Numerical Investigation of Second Sound in Liquid Helium

NUMERICAL INVESTIGATION OF
SECOND SOUND IN LIQUID HELIUM
by
Udo Piram
A Thesis Submitted to the Faculty of the

DEPARTMENT OF ELECTRICAL AND COMPUTER ENGINEERING
In the Graduate College
THE UNIVERSITY OF ARIZONA
In Partial Fulfillment of the Requirements
For the Degree of
DIPLOMINGENIEUR
at the
UNIVERSITY STUTTGART
1991
ACKNOWLEDGMENTS
I would like to thank Dr. François Cellier for guidance and assistance during my studies
at the University of Arizona and the preparation of this paper.
I would also like to thank Prof. Dr. Voß at the University Stuttgart, who organized
the exchange program, which made it possible for me to attend school at the University of
Arizona. I would like to thank the German Academic Exchange Service for the financial
support of the exchange program.
Finally, I would like to express my gratitude to my parents for their support of my
studies in Germany and the USA.
Contents
Contents ii
List of Figures iii
1 Introduction 1
1.1 Mechanical and Electrical Systems . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Thermodynamical Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2 Helium II and Second Sound 4
3 Heat Conduction Modeled With Time–Variant R–C–Networks 7

3.1 R–C–Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.2 Modeling in Terms of Bond Graphs . . . . . . . . . . . . . . . . . . . . . . 9
3.3 Helium II Modeled in Terms of Bond Graphs . . . . . . . . . . . . . . . . . 10
3.4 Numerical Values for the Thermal Parameters . . . . . . . . . . . . . . . . 12
3.4.1 Model 1: Initial Temperature 1.4K . . . . . . . . . . . . . . . . . . 13
3.4.2 Model 2: Initial Temperature 2.1728K . . . . . . . . . . . . . . . . 14
3.5 Simulation and Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
4 Heat Transport and Two Fluid Model 20

4.1 Two Fluid Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.2 Hydrodynamical Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
4.3 Numerical Values for the Thermal Parameters . . . . . . . . . . . . . . . . 23
4.4 Numerical Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
i
CONTENTS ii
4.5 Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

4.6 Simulation and Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
5 Conclusions 29
5.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
5.2 Zusammenfassung . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
A DYMOLA Programs 31
A.1 Model 1: Initial Temperature 1.4K . . . . . . . . . . . . . . . . . . . . . . 32
A.2 Model 2: Initial Temperature 2.178K . . . . . . . . . . . . . . . . . . . . . 37
B ACSL Program 42
Bibliography 50
List of Figures
1.1 Mass–Spring Oscillator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
2.1 Phase Diagram of Helium (schematic) . . . . . . . . . . . . . . . . . . . . . 4

2.2 Specific Heat Capacity of Helium . . . . . . . . . . . . . . . . . . . . . . . 5
3.1 Bond . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.2 0– and 1–Junction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.3 Bond Representation of Heat Conduction . . . . . . . . . . . . . . . . . . . 11
3.4 Subsystem Cell of the Heat Conduction . . . . . . . . . . . . . . . . . . . . 12
3.5 Bond Representation of Heat Conduction with Subsystems . . . . . . . . . 12
3.6 Heat Propagation in Model 1 . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.7 Heat Propagation in Model 2 . . . . . . . . . . . . . . . . . . . . . . . . . 19
4.1 Normal and Superfluid Fractions of Helium II . . . . . . . . . . . . . . . . 21

4.2 Discretization of the Helium Channel . . . . . . . . . . . . . . . . . . . . . 24
4.3 Heat Propagation in Two Fluid Model . . . . . . . . . . . . . . . . . . . . 27
iii
Abstract
At temperatures below 2.17K the phenomenon of thermal waves has been observed in
liquid Helium. This paper describes the reproduction of these waves, called “Second
Sound,” by means of simulation. Two different models for the heat transport in Helium
are tried. First the classical heat diffusion equation is used with temperature dependent
thermal parameters. The second model for heat transport is based on the more complex
Two Fluid Model which was introduced for Helium close to the absolute zero point of
temperature. Simulations with various initial and boundary conditions show that the
first model is not sufficient to describe heat transport in Helium, whereas the second
model gives satisfactory results.
Abstract
Das Phänomen von thermischen Wellen wurde in flüssigem Helium unter 2.17K
beobachtet. Diese Arbeit beschreibt die Reproduktion dieser Wellen, die ,,Second Sound“
genannt werden, durch Simulation. Zwei verschiedene Modelle für den Wärmetransport
wurden angewandt. Als erstes wurde die klassische Wärmeleitgleichung mit tem-
peraturabhängigen thermischen Parametern verwendet. Das zweite Modell für den
Wärmetransport basiert auf dem komplizierteren ,,Zwei–Flüssigkeiten“–Modell, das für
Helium nahe dem absoluten Nullpunkt eingeführt wurde. Simulationen mit verschiede-
nen Anfangs– und Randbedingungen zeigen, daß das erste Modell den Wärmetransport in
Helium nicht zufriedenstellend beschreibt, während das zweite Modell zufriedenstellende
Ergebnisse liefert.
CHAPTER 1. INTRODUCTION 2
stop, but would never reverse its direction, no oscillation is possible. For all times, the
energy of the system is stored in the mass. If we built a chain consisting of dampers and
masses and gave every mass an initial velocity, again the same behavior could be observed;
the masses coming to a stop somewhere, but never changing their direction. Hence, two
different kinds of storage elements are needed for linear oscillations to occur.
The same experiments can also be made with electrical systems, where the capacitor
and the inductance are storage elements and the resistor is the dissipative element.
1.2 Thermodynamical Systems

Thermodynamical Systems are commonly described in terms of temperature and heat
or entropy currents. The thermal capacitance of a material describes the capability of
storing energy in the medium. The overall flow of entropy into an element determines
the increase of temperature, so heat is stored. The relation of entropy flow between two
elements with different temperatures is described as thermal conduction. This flow is a
resistive effect.
We have now one storage element which can be classified as capacitive. There is no
inductive storage element in thermal systems. Let us assume that such an element exists,
this would mean that we can store entropy flow. Therefore, there might be an entropy
flow even without a temperature difference. This is in contradiction to the second law of
thermodynamics which states that a spontaneous heat or entropy flow is always directed
from a point of higher temperature to a point of lower temperature. Hence, if there
is an entropy flow, there must be a temperature difference. Since the previously made
assumption is wrong, there is no thermodynamic inductive element.
In 1.1, we found that there have to be two different kind of storage elements in a
system in order for it to oscillate. Since thermodynamical systems only have a capacitive
storage element, we conclude that they cannot oscillate. However, in 1944 Peshkov was
able to detect temperature waves which he called Second Sound [1, 2]. A tube filled with
Helium is cooled down to a value below a critical temperature Tλ = 2.17K, where it is in
a liquid state. Periodical heating of one end of the tube generates heat waves which can
CHAPTER 1. INTRODUCTION 3
be detected with a thermometer.

This phenomenon seems to contradict the results of the previous paragraph on the first
view. But we must remember that all derivations in 1.1 were done under the assumption of
linear time–invariant systems. The physical properties of Helium in this low temperature
range depend very much on the temperature so that we no longer deal with time invariant
systems. Additionally the equations describing the storage and dissipative elements are
nonlinear.
CHAPTER 2. HELIUM II AND SECOND SOUND 6
published by Vinen describes the basic physical properties of Superfluid Helium also suit-
able for newcomers [6]. A more recent work is Puttermann’s Superfluid Hydrodynamics,
which covers both theory and effects in a complete form [7].
The research in Low Temperature Physics grew almost exponentially, so that there
is now a huge number of papers. Therefore my selection is not at all comprehensive but
rather represents those parts of research I was interested in, namely phenomenological and
quantitative description of thermal behavior and numerical solution of heat transport in
Helium II. Unfortunately I was not able to find a single source which provided me with all
numerical values for physical properties over the Temperature range of interest. So I had
to gather my data from various papers, each describing the measurement of a parameter
in a small temperature range and approximating it through a function. For temperatures
sufficiently smaller than Tλ , I used papers by Kramers [8] and Vinen [9]. Around Tλ the
behavior of the specific heat curve is well documented by Buckingham [10] and Ferrell
[11]. For the parameters describing the thermal conductance, several equations can be
found for either just below or above the critical temperature [12, 13, 14, 15, 16]. I was
able to find papers about numerical analysis of heat propagation in Helium II from two
authors, namely Fiszdon [17] and Gentile [18, 19]. I will later use Gentile’s results for
comparison with my simulations.
Chapter 3
Heat Conduction Modeled With

Time–Variant R–C–Networks
3.1 R–C–Networks
As I mentioned in the introduction, thermodynamical systems have only one energy stor-
age element. The electrical analog of this type of systems is a network consisting merely of
resistors and capacitors. For simplicity, from now on I shall call every system of this type
R–C–network. Since heat transfer is described by a partial differential equation, we are
confronted with a distributed parameter system, whereas classical mechanical or electrical
systems are lumped parameter systems. By dividing the continous space into finite ele-
ments, we obtain ordinary differential equations for the heat conduction, which have the
same form as the equations for mechanical and electrical R–C–networks. As for Helium at
low temperatures the specific heat capacity and the thermal conductance are functions of
the temperature itself, the system is time–variant, even though the temperature doesn’t
explicitly appear in the equations.
To demonstrate that there are differences between time–variant and time–invariant
systems let me give an example. Consider the second–order system
ẋ = A(t)x (3.1)
7
CHAPTER 3. HEAT CONDUCTION MODELED WITH R–C–NETWORKS 8
where ⎛ ⎞
4a −3ae8at
A=⎝ ⎠. (3.2)
ae−8at 0
The eigenvalues of the system can be easily found to be time–independent. They are
λ1 = a (3.3)
λ2 = 3a (3.4)
The solution of this linear time–variant system is
x1 = 1.5(x10 + x20 )e5at − 0.5(x10 + 3x20 )e7at (3.5)

x2 = 0.5(x10 + 3x20 )e−at − 0.5(x10 + x20 )e−3at (3.6)
with the initial condition as ⎛ ⎞

x10
x=⎝ ⎠ (3.7)
x20
If we choose a < 0, both eigenvalues are in the left half of the λ–plane. A linear time–
invariant system is stable if and only if all eigenvalues have negative real parts. However,
from (3.6) we see that the solution is unstable although the eigenvalues of the system are
all negative. We see that we can’t simply transfer a theorem from linear time–invariant
system theory to time–variant systems. If we take the theorem stating that linear time–
invariant systems can only oscillate if they have complex eigenvalues, we can make again
no conclusion about the ability of a nonlinear and time–variant system to oscillate. This
observation encourages us to try to find a model for the heat transfer in Helium II which
has only real eigenvalues but which can be nonlinear and time–variant. By simulating
this system with an experiment it can be detected if the system is capable of oscillating.
We model the heat propagation in Helium II with the normally used heat diffusion
equation letting the specific heat capacity and the thermal conductance be functions of
the temperature. Without reducing generality, we can consider only the one–dimensional
case, i.e. heat conduction through a Helium filled channel with constant variables over
every cross–section.
∂T ∂ 2T
ρcp (T ) = λ(T ) (3.8)
∂t ∂x2
λ = thermal conductivity
ρ = density
cp = specific heat capacity
If we discretize the space into finite elements of length ∆x and if we use the central–
difference formula [21] for the second spatial derivative of T, we get a set of ordinary
differential equations.
dTi λ Ti−1 − 2Ti + Ti−1
= (3.9)
dt ρcp (∆x)2

1 Ti−1 − Ti Ti − Ti+1
= λA − λA (3.10)
A∆xρcp ∆x ∆x
1
= (Q̇in − Q̇out ) (3.11)
A∆xρcp
where
A = cross section of Helium channel

∆x = length of an element
Q̇in = heat current into the element
Q̇out = heat current out of element
3.2 Modeling in Terms of Bond Graphs

The bond graph technique is a tool to graphically represent a physical system. The part
of the graphs giving name to the technique are the bonds that simultaneously connect
two variables of two elements. The two variables that are assigned to each element are
called adjugate variables. One of them is called the “effort” e and the other the “flow” f .
The product of e and f has the dimension of a power, i.e., the bond basically represents
a flow of power. Figure 3.1 shows the bond.
Other basic elements are the junctions of which two different kinds exist, the so–called
“0–junction” and the “1–junction”. At the 0–junction all effort variables are equal, and
the sum of all flow variables adds up to 0. At the 1–junction all flow variables are equal,
and the sum of all effort variables adds up to 0 [Fig 3.2].
3.4.1 Model 1: Initial Temperature 1.4K

The formulas used in this text for the specific heat capacity at constant pressure, cp , are
actually formulas for the specific heat under saturated vapor pressure csat . However, for
T < 2.5K the difference is less than 1%, which can be neglected here [5]. Kramers verified
through experiments an empiric formula found earlier by Hull [8, 20]. They approximated
csat between 0.6 and 1.4K as
J
cp = csat = 100 T 6.2 . (3.14)
kgK
The density of Helium can be considered constant in the temperature range according to
[5]
kg
ρ = 145.5 . (3.15)
m3
The length of the elements and the crosssection of the channel were chosen to be
∆x = 1 · 10−4 m (3.16)
F = 1 · 10−4 m2 . (3.17)
If we define the entropy flow Ṡ of a body as the heat flow Q̇ divided by its temperature
T , which is not the original definition of entropy, but which is convenient and usual in
the bond graph technique [21],
Q̇
Ṡ = (3.18)
T
we find from
dT
∆Q̇ = ρF ∆xcp (3.19)
dt
ρF ∆xcp dT
∆Ṡ = . (3.20)
T dt
If we plug in (3.14 – 3.17) we obtain (in SI–Units)
dT
∆Ṡ = 1.455 · 10−4 T 5.2 . (3.21)
dt
This is the equation describing the modulated capacitance, mC, of the system.
For the heat current as a function of T and dT , London gives an approximation

deduced from his Two Fluid Model [4]. The equation is valid for large heat currents or
large slits as in our case. The formula was later verified through measurements by Vinen
[9]. The heat current density q̇ is given as
1 1 5.6 5.6
dT 3
ρ2 s4λ 3
T T
q̇ = − T 1− (3.22)
dx A Tλ Tλ
where
Tλ = 2.173K. (3.23)
sλ is the specific entropy at T = Tλ , which can be found to be [8]
J
sλ = 1.57 · 103 . (3.24)
kgK
A is the so–called Gorter–Mellink constant, which is a function of temperature but which
we assume to be constant in our case [9, 5].
ms
A = 350 (3.25)
kg
If we replace dT /dx through ∆T /∆x and rewrite the equation for Ṡ, we obtain
1 5.6 5.6
ρ2 s4λ 3
T T 1
Ṡ = −F 1− (∆T ) 3 . (3.26)
A∆x Tλ Tλ
Plugging in numerical values we obtain finally (in SI–Units) an equation describing the
conduction of heat in Helium II
5.6 5.6
T T 1
Ṡ = −154.32 1 − (∆T ) 3 . (3.27)
2.173 2.173
3.4.2 Model 2: Initial Temperature 2.1728K

As a second case, I investigated the heat conduction in the region around Tλ . The specific
heat capacity in the vicinity of this point is described by Ferrell through the empiric
formula [11]:

2 J
cp = 0.626 ln − 2sign(τ ) (3.28)
|τ | gK
T − Tλ
τ =
Tλ
|τ | < 5 · 10−2
For this second case we choose
∆x = 1 · 10−3 m (3.29)
F = 1 · 10−4 m2 (3.30)
kg
ρ = 146.57 3 . (3.31)
m
If we use (3.18) and plug in 3.28 we obtain (in SI–Units)

0.19076 2Tλ dT
∆Ṡ = ln − 2sign(T − Tλ ) (3.32)
T T − Tλ dt
as equation for the capacitance of the system.
For the heat conduction we must consider two different cases, depending on whether
the temperature is above or below Tλ . Below Tλ the heat conduction is described by
Crooks, who gives for the heat flow density q̇ the formula (all in SI–Units) [16]:
1
3.5
sλ gradT 1
q̇ = −ρs sλ T (Tλ − T )− 5.83 (3.33)
263ρn
with
2
ρs = 1.43(Tλ − T ) 3 ρ (3.34)
J
sλ = 1.57 · 103 (3.35)
kgK
Using the numerical values (3.29 – 3.31) we obtain for the entropy flow as function of the
temperature difference
1
Ṡ = −94.9(Tλ − T )0.838 (∆T ) 3.5 . (3.36)
For temperatures above Tλ , Ferrell gives the thermal conductivity of Helium as

1 1
2 3 2 2
λ = λ0 ln (3.37)
τ τ
with
T − Tλ
τ =
Tλ
W
λ0 = 1.36 · 10−3 (3.38)
mK
Using the relation

F ∆T
Ṡ = − λ (3.39)
T ∆x
we obtain
13
−1.36 · 10−4 2Tλ 2Tλ
Ṡ = − ln ∆T. (3.40)
T T − Tλ T − Tλ
The heat conduction for T < Tλ decreases with increasing temperature below Tλ until
it reaches 0 at Tλ . Above that point the conduction increases again but stays by many
orders smaller than below Tλ , which is not surprising since Helium II is known to be
superconductive for heat.
3.5 Simulation and Results

For the simulation I coded the problem in DYMOLA [21, 22]. DYMOLA is not a simula-
tion language but a program generator which supports different simulation languages. It
handles very easily several hierarchically structured layers of subsystems and is also well
suited for systems modeled in bond graphs since the built in “nodes” are similar to the
junctions in bond graphing. I used the VAX/VMS version of DYMOLA, which can gener-
ate programs in DESIRE [23]. DESIRE, which stands for “Direct–Executing–SImulation
in REal time,” is a high–speed direct–executing interactive simulation program with an
on–time output of colored graphs on the screen. It supports different Runge–Kutta meth-
ods and Adams and Gear rules as integration algorithms. The DESIRE programs were
executed on VMS/VAX and on a PC–compatible computer. The integration algorithm
used was Euler forward with a fixed step size of ∆t = 0.5 · 10−7 s for the first model, re-
spectively ∆t = 1.0 · 10−5 s for the second model. Appendix A shows the used DYMOLA
programs.
As heat source I chose for the first model a value of Q̇ = 0.2W for τ = 2 · 10−5 s, which
equals an entropy flow of Ṡ = 0.143W/K. Fig 3.6 shows the temperature of the first 4
space elements over time with an initial temperature of T0 = 1.4K. The heat propagates
in the medium in a dissipative manner, the temperature of an element never exceeds the
temperature of its neighbor to the left. Fig 3.7 shows the results of the second model
with an initial temperature of T0 = 2.1728K just below Tλ . Again the temperatures of
the first 4 elements are shown. The heat flow is Q̇ = 0.1W and the resulting entropy flow
Ṡ = 0.046W/K for τ = 1 · 10−3 s. The heat propagation through Helium in this case is
even slower as before and only the first element exceeds Tλ . For both cases I conducted
simulations with various entropy sources in respect to heat flow and time. I used also
space elements of different lengths between 10−5 m ≤ ∆x ≤ 10−2 m. In none of all these
combinations, a behavior different from the shown diffusive behavior could be observed,
especially did no solution resemble a wave–like heat propagation.
Chapter 4
Heat Transport and Two Fluid

Model
4.1 Two Fluid Model

In an attempt to explain the remarkable properties of Helium II a model was presented
by two different authors in 1941 [24, 25]. The basic idea of the model is that below Tλ
Helium consists of two different liquids which are ideally mixed. One of them is called
the “superfluid component” with the density ρs , the other the “normal component” with
the density ρn . Both partial densities add up to the total density:
ρs + ρn = ρ (4.1)
The fraction of ρs /ρn is a function of the temperature; at Tλ there is no superfluid compo-

nent present whereas for T = 0K the normal component vanishes. Fig 4.1 shows the two
densities plotted over the temperature. Another assumption of the model is that at any
point in space each component has its own velocity called vs respectively vn . Therefore
we can write for the mass current density
J = ρs vs + ρn vn (4.2)
The normal component behaves like an ordinary liquid, it has the viscosity νn . The
superfluid component has a viscosity coefficient equal to zero, it flows without friction.
20
CHAPTER 4. HEAT TRANSPORT AND TWO FLUID MODEL 22
Sound. It can be produced by an oscillating heat source that causes the two components
to oscillate in antiphase.
Let us consider the case of constant heat transfer in a Helium II filled channel. We
apply at one end of the channel a heat source that transforms the superfluid component
into the normal component. This normal component flows away from the source to a
heat sink at the other end of the channel, where it is transformed back to the superfluid
component again. This component flows then back in the other direction. So we have
a steady flow of the two components flowing in opposite directions. This mechanism of
heat transport called superconduction of heat can be interpreted as Second Sound with
the frequency 0.
4.2 Hydrodynamical Equations

The dynamics of Helium II are described by the following conservation equations: con-
servation of mass, conservation of momentum for both components and conservation of
energy. A general form of these equations can be found in [26]. In our case we can
simplify the model in two ways. Without loss of generality we will consider only the one–
dimensional case. Assuming heat transport with a zero net mass flow, we obtain from
(4.4)
ρn
vs = − vn . (4.5)
ρs
Then the normal velocity vn and temperature T describe completely the behavior of
Helium II, and the following equations can be obtained [18]:
3 2
∂T K ∂2T ∂T s ∂vn A(T )ρn ρ A0 ∂vn
= 2
− vn − T+ vn4 + (4.6)
∂t ρcp ∂x ∂x cp ∂x cp ρs cp ∂x
⎛ 2 ⎞
∂vn ρs ∂ 2 v n ∂vn ρ ρs ∂T
= A0 2
− v n ⎝3 + A(T ) ρvn2 ⎠ − s (4.7)
∂t ρn ∂x ∂x ρs ρn ∂x
with the thermodynamic parameters:
K = dissipation
cp = specific heat capacity
s = specific entropy
A(T ) = Gorter–Mellink constant
A0 = viscosity
The two equations are strongly coupled by the presence of ∂vn /∂x in the equation for T
and of ∂T /∂x in the equation for vn . Additionally s, A(T ), cp , ρs and ρn depend on the
temperature and contribute to the coupling of the equations.
4.3 Numerical Values for the Thermal Parameters

For the simulation I followed for a great part the description of Gentile’s numerical in-
vestigations on heat transport in Helium II [18]. The initial temperature and the point
around which the parameters are approximated is T0 = 1.8K. The parameters ρ, K and
A0 are considered to be constant in the used temperature range.
kg
ρ = 145.3 (4.8)
m3
W
K = 5 · 10−4 (4.9)
K
A0 = 2.6 · 10−8 m (4.10)
For the entropy s, the heat capacity cp and the density of the normal component we find
for T > 1.4K [18, 4]:
5.6
T
s = sλ (4.11)
Tλ
J
sλ = 1.57 · 103 (4.12)
kg K
5.6
T
ρn = ρ (4.13)
Tλ
5.6
T
cp = 5.6sλ (4.14)
Tλ
For the Gorter–Mellink constant A(T ) the formula of Crooks was used, which gave good
results in comparison with measured data [16, 5].
A(T ) = 263(Tλ − T )−0.74 (in MKS units) (4.15)


∂Ti+1/2 Ti+1,t − Ti,t
= (4.21)
∂x t
∆x
Ti+1,t + Ti,t
Ti+1/2,t = (4.22)
2 2
∂ vi+1/2 vi+3/2,t − 2vi+1/2,t + vi−1/2,t
= (4.23)
∂x2 t
2∆x2
4.5 Boundary Conditions

At the left end of the channel, at x = 0, the velocity of the superfluid component is
assumed to be zero.
v1 = 0 (4.24)
In the differential equation for ∂v1 /∂t forward–difference formulas have to be used for the
space derivatives instead of central–difference formulas [21]. Using (4.24) we obtain:

∂v3/2 3v3/2,t + 3v2,t − v5/2,t
= (4.25)
∂x t 6∆x
2
∂ v3/2 −2v3/2,t + v2,t
2
= (4.26)
∂x t
∆x2
At the same end, the heat flow q̇0 is introduced into the channel. To obtain an equation
for the temperature T1 of the first space element, an energy balance in the first node
can be made [18]. For this element the dissipation K and the viscosity A0 have been
neglected. Using (4.24) we obtain:
∂T3/2 q̇0 s
= − T2 v2 (4.27)
∂T ρcp 2∆x cp 2∆x
T1 = 2T3/2 − T2 (4.28)
At the right end of the channel, x = l, the boundary conditions are not so important
since the simulation ends as soon as the temperature front reaches this end. We assume
symmetry conditions, so that we can write:

∂v
= 0 (4.29)
∂x x=l

∂T
= 0 (4.30)
∂x x=l
For the second derivatives we will use the same difference formulas with the symmetry
condition:
Tn+i = Tn−i (4.31)

vn+i = vn−i (4.32)
4.6 Simulation and Results

The obtained set of ordinary differential equations was solved using ACSL (Advanced
Continuous Simulation Language) [27]. ACSL is a simulation program for nonlinear
differential equations. It allows a free form of input of the equations since it sorts them
into the required order. It also features a possibility to use submodels which are called
MACROS. As integration algorithms it supports Runge–Kutta, Adams Bashforth, and
Gear formulas.
In order to have only nodes with integer numbers, I renumbered the nodes so that for
space nodes with odd numbers the temperature and for even numbers the velocity of the
normal component is evaluated by simulation. The values in between are calculated as
median of neighbouring nodes. The distance ∆x between two such new nodes was chosen
as
∆x = 2.5 · 10−3 m. (4.33)
The heat input q̇0 has the shape of a rectangular pulse with a value of
W
q̇0 = 3 · 104 (4.34)
m2
and a duration of
τ = 0.5 · 10−3 s. (4.35)
Appendix B shows the ACSL program. I used the Heun algorithm as integration method
with a fixed step size of ∆t = 1 · 10−6 s. The simulation was run for tmax = 4 · 10−3 s.
Fig. 4.3 shows the results of the simulation. Shown is the temperature difference of the
nodes to the initial temperature T0 over time. The distance between two plotted nodes is
5∆x = 12.5·10−3 m. It can be seen that the heat propagates in the medium as a wave since
The distance between the maxima of two neighbouring curves is
∆t = 0.81 · 10−3 s. (4.36)
As the distance between two curves is given as
∆x = 12.5 · 01−3 m (4.37)
we obtain for the propagation velocity

∆x m
v= = 15.4 (4.38)
∆t s
Measurements show for a temperature of 1.8 K a velocity of second sound of u = 19.5 m
s
[28]. The difference in measured and calculated velocity probably originates from the
description of the thermal parameters as functions of the temperature [18], but the sim-
ulation results basically confirm the Two Fluid Model of Helium II.
Chapter 5
Conclusions
5.1 Conclusions
The purpose of this project was to investigate whether the heat propagation in Helium II
can be described with the classical heat diffusion equation and nonlinear parameters. If
this model of Helium had shown the capacity of oscillating, it would have been an example
of a nonlinear oscillating system with only real eigenvalues, as mentioned in chapter 1.
But the results of several simulations with different initial and boundary conditions show
no sign of wave–like heat transport. This shows that the description with the classical but
nonlinear diffusion equation is not sufficient for the heat propagating system in Helium II.
The motivation for the search for nonlinear oscillations was the temperature depen-
dance of the specific heat capacity and heat conduction. But for temperatures below
Tλ the bending of the curves for these parameters is very small. Considering the small
temperature variations, the nonlinear portion of these parameters is not very important.
This motivated a simulation around Tλ where the nonlinearities are significant. But these
nonlinearities are such, that they damp heat transportation even more, the specific heat
capacity has a singularity and goes to infinity at that point and the heat conduction goes
to zero.
The simulation with the Two Fluid Model shows that this theory is sufficient to
describe heat conduction in Helium II. For further research it would be interesting to
29
CHAPTER 5. CONCLUSIONS 30
describe the Two Fluid Model in terms of bond graphs. By showing the energy flow in
the system this could contribute to a better understanding of the thermal behavior of
Helium II.
5.2 Zusammenfassung
Das Ziel dieser Studienarbeit war zu untersuchen, ob die Wärmeleitung in Helium II mit
der klassischen Wärmediffusionsgleichung mit nichtlinearen Parametern beschrieben wer-
den kann. Wäre dieses Modell für Helium schwingungsfähig gewesen, würde es ein Beispiel
darstellen für ein nichtlineares schwingungsfähiges System, das nur Eigenwerte auf der
realen Achse hat, wie in Kapitel 1 erwähnt wurde. Aber die Ergebnisse von mehreren
Simulationen mit verschiedenen Anfangs– und Randbedingungen zeigen keine Anzeichen
einer wellenartigen Wärmeausbreitung. Dies zeigt, daß die Beschreibung mit den klassis-
chen aber nichtlinearen Differentialgleichungen nicht genügend ist für die Wärmeleitung
in Helium II.
Die Temperaturabhängigkeit der spezifischen Wärmekapazität und der Wärmeleitung
war die Motivation für die Suche nach nichtlinearen Schwingungen. Aber für Tempera-
turen unter Tλ ist die Krümmung der Kurven dieser Parameter sehr klein. Berücksichtigt
man weiter die kleinen Temperaturänderungen, so ist der nichtlineare Anteil an diesen
Parametern nicht maßgeblich. Dieser Umstand regte eine Simulation um Tλ an, wo die
Nichtlinearitäten signifikant sind. Aber diese Nichtlinearitäten sind gerade so, daß sie den
Wärmetransport noch mehr dämpfen; die spezifische Wärmekapazität hat an dieser Stelle
eine Singularität und geht gegen unendlich, während die Wärmeleitung nach null geht.
Die Simulation mit dem Zweiflüssigkeitenmodell zeigt, daß diese Theorie die
Wärmeleitung in Helium II zufriedenstellend beschreibt. Für weitere Untersuchungen
wäre es interressant, das Zweiflüssigkeitenmodell mit Hilfe von Bondgraphen darzustellen.
Daß auf diese Weise die Energieflüsse aufgezeigt werden, kann zum Verständnis der ther-
mischen Vorgänge beitragen.
Appendix A
DYMOLA Programs
This appendix contains the DYMOLA programs which were used to simulate the heat
transport with the help of Bond Graphs. It consists each of a main program, a library
file bond.lib and a file exp.ctl which describes the simulation experiment.
31
APPENDIX A. DYMOLA PROGRAMS 32
A.1 Model 1: Initial Temperature 1.4K

model1.dym
@bond.lib
model project
submodel (SF) SF
submodel (cell) c1, c2, c3, c4, c5, c6, c7, c8, c9,c10, ->
c11,c12,c13,c14,c15,c16,c17,c18,c19,c20
submodel (mC) mC(mgamma = 1.455E-4)
node n1, n2, n3, n4, n5, n6, n7, n8, n9,n10, ->
n11,n12,n13,n14,n15,n16,n17,n18,n19,n20,n21
constant T0 = 1.4
output y1, y2, y3, y4
input time, tau
connect SF at n1
connect mC at n1
connect c1 from n1 to n2

SF.F0 = 0.2/T0
SF.time = time
SF.tau = tau
y1 = (c1.e1 - T0) * 1000.
y2 = (c2.e1 - T0) * 1000.
y3 = (c3.e1 - T0) * 1000.
y4 = (c4.e1 - T0) * 1000.
end
bond.lib
model type mRS

cut A(e1/f1), B(e2/-f2)
main cut C[A B]
main path P<A-B>
parameter mtheta = 1.0
constant Tlam = 2.173 , eps =1.E-30
local help1, help2
help1 = (1-(e2/Tlam)**5.6) * (e2/Tlam)**5.6

help2 = (lim(abs(e1)-eps) + eps)**(1/3)
f1 = mtheta * help2 * sgn(e1) * help1
e1 * f1 = e2 * f2
end
model type mC
main cut A(e/f)
parameter mgamma = 1.0
local C
C = mgamma * (e)**5.2
C * der(e) = f
end
model type bond

cut A(x/y), B(y/-x)
main cut C[A B]
main path P<A-B>
end
model type cell

submodel (mRS) mRS(mtheta = 154.32)
submodel (mC) mC(mgamma = 1.455E-4)
submodel (bond) b1, b2, b3
node n1, n2
cut A(e1/f1), B(e2/-f2)

main cut C[A B]
main path P<A-B>
connect b1 from A to n1
connect b2 from n1 to B
connect b3 from n1 to n2
connect mRS from n2 to B
connect mC at B
end
model type SF
main cut A(./-f)
terminal F0, tau, time
f = F0 * swtch(tau - time)
end
exp.ctl
cmodel
simutime 5.E-5
step 0.5E-7
commupoints 101
input 2, time(depend,t), tau(independ,2.E-5)
ctblock
scale = 1
XCCC = 1
label TRY
drunr
if XCCC<0 then XCCC = -XCCC | scale = 2*scale | go to TRY
else proceed
ctend
outblock
OUT
dispt y1, y2, y3, y4
outend
end
A.2 Model 2: Initial Temperature 2.178K

model2.dym
@bond.lib
model project
submodel (SF) SF
submodel (cell) c1, c2, c3, c4, c5, c6, c7, c8, c9,c10
submodel (mC) mC(mgamma = 0.019076)
node n1, n2, n3, n4, n5, n6, n7, n8, n9,n10,n11
constant T0 = 2.1718
output y1, y2, y3, y4
input time, tau
connect SF at n1
connect mC at n1
SF.F0 = 0.1/T0
SF.time = time
SF.tau = tau
y1 = (c1.e1 - mC.Tlam) * 1000.
y2 = (c2.e1 - mC.Tlam) * 1000.

y3 = (c3.e1 - mC.Tlam) * 1000.
y4 = (c4.e1 - mC.Tlam) * 1000.
end
bond.lib
model type mRS

cut A(e1/f1), B(e2/-f2)
main cut C[A B]
main path P<A-B>
parameter mtheta1 = 1.0, mtheta2 = 1.0
constant Tlam = 2.172 , eps =1.E-37
local f1l, f1u, tau
tau = (e2-Tlam)/Tlam
f1l = mtheta1*(abs(e1)+eps)**(1/3.5) * sgn(e1) * ->
(abs(Tlam-e2)+eps)**(0.8381)
f1u = mtheta2 * (2/(abs(tau)+eps))**(1/3) * ->
sqrt(ln(2/(abs(tau)+eps)))*e1/e2
f1 = f1l * swtch(-tau) + f1u * swtch(tau)
e1 * f1 = e2 * f2
end
model type mC
main cut A(e/f)
parameter mgamma = 1.0
constant Tlam = 2.172 ,eps = 1.E-37
local C
C = mgamma * (ln(2*Tlam/(abs(e-Tlam)+eps)) - 2*sgn(e-Tlam))/e
C * der(e) = f
end
model type bond

cut A(x/y), B(y/-x)
main cut C[A B]
main path P<A-B>
end
model type cell

submodel (mRS) mRS(mtheta1 = 94.897, mtheta2 = 1.36E-4)
submodel (mC) mC(mgamma = 0.019076)
submodel (bond) b1, b2, b3
node n1, n2
cut A(e1/f1), B(e2/-f2)

main cut C[A B]
main path P<A-B>
connect b1 from A to n1
connect b2 from n1 to B
connect b3 from n1 to n2
connect mRS from n2 to B
connect mC at B
end
model type SF
main cut A(./-f)
terminal F0, tau, time
f = F0 * swtch(tau - time)
end
exp.ctl
cmodel
simutime 10.E-3
step 10.E-6
commupoints 101
input 2, time(depend,t), tau(independ,1.E-3)
ctblock
scale = 1
XCCC = 1
label TRY
drunr
if XCCC<0 then XCCC = -XCCC | scale = 2*scale | go to TRY
else proceed
ctend
outblock
OUT
dispt y1, y2, y3, y4
outend
end
Appendix B
ACSL Program
This appendix contains the ACSL program used for the simulation of the heat transport
in Helium II with the use of the Two Fluid Model.
helium.csl
PROGRAM helium
INITIAL
MACRO gm(out,tem) $"-----------to calculate Gorter Mellink constant"
out = 10.**(2.1+3.12*alog10(tem)+0.0076/(1-(tem/tlam)))
MACRO END
MACRO temdot(out,temm2,tem,temp2,vm1,vp1)
MACRO REDEFINE v
v = (vm1+vp1)/2
out = k/(rho*5.6*slam) * (tlam/tem)**5.6 *...
(temm2-2*tem+temp2)/(4*dx**2) -...
v*(temp2-temm2)/(4*dx) - tem/5.6 * (vp1-vm1)/(2*dx) +...
(gm(tem)*rho*(1/(1-(tem/tlam)**5.6))*v**4 +...
a0*(tlam/tem)**5.6*((vp1-vm1)/(2*dx))**2)/(5.6*slam)
MACRO END
MACRO vdot(out,vm2,v,vp2,temm1,temp1)
42
APPENDIX B. ACSL PROGRAM 43
MACRO REDEFINE tem

tem = (temm1+temp1)/2
out = ((tlam/tem)**5.6 -1) * ( a0*(vp2-2*v+vm2)/(4*dx**2) -...
slam*(tem/tlam)**5.6*(temp1-temm1)/(2*dx)) -...
v * (3*(vp2-vm2)/(4*dx) +...
gm(tem) * (1/(1-(tem/tlam)**5.6))**2 * rho * v**2)
MACRO END
constant...
tem0 = 1.8, v0 = 0, taut = 0.5E-3, tmax = 4E-3,...
slam = 1.57E+3, tlam = 2.173, a0 =2.6E-8, dx = 2.5E-3,...
k = 5E-4, rho = 0.1453E+3,...
tem2i = 1.8, tem3i = 1.8, tem5i = 1.8, tem7i = 1.8,...
tem81i = 1.8
cinterval cint = 1E-4
algorithm ialg = 3
nsteps nstp = 500
q0 = 3E+4
schedule q0stop .at. taut
END $"-----------------------------------------------of INITIAL"
DYNAMIC
DERIVATIVE
"----------------------------- first two nodes"
temd2 = (q0/(rho*5.6*slam)*(tlam/tem2)**5.6-tem3*v3/5.6)/(2*dx)
tem2 = integ(temd2,tem2i)
tem1 = 2*tem2 - tem3
v1 = 0
vdot2 = ((tlam/tem2)**5.6 -1) * (a0*(v3-2*v2)/dx**2 -...
slam*(tem2/tlam)**5.6*(tem4-tem2)/(2*dx)) -...
v2*( 3*(3*v2+3*v3-v4)/(6*dx) + gm(tem2)*...
(1/(1-(tem2/tlam)**5.6))**2 * rho * v2**2)
v2 = integ(vdot2,0)
"------------------------- nodes in the middle"
tem3 = integ(temdot(tem1,tem3,tem5,v2,v4),tem3i)
v3 = (v2+v4)/2
tem4 = (tem3+tem5)/2
v4 = integ(vdot(v2,v4,v6,tem3,tem5),0)
v5 = (v4+v6)/2
v7 = (v6+v8)/2
v9 = (v8+v10)/2
v11 = (v10+v12)/2
tem12 = (tem11+tem13)/2
v13 = (v12+v14)/2
tem14 = (tem13+tem15)/2
v15 = (v14+v16)/2
tem16 = (tem15+tem17)/2
v17 = (v16+v18)/2
tem18 = (tem17+tem19)/2
v19 = (v18+v20)/2
tem20 = (tem19+tem21)/2
v21 = (v20+v22)/2
tem22 = (tem21+tem23)/2
v23 = (v22+v24)/2
tem24 = (tem23+tem25)/2
v25 = (v24+v26)/2
tem26 = (tem25+tem27)/2
v27 = (v26+v28)/2
tem28 = (tem27+tem29)/2
v29 = (v28+v30)/2
tem30 = (tem29+tem31)/2
v31 = (v30+v32)/2
tem32 = (tem31+tem33)/2
v33 = (v32+v34)/2
tem34 = (tem33+tem35)/2
v35 = (v34+v36)/2
tem36 = (tem35+tem37)/2
v37 = (v36+v38)/2
tem38 = (tem37+tem39)/2
v39 = (v38+v40)/2
tem40 = (tem39+tem41)/2
v41 = (v40+v42)/2
tem42 = (tem41+tem43)/2
v43 = (v42+v44)/2
tem44 = (tem43+tem45)/2
v45 = (v44+v46)/2
tem46 = (tem45+tem47)/2
v47 = (v46+v48)/2
tem48 = (tem47+tem49)/2
v49 = (v48+v50)/2
tem50 = (tem49+tem51)/2
v51 = (v50+v52)/2
tem52 = (tem51+tem53)/2
v53 = (v52+v54)/2
tem54 = (tem53+tem55)/2
v55 = (v54+v56)/2
tem56 = (tem55+tem57)/2
v57 = (v56+v58)/2
tem58 = (tem57+tem59)/2
v59 = (v58+v60)/2
tem60 = (tem59+tem61)/2
v61 = (v60+v62)/2
tem62 = (tem61+tem63)/2
v63 = (v62+v64)/2
tem64 = (tem63+tem65)/2
v65 = (v64+v66)/2
tem66 = (tem65+tem67)/2
v67 = (v66+v68)/2
tem68 = (tem67+tem69)/2
v69 = (v68+v70)/2
tem70 = (tem69+tem71)/2
v71 = (v70+v72)/2
tem72 = (tem71+tem73)/2
v73 = (v72+v74)/2
tem74 = (tem73+tem75)/2
v75 = (v74+v76)/2
tem76 = (tem75+tem77)/2
v77 = (v76+v78)/2
tem78 = (tem77+tem79)/2
v79 = (v78+v80)/2
tem80 = (tem79+tem81)/2
vdot80 = ((tlam/tem80)**5.6 -1) *...
( a0*(v80-2*v78+v76)/(4*dx**2) -...
slam*(tem80/tlam)**5.6*(tem81-tem79)/(2*dx)) -...
v80 *gm(tem80)*(1/(1-(tem80/tlam)**5.6))**2 * rho * v80**2
v80 = integ(vdot80,0)
temd81 = k/(rho*5.6*slam) * (tlam/tem81)**5.6 *...
(tem77-2*tem79+tem81)/(4*dx**2) +...
gm(tem81)*rho*(1/(1-(tem81/tlam)**5.6))*v81**4/(5.6*slam)
tem81 = integ(temd81,tem81i)
v81 = v79
END $"----------------------------------------------- of DERIVATIVE"

DISCRETE q0stop
q0 = 0
END $" of DISCRETE q0stop"
termt (t.gt.tmax .or. (tem81-tem0).gt. 1E-2)
END $" of DYNAMIC"
END $" of PROGRAM"
Bibliography
[1] V. Peshkov, “‘Second Sound’ in Helium II,” J. Phys. USSR, v. 8, p. 381 (1944).
[2] V. Peshkov, “Determination of the velocity of Propagation of the Second Sound in

Helium II,” J. Phys. USSR, v. 10, pp. 389–398 (1946).
[3] E. S. R. Gopal, Specific Heats at Low Temperatures, Plenum Press, New York
(1966).
[4] Fritz London, Superfluids Vol. II, John Wiley & Sons, New York (1954).
[5] K. R. Atkins, Liquid Helium, Cambridge University Press (1959).
[6] W. F. Vinen, “Physical Properties of Superfluid Helium. A General Review,” Inst.

Int. Froid, Int. Inst. Refrig., Saclay, Refrig. Sci. Technol (1981).
[7] S. J. Putterman, Superfluid Hydrodynamics, North–Holland Publishing Company,

Amsterdam London (1974).
[8] H. C. Kramers, J. D. Wasscher, C. J. Gorter, “The Specific Heat of Liquid Helium

between 0.25 and 1.9◦ K,” Physica, v. 18, n. 5, pp. 329–338 (1952).
[9] W. F. Vinen, “Mutual friction in a heat current in liquid helium II, I. Experiments
on steady heat currents,” Proc. Roy. Soc. A, v. 240, pp. 114–127 (1957).
[10] M. J. Buckingham, W. M. Fairbank, “The Nature of the λ–transition in Liquid

Helium,” Progress in Low Temp. Phys., v. 3, pp. 80–112 (1961).
50
BIBLIOGRAPHY 51
[11] R. A. Ferrell, N. Menyhàrd, H. Schmidt, F. Schwabl, P. Szépfalusy, “Fluctuations

and Lambda Phase Transition in Liquid Helium,” Annals of Physics, v. 47, pp. 565–
613 (1968).
[12] M. Archibald, J. M. Mochel, L. Weaver, “Size Dependence of the Thermal Conduc-

tivity of Helium I near the λ Point,” Phys. Rev. Letters, v. 21, n. 16, pp. 1156–1158
(1968).
[13] Guenter Ahlers, “Thermal Conductivity of He I near the Superfluid Transition,”

Phys. Rev. Letters, v. 21, n. 16, pp. 1159–1162 (1968).
[14] Paul E. Dimotakis, “Local temperature measurements in supercritical counterflow

in liquid helium II,” The Physics of Fluids, v. 16, n. 11, pp. 1787–1795 (1973).
[15] Paul E. Dimotakis, “Gorter–Mellink scale, and critical velocities in liquid–helium–II

counterflow,” Physical Review A, v. 10, n. 5, pp. 1721–1723 (1974).
[16] Michael J. Crooks, David L. Johnson, “Temperature Differences and Mutual Fric-
tion in He II Near Tλ ,” Canadian Journal of Physics, v. 49, pp. 1035–1039 (1971).
[17] W. Fiszdon, J. Piechna, W. Poppe, A qualitative numerical analysis of some features

of propagation of counterflow temperature waves in superfluid helium, Max–Planck–
Inst. Strömungsforschung, Göttingen, West Germany, Bericht 9/1988 (1988).
[18] D. Gentile, J. Pakleza, “Numerical solution of transient heat transport through

Helium II,” Numerical Heat Transfer, v. 6, pp. 317–335 (1983).
[19] D. Gentile, J. Pakleza, “An analytical and numerical investigation of heat transport
in superfluid helium,” Arch. Mech., v. 38, pp. 585–593 (1986).
[20] R. A. Hull, K. R. Wilkinson, J. Wilks, Proc. phys. Soc., London A, v. 64, pp. 379
(1951).
[21] François E. Cellier, Continuous System Modeling, Springer–Verlag, New York

(1991).
BIBLIOGRAPHY 52
[22] Hilding Elmqvist, A Structured Model Language for Large Continous Systems,
Ph.D. Thesis, Report CODEN: LUTFD2/(TRFT–1015), Dept. of Automatic Con-
trol, Lund Institute of Technology, Lund, Sweden (1978).
[23] Granino A. Korn, Interactive Dynamic System Simulation, McGraw–Hill, New York
(1989).
[24] L. D. Landau, J. Phys., Moscow, v. 5, pp. 71 (1941).
[25] P. Kapitza, J. Phys., Moscow, v. 5, pp. 60 (1941).
[26] I. M. Khalatnikov, Introduction to the Theory of Superfluidity, Benjamin, New York

(1965).
[27] Advanced Continuous Simulation Language (ACSL), User Guide/Reference Man-

ual, Mitchell and Gauthier, Assoc., Concord, Mass. (1981).
[28] R. J. Donnely, Experimental Superfluidity, Univ. of Chicago Press, Chicago, Ill.

(1967).

Numerical Investigation of Second Sound in Liquid Helium

Uploaded by

Copyright:

Available Formats

Numerical Investigation of Second Sound in Liquid Helium

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Numerical Investigation of Second Sound in Liquid Helium

Uploaded by

Copyright:

Available Formats

NUMERICAL INVESTIGATION OF

SECOND SOUND IN LIQUID HELIUM

A Thesis Submitted to the Faculty of the

List of Figures iii

2 Helium II and Second Sound 4

3 Heat Conduction Modeled With Time–Variant R–C–Networks 7

4 Heat Transport and Two Fluid Model 20

4.5 Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

1.1 Mass–Spring Oscillator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

2.1 Phase Diagram of Helium (schematic) . . . . . . . . . . . . . . . . . . . . . 4

4.1 Normal and Superﬂuid Fractions of Helium II . . . . . . . . . . . . . . . . 21

1.2 Thermodynamical Systems

be detected with a thermometer.

Heat Conduction Modeled With

The solution of this linear time–variant system is

x1 = 1.5(x10 + x20 )e5at − 0.5(x10 + 3x20 )e7at (3.5)

with the initial condition as ⎛ ⎞

A = cross section of Helium channel

3.2 Modeling in Terms of Bond Graphs

3.4.1 Model 1: Initial Temperature 1.4K

For the heat current as a function of T and dT , London gives an approximation

3.4.2 Model 2: Initial Temperature 2.1728K

For this second case we choose

For temperatures above Tλ , Ferrell gives the thermal conductivity of Helium as

Using the relation

3.5 Simulation and Results

Heat Transport and Two Fluid

4.1 Two Fluid Model

The fraction of ρs /ρn is a function of the temperature; at Tλ there is no superﬂuid compo-

4.2 Hydrodynamical Equations

with the thermodynamic parameters:

4.3 Numerical Values for the Thermal Parameters

A(T ) = 263(Tλ − T )−0.74 (in MKS units) (4.15)

4.5 Boundary Conditions

Tn+i = Tn−i (4.31)

4.6 Simulation and Results

The distance between the maxima of two neighbouring curves is

∆t = 0.81 · 10−3 s. (4.36)

As the distance between two curves is given as

∆x = 12.5 · 01−3 m (4.37)

we obtain for the propagation velocity

A.1 Model 1: Initial Temperature 1.4K

connect c13 from n13 to n14

model type mRS

help1 = (1-(e2/Tlam)**5.6) * (e2/Tlam)**5.6

model type bond

model type cell

cut A(e1/f1), B(e2/-f2)

A.2 Model 2: Initial Temperature 2.178K

y2 = (c2.e1 - mC.Tlam) * 1000.

model type mRS

model type bond

model type cell

cut A(e1/f1), B(e2/-f2)

MACRO REDEFINE tem

END $"----------------------------------------------- of DERIVATIVE"

[2] V. Peshkov, “Determination of the velocity of Propagation of the Second Sound in

[5] K. R. Atkins, Liquid Helium, Cambridge University Press (1959).

[6] W. F. Vinen, “Physical Properties of Superﬂuid Helium. A General Review,” Inst.