Simulation and Testing of a Latent Heat Thermal Energy Storage Unit with Metallic Phase Change Material

Abstract

Latent heat thermal energy storage in metallic phase change materials offers a thermal energy storage concept that can store energy at higher temperatures than with sensible thermal energy storage. This may enable the use of high efficiency thermodynamic cycles in CSP applications, which may lead to a reduction in levelised cost of electricity. Eutectic aluminum silicon alloy, AlSi12, is an attractive phase change material because of its moderate melting temperature, high thermal conductivity, and high heat of fusion. A prototype thermal energy storage test rig has been built and tested as to better understand the behavior of latent heat thermal energy storage. A mathematical model was developed to predict the behavior of such a heat storage unit. The model was compared with the behavior of the test rig during discharge. The model proved to simulate the latent heat thermal energy storage with reasonable accuracy. It is recommended that more accurate material property data be obtained and that the thermal energy storage test rig be modified as to improve readings.

Full text

Energy Procedia 49 ( 2014 ) 860 – 869
Available online at www.sciencedirect.com
ScienceDirect
1876-6102 © 2013 J.P. Kotzé. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/3.0/).
Selection and peer review by the scientifi c conference committee of SolarPACES 2013 under responsibility of PSE AG.
Final manuscript published as received without editorial corrections.
doi: 10.1016/j.egypro.2014.03.093
SolarPACES 2013
Simulation and testing of a latent heat thermal energy storage unit
with metallic phase change material
J.P. Kotzéa*, T.W. von Backströma and P.J. Erensa
aMechanical and Mechatronic Engineering, Stellenbosch University, Private Bag X1, Matieland, 7602, South Africa
*jpkotze@sun.ac.za, tel:+27 82 493 8687 fax:+27 21 8084933
Abstract
Latent heat thermal energy storage in metallic phase change materials offers a thermal energy storage concept that can store
energy at higher temperatures than with sensible thermal energy storage. This may enable the use of high efficiency
thermodynamic cycles in CSP applications, which may lead to a reduction in levelised cost of electricity. Eutectic aluminum
silicon alloy, AlSi12, is an attractive phase change material because of its moderate melting temperature, high thermal
conductivity, and high heat of fusion. A prototype thermal energy storage test rig has been built and tested as to better understand
the behavior of latent heat thermal energy storage. A mathematical model was developed to predict the behavior of such a heat
storage unit. The model was compared with the behavior of the test rig during discharge. The model proved to simulate the latent
heat thermal energy storage with reasonable accuracy. It is recommended that more accurate material property data be obtained
and that the thermal energy storage test rig be modified as to improve readings.
©2013 The Authors. Published by Elsevier Ltd.
Selection and peer review by the scientific conference committee of SolarPACES 2013 under responsibility of PSE AG.
Keywords: CSP; PCM; AlSi12; heat transfer
1. Introduction
Currently one of the central goals for concentrating solar power (CSP) is cost reduction. Apart from component
cost reduction, the increase of thermal efficiency of the power block generally underlines all proposed cost reduction
strategies. This generally entails the implementation of high efficiency supercritical CO2 and supercritical or ultra-
supercritical steam cycles. These high efficiency power blocks require source temperatures in excess of 600 to
700°C. Currently this is beyond the maximum operational temperature of established receiver, heat transfer and
thermal energy storage technologies. Kotzé et al. [1] attempts to address all of these limitations proposing the use of
metallic heat transfer fluids and latent heat thermal energy storage (LHTES) in metallic phase change materials.
© 2013 J.P. Kotzé. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/3.0/).
Selection and peer review by the scientific conference committee of SolarPACES 2013 under responsibility of PSE AG.
Final manuscript published as received without editorial corrections.

J.P. Kotzé et al. / Energy Procedia 49 ( 2014 ) 860 – 869 861
LHTES has the ability to store energy isothermally at significantly higher temperatures than possible in sensible
storage media. Most of the proposed phase change materials are salts, with low thermal conductivity [2].
These require large heat transfer surfaces and extensive heat transfer modification of the material, whereas metallic
phase change materials (PCM) have inherently high thermal conductivity and requires no material modification.
This results in simpler heat exchanger configurations.
The application of AlSi12 as a metallic PCM for thermal energy storage (TES) has been proposed by Kotzé et al.
[1], the concept is shown in Figure 1. To demonstrate the working of AlSi12 as a TES solution, a hypothetical
power plant (shown in Figure 2) was proposed, using AlSi12 in every section of the steam generator. Admittedly,
the use of AlSi12 as a TES material is probably limited to a high temperature evaporator for thermodynamic
reasons, but designing an AlSi12 TES unit for every section of a steam generator demonstrated a variety of heat
transfer and process control problems that can be encountered in a steam generator that can easily be extrapolated to
other metallic PCM’s.
NaK in
NaK out
Steam/Water
out
Steam/
Water in
AlSi12
PCM
AlSi12
PCM Housing
Steam/
water
pipes
NaK
Pipes
Cross-section
Fig.1. Illustration of the AlSi12 TES concept
LP-Turbine
G
HP-Turbine
Condenser
O-FWH FW Pump
FW Pump
Superheater
Boiler
Re-heater
Steam drum
NaK to storage
NaK to Recievers
Collector field
Fig.2. Power generating cycle
Kotze et al. [3] presented a systems analysis of the power block, presented in Figure 2. One of the key issues
addressed was that of a moving boundary problem in the PCM. As the TES unit discharges, the heat transfer
characteristics of the heat transfer surfaces of the steam generator changes as the PCM solidifies around the heat
transfer surfaces. This has a major impact on the design and process control of a latent heat TES system. To better
understand design parameters for a latent heat TES using metallic PCMs, it is important to have a validated
numerical simulation of such a heat storage unit (HSU) and the variation of heat transfer parameters.

862 J.P. Kotzé et al. / Energy Procedia 49 ( 2014 ) 860 – 869
Nomenclature
rpi
ri
rm
ro
k
Cp
α
CSP
Inner heat transfer pipe radius
Outer heat transfer pipe radius
Solidification front radius
Outer test section radius
Thermal conductivity
Specific heat capacity
Thermal diffusivity
Concentrating solar power
AlSi12
CFD
DSG
DSC
HSU
ISG
PCM
TES
Aluminum alloy (12% silicon)
Computational fluid dynamics
Direct steam generation
Differential scanning calorimeter
Heat storage unit
Indirect steam generation
Phase change material
Thermal energy storage
2. Problem formulation
As illustrated in Figure 1, the heat storage unit (HSU) concept consists of heat transfer tubes embedded at regular
intervals throughout a pool of phase change material (PCM). The nature of the HSU can take on two forms
depending on the nature of the power cycle. One is as proposed by Kotze et al. [1], with steam generation directly
from storage, referred to as the direct steam generation concept (DSG). The other is a more risky, but significantly
more elegant solution utilizing a liquid metal-water steam generator in the power cycle. This means that there is
only liquid metal flowing through the storage bank, resulting in a much simpler analysis. This is referred to as the
indirect steam generation concept (ISG).
In the DSG concept the liquid metal heat transfer pipes (heat input) and the steam/water heat transfer pipes (heat
output) are all embedded at regular intervals throughout the melt. This may lead to a very complex thermal
distribution throughout the PCM which is highly dependent on the exact geometry of both the liquid metal and the
steam/water heat exchange pipes. To simplify the analysis, only discharge conditions will be considered for the DSG
concept, since a computational fluid dynamics (CFD) solution is necessary for charge conditions. On the other hand,
in the ISG concept, the same heat exchange surface is used to charge and discharge the HSU, resulting in a fairly
simple heat transfer problem from a geometrical point of view.
For simplification, the entire volume of PCM is discretized into hexagonal cylinders around each heat exchange
pipe as illustrated in Figure 3. As the HSU discharges, the PCM solidifies in cylinders around the heat transfer pipes,
eventually the cylinders will grow into each other. This will create a situation where the heat transfer characteristics
of the heat exchange surfaces are nonlinear, and heat transfer rates will decrease rapidly. Therefore the area between
the cylinders is treated as a dead volume. It can be geometrically proven that if the cylinders expand to the point
where they touch each other, 9.3% of the total volume of PCM is still liquid. This volume is ignored in the analysis
and is considered to be construction material.
The heat transfer model is built on the following primary assumptions:
‚ Because the PCM can be considered as isotropic, the thermo-physical properties of the liquid or the solid
phase is constant within the operational range of the HSU.
‚ Volumetric expansion during phase change is negligible.
‚ Conduction in the axial direction is negligible due to the isothermal state of the LMTES unit, making the
heat transfer in the PCM one dimensional
‚ Natural convection at the solid-liquid interface is negligible due to the high thermal conductivity of
metallic PCMs (181 W/m.K), making thermal conduction the dominant heat transfer mechanism.
‚ Perfect thermal contact between the PCM and heat transfer pipes.
Thus, the model can be simplified to a two dimensional conduction problem described in Figure 3.
The heat transfer problem is essentially three concentric cylinders. The liquid phase is treated as a solid, due to
the exceptionally low Prandl number of molten aluminium (0.000021244). One of the key aspects of this problem is
the moving boundary, rm. As the HSU charge and discharge, the boundary moves. The temperature at this boundary
is equal to the melting point of the PCM, Tm.
Hoshi et al.[4] considered a similar heat transfer problem in an effort to investigate the importance of thermal

J.P. Kotzé et al. / Energy Procedia 49 ( 2014 ) 860 – 869 863
conductivity of a PCM on the performance of a latent heat HSU. Using the one dimensional heat conduction
equation along with appropriate boundary conditions and an energy balance on the moving boundary, they were able
to predict the movement of the moving boundary. This model proved effective, and a similar model was used by He
et al. [5] to predict the performance of a HSU of a significantly different design to that of Kotzé et al.[1]. The
problem can be described by a set of differential equations.
Fig.3. – Discretization and two dimensional models for charging and discharging
The conduction problem can be described using the one dimensional conduction equation for cylinders:
絞劇
絞建 噺糠絞態劇
絞堅態髪糠
堅
絞劇
絞堅
The boundary conditions are:
Convective boundary condition on the inside of the heat transfer pipe (rpi):
伐倦畦 項劇盤堅椎沈建匪
項堅 噺芸
墜沈鎮
Where Qoil is the heat removed with the oil.
Interface boundary conditions between the PCM and heat transfer pipe (ri):
劇椎岫堅沈岻噺劇
牒寵暢岫堅沈岻
And:
伐倦椎
項劇
椎盤堅沈建匪
項堅 噺倦
牒寵暢
項劇牒寵暢岫堅沈建岻
項堅
Interface boundary condition between the solid/liquid interfaces:
劇鎮岫堅
陳岻噺劇
鎚岫堅
陳岻
And:

864 J.P. Kotzé et al. / Energy Procedia 49 ( 2014 ) 860 – 869
伐倦牒寵暢ｅ鎚墜鎮沈鳥
項劇
椎盤堅沈建匪
項堅 噺倦
牒寵暢ｅ鎮沈槌通沈鳥
項劇牒寵暢岫堅沈建岻
項堅
The outer boundary conditions are considered adiabatic during discharge.
The authors who did analytical work on similar Stefan problems used the finite difference method to solve the
conduction problem, but it fails to take into account the solidification process on solid-liquid interface [4] [5]. The
problem was rather solved using an enthalpy tracking method. The cylinder is discretized in cylindrical shells and
the internal energy and enthalpy of each element are calculated using the initial conditions. The enthalpy is
determined through enthalpy/temperature relationships that are determined by linearizing the enthalpy/temperature
graph for AlSi12. The conduction between elements is calculated using a resistance model, derived from the one
dimensional conduction equation. The temperature of each node is evaluated explicitly and obtained from the
enthalpy/temperature graph. Accordingly the stability of the model had to be checked. Grid independence and
stability was verified, and a Δr=0.001m, and Δt=0.001s was chosen for the simulation. The model was implemented
in Matlab.
3. Experiment
To validate the concept and to give a basis to compare a simulation with, an experimental setup was built. The
test section is a 1m long cylinder filled with AlSi12 alloy with a single heat transfer pipe through the middle of the
cylinder. Due to laboratory safety restrictions, it was not possible to use liquid metals to heat the AlSi12 through the
internal heat transfer pipe, but heat has to be added to the AlSi12 through the outer cylinder walls using electrical
band heaters. Therefore only discharge conditions could be tested. This is acceptable since the assumptions for
charging and discharging are exactly the same. The geometry of the cylinder is described in Figure 4 and the detail
geometry is given in Table 1. Note that the positions of probes 2 and 3 are unusual; this is because the probes bent
during the casting process.
The internal cooling pipe is cooled using ISO100 quenching oil. This has been selected as it yields comparable
heat transfer rates to that which was predicted for high pressure steam, and because it does not thermally shock the
internal cooling pipe when it is introduced to the test section. The oil is cooled with water from a cooling tower in a
plate heat exchanger. The process diagram is shown in Figure 5 with a photo of the test rig in Figure 6. The heat
removed from the test section is measured by measuring the flow rate, inlet- and outlet temperatures of the cooling
oil.
Fig.4. Test section

J.P. Kotzé et al. / Energy Procedia 49 ( 2014 ) 860 – 869 865
Table 1 - Test section
Cylinder geometry
Outer cylinder
Inside diameter
398
mm
Outside diameter
408
mm
Heat transfer pipe
Inside diameter
24.4
mm
Outside diameter
33
mm
Length in contact with AlSi12
1270
mm
Volume of AlSi12
0.1533
m3
Mass of AlSi12
408
kg
Thermocouple placement from the centre
Probe 1
30
mm
Probe 2
61
mm
Probe 3
76
mm
Probe 4
135
mm
Probe 5
180
mm
Dump tank
Water pump P
TC
P
TC
Gate valve
F
Header tank
(Oil)
Vapour
condenser Test section
Ball Valve
Ball Valve
Ball Valve
Solenoid valve
One way valve
One way valve
Cooling
tower
Heaters
Flow
meter
TC
F
Flow meter
TC
Test
section
cooling
loop
Primary
oil
coolant
loop
Water
cooling
loop
Fig.5. - Test rig process diagram

866 J.P. Kotzé et al. / Energy Procedia 49 ( 2014 ) 860 – 869
Fig. 6 – The experimental setup
4. Material properties
To accurately correlate the numerical analysis to the experiment, it is important that the material properties of the
AlSi12 and the quenching oil are accurately determined. Some of the thermo physical properties could not be
measured accurately and literature values had to be used. The results are shown in Table 2.
Table 2 - Thermo physical properties of materials in the experiment
Thermophysical properties of AlSi12
Source
Density
2661
kg/m3
[5]
Specific heat
0.939
kJ/kg.k
[5]
Heat of fusion
462
kJ/kg
Measured: DSC
Phase change temperature
577
°C
[5] / Measured: DSC
Thermal conductivity
181
W/m.K
[5]
Thermophysical properties of Mild steel
Density
7854
kg/m3
Specific heat
1.169
kJ/kg.k
Thermal conductivity
30
W/m.K
ISO 100 quenching oil
Density at 60 °C
890
kg/m3
Measured: Lab
Specific heat at 60 °C
1.950
kJ/kg.K
Measured: MDSC
Kinematic viscosity at 60 °C
20.2
mm2/s
Measured: ASME1321

J.P. Kotzé et al. / Energy Procedia 49 ( 2014 ) 860 – 869 867
5. Results and comparison
The test section was first heated until all of the AlSi12 had melted and stabilized at 620°C. The oil loop was pre-
heated until the oil was at operational temperature (35°C), after which the oil were directed through the test section
and with the water cooling loop turned on. The system was kept running until the test section was discharged
completely and cooled down to 400°C. The oil and water inlet and outlet temperatures, coolant flow rates, and the
temperature of the internal probes were measured throughout discharge.
At the beginning of the test the temperature of the entire melt dropped down to the phase change temperature
relatively quickly as the heat transfer rate was in the excess of 100kW due to the high temperature difference
between the melt and the oil passing through the inner heat transfer pipe. During phase change the heat transfer from
the melt to the oil was 36kW on average (see Figure 7). The temperature of Probe 1 dropped off quickly through the
melting point because of the small volume of PCM between it and the heat transfer pipe. Figure 7 clearly shows how
the solidification front travels past the probes, from probe 1 to probe 5. Notably 2 and 3 solidifies close to each other
in time because they are close to each other. As the solidification front moves out, the volume of PCM increases,
and it is notable that probe 5 stays on the phase change temperature of 577°C throughout discharge. The thermal
gradient from the solidification front to the inner pipe is relatively low due to the high thermal conductivity of the
AlSi12 PCM, this can be seen by the small temperature difference between probe 1, which has been discharged first
and probe 5 which is discharged last and how long probe 1 remains essentially isothermal throughout the entire
discharge process. The test section remained in phase change discharge for 78.5 minutes, and integrating the power
output yields that the energy removed from the test section during phase change was 169 MJ, correlating well with
the measured heat of fusion taking losses into account.
The nodes corresponding to the probes were simulated. The simulation results are shown in Figure 8. The
experimental and simulated data are plotted together in Figure 9. It can be seen that the temperature profiles of the
simulation matches the experimental data closely. With further data analysis of the simulation data, it can be seen
that the solidification front is not a sharp boundary but rather a zone where the enthalpy falls within the boundaries
of the latent heat discharge. The inflictions on the graphs indicate the moment when the enthalpy of the PCM
surrounding the probe falls below the latent region. The occurrence of these inflictions is matched well by the
model, indicating that the model is predicting the position of the solidification front well. The movement of the
solidification front through the melt is shown in Figure 10.
As soon as the entire storage unit discharges, the temperatures of all the probes fall rapidly as all of the PCM is in
sensible mode. The curve with which this transition occurs is not matched well, and further testing will be done to
investigate this discrepancy.
Fig.7.–Variation of temperature against time of various radial positions - Experimental
10
20
30
40
50
60
70
80
350
400
450
500
550
600
650
0 1000 2000 3000 4000 5000 6000
Thermal output (kW)
Temperature (°C)
Time (s)
Probe 1 (°C)
Probe 2 (°C)
Probe 3 (°C)
Probe 4 (°C)
Probe 5 (°C)
Power output (kW)

868 J.P. Kotzé et al. / Energy Procedia 49 ( 2014 ) 860 – 869
Fig.8. – Variation of temperature against time of various radial positions –Analytical
Fig.9. - Comparison between the experiment and simulation results
Fig.10. - Movement of solidification front through melt in discharge
10
20
30
40
50
60
70
80
350
400
450
500
550
600
650
0 1000 2000 3000 4000 5000 6000
Thermal output (kW)
Temperature (°C)
Time (s)
Sim Probe 1 (°C)
Sim Probe 2 (°C)
Sim Probe 3 (°C)
Sim Probe 4 (°C)
Sim Probe 5 (°C)
Sim Power output (kW)
350
400
450
500
550
600
650
0 1000 2000 3000 4000 5000 6000
Temperature (°C)
Time (s)
Probe 1 (°C)
Probe 2 (°C)
Probe 3 (°C)
Probe 4 (°C)
Probe 5 (°C)
Sim Probe 1 (°C)
Sim Probe 2 (°C)
Sim Probe 3 (°C)
Sim Probe 4 (°C)
Sim Probe 5 (°C)
0
0.05
0.1
0.15
0.2
0.25
0 1000 2000 3000 4000 5000 6000
rm (m)
Time (s)

J.P. Kotzé et al. / Energy Procedia 49 ( 2014 ) 860 – 869 869
6. Conclusions and recommendations
Latent heat thermal energy storage in metallic phase change materials offers high temperature, isothermal energy
storage. The higher storage temperatures may lead to a reduction in LCOE through the use of higher efficiency
power blocks. Kotze et al. [1] proposed the use of metallic phase change materials along with metallic heat transfer
fluids as a storage concept and identified AlSi12 as a good candidate metallic PCM for research purposes. To prove
the concept and to evaluate the heat transfer analysis, a prototype LHTES unit was built and tested. It has a unique
construction enabling the measurement of the solidification front of the PCM through discharge. The data obtained
from this test is presented and it shows that the test rig works well within designed parameters.
A heat transfer model of the moving boundary problem is presented. The model is solved using an enthalpy
tracking method rather than a finite difference method. This model is used to predict the performance of a large
thermal energy storage system [1], and has been implemented on a model representing the test setup for validation.
The results show that trends could be matched to a reasonable degree; the results will be improved with better
materials testing and model refinement.
Acknowledgements
The authors would like to thank the Department of Science and Technology of South Africa through the Solar
thermal spoke fund, the Stellenbosch University Hope project, the South African National research foundation, and
the center for renewable and sustainable energy studies (CRSES) for funding the resources to perform this work and
present it at SolarPACES. The advice and support of my co-authors, colleagues, and my father, JCB Kotzé is also
greatly appreciated.
References
[1] Kotzé JP, von Backstrom TW, Erens PJ,High temperature thermal energy storage utilizing metallic phase change materials and metallic heat
transfer fluids.: ASME: Journal of Solar Energy Engeneering, Vol. 135, 035001 pp.1-6
[2] Kenisarin MM,High Temperature Phase Change Materials for Thermal Energy Storage. 2009.11.011, Tashkents : Renewable and Sustainable
Energy Reviews, 2010.
[3] Kotzé JP, von Backstrom TW, Erens PJ,Evaluation of a latent heat thermal energy storage system using AlSi12 as a phase change material.,
Marrakesch : SolarPACES, 2012.
[4] Hoshi, Akira, et al., Screening of high melting point phase change materials (PCM) in solar thermal concentration technology based CLFR. et
al. 2005 , Solar Energy 79, pp. 332-339.
[5] He, Qiao and Zhang, WennanA study on latent heat storage exchangers with the high-temperature phase-change material.. 2001, International
Journal of Energy Research, pp. 25:331-341.