nu
kL TL TF kS kL
The analytical equation of Eq. (7.31) at the initial condition
s 0; ~d 1 7:32
is as follows:

Z~d ~kC ~d ln ~d s ~d ~d ~d 1 ~d0 ~d
s h id ~d; 7:33
~kC B Nu ~kC ~d ln ~d s ~d ~d ~d 1 ~d0 ~d
1
presenting the time of the solidied thickness development ~d [ 1 on a cylindrical

surface. The numerical calculation
of integral of the above formula requires the
~
knowledge of function s d , dening of which was the subject of the above
considerations (Eqs. 7.23, 7.28, and Fig. 7.14.).
Fig. 7.15 Development of C

both the solidied layer and
contact layer (Weigand and
Lipnicki 2016)

C
C
The development of both the solidied layer d and contact layer dC in the
solidication process according to two theoretical models, model of contraction1
and model of deformation2, is presented in Fig. 7.15.
Model-2 is more realistic, especially for long solidication times when the
solidied layer is already of adequately considerable thickness, and the role played
by thermal deformations is visible. When dealing with model-2, there occur some
calculation difculties and that is why the beginning of the curves is not shown.
However, the beginning of solidication (for small times) is well described by
model-1, the model of contraction.
The presented models of the contact layer allow to dene the relation between
the width of the contact layer and that of the solidied one. As it can be seen from
the presented analysis, the importance of the contact layer in the phase-change
process is signicant. Despite a very small width of the contact layer, its effect on
the general heat resistance is considerable. In Figs. 7.16 and 7.17, the impact of
Fig. 7.16 Development of

the layer solidied on a
1.0
rectangular wall for B = 42
and Nu = 5 (Weigand and 0.8
Lipnicki 2016) perfect contact
0.6
0.4
k C = 0 .0 0 8 7
0.2
0
0 0.002 0.004 0.006 0.008 0.01
Fig. 7.17 Development of

the layer solidied on a 2
cylindrical wall for B = 42 perfect contact
and Nu = 5 (Weigand and 1.5
Lipnicki 2016)
1
k C = 0 .0 0 8 7
0.5
0
0 0.002 0.004 0.006 0.008 0.01
thermal resistance on the contact layer on the solidication time is presented. In the
case of a perfect contact of a liquid solidifying on a cold wall (kC 1), the liquid
solidication is the fastest. When the contact layer is omitted in the solidication
process, the calculation results may be incorrect. It was proved that contact layer
width depends on that of the solidied layer and changes during the solidication
process. Based on the kinetic gas theory, it was shown that the gap thermal
resistance depends on Kundsen number which changes together with the increase in
the gap thickness.
Chapter 8
Phase Heat Accumulator
The model of a phase heat accumulator presented in this chapter is based on the
description of both the heat flow and the construction of this unit (Lipnicki and
Krasowski 2011, 2014patent). The proposed solution provides an effective and
fast transfer of the earlier accumulated heat of the phase change (latent heat) from
the inside of a PCM to the heated object by means of a heat compressor
pump. A working medium of the heat compressor pump is a low-boiling liquid.
8.1 Selected Review of the Units Used for Heat Storage
Heat accumulators serve to store the heat in the period of produced heat surplus
occurring, for example in heat and power plants, and to release the heat in the
periods of its deciency. The most common heat accumulators also named heat
containers can be divided into two basic groups:
Containers using the capacity heat and
Containers using the phase-change heat.
The basic drawback of capacity containers are their big sizes. Because of rela-
tively low specic heat of applied agents (water is much used), big sizes of
accumulators indirectly generate high investment costs.
Big sizes of the heat storage unit can be eliminated by using materials charac-
terized by the phase change. The phase-change heat is much higher than the specic
heat of materials. However, the application of phase change, evaporationcon-
densation, despite considerable changes in enthalpy, turned out to be impractical
because of a considerable change in material volume during the transition phase.
Most practical are phenomena of the solid phase change into the liquid one and in
reverse order, i.e. melting and solidifying. The use of PCM provides a lot of
possibilities for both storage and release of heat of temperature constant during the

DOI 10.1007/978-3-319-53432-9_8
120 8 Phase Heat Accumulator
accumulator loading and unloading. Presently, a great number of phase-change

materials (Annex.) that make possible the heat accumulation within different,
desired limits of temperature are known. During the phase-change heat accumu-
lation, an important role is played by the outer geometry of materials that affects the
velocity of the process of heat accumulation and release. PCMs are usually poor
heat conductors and that is why the heat transfer through a PCM is impeded. A vital
parameter of a heat accumulator is the size of the outer surface of the material, a
derivative of both its volume and form. A larger outer surface causes an increase of
heat energy, but the expansion of the surface is also limited.
Descriptions of PCMs, melting and solidication, are known in the literature.
Weaver and Viskanta (1986), Betzel and Beer (1988), Wang et al. (1999),
Sugawara et al. (2008), Lipnicki et al. (2001), theoretically and experimentally,
investigated the solidication and melting of PCMs of different geometrical
arrangements. They also researched the mass and heat transport in a PCM during
the convection phase change.
Weaver and Viskanta (1986) analysed the phase-change materials in porous
structures, submerging copper and aluminium balls in a PCM to improve the
thermal conduction effectiveness of the material.
The most effective solutions can be found in the patent solutions (Prusinski and
Johnson 1979; DeVogel and Gorges 1988; Colvin et al. 1989; Weber and Rummel
1996; Ival 2000; Lukas et al. 2002). In these solutions, the role of the phase-change
heat, its great value and the stability of temperature or its slight change in the
process were appreciated. The heat accumulators were used in specic practical
cases. The construction presented in patents aims to improve the solution which
would at the same time combine the phase change (solidication and melting) with
exploiting the thermodynamic equipment working along anticlockwise circulation
(heat pump and refrigerator). Combination of those two phenomena distinctly
increases the effectiveness of the heat transport process and allows to release earlier
accumulated heat effectively and with much power. The idea of heat transport
according to the invention (Lipnicki and Krasowski 2014) relays on the heat
absorption from the inside of the material and at the same time omitting its outer
surface. Thanks to the proposed method, the disadvantageous effect of low thermal
conductivity of PCMs on the heat transport is radically reduced.
8.2 Cooperation of the Heat Accumulator Storage

with the Heat Pump
In the discussed heat accumulator two simultaneously occurring phenomena are

applied: PCM phase changes and anticlockwise thermodynamic circulation are
realized (patent, Lipnicki and Krasowski 2014). A diagram of the unit and the
method of operating is presented in Figs. 8.1, 8.2 and 8.3.
8.2 Cooperation of the Heat Accumulator Storage with the Heat Pump 121
element
10 12
10 9 akumulatora
11 8
( 2L ) ( 2L )
( 1L ) ( 1L )
6
16 6
2
q q0 6 q
6
3 4
13 13
1
( 3L )
( 3L )
( 4L ) ( 4L )
1 7 15 14
14
Fig. 8.1 Diagram of a heat accumulator (Lipnicki and Krasowski 2014)
Fig. 8.2 Thermodynamic p

cycle of the heat pump on
Tsk q
Mollier chart pi 2L
3L
4L 1L
Tp
0
q0
x=
1
x=
Fig. 8.3 Repeatable element H/2

of the heat accumulator
PCM
czynnik
niskowrzcy
TF
Tp
front
krzepnicia q0
d dt

An example of the heat accumulator was graphically presented in Fig. 8.1,

where its cooperation with a heat pump was shown. The heat accumulator is in a
form of a closed container1 with outer surface covered with isolation mantle2
made of a good heat conductor. The boilers3 covered with PCM (phase-change
material)4 which are the repeatable components of the accumulator are placed
inside the container as close to each other as possible.
Inside the boilers, there flows a low-boiling thermodynamic agent6. Every
boiler is tted with inlet7 and exit8 pipes, both going out of the container. All
exit pipes are connected with the exit tube9 joining together the entries of two
compressors10 mounted on the common shaft11 with electric drive12. The
exits of compressors are connected to condensers13, which in turn are coupled
with inlets of proper throttle valves14. The exits of the throttle valves are con-
nected to the inlet tube15 joined by boiler inlet pipes. The boilers, compressors,
condensers and throttle valves connected by pipes are elements of the heat pump
securing the thermodynamic anticlockwise circulation by using a low-boiling
thermodynamic agent.
The accumulator heat is stored in the PCM4. During heat absorption from the
accumulator, the PCM is subjected to solidication process, and during heat
delivery, the material undergoes the process of melting. The flow of a low-boiling
thermodynamic agent in the pump system is forced by compressors (Fig. 8.2). The
solidication temperature of PCM is slightly higher than that of evaporation of the
low-boiling thermodynamic agent, TF [ Tp (Fig. 8.3). When flowing through the
condenser, the low-boiling thermodynamic agent absorbs the solidication heat of
the phase-change material and vaporizes inside the boilers. After passing through
the compressors, the low-boiling thermodynamic agent under high pressure flows
through the condensers where it gets liqueed in temperature Tsk . The condensers
are in the form of heat exchangers where the low-boiling thermodynamic agent
releases its heat to the secondary space heating circuit. From the condensers, the
low-boiling thermodynamic agent flows to the throttle valves where the pressure is
reduced to that in boilers. The energy used to compress the low-boiling thermo-
dynamic agent by means of the compressors is slightly in relation to the amount of
heat absorbed from the condensers.
The process of the accumulator loading is realized by melting the PCM by
means of a hot medium supplied from the outside of the heating coil16 of the
temperature higher than that of PCM melting.
The method of releasing the accumulated phase-change heat which relies on the
use of the phenomena of material solidication and simultaneous vaporization of
the thermodynamic agent generating the thermodynamic change in anticlockwise
circuit is characterized by the fact that the low-boiling thermodynamic agent that
accumulates the flow heat vaporizes inside the boilers placed within the PCM.
8.3 Example of Calculations 123
8.3 Example of Calculations
The n-octacosane C28H58 of the following thermophysic parameters: solidication

heat, L 233:3 kJ/kg; solidication temperature, TF 62:1 C; thermal conduc-
tivity coefcient, k 0:267 W/(m K); density, q 900 kg/m3 ; heat diffusion
coefcient, js 1:55 107 m2 =s; specic heat, c 1:92 kJ/(kg K); solidied
layer thickness, H 0:05 m; butting face of the element, Fi 1 m2 was used as
the PCM to illustrate the functioning of the heat accumulator.
As the thermodynamic agent for the heat pump, the low-boiling liquidsymbol
1R134awas applied of which thermodynamic properties were presented accord-
ing to Mollier diagram (Fig. 8.4). For the considered thermodynamic agent, the
boiling temperature Tp 50 C pp 13:17 bar and condensation temperature
Tsk 70 C psk 21:13 bar were assumed. In addition, the coefcient
1=Bi 0:01 dening the cooling conditions of PCM, the coefcient
1=BiCON 0:24 dening the contact thermal resistance and Stefan number Ste
0:0996 were accepted. According to diagram pi, enthalpies for agent R134a in
proper points of the thermodynamic diagram presented in Figs. 8.2 and 8.4 are
(1L)i1L 421:4 kJ=kg, (2L)i2L 428 kJ=kg, (3L)i3L 304 kJ=kg and
(4L)i4L 304 kJ=kg.
The thickness of a solidied layer and the velocity of its development according
to earlier obtained Eqs. (4.6) and (4.7) are dened by formulae
Fig. 8.4 Theoretical Linde cycle on Mollier diagram

s

2
~d 1 1 1 1 d~d 1
2s;
r 2 ;
Bi BiCON Bi BiCON ds
Bi BiCON 2s
1 1
6:1
where non-dimensional thickness, non-dimensional time, Stefan number, Fourier

number, heat transfer resistance and contact layer resistance are dened as follows:

~ d c TF T p js t 1 ks 1 ks
d ; s Ste Fo; Ste ; Fo 2 ; ;
H L H Bi hH BiCON hCON H
Figure 8.5 presents the thickness

. development of a solidied layer and the
~
velocity of the development dd ds, in time s. Based on the above, the heating
power and total phase transition heat can be dened for every element of the
accumulator, and those are of following values:
in the initial moment:
q L js d~d
Q_ 0 Ste Fi
H ds
900 233:3 1:55 107 1
0:0996 1 p kW
0:05 0:25 2 0
2
0:259 kW;
Fig. 8.5 Thickness and d

development velocity of a
d
solidied layer
5.0 0.5
4.0 d 0.4
d
3.0 0.3
2.0 0.2
1.0 0.1
0
0 0.05 0.1 0.15 0.2 0.25
8.3 Example of Calculations 125
in the nal moment:
q L js d~d
Q_ s 0:25 Ste Fi
H ds
900 233:3 1:55 107
0:0996 1
0:05
1
p kW
0:252 2 0:25
0:086 kW;
total heat released during a full solidication cycle in one element:
H 0:05
Qci q L Fi 900 233:3 1 kJ 5:25 MJ:
2 2
Medium heating power of the element during its complete solidication,
H2 0:052
tc sc 0:25s 40; 484 s;
Ste js 0:0996 1:55 107
is
_ i Qci 5250 kW 0:130 kW:

Q
tc 40; 484
The solidication heat stream transferred through an evaporating thermody-

namic agent in a boiler (Figs. 8.2 and 8.4) is equal to the solidication heat stream
_ i 2 0:13kW 0:26 kW:

Q_ 0 m_ q0 m_ i1L i4L 2 Q
The mass stream of the coolant flowing through the heat pump is

2Q_ i 0:260 kg kg
m_ 0:00221 :
i1L i4L 421:4 304 s s
Heat stream released in condenser per one element
Q_ i m_ q m_ i2L i3L 0:00221 428 304kW 0:274 kW;
and the thermodynamic agent compression power per one element
Ns m_ ls m_ i2L i1L 0:00221 428 421:4kW 0:014 kW:

Upon the presented above example of calculations, it can be seen that the
medium heating power of the proposed accumulator, consisting of 50 elements in
time of heat deciency (40,484 s), is approximately 13.7 kW. On the other hand,
the demand for medium electric power for compressor drive of a heat pump cycle is
approximately 0.7 kW. Because of the small difference between the condenser
temperature and that of a heat pump vaporizer, the coefcient of heat pump per-
formance (COP) is very high
Q_ i 0:274
COP 19:6:
Nsi 0:014
The presented conception of equipment consisting of a heat accumulator using

the solidication heat and heat compressor pump for heating is very advantageous
in regard to power engineering. The equipment is of high energetic performance
(COP + 19.6), and the transported heat stream (heating power) at a particular size
of the accumulators is very large.
Summary
A simplied theory of the solidication process occurring in arrangements of dif-

ferent geometry is presented. The theory is applied when calculating the thermal
capacity and velocity of the heat storage and release. In the investigated heat
exchangers, the exploitation of both latent heat and melting heat for so-called PCMs
is analysed. Moreover, the signicance and the role played by latent heat in the case
of alloys subjected to casting as well as to diffusion soldering are described.
Special attention has been given to the signicant role of the contact layer, i.e. air
gap that is formed in the technologies is analysed. This layer between the solidi-
fying material and the wall of the container (channel) or the base is of considerable
thermal resistance in the total heat transfer balance of the systems discussed.
The proposed simplied method of analysing the heat flow during crystallization
occurring in the investigated systems allows to estimate
the kinetics of thickening of the solidication layer
the nal thickness of the solidication layer at assigned conditions of
solidication
the amount of heat accumulated as a result of latent heat emission
the air gap kinetics of thickening during solidication
the contact layer resistance in relation to time.
The proposed theory is also useful when describing solidication occurring on
the border of the base and liquid metal, for example copper/liquid tin, nickel/liquid
tin and others. However, in this case, the contact layer (air gap) is replaced by
crystallization products, i.e. reactive sublayers. These products are usually
inter-metallic phases/compounds appearing there thanks to diffusion and reactions
accompanying the crystallization that usually proceeds in metastable conditions.
The reactive sublayers created as a result of the base/fluid interface can be also
analysed by means of the proposed theory and by treating those sublayers as
additional heat resistance, i.e. as a contact layer.
The analysis leads to the conclusion that in the solidication process occurring
near to a heat-consuming wall, a contact layeran air gap of width dC in the case of
a rectangular wall and dC R in the case of a cylindrical one is formed. The width
of the contact layer directly depends on that of the solidied layer, d, (Chap. 7,

DOI 10.1007/978-3-319-53432-9
128 Summary
Table 7.2.). Within the contact layer, there appears some difference in temperatures,
T TW and heat flow resistance, 1=hCON . The thermal resistance of the contact
layer was dened based on the theory of gas kinetics, and so for air contained in this
contact layeran air gap. It was shown that thermal resistance increases propor-
tionally to Knudsen number Kn (Table 7.2.). However, the prevailing factor
influencing the rise of the contact layer heat resistance is the increase of its
thickness. These calculations have been experimentally veried to a satisfying
degree by observing the solid/liquid interface displacement.
The original research method proposed in this paper is based on indirect cal-
culation of the contact layer heat resistance, i.e. on measuring the thickness of the
solidied layer (at conditions proposed for a rectangular wall, dC =d 0:000163,
Eq. (7.16). It should be pointed that the measurement of the solidied layer
thickness is very precise and that decides on the accuracy of estimation of the
contact layer thickness and hence resulting heat resistance (Table 7.2, Fig. 7.12).
The suggested way of estimation of the contact layer heat resistance is
competitive/more accurate in comparison with other methods known in the litera-
ture of the subject.
The kinetics of thickening of layers solidifying on two types of geometric
arrangements, i.e. rectangular and cylindrical ones, is described in this paper both
with and without the consideration of free or forced convection. When comparing
both descriptions, it can be stated that at an increase of the wall curvature, the
surface of the heat flow diminishes when approaching the wall, thus inhibiting the
latent heat flow. In addition, the bulk of solidied PCM and so the amount of
accumulated heat are greater in the case of a rectangular base than a cylindrical one
(Figs. 4.3, 5.3, 7.16 and 7.17).
The presented analyses (Table 7.2) show that the heat resistance of the contact
layer is of special signicance in the initial period of solidication, as being then the
lowest.
In the case of diffusion interconnections, the thermal resistance of reactive sub-
layers depends on the type of products of crystallization occurring on the solid/liquid
metal border. On the copper/fluid tin border, the thermal conductivity (converse to
thermal resistance) is a lot higher than in a nickel/fluid tin interconnection (Fig. 7.5).
This is because some of the main parameters affecting the thermal resistance changes
are different thermal conductivities and different thermal expansions of particular
elements of the interconnection.
The conducted analyses suggest the following future research:
1. to develop the current simplied theory into one, including all the occurring
phenomena to create some more complete models for systems that are more
complex than those analysed in this work
2. to nd the connection between the gas kinetics in the contact layer (air gap) and
the thermal resistance coefcient for different geometric arrangements of PCMs
to be able to dene the thermal resistance of that layer more precisely
3. to conduct additional experimental research on thickness kinetics of a contact
layer (air gap) by observing the latter with a camera of great enlargement capacity.
Annexe 1 (Zalba i in., 2003; Domaski,
Moszyski, 1983; and Others)
PCM material Solidication temperature Solidication heat Heat conductivity

C kJ/kg (W/m deg)
Organic compounds
C6H12O6fructose 141.3 174.2
HOCH2COOH 62.9 109.3
C28H58noctacosane 62.1 253.5 0.267
C17H34 41.3 195.4
C24H48tetracosane 41.0 216.3
C24H38oktadecane 28.0 144.3
C17H36 21.9 214.0
CH3COOH 16.8 200.0
HOCH2CH2OH 12.6 146.5
Humphrey parafn C16 14.0 20.0 200.4
Wax 45.7 47.0 209.3
Sun oil P116
Hydrates
MgNO3 2 6H2 O 90.1 169.0
ZnNO3 2 4H2 O 90.1 159.1
NaOH H2 O 64.5 272.1
FeNO3 2 6H2 O 60.1 125.6
Na2 B4 O7 10H2 O 60.6 283.8
CdNO3 2 4H2 O 59.7 106.3
Na2 C2 H3 O2 3H2 O 57.9 265.2
CaCl2 6H2 O 29.4 173.0
MnNO3 2 6H2 O 26.5 140.4
CuNO3 2 6H2 O 24.6 123.0
NaCrO4 10H2 O 16.3 171.6
Na2 SO4 10H2 O 32.5 250.0
Na2 S2 O3 5H2 O 50.0 209.0
KF 4H2 O 18.0 231.0
Alloys
(continued)

DOI 10.1007/978-3-319-53432-9
130 Annexe 1 (Zalba i in., 2003; Domaski, Moszyski, 1983; and Others)
(continued)
PCM material Solidication temperature Solidication heat Heat conductivity
C kJ/kg (W/m deg)
60Al=40Mg 450.0 349.0
50Mg=50Si 1000.0 873.0
33Mg=67Zn 604.0 260.0
85Al=15Cn 548.0 302.0
46Mg=54Zn 340.0 302.0
52CaCl=48NaCl 500.0 536.0 TJ/m3
52Bi=26Pb=22In 70.0 30.2
49Bi=21In=18Pb=12Sn 58.0 90.9
Salts and their mixtures
KCl 774.0 342.0
NaCl 810.0 493.0
KNO3 323.0 150.5
LiNO3 250.0
NaNO3 321.0 225.0
31Na2 SO4 =13NaCl=16KCl=40H2 O 4.0 235.0
37Na2 SO4 =17NaCl=46H2 O 18.0 233.0
79AlCl3 =17NaCl=4ZnCl2 93 235.0
60AlCl3 =26NaCl=14KCl 103 214.0
46LiF=44NaF=10MgF2 634.6 847.0
Metals and other elements
Feiron 1535.0 270.0
Alaluminium 660.0 395.0
Mgmagnesium 651.0 297.0
Znzinc 420.0 114.0
Nasodium 98.0 116.0
Gegermanium 957.0 476.0 965.0?
Sisilicon 1420.0 1639.0 3605.0?
Inorganic compounds
Substance Solidication Solidication Heat conductivity Density; kg/m3
temperature heat kJ/kg W/m deg
C
H2O 0 333 0.612 998 (liquid, 20 C)
(Liquid 20 C) 917 (solid body, 0 C)
LiCIO3 3H2 O 8.1 253 1720
ZnCl2 3H2 O 10.0
K2 HPO4 6H2 O 13.0
NaOH 3H2 O 15.0
Na2 CrO4 10H2 O 18.0
KF 4H2 O 18.5 231 1447 (liquid 20 C)
1455 (solid body, 18 C)
(continued)
Annexe 1 (Zalba i in., 2003; Domaski, Moszyski, 1983; and Others) 131
(continued)
C
MnNO3 2 6H2 O 25.8 125.9 1738 (liquid, 20 C)
1795(solid body, 5 C)
CaCl2 6H2 O 29 190.8 0.540 1562 (liquid, 32 C)
LiN03 3H2 O 30 296
Na2 SO4 10H2 O 32 254 0.544 1485
Na2 CO3 10H2 O 3236 246.5 1442
CaBr2 6H2 O 34 115.5 1956 (liquid, 35 C)
Na2 HPO4 12H2 O 35.5 265 1522
ZnNO3 2 6H2 O 36 146.9 0.464 1828 (liquid, 36 C)
KF 2H2 O 41.4
KCH3 COO 3=2H2 O 42
K3 PO4 7H2 O 45
ZnNO3 2 4H2 O 45.5
CaNO3 2 4H2 O 47 (42.7)
Na2 HPO4 7H2 O 48
Na2 SO3 5H2 O 48 201 1600 (solid body)
ZnNO3 2 2H2 O 54
NaOH H2 O 58
NaCH3 COO 3H2 O 58 264 1450
CdNO3 2 4H2 O 59.5
FeNO3 2 6H2 O 60
NaOH 64.3 221.6 1690
Na2 B4 O7 10H2 O 68.1
Na3 PO4 12H2 O 69
Na2 P2 O7 10H2 O 70 184
BaOH2 8H2 O 78 265.7 0.653(liquid 85.7 C) 1937 (liquid 84 C)
AlKSO4 2 12H2 O 80
KalSO4 2 12H2 O 85.8
Al2 SO4 3 18H2 O 88
AlNO3 3 8H2 O 89
MgNO3 2 6H2 O 89 162.8 0.490 (liquid 95 C) 1550 (liquid 94 C)
NH4 AlSO4 6H2 O 95 269
Na2 S 11=2H2 O 97.5
CaBr2 4H2 O 110
Al2 SO4 3 Br2 16H2 O 112
MgCl2 6H2 O 117 168.6 0.570 (liquid 120C) 1450 (liquid120C)
MgNO3 2H2 O 130
(continued)
132 Annexe 1 (Zalba i in., 2003; Domaski, Moszyski, 1983; and Others)
(continued)
C
NaNO3 307 172 0.5 2260
KNO3 333 266 0.5 2110
KOH 380 149.7 0.5 2044
MgCl2 714 452 2140
NaCl 800 492 5 2160
Na2CO3 854 275.7 2 2533
KF 857 452 2370
K2CO3 897 235.8 2 2290
References
Alexiades V., Solomon A. D.: Mathematical Modeling of Melting and Freezing Processes,
Hemisphere Publishing Corporation, 1993.
Artyukin E.A., Nenarokomov A.V., Deriving the thermal contact resistance from the solution of
inverse heat-conduction problem, Journal of Engineering Physics, 495499, 1984.
Azouni M. A., Evolution of non linear flow structure with aspect ratio, Int. Com. Heat Mass
Transfer 14 (1987) 44745.
Beck J.V., Determination of optimum, transient experiments for thermal contact conductance,
International Journal of Heat and Mass Transfer, Vol. 12, pp. 621633, 1969.
Benard H., Ann. Chimie Phys., vol. 23, p. 61, 1901.
Betzel T., Beer H.: Solidication and melting heat transfer to an unifexed phase change material
(PCM) encapsulated in a horizontal concentric annulus. Wrme- und Stoffbertragung, 22,
335344, 1988.
Bisshopp F.E., On two-dimensional cell patterns, Journal Of Mathematical Analysis And
Applications, 1, 373385, 1960.
Boussinesq J., Theorie Analytique de la Chaleur, Grauthier-Villars, Paris, 2,172, (1903).
Braszczyski J., Teoria procesw odlewniczych, PWN, Warszawa, 1989.
Bydaek A., Lipnicki Z., Charakterystyka krzepnicia metalu w warunkach wypeniania formy,
Krzepnicie Metali i Stopw, nr 15, s. 7784, 1990.
Bydaek A., Lipnicki Z. Bydaek A., Krzepnicie stopu cyna-ow na pytkach miedzianych w
warunkach dynamicznych, Archiwum Odlewnictwa, nr 14, s. 7681, 2004.
Carslaw H.S, Jaeger J.C.: Conduction of heat in solid, Oxford University Press, 1959.
Cebeci T., Bradshaw P., Physical and computational aspects of convective heat transfer,
Springer-Verlag, NewYork, 1984.
Cengel Y.A., Heat Transfer, A Practical Approach, Mc Graw-Hill, 2003.
Chandrasekhar S., Hydrodynamic and hydromagnetic stability, University of Chicago, Oxford at
the Clarendon Press, 1961.
Cheung F.B., Epstein M., Solidication and melting in fluid flow, Advances in Transport
Processes, A.S.Mujumder and R.A. Mashelkar eds., Wiley Eastern Limited, New Delhi, Vol. 3,
35117, 1984.
Chworinow N.: Krystalizacja i nieodnorodnost stali, Masgiz, Moskwa, 1958.
Crank J., Free and moving boundary problems, Clarendon Press, Oxford, 1984.
Davis S. H., Mueller U., Dietsche C., Pattern selection in single-component systemscoupling
Benard convection and solidication, J. Fluid Mechanics 144 (1984) 133151.
Domaski R., Moszyski J.R., Moliwoci i problemy magazynowania energii cieplnej, Biuletyn
Informacyjny Instytytu Techniki Cieplnej Politechniki Warszawskiej, Nr 62, 2958, 1983.
Edler W., Laminar free convection in a vertical slot, J. Fluid Mechanics 23(1965) 7798.
Edler W., Turbulent free convection in a vertical slot, J. Fluid Mechanics 23 (1965) 99111.
Epstein M., The growth and decay of a frozen layer in forced flow, Int. J. Heat Mass Transfer 19,
12811288, 1976.

DOI 10.1007/978-3-319-53432-9
134 References
Flemings M.C., Solidication Processing, Mc. Graw-Hill Inc., 1974.

Fukusako S., Yamada M., Solidication of pure liquids and mixtures inside ducts and over
external bodies. Applied Mechanics, vol. 47, No 12, part 1, pp. 588621, 1994.
Furmaski P., Floryan J. M., Wall effects in heat conduction through heterogeneous material,
International Journal of Heat and Mass Transfer, 37, 19451955, 1994.
Furmaski P., Winiewski T.S., Thermal contact resistance and other thermal phenomena at
solid-solid interfce, Institute of Heat Engineering, Warsaw University of Technology, Warsaw,
2002.
Hu H., Argyropulos S. A., Mathematical modelling of solidication and melting: a review,
Modeling Simul. Mater Sci. Eng., 4, 371396, 1996.
Ignaszak Z., Lipnicki Z., Bydaek A., Solidication of liquid metal flowing below a cold plate,
Fifth International Pamir Conference on Fundamental and Applied MHD, vol. 2, II131II136,
Ramatuelle, France, 2002.
Jaworski M.: Pomiary temperatury przemiany fazowej i stopnia przechodzenia materiaw
stosowanych do akumulacji ciepa (PCM). Biuletyn Instytutu Techniki Cieplnej Politechniki
Warszawskiej, Nr 88, 312, 2001.
Jeffreys H., The stability of a layer of fluid heated below, Phil. Mag. 2, 83344, 1926.
Kays W., Crawford M., Weigand B., Convective heat transfer and mass transfer, Mc Graw Hill,
Higher Education, 2005.
Kowalski S.J., Lipnicki Z., Natural convective flow between an array of vertical cylinders, Nuclear
Engineering and Design, no 113, 1989.
Kumar R., Kalam M.A., Laminar thermal convection between vertical coaxial isothermal
cylinders, J. Heat Mass Transfer vol. 34, no 2, 513524, 1991.
Krasowski T., Lipnicki Z., Koncepcja wsppracy bloku gazowo-parowego elektrociepowni
``Zielona Gra'' z zasobnikiem ciepa, Oczyszczanie ciekw i przerbka osadw ciekowych.
T. 3. Uniwersytet Zielonogrski, Zielona Gra, s. 299308, 2009.
Lapadula C., Mueller W.K., Heat conduction with solidication and convection boundary
condition at the freezing front, Int. J. Heat Mass Transfer 9, 702705, 1966.
Lienhard IV J.H., Leinhard V J.H., A heat transfer textbook, Phlogiston Press, Cambridge,
Massachusetts, U.S.A,, 2004.
Lipnicki Z., Dynamical solidication of the liquid flowing in a cooling cylindrical channel,
Interactive Dynamics of Convention and Solidication: proceedings of the NATO Advanced
Study Institute Dordrecht : Kluwer Academic Publishers,(NATO Science Series E; Vol.
219)s. 237239, 1992.
Lipnicki Z., Sobich J., Dynamical solidication of not over heat matal ling in a cooling
cylindrical channel, Solidication of Metals and Alloys, vol. 19, pp. 173180, 1994.
Lipnicki Z., Sobich J., Zatverdevanie zidkometalliceskogo potoka v magnitnom pole, Magnitnaja
Gidrodinamika T. 31, no 34, s. 303306, 1995.
Lipnicki Z., Sobich J., The solidifying liquid flow in a cooling plane channel forced by the
changeable pressure, Applied Mechanics and Engineering, Vol. 1, no 2, s. 307316, 1996.
Lipnicki, Z., Weigand B., An experimental investigation of the freeze shut of a convectively cooled
parallel plate channel, Fifth Int. Symp. Thermal Eng. & Sci. for Cold Regions, S. 270275,
Ottawa, Kanada, 1996.
Lipnicki Z., Krzepnicie cieczy przepywajcej w kanale chodzcym, wydanie drugie rozszerzone
z elementami aplikacji odlewniczych, Wydawnictwo Politechniki Zielonogrskiej, Zielona
Gra, 1999.
Lipnicki Z., Dynamical solidication of metal ling a cooled cylindrical channel, Int. Commun
Heat and Mass Transfer 27(5), 689705, 2000.
Lipnicki Z., Ignaszak Z., Lechw H., Przemiana fazowa w materiaach w zastosowaniu do
projektowania akumulatorw ciepa, Archiwum Odlewnictwa, Nr1, 113125, 2001.
Lipnicki Z., Role of the contact layer between liquid and solid on solidication process, Int.
J. Heat and Mass Transfer 46, 21492154, 2003.
References 135
Lipnicki Z., Bydaek A.m SEM and EDX of microstructure of the contact layer in a solidication
process, Microscopy and Analysis no 96, s. 56, 2005.
Lipnicki Z., Weigand B., Bydaek A., On the effect of a variable thermal contact resistance on the
solidication process, Archives of Metallurgy and Materials 50, 10551064, 2005.
Lipnicki Z., Bydaek A., The contact layer between wave tin and copper in a solidication
process, Fundamental and Applied MHD, 7th International pamir Conference, proceedings and
COST P17 Annual Workshop 2008. Presqu'Ile de Giens, Francja, Vol. 2, s. 911915, 2008.
Lipnicki Z., Weigand B., Natural convection flow with solidication between two vertical lled
with a porous medium, Heat and Mass Transfer, 44, 14011407, 2008.
Lipnicki Z., Kuczma M., Numerical solution of the problem of variable thermal contact
resistance, Computer Methods in MechanicsCMM 2009,18th international conference, short
papers. Zielona Gra, 2009.
Lipnicki Z., Stasiek ., Weigand B., An experimental study of asymmetric solidication in a
convectively cooled parallel plate channel, Thermodynamics of Phase Changes -
EUROTHERM-84. Namur, Belgia, [4] CD-ROM, 2009.
Lipnicki Z., Krasowski T., Ukad pracy akumulatora ciepa, patent P. 391082, date of notication:
17.01.2010, received date: 30.04.2014.
Lipnicki Z., Weigand B., Influence of thermal boundary layer on the contact layer between a
liquid and a cold plate in a solidication process, Heat and Mass Transfer, 47, 16291635,
2011.
Lipnicki Z., Opr cieplny warstwy przejciowej w procesie krzepnicia, Zielona Gra, 2012.
Lipnicki Z., Weigand B., An experimental and theoretical study of solidication in a
free-convection flow inside a vertical annular channel, Int. J. Heat and Mass Transfer 55,
655664, 2012.
Lipnicki Z., Sasse D., Weigand B., Experimental and analytical investigation of the solidication
around cooled cylinders subjected to free convection,, Int. J. Heat and Mass Transfer 78,
321329, 2014.
Lipnicki Z., Panto K., Weigand B., Role of the contact layer in continuous casting of thin metal
rods, Archives of Metallurgy and Materials, 59, 167172, 2014.
Lipnicki Z., Panto K., Role of the continuous casting forms on the shape of the solidied crust,
Archives of Metallurgy and Materials, 60, 25532557, 2014.
oc . ., aap opa co, ocyapceoe aeco
aeco o-Maeaeco epayp, Moca, 1962.
Longa W., Krzepnicie odleww, Wydawnictwo lsk, Katowice, 1985.
Loulou T., Artyukhin E. A., Bardon J.P., Solidication of molten tin drop on a nickel substrate,
10th Int. Heat Transfer Conference, Brighton, UK, Vol. 4, pp. 7378, 1998.
Loulou T., Artyukhin J.P., Bardon J. P., Estimation of thermal contact resistance during the rst
stages of metal solidication process: I experiment principle and modelisation, Int. J. Heat
and Mass Transfer 42, 21192127, 1999.
Loulou T., Artyukhin J.P., Bardon J. P., Estimation of thermal contact resistance during the rst
stages of metal solidication process: II experimental setup and results, Int. J. Heat and Mass
Transfer 42, 21292142, 1999.
Madejski J., Solidication in flow through channels and into cavites, International Journal of Heat
and Mass Transfer, vol. 19, pp. 13511356, 1976.
Madhusudana C.V., Thermal contact conductance, Springer-Verlag, New York Inc. 1996.
Mehling H., Cabeza L. F., Heat and cold storage with PCM, Springer-Verlag Berlin Heidelberg,
2008.
Mochnacki B., Suchy J. S., Modelowanie i symulacja krzepnicia odleww, PWN, Warszawa,
1993.
Mutwil J., Bydaek A., On dynamic solidication of metals and alloys, Simulation Designing and
Control of Foundry Processes, FOCOMP 86, Krakw, pp. 339349, 1986.
Neumann S. O., Beer H., Morphology of ice layers in turbulent straight walled diffusers flows, 7th
Int. Symposium on Therm. Eng. Sci. for Cold Regions, Seoul Korea, 2001.
136 References
Nishida Y., Droste W., Engler S., The air-gap formation at the casting-model Interface and t
transfer mechanism through the gap. Metall. Trans. B17, 833844, 1986.
Pellew A., Southwell R.V., On maintained convection motion in a fluid heated from below, Proc.
Roy. Soc. (London) A, 176, 31243, 1940.
Prashant V., Varun, Singal S.K., Review of mathematical modeling on latent heat thermal energy
storage systems using phase-change material, Renewable and Sustainable Energy Reviewes,
12, 9991031, 2008.
Prosnak W.J., Mechanika pynw, PWN, Warszawa, 1970.
Ragone D.V., Adams C.M. Jr., Taylor H.F., Some factors affecting fluidity of metals, Transactions
of the American Foundrymans Society 64, pp. 640652, 1956.
Rayleigh J. W. S., On convection currents in a horizontal layer of fluid when the higher
temperature is on the under side, Phil, Mag. 32, 529-46 rwnie Scientic Papers, 6, 43246,
Cambrige, England, 1916.
Rubinstein L.I., Problem Stefana, Riga, Zvajgznie, 1967 (The Stefan Problem, American
Mathematical Society, 1971, translated from Russian by Solomon A.D.).
Sahai V., Predicting interfacial contact conductance and gap formation of investment cast
alloy718, Journal of Thermophysics and Heat Transfer Vol. 12, No. 4, pp. 562566, 1998.
Saviano J. M., Siegel R., An analytical solution for solidication of a moving warm liquid onto an
isothermal cold wall, Int. J. Heat Mass Transfer 12, 803809, 1969.
Shamsunder N., Sparrow E. M., Analysis of multidimensional conduction phase change via the
enthalpy model, Journal of Heat Transfer, 97, 333340, 1975.
Smirnow W.I., Matematyka Wysza t. IV, rwnania o pochodnych czstkowych, zagadnienia
brzegowe, Warszawa, PWN, 1962.
Sugawara M., Komatsu Y., Beer H.: Melting and freezing around a horizontal cylinder placed in a
square cavity. Heat Mass Transfer, 45:8392, 2008.
Talmon Y., Davis H.T., Scriven L.E., Progressive Freezing of Composites Analyzed by Isotherm
Migration Methods, AIChE Journal, 27, 928937, 1981.
Viskanta R., Phase-change heat transfer in Solar heat storage, Latent Heat Transfer (Edited by G.
A. Lane), CRS Press, Int. Bocaratod, Florida, pp. 153221, 1983.
Viskanta R., Heat transfer during melting and solidication of metals, J. of Heat Transfer 110,
12051219, 1988.
Wang X. G., E.F. Matthyes E.F., Interfacial thermal contact during rapid solidication on
substrate, Proceedings of 10th Int. Heat Transfer Conference, Brighton, UK 4, 169174, 1994.
Wang Y., Amiri A., Vafai K., An experimental investigation of the melting process in a
rectangular enclosure, International Journal of Heat and Mass Transfer 42, 36593672, 1999.
Weaver J.A., Viskanta R., Melting of frozen porous media contained in a horizontal or a vertical
cylindrical capsule, Int. J. Heat Mass Transfer 29, 19431951, 1986.
Weigand W., Braun J., Neumann S.O., Rinck K.J., Freezing in forced convection flows inside
ducts: A review, Heat and Mass Transfer 32, 341352, 1997.
Weigand B, Arnal M. Lipnicki Z., An analytic solution prescribing the time-dependent formation
of a solid crust inside a convectively cooled plane channel, Advances in cold-region thermal
engineering and sciences : technological, environmental and climatological impact :
proceedings of the 6th International Symposium held in Darmstadt, Berlin, Springer,
(Lecture Notes in Physics; 533) - s. 147158, 1999.
Weigand B., Lipnicki Z, Development of the contact layer and its role in the phase change
process, International Journal of Heat Mass Transfer, 93, 10821088, 2016.
Wnuk R., Akumulacja ciepa z przemian fazow zoa w sonecznych systemach grzewczych,
Techniczne, Ekologiczne i Ekonomiczne Aspekty Energetyki Odnawialnej, Warszawa, 2001.
Zalba B., Marin J.M., Cabeza L.F., Mehling H., Review on thermal energy storage with phase
change: materials, heat transfer analysis and applications, Applied Thermal Engineering 23,
251283, 2003.
References 137
Patents:
Prusinski R. C., Johnson T. E, Heat absorbing panel, European Patent Ofce Nr 0003442, 1979.
DeVogel N.Gorges F., J.: Thermo container wall, European Patent Ofce Nr 0 271 118, 1988.
Weber R. M., Rummel Kerrin A: Heat sinks, European Patent Ofce Nr 0 732 743, 1996.
Lukas H., Gasper C.G., Logic J. A., Larrabee S., R., Rogers C. J., Voss M. G., Meissner A. P.,
Frietkiewicz B. J., Latent heat storage device, European Patent Ofce Nr 1 195 571, 2002.
Ival O. S., Water heating unit with integral thermal energy storage, United States Patent Nr 6 047
106, 2000.
Colvin et al.: Thermal energy storage apparatus using encapsulated phase change material,
United States Patent Nr 4 807 696, 1989.
Lipnicki Z., Krasowski T., Ukad pracy akumulatora ciepa, Polska, Nr 39 1082, 2014.

Dynamics of Liquid Solidification - Thermal Resistance of Contact Layer

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Dynamics of Liquid Solidification - Thermal Resistance of Contact Layer

Hochgeladen von

Copyright:

Verfügbare Formate

Mathematical Engineering

More information about this series at http://www.springer.com/series/8445

ISSN 2192-4732 ISSN 2192-4740 (electronic)

Springer International Publishing AG 2017

Printed on acid-free paper

This Springer imprint is published by Springer Nature

5 Solidication in an Annular Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

Zygmunt Lipnicki is the professor of the Institute of

solidication or melting process in heat-storing units, it is possible to reduce the

1.1 The General Solidication Problem

Springer International Publishing AG 2017 1

1.2 Equations Describing the Solidication Problem

In the above equation, an additional constant coefcient of liquid heat conduc-

TL x; y; z; tjF Ts x; y; z; tjF TF ; 1:6

Fig. 1.1 Solidication of a liquid in a space

L iLF iSF 1:8

where T0 is the temperature of the environment, h means coefcient of heat

As it results from the principle of the energy conservation, a change in the

compatible with the condition (1.7).

2.1 Heat Storage Based on the Solidication Process

Heat storage is a subject of considerable importance in power engineering. The

Springer International Publishing AG 2017 7

reservoir PCM - liquid

Fig. 2.1 Heat storage and release

Fig. 2.2 PCM solidication processes

solidication is accompanied by the generation of heat Q12 and flow of entropy

2.1.1 Multidimensional Solidication Heat Transfer

In practice, geometric shapes of applied PCMs are complex and multidimensional

2.1.2 Solutions of the Practical Solidication Problem

Fig. 2.3 Element of heat

The flow of heat in single elements can be modelled as one-dimensional heat

with Tthe temperature and kS the coefcient of heat conductivity through a

2.2 The Geometry of Outer PCMs

(a) (b) (c) R1

Fig. 2.4 Examples of outer geometry of PCMs

2.3 The Criterion for the Selection of Accumulator Design

(a) (b) (c) (d)

Fig. 2.5 Exemplary models of solidication of a flat liquid layer

condition of heat absorption,

3.1 Equations Describing the Problem with

A liquid PCM subjected to solidication in a reservoir is treated as a Newton liquid

where the macroscopic parameter of heat transmission h describes the amount of

with temperature Ts of the solidied layer.

Springer International Publishing AG 2017 15

for Cartesian x, y and z system

for cylindrical r; u; z system

for spherical r; u; / system

Fig. 3.1 Model of H

Thus, the average temperature in the solidied layer will be equal to

and the change of heat accumulated in the solidied layer will be

where T0 is the temperature of the environment, h0 is the coefcient of heat transfer

3.2 Strict and Approximate Solution

In order to conduct a comparative analysis of solutions of the solidication prob-

3.2.1 Strict Solution of the One-Dimensional

After introducing denotations

Fig. 3.2 Solidication of a TF

Equation (3.9) takes the following form

1 hS j~x0 0; s [ 0; 2 hS j~x~d 1; s [ 0; 3 hS js0 1; ~x [ 0;

where C1 and C2 are integration constants.

Finally, the integration constants determined by the boundary conditions 1 and

Next, by means of boundary condition 4, following equation is obtained

which denes solidication constant K in Stefan number Ste function.

For a perfect contact (absence of heat flow resistance), BiCON ) 1, hw ) 0, i.e.

1 hS j~x0 0; s [ 0; 2 hS j~x~d 1; s [ 0; 3 hS js0 1; ~x [ 0;