Beruflich Dokumente
Kultur Dokumente
ZygmuntLipnicki
Dynamics
of Liquid
Solidification
Thermal Resistance of Contact Layer
Mathematical Engineering
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).
Dynamics of Liquid
Solidication
Thermal Resistance of Contact Layer
123
Zygmunt Lipnicki
University of Zielona Gra
Zielona Gra, Lubuskie
Poland
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
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, 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
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.
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
@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
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
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)
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 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
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
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
.
Q
intermedium
T T
1 2 TL 1
TF 2
TS
Q1 2 Q 1 2
S S
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.
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).
10 2 Solidication of PCMs
n2
solid layer
n1 q2
l
q1
PCM
2
cooling liquid
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
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
H R
(d)
L L
L R
R2
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).
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
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
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,
14 2 Solidication of PCMs
r2 Ts x; y; z 0; 3:2
@2 @2 @2
r2 ;
@x2 @y2 @y2
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
0
d
T T T TF : 3:4
d
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
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
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
0
contact layer q T TW
3.2 Strict and Approximate Solution of the One-Dimensional 19
@hS 1 @ 2 hS
: 3:10
@s Ste @~x2
The limiting conditions (boundary and initial ones) for the solidied layer are as
follows:
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
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
1
C1 0 and C2 :
erf K
20 3 The Model of Solidication of a Liquid with the Contact Layer
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
Ste
p K exp K 2 erf K; 3:15
p
~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
3.2 Strict and Approximate Solution of the One-Dimensional 21
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
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 ;
Fig. 3.3 Distribution of temperature in the solidied layer and the cold wall
22 3 The Model of Solidication of a Liquid with the Contact Layer
- 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
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
3.2 Strict and Approximate Solution of the One-Dimensional 23
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
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
24 3 The Model of Solidication of a Liquid with the Contact Layer
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
cooling liquid
l
insulation
PCM element xb
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
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
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
~d 0; s 0: 4:5
s
2
~ 1 1 1 1
d 2s 4:6
Bi BiCON Bi BiCON
28 4 Solidication on a Rectangular Geometrics
(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
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.
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).
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
30 4 Solidication on a Rectangular Geometrics
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
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
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
@u @u ddl @T @T ddt
ii ; hh 1 and ii ; hh 1:
@y @x dx @y @x dx
where h and ~
u are presented by expressions in square brackets.
32 4 Solidication on a Rectangular Geometrics
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
~
~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
4.2 Forced Convection Effects 33
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
# @ ~d
Nu ; 4:22
~d @s
N u / Re
1
Pr
0
0 1 2 3 4 5
34 4 Solidication on a Rectangular Geometrics
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
~
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
4.2 Forced Convection Effects 35
(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
(Lipnicki and Weigand 2011)
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
36 4 Solidication on a Rectangular Geometrics
(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)
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
4.2 Forced Convection Effects 37
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)
0.06
steady state =0.01
0.04
=0.013
0
=0.001 x
0 0.2 0.4 0.6 0.8 1.0
38 4 Solidication on a Rectangular Geometrics
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.
Cu
d H
TF cold
plate
frozen T
layer
T
TW adiabatic
liquid . surface
q
0 x
Rys. 2.12
40 4 Solidication on a Rectangular Geometrics
dd TF T TF T
qs L t h T 1 T F k s ; ks hCON T TW : 4:33
d d d
L dd h
T TW qs t T1 TF :
hCON d hCON
d x T T0 TW T0
s Fo Ste; ~d ; ~x ; h ; hW ;
H H T F T0 TF T0
t js c s TF T 0 H hCON
Fo ; Ste ; BiCON ;
H2 L ks
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
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
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
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
42 4 Solidication on a Rectangular Geometrics
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
1 1
b ~2 ~dmax
2 : 4:41
dmax ~kSte ~kSte
1
1
2 BiCON 2 ~
j BiCON ~
j
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
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
@w
w0; s 0 and w~x; 0 0; 1; s 0: 4:46
@~x
X
1 j~ 2
w~x; s An eStekn s ebs sinkn ~x: 4:47
n1
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
2b
An ~
j
:
kn kn2 Ste b
Finally, the exact analytic solution of Eq. (4.35) takes the form of innite series
(Lipnicki 2003)
44 4 Solidication on a Rectangular Geometrics
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
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
4.2 Forced Convection Effects 45
0.4
1/BiCON = 1 1/BiCON = 2
0.2
0
0 0.1 0.2 0.3 0.4 0.5
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.
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
Fig. 4.16 Temperature distribution and development of the solidied layer for different thermal
resistances of the contact layer (Lipnicki 2003)
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
PCM
frozen
layer Tf
T1
TF
warm
T wall
TW
contact u2
layer liquid
u1
cold x
wall
H
2
@T @T @ T @2T
u1 u2 j 2 ; 4:50
@x @y @x2 @y
@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.
48 4 Solidication on a Rectangular Geometrics
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
d 2 ~u
Ra h P;
d~x2 4:53
d2h
0:
d~x2
Z1
~ud~x 0: 4:54
~d
~x ~d; ~u 0; h 0 i ~x 1; ~u 0; h 1: 4:55
h C1~x C2 ; 4:56
1 ~d
h ~x : 4:57
1 ~d 1 ~d
!
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
50 4 Solidication on a Rectangular Geometrics
@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).
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
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
dT
t ar2 T: 4:64
d
52 4 Solidication on a Rectangular Geometrics
~ 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
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
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)
DW ~z 2p 2p
~u1 sin ~x;
a2 ~l ~l
2p
~u2 0; a ;
~l
2p
~u3 W ~z cos ~x;
~l
54 4 Solidication on a Rectangular Geometrics
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 Free Convection Effect 55
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
56 4 Solidication on a Rectangular Geometrics
RaH gbT1 jm
T2 H 3
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.
4.3 Free Convection Effect 57
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
q L dd h
T TW s Tf TF 4:74
hCON dt hCON
~ 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
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
0.2
0
0 0.2 0.4 0.6 0.8 1
0.2
0
0 0.2 0.4 0.6 0.8 1
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.
4.3 Free Convection Effect 59
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
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 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
0.4
Nu =1, A=0.5
Nu =2, A=0.5
0.2
0
0 1 2 3
4.3 Free Convection Effect 61
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
g b Tf TF H d3
RaHd :
mj
62 4 Solidication on a Rectangular Geometrics
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
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.
PCM
PCM
PCM
PCM
PCM
PCM
TF
TO
TW
cooling frozen
liquid layer
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
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
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
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
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
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
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.
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
68 5 Solidication in an Annular Space
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
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
70 5 Solidication in an Annular Space
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 :
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
q L dd h
T TW s T f TF : 5:15
hCON dt hCON
5.2 Solidication of a Superheated Liquid 71
~d d ; ks TF TF
B ; Nu
hR
;
R kL Tf TF kL
5:16
hCON R
BiCON ; s SteFo;
ks
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
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
Z~d Z~d
B dd B dx
2 : 5:21
Nu2 A d ln d Nu Aexx
1 1
72 5 Solidication in an Annular Space
(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)
5.2 Solidication of a Superheated Liquid 73
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
1
ds d~d: 5:24
BBiCON Nu
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)
74 5 Solidication in an Annular Space
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.
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)
76 5 Solidication in an Annular Space
. .
water g
test
section inner copper
ice layer cylinder
water flow
ilumination
Fig. 5.10 Components of the annular channel, a two sleeves; b top view (Lipnicki and Weigand
2012)
5.3 The Experiment 77
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.
ice layer
subsection 1
g
water
H subsection 2
subsection 3
R
78 5 Solidication in an Annular Space
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
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.
80 5 Solidication in an Annular Space
(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)
(Lipnicki and Weigand 2012)
(a) (b)
1.06 1.06
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)
5.3 The Experiment 81
(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
(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)
0.3
0.2
0.1
TW , 0C
0
-12 -10 -8 -6 -4 -2 0
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 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.
cooling medium
solidified liquid(PCM) front of stream
PCM
cylindrical channel
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
86 6 Solidication of a Liquid Flowing into the Channel
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
Ur 0 ; t Ct r 2 r 02 ; 6:6
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
d 2
~r ~x w P ~r 2 2 C 8 ~x w 6:9
d~t
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.
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
88 6 Solidication of a Liquid Flowing into the Channel
TO
y solid
TF layer
0
r(t) R x
liquid
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
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
Zr
QF 2 p r dr q L 1; 6:15
R
where
4 Ste b
a p 6:17
p Pr
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
where
16 2 16 a2
u ~r ln ~r 2
a2 a2
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.
92 6 Solidication of a Liquid Flowing into the Channel
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
Zt Zt
Q hCH 2 p R T T0 dt 1 hCON 2 p R TF T dt 1:
0 0
6:23
hCH T0 hCON TF
T 6:24
hCH hCON
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
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
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:
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.
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
Springer International Publishing AG 2017 97
Z. Lipnicki, Dynamics of Liquid Solidication, Mathematical Engineering,
DOI 10.1007/978-3-319-53432-9_7
98 7 Role of the Contact Layer in a Solidication Process
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
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
100 7 Role of the Contact Layer in a Solidication Process
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
40
thermal conductivity
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
~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
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
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
(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.
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
104 7 Role of the Contact Layer in a Solidication Process
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)
7.2 A Theoretical Analysis of Thermal Contact Resistance 105
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.
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
0 C d
7.3 Development of the Contact Layer During the Solidication 107
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
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.
108 7 Role of the Contact Layer in a Solidication Process
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
dC qF qW q
s 2 W 1 : 7:15
d qW qF qF qF
dC
s 1:63044 103 : 7:16
d
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
0.4 0.004
C
0.2 0.002
0
0 0.002 0.004 0.006 0.008 0.01
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
110 7 Role of the Contact Layer in a Solidication Process
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
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
7.3 Development of the Contact Layer During the Solidication 111
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.
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.
112 7 Role of the Contact Layer in a Solidication Process
R
C
cylinder
TL
TF
TW
T
q
Zd
m 2p qS d0 dd0 p d2 R2 qF ; 7:21
dC
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).
7.3 Development of the Contact Layer During the Solidication 113
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.
114 7 Role of the Contact Layer in a Solidication Process
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)
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
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
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
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
0.4
k C = 0 .0 0 8 7
0.2
0
0 0.002 0.004 0.006 0.008 0.01
0
0 0.002 0.004 0.006 0.008 0.01
7.3 Development of the Contact Layer During the Solidication 117
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.
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
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
4L 1L
Tp
0
q0
x=
1
x=
czynnik
niskowrzcy
TF
Tp
front
krzepnicia q0
d dt
122 8 Phase Heat Accumulator
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
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;
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
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;
H 0:05
Qci q L Fi 900 233:3 1 kJ 5:25 MJ:
2 2
H2 0:052
tc sc 0:25s 40; 484 s;
Ste js 0:0996 1:55 107
is
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
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
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)
(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)
Substance Solidication Solidication Heat conductivity Density; kg/m3
temperature heat kJ/kg W/m deg
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)
2195 (solid body, 24 C)
Na2 HPO4 12H2 O 35.5 265 1522
ZnNO3 2 6H2 O 36 146.9 0.464 1828 (liquid, 36 C)
1937 (solid body, 24 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)
Substance Solidication Solidication Heat conductivity Density; kg/m3
temperature heat kJ/kg W/m deg
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.
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.