Applied Energy 189 (2017) 506–519

Contents lists available at ScienceDirect

Applied Energy
journal homepage: www.elsevier.com/locate/apenergy

Experimental characterization and simulation of a hybrid sensible-latent

heat storage
Christoph Zauner a,⇑, Florian Hengstberger a, Benjamin Mörzinger a, Rene Hofmann a,b, Heimo Walter b
a
AIT Austrian Institute of Technology GmbH, Energy Department, Sustainable Thermal Energy Systems, Giefinggasse 2, 1210 Wien, Austria
b
Vienna University of Technology, Institute for Energy Systems and Thermodynamics, Getreidemarkt 9/BA, 1060 Wien, Austria

h i g h l i g h t s

A new versatile sensible-latent heat storage concept was designed.

An inverted shell-and-tube heat exchanger with the PCM inside the tubes was used.

We characterized a prototype containing 208.2 kg HDPE and 515.1 kg thermal oil.

Two computer models were developed showing excellent agreement with experiments.

A cost calculation was performed and two promising applications were described.

a r t i c l e i n f o a b s t r a c t

Article history: Versatile and economically competitive thermal energy storages are necessary to fulfill the widely differ-
Received 21 September 2016 ing requirements for storages applied in renewable energy systems, process heat, district heating, power
Received in revised form 12 December 2016 generation and domestic heating. We present the concept of a hybrid sensible-latent heat storage based
Accepted 14 December 2016
on an adapted commercial shell-and-tube heat exchanger. The phase change material (PCM) is encapsu-
lated within the tubes and thermal oil serves as sensible heat storage as well as the heat transfer medium.
We designed and built a prototype using high density polyethylene (HDPE) as PCM and characterized the
Keywords:
storage on a dedicated test rig at AIT. Energy capacities and power profiles are presented for different
Thermal storage
PCM
mass flows and (dis)charging temperatures. Two physical models were developed and implemented
Experimental storage characterization using the Modelica language. Dymola was used to simulate the behavior of the prototype storage. Very
Simulation good agreement was achieved between simulation and experiment. Using the models, we studied the
Costs heat transfer within the storage in detail, which enabled us to present how to adapt the storage geometry
and PCM properties to cover a broad range of applications. We discuss storage costs and calculate mate-
rial costs per stored kilowatt-hour for different PCM-thermal oil volume ratios as a function of the tube
outer diameter. Finally, we highlight the main advantages and design freedoms of our concept and
describe concrete application scenarios in district heating and process heat.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction
Energy storages are necessary to increase the share of intermit-
tent renewable energy systems, increase energy efficiency and thus
mitigate CO2 emissions [1,2]. A substantial fraction of the energy
demand is thermal energy. Process heat is required in many indus-
tries to produce goods and building climatization in omnipresent.
Heating networks in industries as well as urban regions distribute
this heat generated by central power plants or decentralized heat-
⇑ Corresponding author.
E-mail address: christoph.zauner@ait.ac.at (C. Zauner).
URL: http://www.ait.ac.at (C. Zauner).
ing units. In most cases the heat demand is not constant, but var-
ious peaks are present. If renewable intermittent energy sources
are utilized, there is often a big mismatch between heat supply
and demand. Thermal energy storages are necessary to bridge this
temporal mismatch and are utilized for peak shifting and demand
side management [3–5]. In addition, they enable waste heat recov-
ery in different industrial processes and sectors, which is reviewed
by Miro et al. [6] including mobile thermal storage [7,8]. Thermal
energy storages are also investigated in combinations with heat
pumps to increase their efficiency [9,10] and offer concentrating
solar power plants an advantage over photovoltaic systems by
allowing electricity production when the sun does not shine [11–
13]. Recently thermal energy storage have also been investigated

http://dx.doi.org/10.1016/j.apenergy.2016.12.079
0306-2619/Ó 2016 Elsevier Ltd. All rights reserved.
 C. Zauner et al. / Applied Energy 189 (2017) 506–519
Nomenclature
a heat transfer coefficient (W/m2 K)
k thermal conductivity (W/m K)
l dynamic viscosity (kg/m s)
q density (kg/m3)
A surface area (m2)
D diameter of tube (m)
h phase change enthalpy (kJ/kg)
l length of tube (m)
m mass (kg)
Q heat (kW h)
r radius (m)
T temperature (°C)
t time (s)
v velocity (m/s)
cp constant pressure specific heat (J/kg K)
m _ mass flow rate (kg/s)
Q_ heat transfer rate (kW)
q_ heat flux (W/m2)
Gz Graetz number (–)
in micro-CHP plants to optimize costs and revenues [14] and to
optimize the performance of a bank of chillers [15].
Latent heat storages offer large energy densities without high
pressures as in steam accumulators or pressurized water tanks.
This technology is a promising candidate, especially for tempera-
tures above 100 °C. Many different storage concepts have been
investigated for a wide range of application such as concentrated
solar power [16–21], industrial waste heat recovery [22–28] and
solar cooling [29,30].
Among the different investigated concepts [31] shell-and-tube
heat exchangers offer an interesting opportunity for latent heat
storages. In most cases a configuration was chosen, where the
phase change material is located on the shell side and the heat
transfer fluid is flowing within the tubes. There have been many
theoretical studies: Tay et al. [32] and Fornarelli et al. [33] used
computational fluid dynamics (CFD) to study melting, Pirasaci
and Goswami [34] and Tehrani et al. [35] analyzed different stor-
age geometries for power plants. Bai et al. [36] focused on the eco-
nomics. The effect of introducing fins to enhance heat transfer was
investigated by different authors [37,38] and the effect of the PCM
thermophysical properties was studied in [39]. Experimental stud-
ies were performed studying the phase change process in single
tubes [40–43]. Lab-scale storages were built and characterized by
Lopez et al. [44] and Tay et al. [45].
Almost no work was done on the inverted configuration, where
the PCM is inside the tubes and the heat transfer medium is flow-
ing on the shell side, although such a concept offers interesting
opportunities. Especially, since it is easy to vary the latent and sen-
sible energy contributions over a wide range and to realize hybrid
sensible-latent heat storages using this configuration. In such stor-
ages, one can benefit from both, the high energy density of PCMs
and the high powers of the sensible storages. As is emphasized in
[46,47] a big advantage is that the PCM allows the reduction of
the quick drop of the outlet temperature of sensible storages dur-
ing discharging. In addition, Geissbühler et al. [48] show that it is
also possible to reduce the fast power decrease of sensible storages
by including PCMs. Furthermore, inverted shell-and-tube heat
exchangers enable the macro-encapsulation of the PCM within
the tube in an efficient way and generally have better heat transfer
characteristics on the heat transfer fluid side (cf. [49]). The oppor-
tunities of inverted shell-and-tube sensible-latent heat storage are
discussed in more detail in the next section.
507
Pr Prandtl number (–)
Re Reynolds number (–)
Subscripts
amb ambient
avg average
c center
cryst crystallization
f phase change front
i inner
in inlet
melt melting
o outer
out outlet
st steel
t tube
vol volume
To our best knowledge only Zhang, Ma and Xiao investigated a
PCM storage that is related to the inverted shell-and-tube configu-
ration. They presented fluid dynamics simulations [50,51] and
built and characterized a lab-scale storage [52]. They encapsulated
an eutectic mixture of NaNO3 and KNO3 and a nickel foam to
improve heat transfer into stainless steel containers. These con-
tainers were immersed in the heat transfer fluid (thermal oil)
and placed inside an insulated tank.
In this work, we present for the first time a fully integrated solu-
tion of a hybrid sensible-latent heat storage based on an inverted
shell-and-tube configuration. We redesigned a shell-and-tube heat
exchanger, encapsulating the PCM within the tubes. The thermal
oil on the shell side, simultaneously serves as a sensible heat stor-
age and heat transfer medium. The whole storage was produced by
an industrial company using professional and standardized manu-
facturing techniques. In that way, we ensure high prototype qual-
ity, scalability for later commercial storages and a short time-to-
market enabling CO2 savings quickly. It will be possible to manu-
facture storages of several meters in diameter and length contain-
ing tons of PCM with the very same production methods.
2. Experimental investigation
2.1. Storage design
As outlined in the introduction, most authors investigating shell
and tube heat exchanger in PCM storages placed the PCM on the
shell side and the heat transfer fluid flows within the tubes
whereas in our design the situation is inverted. Depending on the
needs of the application (energy density, power profile), one design
might be superior to the other.

The volume share of PCM in our configuration is limited to a
maximum of 90% in a densely packed staggered tube arrange-
ment. Necessary gaps between the tubes for heat transfer fluid
flow reduce the PCM volume further (cf. Fig. 2).

The number and length of weld seams depends on the storage
capacity and power profile. As this can be a major cost driver,
a careful assessment has to be performed for each case.

In our design, as the heat transfer fluid is outside of the cylindri-
cal tubes containing the PCM, there is a large heat transfer area
in the beginning. It decreases as the phase front moves towards
508 C. Zauner et al. / Applied Energy 189 (2017) 506–519
the center of the tubes. In the conventional configuration, the
situation is inverted. Again, one has to assess from case to case,
which geometry meets the required power profiles for storage
charging and discharging best.

The heat transfer coefficients between the tubes separating the
PCM and the heat transfer fluid are very different for both con-
figurations. Especially, if the phase front is close to the tubes,
this heat transfer is likely to be the limiting factor regarding
storage power. In our design, turbulent flow can be achieved
more easily which is advantageous.

In our design, one can greatly vary oil and PCM volume frac-
tions. The sensible energy stored within the oil, can be (dis)
charged very quickly allowing high storage power, while the
PCM part may satisfy more constant power requirements.

By employing the high energy density of the PCM, the fast
decrease of outlet temperature and power of sensible storages
can be reduced.
In order to test the capabilities of such an inverted shell and
tube hybrid sensible-PCM storage, we first selected appropriate
phase change and sensible storage materials.
We presented in previous work [53] that polymers are a very
promising PCM class because several commercially available types
show high enthalpies, different phase change temperatures up to
300 °C, promise to be long-term stable, are produced on a large
scale and can be compounded with additives to increase their ther-
mal conductivity. We found high density polyethylene (HDPE) to
be most suitable and selected INEOS Rigidex HD6070EA after
detailed thermophysical characterization. We furthermore showed
that this material has excellent cyclic stability, and thus chose it
also for the prototype storage in this contribution. In Fig. 1 we
show its main thermophysical properties.
As a sensible storage medium water and thermal oils are widely
used. We selected the thermal oil Marlotherm SH due to the fol-
lowing advantages over pressurized water:

A higher boiling point and consequently lower pressure enables
thinner tank walls, which reduces material costs.

In addition, a design is feasible to remain in the lowest safety
category for pressure vessels, which saves costs for certificates
and maintenance.
Fig. 1. Thermophysical data for HDPE INEOS Rigidex HD6070EA. The curves for
thermal diffusivity and thermal conductivity are interpolated between solid and
liquid regions, indicated by the yellow box, as no measurements can be performed
within the melting range. The volume and density are normalized to their values at
20 °C. (For interpretation of the references to color in this figure legend, the reader
is referred to the web version of this article.)
The geometry of the tank and the shell and tube heat exchanger
was chosen such that the contributions of the oil and PCM to the
total energy of the storage were about equal. The storage peak
power was designed to be less than 100 kW for the envisioned
characterization experiments, which is the limit of the cooling
power of our test rig. Also, we wanted to build a prototype that
already included all relevant features of a future industrial product.
These design criteria could be met by employing 19 PCM tubes,
each with a length of 3 m, in a staggered arrangement as depicted
in Fig. 2. The volume share of PCM to the total tank volume was
27%.
Four baffles separate the tank volume into five compartments
(Fig. 3) with one connection to the piping system in the top and
bottom compartment, respectively.
We insulated the tank with three 10 cm thick layers of mineral
wool (Fig. 4).
The final storage contained 208.2 kg of PCM and 515.1 kg of
thermal oil. It was characterized using a dedicated test rig at AIT
(Fig. 5).
2.2. Storage characterization
We used a thermostat (Lauda ITH350) connected to our in-
house cooling system to characterize the storage. The inlet and
outlet temperatures were measured with Pt100 resistance ther-
mometers and the mass flow was recorded with a clamp-on ultra-
sonic flow meter (Flexim Fluxus F601). In addition, we inserted
thermocouples into several PCM tubes reaching until the center
of the bottom and top compartment, respectively (Figs. 6 and 7).
We characterized the energy and (dis)charging power of the
storage by changing the inlet temperature of a thermally stabilized
storage in a stepwise manner. The mass flow was kept constant
within each experiment. Fig. 8 shows the results of a measurement
where the storage was discharged from 150 °C to 110 °C with a
mass flow of 0.7 kg/s. In the upper plot one observes that the dis-
charging power peaks at 55 kW and decreases rapidly within the
first 1500 s. The response of the thermal oil is faster, whereas the
PCM part can be used to deliver power more steadily as will be
analyzed in more detail in the simulation section below. The lower
plot in Fig. 8 shows the temperatures measured by the thermocou-
ples within various PCM tubes at different positions. One can
clearly observe the plateaus when the HDPE is crystallizing. As
expected, the top compartment is fully crystallized before the bot-
tom one. However within each compartment, no crystallization
pattern can be deduced reliably. Most likely, the thermocouples
could not withstand the mechanical force induced upon PCM crys-
tallization and were moved away from their initial mounting
positions.
76,50
0
11
30°
110
610
Fig. 2. Geometry and arrangement of the 19 PCM tubes within the tank. Four baffles
separate the tank into five compartments.
 C. Zauner et al. / Applied Energy 189 (2017) 506–519 509

Fig. 3. Tube bundle featuring the PCM tubes as well as separator baffles.

Fig. 5. The hybrid latent-sensible heat storage was connected to the thermostat.
Mass flow and temperatures were measured to characterize storage power and
energy capacity. For energy balancing the Pt100 sensors closest to the inlet and
outlet were used.

Fig. 4. Three layers of insulation with a total thickness of 30 cm were mounted

around the tank reducing the heat losses for a storage heated up to 150 °C to less
than 200 W.

In the next experiment, we investigated the influence of an

increased mass flow. The storage was again discharged from
150 °C to 110 °C, the mass flow doubled to 1.4 kg/s. As expected
(Fig. 9), the initial power peak, which is mostly related to the sen-
sible energy of the thermal oil increases. This fast part of the stor-
age can be utilized in applications where the power demand peaks
often and quickly. In contrast, the longer and constant power tail
related to the PCM is not influenced greatly, which indicates that
the heat transfer from the tubes into the oil is not a limiting factor
in the investigated flow range. This part of the storage is more suit-
able for base load conditions.
Another parameter affecting the storage characteristics is the
inlet temperature. In Fig. 10 we compare discharging the storage
from 150 °C to 110 °C against from 150 °C to 95 °C with a mass
flow of 0.7 kg/s. Because INEOS Rigidex HD6070EA crystallizes
Fig. 6. Thermocouples were inserted into the PCM tube to record the HDPE
around 125 °C, this corresponds to a doubling of the temperature temperature during operation.
difference between the inlet set point temperature, T in;set , and the
average crystallization temperature. In this case not only the peak
power but also the discharging power related to HDPE is increased. Next, we also varied the initial storage temperature between
This can be attributed to a faster crystallization (cf. PCM tempera- 140 °C, 150 °C and 160 °C and discharged to 95 °C with a mass flow
tures in the top compartment shown in the central plot) as well as of 0.7 kg/s. Especially, the power peak in the beginning is affected.
a larger sensible PCM-energy. The increase in total discharging Since the driving temperature difference between the inlet
energy is quantified in the lower plot for both cases. temperature and the PCM crystallization temperature was not
510 C. Zauner et al. / Applied Energy 189 (2017) 506–519

Fig. 7. Left: Positions and enumeration of the thermocouples reaching into the center of the top compartment. Right: Positions and enumeration of the thermocouples
reaching into the center of the bottom compartment.

Fig. 8. Temperatures, power profiles and discharging energy for a mass flow of 0.7 kg/s.

changed, the PCM power tail is unaffected. This is also reflected in
the PCM temperatures. The total energy capacity is increased by
additional sensible heat stored in the oil and HDPE (see Fig. 11).
In realistic applications it is possible that the storage is not fully
charged before being discharged again. Due to the intricate heat
transfer fluid flow pattern in our design it is a priori not clear
how the storage will behave under such circumstances. Thus, we
prepared the storage in two configurations:

fully charged (f.c.): the storage was charged circulating 150 °C
heat transfer fluid through the storage until all PCM was at
150 °C. Thus, all the PCM is in its liquid state.

partially charged (p.c.): the charging process circulating 150 °C
heat transfer fluid through the storage was interrupted when
the PCM was in the transition region. Some parts of the PCM
in the storage are already in the liquid state while other parts
are still solid as indicated by our thermocouples.
Fig. 12 shows the comparison of discharging a fully and a
partially charged storage. The power profile and outlet tempera-
ture behave quite similar, which is especially evident from the
totally extracted energy. Although the PCM temperatures within
various tubes may differ, this is averaged out over the whole
storage. This fact is very beneficial because it makes the storage
behavior predicable even when discharging a partially charged
storage.
As a final step, we investigated the behavior of the storage for
a series of partial charging and discharging cycles. Initially the
storage was fully charged up to 150 °C. Using a constant heat
transfer fluid mass flow of 0.7 kg/s, we discharged the storage
until some of the PCM was crystallizing and then started to
reheat for some time. This cycle was repeated several times.
The resulting non-trivial power and energy profiles as well as
PCM temperatures are shown in Fig. 13. It is important to note
that due to the limited heating power of our thermostat we
 C. Zauner et al. / Applied Energy 189 (2017) 506–519 511

Fig. 9. Temperatures, power profiles and discharging energy for different mass flows.

Fig. 10. Temperatures, power profiles and discharging energy for different inlet temperatures.

were also limiting the charging power of the storage to around
40 kW. Of course, the storage charging power can exceed
40 kW, but we cannot characterize it with our equipment.
The characteristic figures of our storage extracted from various
measurements are summarized in Table 1. The respective contri-
butions of PCM and thermal oil as well as the storage energy den-
sities have been estimated.
3. Simulation
3.1. Description of the storage models
The goal of our storage modeling efforts was to have an easy
design tool that can be used for quick dimensioning purposes. Also,
we would like to apply our model in simulations of larger energy
512 C. Zauner et al. / Applied Energy 189 (2017) 506–519

Fig. 11. Temperatures, power profiles and discharging energy for different initial storage temperatures.

Fig. 12. Temperatures, power profiles and discharging energy compared for discharging a fully charged (f.c.) and a partially charged (p.c.) storage. The mass flow was 0.7 kg/s.

systems. Therefore, we used a node modeling approach instead of
3-dimensional fluid dynamics. The Modelica language [54] was
used to implement the physical equations and Dymola [55] to per-
form the simulations.
We implemented two different approaches to model the PCM
filled tubes:

Stefan-model: the Stefan-problem is formulated for phase
change storages as described in [56].

cp ðTÞ-model: a lumped capacitance model that uses a function
for the specific heat that includes the phase change enthalpy.
The schematic of the Stefan-model is shown in Fig. 14. We
exploit the radial symmetry of the tubes and use temperatures
nodes as indicated by the green1 dots. The heat transfer between
the nodes is modeled according to
Q_ ¼ af ðrÞDT ð1Þ
 C. Zauner et al. / Applied Energy 189 (2017) 506–519 513

Fig. 13. Temperatures, power profiles and energy profile for a series of partial storage charging and discharging experiments. The mass flow was 0.7 kg/s.

Table 1
Measured storage characteristics evaluated at different time intervals from the beginning of discharging, tdischarge . The sensible energy stored was estimated according to
Q ¼ mcp ðT charged  T discharged Þ for steel, thermal oil and PCM. The latent part is the difference between the total energy, Q, and all sensible contributions.

T charged T discharged m_ tdischarge Q qv ol qmass;PCM Q l;PCM Q s;PCM Q s;st Q s;oil

(°C) (°C) (kg/s) (h) (kW h) (kW h/m3) (kW h/kg) (kW h) (kW h) (kW h) (kW h)
150 110 1.4 5 29.9 34.1 0.144 9.2 4.7 4.5 11.2
150 110 0.7 5 29.6 33.8 0.142 8.9 4.7 4.5 11.2
150 110 1.4 2.5 28 31.9 0.134 7.3 4.7 4.5 11.2
150 110 0.7 2.5 26.9 30.7 0.129 6.2 4.7 4.5 11.2
150 90 0.7 2.5 40.3 46.0 0.194 9.5 7.3 6.8 16.5
160 95 0.7 2.5 42.5 48.5 0.204 8.7 8.1 7.3 18.1
150 95 0.7 2.5 37.1 42.3 0.178 8.8 6.7 6.2 15.2
140 95 0.7 2.5 32.5 37.1 0.156 9.6 5.3 5.1 12.3

Depending on the heat transfer mode (conduction, convection)
and the involved media (PCM, steel, oil), appropriate heat transfer
coefficients, a, are used. The function f ðrÞ is a geometrical factor
taking care of the cylindrical geometry. The driving force for the
heat transfer is the temperature gradient, DT.
Since our polymer is highly viscous2 we can neglect convective
heat transfer within the PCM and the heat transfer between the
phase front and the inner tube wall was modeled by
rt;i
Q_ ¼ 2plt kPCM ln ðT PCM;f  T t;i Þ; ð2Þ
r PCM;f
where lt is the length of the tube within one compartment, rt;i is the
tube inner radius, r PCM;f is the phase front radius, T PCM;f is the phase
change temperature and T t;i is the inner tube wall temperature.
In the Stefan-model, the phase front is moving and it’s position,
r PCM;f ðtÞ, is dynamically calculated via
dr PCM;f
Q_ ¼ 2plt hPCM qPCM ðrt;i  rPCM;f Þ ð3Þ
dt
This phase change equation is triggered as soon as the PCM
temperature, T PCM , reaches the phase change temperature of the
2
The melt flow index is 7.6 g/10 min at 190 °C and 2.16 kg load.
material. Otherwise, the PCM is modeled by a single lumped capac-
ity located at an intermediate position. We found that
rPCM;sensible ¼ 0:83r t;i gave the best fit to the experiments:
dT PCM
Q_ ¼ mPCM cp;PCM ð4Þ
dt
Contrary to this Stefan-model, in the cp ðTÞ-model the PCM is
resolved radially in more detail as indicated in Fig. 15. More nodes
are introduced and each node is described by a lumped capacity
similar to Eq. (4). However, in this case the employed thermal
capacitance function includes also the latent heat derived from
thermophysical measurements [53] in addition to the sensible heat
and was modeled by
hPCM
cp;PCM ðTÞ ¼ 1913:3 þ 4:3923T þ pffiffiffiffiffiffiffi
r 2p !
 2
1 T PCM  T phch
 exp  ð5Þ
2 r
Concerning the heat transfer from the tubes to the heat transfer
fluid, we employed correlations for shell and tube heat exchangers
from literature. We implemented equations from VDI Heat Atlas
[57,58], treating tube bundles in different alignments. Empirical
514 C. Zauner et al. / Applied Energy 189 (2017) 506–519
Fig. 14. In the Stefan-model the phase front of the PCM is moving. Thermal
conduction occurs within the PCM and the steel tubes. Heat transfer correlations for
shell-and-tube heat exchangers are used from literature to describe the energy flow
between the steel tubes to the thermal oil.
Fig. 15. In the cp ðTÞ-model the PCM is resolved in more detail. The latent energy
stored within the PCM is modeled by extending the specific heat function by a
latent heat term (cf. Eq. (5)).
correction factors taking into account several compartments sepa-
rated by baffles, small tube numbers, temperature dependence of
physical properties, and baffle-leakage were also included.
The energy stored by the sensible oil part of the storage was
also modeled using a lumped capacitance model employing an
analogous equation for the thermal oil as for the PCM (cf. Eq. (4)).
3.2. Simulation of the lab-storage
In order to validate the simulations, we compared them to
experimental measurements. Fig. 16 shows the discharging power,
outlet temperature and stored energy when discharging the stor-
age from 150 °C to 110 °C with a mass flow of 0.7 kg/s. Very good
agreement was achieved between the simulations and the experi-
mental data.
Using the computer models, it is possible to analyze the contri-
butions of the different storage media (oil, PCM, steel tubes) to the
total storage power and energy capacity (Fig. 17).
The sensible oil part of the storage reacts quickest and is mainly
responsible for the power peak at the beginning. The PCM contri-
bution rises until it reaches the maximum around 800 s. The solid-
ifying PCM delivers slightly decreasing power, whereas the
sensible energy stored within the oil is completely discharged after
about 2000 s. Both storage media (HDPE, Marlotherm SH) are
responsible for about half of the totally stored energy. The sensible
energy stored in the steel tubes is small compared to the PCM and
the thermal oil. From this figure one can also see that the PCM flat-
tens the steep power decrease of the sensible oil storage (compare
the total power profile to the power profile of the oil part alone).
For a given mass flow, the outlet temperature is affected likewise
by virtue of the storage power equation, Q_ ¼ mc _ p ðT out  T in Þ. This
is an advantage of hybrid sensible-latent storages.
As described above, in the Stefan-model, crystallization and
thus phase front movement sets in as soon as the temperature of
the single lumped PCM capacity reaches the crystallization tem-
perature. Due to numerical drawbacks, five unphysical power
peaks arise, one for each compartment starting to crystallize. The
cp ðTÞ-model possesses a better radial resolution, which is why no
such peaks occur. However, because in the Stefan-model, the peaks
diminish quickly, the simulations still yield a reasonable total stor-
age power and energy behavior.
3.3. Variation of storage parameters
Since both models predict nearly identical curves and the
Stefan-model is much faster we used it to study the influence of
various PCM and geometry parameters. First, we doubled the
PCM thermal conductivity3 in five steps from 0.4 W/mK up to
6.4 W/mK for the same simulation as before (Fig. 16). The results
are shown in Fig. 18. As expected, the sensible oil part of the storage
is only slightly affected, but the PCM-power is considerably
increased and the time needed to discharge the storage is reduced.
The dependency is non-linear due to the cylindrical geometry of
the PCM. At higher thermal conductivities another heat transfer than
conduction through the PCM becomes the limiting factor4, which is
why the power cannot be increased significantly by further increas-
ing the PCM thermal conductivity.
In order to increase the share of PCM in the total storage, one
can decrease the spacing between the tubes and the tank diameter
accordingly. A minimal tube to tube distance of 1 cm guarantees
reasonable fluid flow and is still compatible with the current weld-
ing process joining the tubes and tube sheets. The tank diameter
was reduced from 585 mm to 442 mm. In the resulting storage,
the oil power peak lasts shorter and the PCM is the main contrib-
utor to the total energy capacity (Fig. 19), which is reduced due to
the smaller storage volume.
If an application demands higher storage power one can again
increase the PCM thermal conductivity. A value of 1.6 W/mK is suf-
ficient to reduce the total storage discharging time to roughly
90 min. The total storage capacity of this configuration is
20 kW h with a share of 80% contributed by the HDPE (Fig. 20).
4. Storage costs
The total costs of our storage can be divided into manufacturing
and material costs. Both are very much dependent on the design
3
It is well known how to compound polymers with thermally conductive fillers
such as metal powders or graphite to increase their thermal conductivity.
4
Although not generally true, in most situations the limiting heat transfer after
PCM conduction is the heat transfer from the tube walls into the thermal oil.
 C. Zauner et al. / Applied Energy 189 (2017) 506–519 515

Fig. 16. Temperatures, power profiles and discharging energy for different mass flows.

Fig. 17. Comparison of the contributions of the oil, PCM and steel tubes to the total discharging power and storage capacity.

details of the storage, the cost structure of the manufacturer and
the number of storages of a particular order (cf. economy of scale).
Different manufacturers have different machinery and labor costs
depending on their available equipment, available capacities fac-
tory location and profit policy. Thus, unfortunately we are not able
to provide reasonable data on the manufacturing costs.
However, it is possible to estimate material costs of our storages
for different designs. The main material costs occur for the shell-
and-tube heat exchanger, the thermal oil and the PCM. The shell-
and-tube heat exchanger consists of steel tubes containing the
PCM, the container and the baffles. The container and the baffles
are usually produced by the storage manufacturer themselves from
flat steel sheets. The steel sheets and the tubes are bought from sup-
pliers. For our cost calculation, we used the online prices for orders
larger than 5000 kg of the supplier Salzgitter-Mannesmann [59],
which provides a reasonable estimate for sufficiently large shell-
and-tube heat exchanger manufacturers. The price for high density
polyethylene depends on grade and availability [60]. For our pro-
duct, INEOS Rigidex HD6070EA, the price is 1 EUR/kg for large quan-
tities. Also, the prices of thermal oil vary. For our type, Marlotherm
SH, 1 EUR/kg can be assumed for large quantities.
In the upper plot of Fig. 21 we show the prices of steel tubes
made of 1.0345 (as used in our storage) for different tube diame-
ters. The material costs for the steel sheets for the container and
baffles (1.0345, 1.0425) are negligible compared to the costs of
the tubes.
516 C. Zauner et al. / Applied Energy 189 (2017) 506–519

Fig. 18. Comparison of the power and energy profiles for increased PCM thermal conductivities.

Fig. 19. Comparison of the power and energy profiles for different tube-tube spacings.

In our storage design, we can greatly vary the number and diam-
eter of PCM-containing tubes and thus the volume ratio of PCM to
thermal oil, V PCM =V oil . For example, one can use a larger number of
small tubes or smaller number of large tubes to achieve a certain
ratio. Of course, also the total storage capacity is influenced by this
choice. To compare the costs of different configurations, we calcu-
lated the material costs per stored kilowatt-hour for different
PCM-thermal oil volume ratios as a function of the tube outer diam-
eter. We assumed that the storage was heated from 100 °C to 150 °C.
The result is illustrated in the lower plot of Fig. 21. The general trend
shows that the material costs per kilowatt-hour decrease upon
increasing the tube diameter independent of the ratio of PCM to
thermal oil. Furthermore, the cost reduction is larger for storage con-
figurations with larger PCM fractions. Also, increasing the share of
PCM reduces material costs per kilowatt-hour which reflects the
higher energy density of the PCM. However, this effect is non-
linear and decreases as one increases V PCM =V oil .
As already mentioned in the beginning, the total storage costs
also have to include the manufacturing costs. However, the mate-
rial costs calculated here provide a lower bound and allow to esti-
mate if a storage of our kind can be economically feasible for a
specific application.
 C. Zauner et al. / Applied Energy 189 (2017) 506–519 517

Fig. 20. Comparison of the power and energy profiles for different tube-tube spacings and increased PCM thermal conductivity.

Fig. 21. Material costs (steel tubes, PCM, thermal oil) per stored kilowatt-hour for different PCM-thermal oil volume ratios as a function of the tube outer diameter. We
assumed that the storage was heated from 100 °C to 150 °C.

5. Applications Our hybrid sensible-latent storage concept is capable of fulfill-

Thermal energy storages are necessary for many applications in
renewable energy, process heat, district heating, power generation
and domestic heating. Therefore, also the requirements regarding
storage capacity, temperature levels and power profiles for charg-
ing and discharging as well as economic boundary conditions vary
greatly.
ing a wide range of requirements because many different configu-
rations with different storage properties are possible:

small diameter tubes: even for low conductive PCMs, the latent
energy can be transferred to the heat transfer fluid effectively
(cf. Eq. (2)). This allows high power storages. However, many
tubes are required if a large PCM share is desired.
518 C. Zauner et al. / Applied Energy 189 (2017) 506–519

large diameter tubes: A high PCM share is possible with a small
number of tubes and thus weld seems, which is cost-effective. It
should be noted that due to our cylindrical geometry larger PCM
volume fractions than for densely packed spheres (like in
packed PCM beds) are possible. However, as is shown in
Fig. 18 one might need higher PCM thermal conductivities if
high storage powers are required.

fewer number of tubes: Using fewer tubes, one can increase the
share of sensible energy in the storage. This part of the storage
can charged/discharged quickly, while the remaining PCM
delivers more constant power. If e.g. the sensible part of the
storage is already discharged and suddenly excess waste/solar
heat has to be stored, it can be quickly injected into the sensible
part of the storage (cf. Fig. 13).
In our ongoing projects, we investigate the application of our
storage concept for district heating networks and industrial pro-
cesses. Although, we are still in an early research phase, advan-
tages of our hybrid sensible-latent thermal storages are
becoming evident:

district heating: In many district heating networks, there is a
peak power demand in the morning before people leave their
homes. This is followed by a period requiring lower base load.
Our storage is ideally suitable for such an application. It can
be charged during the night when the load is low. When the
morning peak load occurs, the storage delivers high discharge
power mainly via the sensible part. Afterwards, during the
day the PCM delivers the base load.

Aluminum die casting [61]: Waste heat is generated within the
casting process when the molten aluminum injected into the
casting mold is solidified and cooled. Being a batch process
the waste heat is also generated discontinuously. In addition
the waste heat amount depends very much on the casting pro-
duct and casting frequency. Because our storage characteristics
can be adjusted readily by varying the tube geometry, one can
build a suitable storage for many different waste heat profiles.
6. Summary
In the first section, we describe the storage concept: the PCM is
encapsulated in the tubes of a shell-and-tube heat exchanger and
the sensible storage medium on the shell side also serves as heat
transfer medium. Next, we present the design and construction
of a prototype storage where we use high density polyethylene
(HDPE) as phase change material. Geometric details of the heat
exchanger are described and intermediate steps of the manufactur-
ing process are shown. The storage prototype was installed in a
dedicated test rig at the AIT. We present the results of a one month
measurement campaign comprising more than 40 heating and
cooling cycles where the energy capacities, power profiles and
PCM temperatures were characterized. The mass flows as well as
initial and final temperatures were varied to study the storage
behavior in detail. In addition we investigate the behavior of the
storage for partial charging and discharging. The characteristic fig-
ures of the storage are summarized in a table for different situa-
tions. In order to be able to quickly design appropriate storages
for many different applications a fast computer model is necessary.
We implemented two different node models in Modelica and per-
formed the simulations in Dymola. The models were compared to
the experimental results and showed very good agreement and
demonstrated that the simulations are accurate and fast enough
to fulfill our requirements. Thus, we used to model to analyze
the heat transfer phenomena and energy contributions of the oil
and PCM parts in more detail. This served as a basis to know which
parameters have to be altered in order to be able to fulfill different
requirements. Finally, we exemplarily present the influence of
smaller tube-tube spacings and increased PCM thermal conductiv-
ity. Next, we discuss the costs of our storage design. We calculate
the material costs per stored kilowatt-hour for different PCM-
thermal oil volume ratios as a function of the tube outer diameter.
In the last section, we discuss the results of our investigations with
special regard to real applications. The main advantages and design
freedoms are highlighted and concrete application scenarios in dis-
trict heating and process heat are described.
Author contributions and acknowledgments
The Austrian Research Promotion Agency (FFG) is gratefully
acknowledged for funding this work under Grant No. 838669 (Stor-
eITup!) and under Grant No. 848914 (StoreITup-IF). C.Z. conceived,
submitted and leads the projects, designed and characterized the
storage, developed the models and performed the simulations with
the great help of B.M. and F.H. Christof Group built the storage. B.
M. performed his master thesis supervised by C.Z., F.H. and H.W. at
AIT. H.W. and R.H., who is the PhD supervisor of C.Z., are gratefully
acknowledged for valuable discussions regarding storage design,
modeling, costs and characterization. The paper was written by
C.Z. and commented by all authors.
References
[1] Arce P, Medrano M, Gil A, Oró E, Cabeza LF. Overview of thermal energy storage
(TES) potential energy savings and climate change mitigation in Spain and
Europe. Appl Energy 2015;88:2764–74.
[2] Cabeza LF, Miró L, Oró E, de Garcia A, Martin V, Kronauer A, et al. CO2
mitigation accounting for thermal energy storage (TES) case studies. Appl
Energy 2015;155:365–77.
[3] Arteconi A, Ciarrocchi E, Pan Q, Carducci F, Comodi G, Polonara F, et al. Thermal
energy storage coupled with PV panels for demand side management of
industrial building cooling loads. Appl Energy 2016. doi: http://dx.doi.org/
10.1016/j.apenergy.2016.01.025 [in press, corrected proof].
[4] da Cunha JP, Eames P. Thermal energy storage for low and medium
temperature applications using phase change material – a review. Appl
Energy 2016;177:227–38.
[5] Arteconi A, Hewitt NJ, Polonara F. State of the art of thermal storage for
demand-side management. Appl Energy 2012;93:371–89.
[6] Miró L, Gasia J, Cabeza LF. Thermal energy storage (TES) for industrial waste
heat (IWH) recovery: a review. Appl Energy 2016;179:284–301.
[7] Guo S, Zhao J, Wang W, Yan J, Jin G, Zhang Z, et al. Numerical study of the
improvement of an indirect contact mobilized thermal energy storage
container. Appl Energy 2016;161:476–86.
[8] Guo S, Zhao J, Wang W, Yan J, Jin G, Wang X. Techno-economic assessment of
mobilized thermal energy storage for distributed users: a case study in China.
Appl Energy 2016. http://dx.doi.org/10.1016/j.apenergy.2016.08.137 [in press,
corrected proof].
[9] Renaldi R, Kiprakis A, Friedrich D. An optimisation framework for thermal
energy storage integration in a residential heat pump heating system. Appl
Energy; 2016. http://dx.doi.org/10.1016/j.apenergy.2016.02.067 [in press,
corrected proof].
[10] Alimohammadisagvand B, Jokisalo J, Kilpeläinen S, Ali M, Sirén K. Cost-optimal
thermal energy storage system for a residential building with heat pump
heating and demand response control. Appl Energy 2016;174:275–87.
[11] González-Portillo LF, Munoz-Antón J, Martínez-Val JM. An analytical
optimization of thermal energy storage for electricity cost reduction in solar
thermal electric plants. Appl Energy 2016. http://dx.doi.org/10.1016/j.
apenergy.2016.10.134 [in press, corrected proof].
[12] Jacob R, Belusko M, Fernández AI, Cabeza LF, Samana W, Bruno F. Embodied
energy and cost of high temperature thermal energy storage systems for use
with concentrated solar power plants. Appl Energy 2016;180:586–97.
[13] Wagner SJ, Rubin ES. Economic implications of thermal energy storage for
concentrated solar thermal power. Renew Energy 2014;61:81–95.
[14] Diaz G, Moreno B. Valuation under uncertain energy prices and load demands
of micro-CHP plants supplemented by optimally switched thermal energy
storage. Appl Energy 2016;177:553–69.
[15] Soler MS, Sabaté CC, Santiago VB, Jabbari F. Optimizing performance of a bank
of chillers with thermal energy storage. Appl Energy 2016;172:275–85.
[16] Xu CCB, Li P. Application of phase change materials for thermal energy storage
in concentrated solar thermal power plants: a review to recent developments.
Appl Energy 2015;160:286–307.
[17] Liu M, Tay NHS, Belusko M, Bruno F. Investigation of cascaded shell and tube
latent heat storage systems for solar tower power plants. Energy Procedia
2015;69:913–24.
 C. Zauner et al. / Applied Energy 189 (2017) 506–519
[18] Nithyanandam K, Pitchumani R. Design of a latent thermal energy storage
system with embedded heat pipes. Appl Energy 2014;126:266–80.
[19] Laing D, Bahl C, Bauer T, Lehmann D, Steinmann W-D. Thermal energy storage
for direct steam generation. Sol Energy 2011;85:627–33.
[20] Bayón R, Rojas E, Valenzuela L, Zarza E, León J. Analysis of the experimental
behaviour of a 100 kwth latent heat storage system for direct steam
generation in solar thermal power plants. Appl Therm Eng 2010;30:2643–51.
[21] Laing D, Bauer T, Steinmann W-D, Lehmann D. Advanced high temperature
latent heat storage system – design and test results. In: Effstock – the 11th
international conference on thermal energy storage; 2009.
[22] Merlin K, Soto J, Delaunay D, Traonvouez L. Industrial waste heat recovery
using an enhanced conductivity latent heat thermal energy storage. Appl
Energy 2016;183:491–503.
[23] Vogel J, Felbinger J, Johnson M. Natural convection in high temperature flat
plate latent heat thermal energy storage systems. Appl Energy
2016;184:184–96.
[24] Pointner H, Steinmann W-D. Experimental demonstration of an active latent
heat storage concept. Appl Energy 2016;168:661–71.
[25] Johnson M, Vogel J, Hempel M, Dengel A, Seitz M, Hachmann B. High
temperature latent heat thermal energy storage integration in a co-gen plant.
Energy Procedia 2015;73:281–8.
[26] Zipf V, Neuhäuser A, Willert D, Nitz P, Gschwander S, Platzer W. High
temperature latent heat storage with a screw heat exchanger: design of
prototype. Appl Energy 2013;109:462–9.
[27] Laing D, Bauer T, Breidenbach N, Hachmann B, Johnson M. Development of
high temperature phase-change-material storages. Appl Energy
2013;109:497–504.
[28] Chui JN, Meany BH, Martin V. Industrial surplus heat utilization through
mobile thermal energy storage with enhanced operating strategy. In:
Greenstock – the 13th international conference on thermal energy storage;
2015.
[29] Gil A, Oró E, Castell A, Cabeza LF. Experimental analysis of the effectiveness of a
high temperature thermal storage tank for solar cooling applications. Appl
Therm Eng 2013;54:521–7.
[30] Gil A, Oró E, Miró L, Peiró G, Ruiz A, Salmerón JM, et al. Experimental analysis
of hydroquinone used as phase change material (PCM) to be applied in solar
cooling refrigeration. Int J Refrig 2014;29:95–103.
[31] Agyenim F, Hewitt N, Eames P, Smyth M. A review of materials, heat transfer
and phase change problem formulation for latent heat thermal energy storage
systems (LHTESS). Renew Sustain Energy Rev 2010;14(2):615–28.
[32] Tay NHS, Belusko M, Liu M, Bruno F. Investigation of the effect of dynamic
melting in a tube-in-tank PCM system using a CFD model. Appl Energy
2015;137:738–47.
[33] Fornarelli F, Camporeale SM, Fortunato B, Torresi M, Oresta P, Magliocchetti L,
et al. CFD analysis of melting process in a shell-and-tube latent heat storage
for concentrated solar power plants. Appl Energy 2016;164:711–22.
[34] Pirasaci T, Goswami DY. Influence of design on performance of a latent heat
storage system for a direct steam generation power plant. Appl Energy
2016;162:644–52.
[35] Tehrani SSM, Taylor RA, Saberi P, Diarce G. Design and feasibility of high
temperature shell and tube latent heat thermal energy storage system for solar
thermal power plants. Renew Energy 2016;96:120–36.
[36] Bai F, Wang Y, Wang Z, Sun Y, Beath A. Economic evaluation of shell-and-tube
latent heat thermal energy storage for concentrating solar power applications.
Energy Procedia 2015;69:737–47.
[37] Al-Abidi AA, Mat S, Sopian K, Sulaiman MY, Mohammad AT. Internal and
external fin heat transfer enhancement technique for latent heat thermal
energy storage in triplex tube heat exchangers. Appl Therm Eng
2013;53:147–56.
[38] Rathod MK, Banerjee J. Thermal performance enhancement of shell and tube
Latent Heat Storage Unit using longitudinal fins. Appl Therm Eng
2015;75:1084–92.
519
[39] Tao YB, Carey VP. Effects of PCM thermophysical properties on thermal storage
performance of a shell-and-tube latent heat storage unit. Appl Energy
2016;179:203–10.
[40] Wang Y, Wang L, Xie N, Lin X, Chen H. Experimental study on the melting and
solidification behavior of erythritol in a vertical shell-and-tube latent heat
thermal storage unit. Int J Heat Mass Transf 2016;99:770–81.
[41] Hosseini MJ, Rahimi M, Bahrampoury R. Experimental and computational
evolution of a shell and tube heat exchanger as a PCM thermal storage system.
Int Commun Heat Mass Transf 2014;50:128–36.
[42] Longeon M, Soupart A, Fourmiguré J-F, Bruch A, Marty P. Experimental and
numerical study of annular PCM storage in the presence of natural convection.
Appl Energy 2013;112:175–84.
[43] Wang W-W, Wang L-B, He Y-L. The energy efficiency ratio of heat storage in
one shell-and-one tube phase change thermal energy storage unit. Appl
Energy 2015;138:169–82.
[44] López-Navarro A, Biosca-Taronger J, Corberán JM, Penalosa C, Lázaro A, Dolado
P, et al. Performance characterization of a PCM storage tank. Appl Energy
2014;119:151–62.
[45] Tay NHS, Bruno F, Belusko M. Experimental investigation of dynamic melting
in a tube-in-tank PCM system. Appl Energy 2013;104:137–48.
[46] Geissbühler L, Kolman M, Zanganeh G, Haselbacher A, Steinfeld A. Analysis of
industrial-scale high-temperature combined sensible/latent thermal energy
storage. Appl Therm Eng 2016;101:657–68.
[47] Zanganeh G, Khanna R, Walser C, Pedretti A, Haselbacher A, Steinfeld A.
Experimental and numerical investigation of combined sensible latent heat for
thermal energy storage at 575 °C and above. Sol Energy 2015;114:77–90.
[48] Geissbühler L, Kolman M, Zanganeh G, Haselbacher A, Steinfeld A. Multi-
layered solid-PCM thermocline thermal storage concept for CSP plants.
Numerical analysis and perspectives. Appl Energy 2015;142:337–51.
[49] V. Gesellschaft Verfahrenstechnik und Chemieingenieurwesen, VDI Heat Atlas.
Springer.
[50] Xiao X, Zhang P. Numerical and experimental study of heat transfer
characteristics of a shell-tube latent heat storage system: Part I – charging
process. Energy 2015;79:337–50.
[51] Xiao X, Zhang P. Numerical and experimental study of heat transfer
characteristics of a shell-tube latent heat storage system: Part II –
discharging process. Energy 2015;80:177–89.
[52] Zhang P, Ma F, Xiao X. Thermal energy storage and retrieval characteristics of a
molten-salt latent heat thermal energy storage system. Appl Energy
2016;173:256–71.
[53] Zauner C, Hengstberger F, Etzel M, Lager D, Hofmann R, Walter H.
Experimental characterization and simulation of a fin-tube latent heat
storage using high density polyethylene as PCM. Appl Energy
2016;179:237–46.
[54] Modelica Association. Modelica – a unified object-oriented language for
systems modeling. <www.modelica.org>.
[55] Dassault systems; 2016. <www.dymola.com>.
[56] Mehling H, Cabeza LF. Heat and cold storage with PCM. Springer; 2010.
[57] Gnielinski V. VDI heat atlas: heat transfer in cross-flow around single rows of
tubes and trough tube bundles, vol. 2010. Springer; 2010 [chapter G7].
[58] Gaddis ES, Gnielinski V. VDI heat atlas: shell-side heat transfer in baffled shell-
and-tube heat exchangers, vol. 2010. Springer; 2010 [chapter G8].
[59] Salzgitter Mannesmann Stahlhandel Gmbh. Düsseldorf, Germany. <https://
shop.salzgitter-mannesmann-stahlhandel.com/>.
[60] Kunststoff Information Verlagsgesellschaft mbh. Bad Homburg, Germany.
<https://rohstoffe.kunststoffweb.de/en/index.html>.
[61] Schmalisch B, Weber B, Harrison A, Gradinger R, Zauner C. Additives Fertigen
mit der MPA-Technologie am Beispiel eines Druckgusswerkzeuges, 9.
Ranshofner Leichtmetalltage; 2016.