r
T
r
ThermaComp2009, September 810,2009, aples, Italy 2
0 =

\

+
V
t
(1)
( ) ( ) ( )
r ref
r
r r r
S sin T T g
v
r r
v
v
r
p
r
v
v V v
t
+
=

\

+
2 2
2
2
(2)
( ) ( ) ( )
S cos T T g
v
r r
v
v
p
r r
v v
v V v
t
ref
r
r
+ +
+ +
= +

\

+
2 2
2 1
(3)
( ) ( ) ( ) H h H , L V H
t
T h V h
t
+ =
(
\

+
=

\

+
(4)
where S
r
and S
are Darcy like source terms that nullify velocities in the solid phase. The
jump of total enthalpy H on the phase change temperature is converted to linear function
H in the narrow temperature interval. Details on numerical integration of equations (1)
(4) are given in [2].
3. RESULTS AND DISCUSSIONS
A number of numerical calculations are performed in order to estimate the influence of inner pipe
temperature on the shape of icewater interface, the temperature and flow fields in the liquid phase,
local and average heat transfer coefficients. Annulus dimensions are r
i
=12.7 mm and r
o
=50.8 mm
and ice in the annulus is initially on the phase change temperature in all calculations.
The temperature and flow fields in several time instances for the case of melting with T
w
=4
o
C are
given on Figure 2. At the beginning, conduction is dominant heat transfer mechanism and icewater
interface has the shape of circle with the center on the axis of annulus. As melting continues,
convection in the liquid phase intensifies which results in isotherms deformations. In this case,
water in the vicinity of inner pipe has the highest density and therefore flows downwards along the
inner pipe, while the water in the vicinity of ice surface flows upwards.
FIGURE 2. Temperature field (left) and flow field (right) for T
w
=4
o
C
Local heat transfer coefficients as a
function of angle are given on Figure 3.
It is evident that heat transfer is more
intense on the top of the heat source
(higher values) since relatively cold
water, previously cooled down during the
flow along the ice surface comes into
contact with heat source. On the ice
surface there is an opposite situation
relatively warmer water comes in contact
with ice surface below heat source and
consequence is pear shaped melted region
facing downwards.
0 30 60 90 120 150 180
0
50
100
150
200
250
300
Angle
L
o
c
a
l
h
e
a
t
t
r
a
n
s
f
e
r
c
o
e
f
f
i
c
i
e
n
t
s
h
[
W
/
m
2
K
]
f[
o
]
172 min
30 min
90 min
213 min
133 min
FIGURE 3. Local heat transfer coefficients for T
w
=4
o
C

30 min

133 min
214 min
i
r f w
l
r
T
T T
h 
\

=
+ clockwise
 counterclockwise
ThermaComp2009, September 810,2009, aples, Italy 3
In all cases when T
w
>4
o
C, isotherm T=4
o
C, i.e. density extreme is somewhere between the surface
of inner pipe and ice surface, and two distinct type of flow cells appears in the melted region. The
temperature and flow fields in several time instances for T
w
=8
o
C are given on Figure 4. Icewater
interface retains circular shape over a relatively longer period of time. As melting continues, flow in
the outer vortex, next to the icewater interface intensifies and occupies most of the melted region
and as a result moving of interface is determined by the flow in the outer vortex and melting is more
intense under the heat source.
FIGURE 4. Temperature field (left) and flow field (right) for T
w
=8
o
C
Local heat transfer coefficients from the
pipe surface to the water for t=67 min
(Figure 5) have minimum values for
100
o
, i.e. on the contact of inner and
outer vortex where temperature gradients
have minimum. For <100
o
local heat
transfer coefficients are relatively smaller
since there is a clockwise circulation in
inner vortex down the isotherm T=4
o
C and
then along the surface of cylinder T
w
=8
o
C.
For >100
o
C local heat transfer
coefficients are relatively higher since
there is a counterclockwise circulation of
water in outer vortex, along the ice surface
T
f
=0
o
C and then down the warm surface of cylinder T
w
=8
o
C.
For T
w
=13.6
o
C (Figure 6) at the very beginning of process, melted region takes pear shape facing
upward. Melting is more intense above heat source, and retarded over a quite long period under the
cylinder where heat transfer is conduction dominated.
FIGURE 6. Temperature field (left) and flow field (right) for T
w
=13.6
o
C
Average values of heat transfer coefficients from inner pipe surface to the water for T
w
4, 6 and 8
o
C
are given on Figure 7. Initial high values are characteristic for transient conduction dominated heat
0 30 60 90 120 150 180
0
50
100
150
200
250
300
Angle
L
o
c
a
l
h
e
a
t
t
r
a
n
s
f
e
r
c
o
e
f
f
i
c
i
e
n
t
s
[
W
/
m
2
K
]
f[
o
]
FIGURE 5. Local heat transfer coefficients for T
w
=8
o
C
+
67 min
16 min
157 min
i
r f w
l
r
T
T T
h 
\

=

8 min 17 min
41 min
+

+
+ clockwise
 counterclockwise
+ clockwise
 counterclockwise
+

ThermaComp2009, September 810,2009, aples, Italy 4
transfer. After 16000 s for T
w
=4
o
C and 12000 s for T
w
=8
o
C water comes into contact with outer
insulated pipe of annulus and therefore there is a drop of average heat transfer coefficients.
Variation of molten volume ratio V/V
o
(Vvolume of molten region, V
o
volume of heat source) with
time is given on Figure 8, together with experimental data of White et al. [6] for T
w
4, 8 and 10
o
C.
0 5000 10000 15000 20000
0
50
100
150
200
250
300
350
400
450
500
Time [s]
A
v
e
r
a
g
e
h
e
a
t
t
r
a
n
s
f
e
r
c
o
e
f
f
i
c
i
e
n
t
s
h
[
W
/
m
2
K
]
T
w
=4
o
C
T
w
=8
o
C
T
w
=12
o
C
0 1000 2000 3000 4000 5000 6000 7000 8000
0
0.5
1
1.5
2
2.5
3
3.5
4
Time [s]
R
a
t
i
o
V
/
V
o
4
o
C
6
o
C
8
o
C
9
o
C
10
o
C
FIGURE 7. Average heat transfer coefficients FIGURE 8. Variation of molten volume with time
4. CONCLUSIONS
Isothermal phase change, such as ice melting, can be successfully simulated with proposed single
domain enthalpy based model. Performed calculations exhibit qualitatively good agreement with
available experimental [6] and numerical results [1] where moving grid temperature based method
is used. Temperature of inner cylinder wall strongly affects the flow in liquid and shape of solid
liquid interface unlike the freezing process [7] where convection influence is less pronounced. For
inner pipe temperatures under or equal 4
o
C water with the highest density is next to the inner wall,
and flow is directed to the bottom of annulus causing the pear shape melted region on the bottom of
annulus, and highest heat transfer coefficient on the top of annulus. With inner wall temperature T
w
above 4
o
C, isotherm with highest water density moves in the liquid region, and separates two
distinct cells with opposite flow directions. Higher wall temperatures move maximum density
isotherm closer to the phase change interface and causes supremacy of clockwise flow in liquid and
forming of pear shape on the top side of annulus. Average heat transfer coefficients on the inner
pipe wall change during the melting process and can be obtained by integrating local values around
the circumference of inner pipe.
REFERENCES
[1] C.J Ho, S. Chen, Numerical simulation of melting of ice around a horizontal cylinder,
International Journal of Heat and Mass Transfer, 29/9, 13591369, 1986.
[2] V.R. Voller, C. Prakash, A fixed grid numerical methodology for convection diffusion mushy
region phase change problems, International Journal of Heat and Mass Transfer, 30/8, 1709
1719, 1987.
[3] L.S. Yao, J. Prusa, Melting and freezing, Advances in Heat Transfer, 19, 195, 1989.
[4] A. Saito, Recent advances in research on cold thermal energy storage, Journal of refrigeration,
25, 177189, 2002.
[5] S.V. Patankar, umerical Heat Transfer and Fluid Flow, Hemisphere, New York, 1980.
[6] D. White, R. Viskanta, W. Leidenfrost, Heat transfer during melting of ice around a horizontal,
isothermal cylinder, Experiments in Fluids, 4, 171179, 1986.
[7] E. Tombarevi, I. Vuanovi, Influence of inner pipe wall temperature on freezing of water in a
horizontal cylindrical annulus, Proceedings of EUROTHERM Seminar r.84Thermodynamics
of phase change, Namur, Belgium, 2009.
2
0
2
1
hd h