Applied Energy 360 (2024) 122895

Contents lists available at ScienceDirect

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

Exergy efficiency and thermocline degradation of a packed bed thermal

energy storage in partial cycle operation: An experimental study
Paul Schwarzmayr ∗, Felix Birkelbach, Heimo Walter, René Hofmann
Institute for Energy Systems and Thermodynamics, TU Wien, Getreidemarkt 9, Vienna, 1060, Austria

ARTICLE INFO ABSTRACT

Keywords: To enable the exploitation of the industry sectors’ huge waste heat potential this study investigates the
Packed bed thermal energy storage utilization of packed bed thermal energy storage systems for the waste heat recovery in the iron and steel
Waste heat recovery industry. The main goal is to assess the partial cycle operation of a packed bed thermal energy storage
Energy/exergy efficiency
with industrial exhaust gas that is contaminated with high amounts of metal powder. A continuously rising
Thermocline degradation
accumulation of powder particles inside the packed bed requires the storage to be charged from the bottom
Partial cycle operation
Iron/steel industry
in order to facilitate the removal of powder hold-up. Therefore, investigations focus on the dependency of the
thermal performance of the storage on the flow direction of heat transfer fluid during charging/discharging.
Exergy efficiency and thermocline degradation are evaluated as key performance indicators for different flow
directions of the heat transfer fluid using a lab-scale test rig. Results show that, compared to charging from
the top, a slightly faster thermocline degradation occurs for a storage that is charged from the bottom. Still,
energy and exergy efficiencies in partial cycle operation are well above 85% and 80% respectively regardless
of the heat transfer fluid flow direction. The thermal power rate during discharging is stable between 74%
and 93% with respect to the maximum input rate for both charging the storage from the top and from the
bottom. All in all, packed bed thermal energy storage systems are found to be suitable for waste heat recovery
in industrial processes. Only a marginal deterioration of the thermal performance of the storage has to be
expected if the storage is required to be charged from the bottom.

1. Introduction systems also have some interesting qualities from a technical point of
view. Since PBTES systems take advantage of an effect called thermal
A packed bed thermal energy storage (PBTES) is a sensible type of stratification [15], they typically reach higher exergy efficiencies than
thermal energy storage (TES) that uses a packed bed of solids as heat conventional sensible heat TES systems [16,17] like Ruths steam stor-
storage material, a gas (or liquid [1]) as heat transfer fluid (HTF) [2,3] age [18] and hot water tanks. Thermal stratification means, that for a
and is capable of storing high-temperature heat. The fact that the HTF partially charged PBTES the storage volume is divided into a hot and
in a PBTES gets in direct contact with the storage material leads to a cold zone that are separated by a thin volume-slice which is called
an enhanced heat transfer and therefore high power rates compared to ‘‘thermocline’’ [19]. To charge the PBTES, hot HTF passes through the
other storage technologies that use a gaseous medium as HTF [4,5]. storage entering from the hot end. This expands the volume of the hot
Moreover, there is no need for heat exchanger tubes as it would be zone. To recover the stored heat, cold HTF passes through the packed
in other sensible and/or latent heat TES systems [6]. This reduces
bed entering from the cold end, which pushes back the thermocline
investment costs and increases the robustness of PBTES systems against
and expands the cold zone [20]. In a well designed PBTES [21,22] the
aggressive/abrasive HTFs.
thermocline is as thin as possible which allows to recover heat at nearly
As the HTF can be any type of gas and the storage material can be
the same temperature at which it was stored for most of the discharging
any type of solids (rocks/gravel [7], slag [8], glass spheres [9], porce-
period [23,24]. This effect leads to higher exergy efficiencies compared
lain spheres [10], ceramic spheres [11], alumina spheres [12]) PBTES
systems are usually cost-efficient and versatile in its application [13]. to TES technologies that do not take advantage of thermal stratification.
Furthermore, PBTES systems use one non-pressurized vessel as storage Based on these qualities a potential application of PBTES systems
tank which further reduces investment costs compared to other storage is waste heat recovery in the energy-intensive industry [25]. In their
technologies [14]. In addition to these economic advantages, PBTES study on the selection of TES options for utilization of industrial waste

∗ Corresponding author.
E-mail addresses: [email protected] (P. Schwarzmayr), [email protected] (R. Hofmann).

https://doi.org/10.1016/j.apenergy.2024.122895
Received 19 September 2023; Received in revised form 2 February 2024; Accepted 17 February 2024
Available online 21 February 2024
0306-2619/© 2024 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
P. Schwarzmayr et al. Applied Energy 360 (2024) 122895

Nomenclature h Hot
htf Heat transfer fluid
Acronyms 𝑖 Sample index
𝑗 Volume element index
ASU Air supply unit
loss Losses to the surrounding
EU European Union
max Maximum
FWD Forward
min Minimum
HTF Heat transfer fluid
mix Mixed
LD Linz-Donawitz
p Pumping
PBTES Packed bed thermal energy storage
Q Heat
REV Reverse
ref Reference (𝑇ref = 303.15 K,
RTD Resistance temperature detector
𝑝ref = 101 325 Pa)
SOC State of charge
sat Saturation (losses)
TES Thermal energy storage
str Stratified
Roman symbols
Superscripts
𝐴 Volume-specific surface area of the packed
+ Upper boundary of the SOC
bed in m2 m−3
− Lower boundary of the SOC
𝐵 Exergy in J
𝐵̇ Exergy in-/output rate in W
𝐵̂̇ Maximum gross exergy in-/output rate in
W heat, Manente et al. [26] state, that the integration of PBTES systems
 Set of sample indices for charging can significantly increase the efficiency and process steam output of in-
𝑐 Specific heat capacity of storage material dustrial energy systems. By storing excess heat from industrial processes
J kg−1 K −1 during periods of low heat demand and releasing the stored heat during
𝑐𝑝 Specific heat capacity at constant pressure periods of high heat demand, PBTES systems can contribute to the
for HTF in J kg−1 K −1 exploitation of the industry sector’s enormous waste heat potential. The
𝑀 Momentum of energy in J m theoretical waste heat potential in the European Union (EU) industry
𝑚 Mass in kg was estimated to be 918 TW h in 2014 [27] which is nearly 8% of
𝑀𝐼𝑋 𝑀𝐼𝑋 number the EU’s annual final energy consumption. As stated by Papapetrou
𝑝 Pressure in Pa et al. [28], approximately one third of this theoretical waste heat
potential (304 TW h) is technically recoverable. By acting as a substitute
𝑄 Heat in J
for heat generated from burning natural gas, the utilization of this
𝑄̇ Heat/energy in-/output rate in W
waste heat potential could significantly reduce annual CO2 emissions.
𝑄̂̇ Maximum gross heat/energy in-/output
The main barriers of implementing waste heat recovery in the
rate in W energy-intensive industry are of economic and technological nature.
 Set of sample indices for recovery/ On the one hand, waste heat recovery systems should be cost efficient
discharging and on the other hand they have to operate under very challenging
 Set of temperature sensors inside the conditions which makes them expensive. Table 1 provides a summary
packed bed  = [1, 9] of recent publications in which the improvement of high-temperature
𝑆 Entropy in J K −1 industrial waste heat recovery using a TES was investigated. As indus-
𝑆𝑂𝐶 State of charge trial waste heat is almost always available in form of hot exhaust gas
𝑇 Temperature in K that is contaminated with high amounts of dust, waste heat recovery
𝑈 Internal energy in J systems need to be designed to operate with gas-powder two phase
𝑊 Work in J flow as HTF. Powder that passes through the system may accumulate
in the packed bed and could lead to a degradation of the thermal
𝑥 Vertical distance of temperature sensors to
performance or even a clogging of the voids between the packed bed
cold end of the storage in m
particles of the storage material [29]. In their studies on utilizing TES
Greek symbols systems for the waste heat recovery in the iron and steel industry
Keplinger et al. [30] and Ortega-Fernández et al. [31] tackled this
𝛼 Heat transfer coefficient between packed
issue by placing a gas-to-gas heat exchanger between the waste heat
bed and HTF in W m−2 K −1
source and the storage (see Fig. 1). Similar studies were conducted
𝜅 Isentropic exponent of HTF (𝜅 = 1.4) by Slimani et al. [32] who investigated a horizontal flow PBTES that
𝜂 Efficiency uses slag bricks as storage material. For detailed information about
𝜈 Specific volume of HTF in m3 kg−1 the design and analysis of horizontal flow PBTES systems the authors
refer to Odenthal et al. [33] Soprani et al. [17]. Since gas-to-gas heat
Subscripts
exchangers suffer from low exergy efficiencies and the lifetime of a
B Exergy heat exchanger that is operated with a gas-powder two phase exhaust
c Cold gas will be short, this approach is not considered efficient from both
E Energy a technological and an economic point of view. Therefore, the authors
exp Experimental of this study propose the direct use of gas-powder two phase exhaust
gas from a steel making process as HTF in a PBTES as it is shown in
gen Generation
Fig. 1. They gray lines and components (gas-to-gas heat exchanger)
in this figure represent the system as it was proposed by Keplinger

2
P. Schwarzmayr et al. Applied Energy 360 (2024) 122895

Table 1
Summary of studies that consider PBTES systems for waste heat recovery in the iron and steel industry.
Ref. Objective Approach Aspects and results relevant for this work
Keplinger et al. [30] Improve electric arc furnace waste heat Dynamic simulation • Hot water stratified storage is integrated via a hot-gas heat exchanger
recovery via thermocline TES integration study • Thermocline storage integration enables decoupling of waste heat
source and from steam production
• Maximum productivity and economic efficiency can be reached
Ortega-Fernández et al. [31] Waste heat recovery from electric arc Case study • High temperature filter and flue-gas heat exchanger are used to
furnace exhaust gases using a TES protect PBTES from gas-powder two phase flow
• Operation planning to synchronize waste heat source and PBTES is of
high importance
• System efficiency of around 65% is obtained
Slimani et al. [32] Horizontal PBTES for waste heat Numerical and • Drop-out box and flue-gas heat exchanger are used to protect PBTES
recovery in the iron and steel industry economic study from gas-powder two phase flow
• Drop-out box and heat exchanger alone have a combined thermal
efficiency of 45%
• PBTES waste heat recovery using slag as storage material is
economically and ecologically attractive
Touzo et al. [34] PBTES for high-temperature waste heat Experimental and • The issue of gas-powder two phase flow was not investigated
recovery numerical analysis • Performance of commercial-scale PBTES was validated
• During cyclic operation the PBTES reaches a thermal efficiency of up
to 90%

et al. [30] and Ortega-Fernández et al. [31]. The colored lines illustrate a PBTES is considered. For a PBTES that is to be used for the waste
the integration of the PBTES for the direct use of industrial exhaust gas heat recovery in a steel producing process, standby periods will be
as HTF in a PBTES. This approach significantly reduces investment costs seldom and cycle durations will be short (40 min to 60 min). There
(no heat exchanger and less piping) and increases the exergy efficiency are no studies that compare the performance of a PBTES in partial
of the whole system. However, the challenges that arise when using cycle operation for different flow directions of the HTF available to the
gas-powder two phase exhaust gas as HTF in a PBTES are still in place. authors. The remainder of this introductory section consist of a concise
In one of their previous studies [29] the authors already investi- literature review of existing partial cycle studies and an elaboration of
gated effects like pressure drop and powder hold-up in a PBTES that the research gap this paper aims to address. A short summary of this
is operated with a gas-powder two phase exhaust gas from a steel literature review is also provided in Table 2.
producing process. The results of this study revealed that the proposed Comprehensive reviews considering experimental investigations on
approach is viable for the waste heat recovery from gas-powder two the partial cycle operation of PBTES systems were provided by Esence
phase exhaust gas if certain measures are taken. In [29] it was found et al. [35] and Gautam et al. [2,3]. In their papers on packed bed
that 98% of the powder that is introduced into a PBTES with gas-powder thermocline storage performance Bayón et al. [37], Biencinto et al. [38]
two phase exhaust gas deposits inside the storage and is concentrated and Wang et al. [39] point out the importance of the operation strat-
at the packed bed’s surface at which the HTF enters the packed bed. egy (partial extraction of the thermocline at the end of charing and
Therefore it was suggested to switch the operating conditions of a discharging) on the performance of a PBTES in dynamic operation.
PBTES so that it is charged from the bottom and discharged from Zavattoni et al. [40,41] conducted experiments considering the thermal
the top. This way the powder hold-up inside the storage would be stratification under charge/discharge cyclic conditions on their small-
concentrated at the bottom surface of the packed bed where it can scale TES test rig and found that it takes up to 30 cycles until a steady
be easily removed via knocking/trembling mechanisms. Therefore one state operation is established. Mertens et al. [42] investigated a PBTES
major adaption that has to be made compared to PBTES systems that for solar applications and highlighted that preheating is essential for
operate in modern concentrated solar power (CSP) plants is the flow the long term performance of a PBTES in partial cycle operation. They
direction of the HTF during the charging and discharging periods. Since also state, that although this leads to an additional energy consumption
the natural choice of the HTF flow direction through a PBTES during during initial cycles, it also provides additional flexibility during the
charging is from top to bottom [35] only one study that investigates the long term operation. Zanganeh et al. [43] numerically investigated an
thermal performance of a PBTES that is charged from bottom to top is industrial scale gas–solid PBTES and stated, that the long-term perfor-
available [36]. However, in this study only the standby operation of mance of a PBTES is defined by its stabilized state. They found, that the
duration of the stabilizing phase, both in terms of time and number
of cycles can be controlled via operating parameters. Similar investi-
gations were conducted by Bruch et al. [23] and Johnson et al. [44].
Bruch et al. experimentally studied an oil-rock PBTES in partial cycle
operation and came to similar conclusions as Zanganeh et al.. Especially
for short cycles, they found that thermocline degradation during partial
cycle operation reduced the effective storage capacity by about 10%.
The same conclusions were drawn by Cascetta et al. [45] who investi-
gated a gas–solid PBTES under partial cycle operation. The numerical
study of McTigue et al. [46] provides evidence on decreased efficiencies
for long partial cycles especially if perturbations of cycle durations are
involved. In their latest study on a pilot-scale oil-rock PBTES, Bruch
et al. [24] investigated the impact of end-of-discharge criteria on the
utilization rate of the storage capacity. They found, that by lowering the
minimum discharge temperature, the utilization rate can be increased
significantly. Since all these existing studies investigate the partial cycle
Fig. 1. Comparison of literature approach [30,31] and the system proposed in the operation for cut-off temperature-controlled cycles, the application of
present study. their results to a PBTES that is to be used for waste heat recovery in the

3
P. Schwarzmayr et al. Applied Energy 360 (2024) 122895

Table 2
Summary of studies carried out to investigate the partial cycle operation of a PBTES.
Ref. Objective Approach Results
Bayón et al. [37] Exergy efficiency of a PBTES in Numerical study • Storage efficiency decreases during consecutive charge/discharge cycles
partial cycle operation • This strong decrease can be relieved via partial thermocline extraction
• Without thermocline extraction, long term energy efficiency is below 50%
Bruch et al. [23] Efficiency of an oil/rock PBTES in Numerical and • Valuable part of storage tank and/or thermal efficiency reduce during partial
partial cycle operation experimental study cycle operation
• Long-term efficiency in partial cycle operation is defined by a stabilized
state which is reached after 10 cycles
• Thermocline control during long-term partial cycle operation is important
Bruch et al. [24] Cycling behavior of a pilot-scale Experimental study • Stabilized state in partial cycle operation is independent on initial conditions
oil/rock PBTES • Proper definition of end-of-cycle condition is important for efficient storage
operation
• Energetic efficiencies of 88% to 92% were observed
Cascetta et al. [45] Thermal behavior of a PBTES in Experimental study • Thermocline degradation during partial cycle operation is significant
repeated cycle operation • Stabilized state is reached after 7 cycles
• Stabilized state can be controlled by design and operational parameters
(mass flow, temperature threshold, aspect ratio)
Johnson et al. [44] Exergy efficiency of a PBTES for Numerical study • Exergy losses during partial cycle operation are cause by thermocline
ten charge–discharge cycles degradation and heat losses to the surrounding
• Exergy efficiency in the stabilized state is 80% to 84%
• To improve the partial cycle performance of PBTES, the focus should be on
minimizing thermocline degradation
Mertens et al. [42] Dynamic simulation of a PBTES Numerical study • Stabilized state is reached after 5 cycles
for CSP applications • Stabilized state energy efficiency is observed to be 95%
• Air/rock PBTES systems are an efficient and cost-effective alternative to
molten-salt TES systems
Schwarzmayr et al. [36] Efficiency of a PBTES in standby Experimental study • Impact of HTF flow direction on thermocline degradation was investigated.
operation • For long standby periods, charging the PBTES from the bottom leads to
significantly higher exergy losses
• If standby periods are short, charging a PBTES from the bottom does not
lead to additional energy or exergy losses
Zanganeh et al. [43] Testing of a pilot-scale PBTES for Numerical and • Thermocline degradation could be observed during partial cycle operation
CSP applications experimental study • Stabilized state thermal efficiency was observed to be 95%
Zavattoni et al. [40] Analysis of a PBTES under Numerical (CFD) study • Stabilized state is reached after 8–10 cycles
charge/discharge cyclic conditions • Stabilized state thermal efficiency was observed to be 95%
• Pumping energy accounts for 7% of the input energy during charging

iron and steel industry is limited. When integrated into an industrial of the measurement data. Experimental results as well as a summary
energy system, a TES has to follow the schedule of the upstream and interpretation of the main findings are provided in Section 4. These
industrial process. Therefore, a time-controlled operation of the TES main findings include reference values for energy/exergy efficiencies
is more appropriate in this case. Under such operating conditions, end- and maximum power rates for different operating conditions of the test
of-charge and end-of-discharge criteria as defined by Bruch et al. [24] rig. For convenience, some additional abbreviations will be used in the
may not be satisfied for any cycle which has a negative impact on rest of this paper. Charging the PBTES from the top and discharging it
the exergy efficiency of the PBTES. Furthermore, as delineated in the from the bottom will be referred to as the FWD-mode, whereas charging
previous paragraph, none of the mentioned studies (except [36], where the PBTES from the bottom and discharging is from the top will be
the standby operation was studied) investigated a PBTES that is charged referred to as the REV-mode.
from the bottom and discharged from the top.
Therefore the goal and novelty of this work is the experimental 2. Material and methods
investigation of the thermal performance of a PBTES in time-controlled
partial cycle operation with a HTF charging flow direction from bottom 2.1. Experimental setup
to top. The motivation for this adapted operating strategy (shown in
Fig. 1) – which stands in contrast to favorable operating conditions in For the evaluation of the exergy efficiency and thermocline degra-
terms of mitigating thermocline degradation – is the accumulation of dation of PBTES systems in partial cycle operation a lab-scale test rig
powder hold-up in a PBTES that is charged with gas-powder two phase of a vertical gas–solid PBTES was used. The main part of the test rig
exhaust gas. Key performance indicators such as exergy efficiency, ther- – the storage tank – is a vertical, slightly conical steel vessel that
mocline degradation and thermal power rates will be used to evaluate is filled with 891 kg of storage material. The conical design of the
the proposed operating strategy. The results provided in this study will storage tank reduces thermo-mechanical stresses (thermal ratcheting
support industrial decision makers to assess the potential and economic see e.g. [47]) in the walls. LD-slag, a by-product from the iron and
benefits of a PBTES system for the waste heat recovery in their energy steel industry, is utilized as storage material. The choice of LD-slag as
intensive processes. storage material is substantiated both by its low costs and excellent
The remainder of this paper is divided into four sections. The most heat transfer properties. The geometric shape of the slag particles leads
important information and properties of the materials and the test to an enhanced heat transfer between HTF and storage material and
rig that are used for the experimental investigations are provided in an even perfusion of the packed bed. Both have a positive impact on
Section 2. At the end of Section 2 the experimental procedure as well the thermal performance (power rate and storage capacity) of the TES.
as some important experimental operating parameters are summarized. The whole test rig (storage tank and piping) is equipped with multiple
Section 3 includes the main equations, uncertainty estimations and con- layers of thermal insulation to reduce heat losses. In Fig. 2 photographs
ceptional considerations that are necessary to understand to evaluation of the storage material and the test rig with and without thermal

4
P. Schwarzmayr et al. Applied Energy 360 (2024) 122895

Fig. 2. PBTES test rig at the laboratory of TU Wien: with thermal insulation (left),
storage material (center), without thermal insulation (right).
Source: Reprinted from [36].

insulation are shown. Detailed information about the test rig geometry,
the utilized materials, the insulation and the operating conditions for
the experiments are provided in Table 3.
In order to measure the thermal behavior of the experimental setup,
the test rig is equipped with various temperature sensors. Two sensors
are placed in the pipes that are connected to the top and bottom of the Fig. 3. Simplified process flow diagram of the PBTES test rig at the laboratory of TU
storage tank to measure the temperature of the HTF that enters/exits Wien.
the packed bed (𝑇B and 𝑇T ). The temperature distribution inside the Source: Reprinted from [36].
packed bed is measured with nine sensors that are evenly distributed
along the center axis of the storage tank. To guarantee accurate mea-
surements, four-wire resistance temperature detectors (RTD) class AA Furthermore this effect may be even worse if the TES is designed to
(or 1/3 DIN) PT100 sensors are utilized. The measurement uncertainty be operated in REV-mode.
of the temperature sensors including all steps of signal conversion is In order to close the research gap described at the end of the
less than ±0.6 °C for a measured value of 300 °C. To estimate exergy introduction (Section 1), multiple experiments were conducted as part
losses due to pumping work a piezoresistive differential pressure sensor of this study. For a better understanding of the experimental procedure,
is used to measure the pressure drop 𝛥𝑝 between the packed bed in- Fig. 4 provides a schematic figure of the evolution (degradation) of
and outlet pipes. This sensor was calibrated to an accuracy of ±0.06% the temperature profiles inside a PBTES during partial cycle operation
of full scale before the experiments. Fig. 3 shows a simplified process in FWD-mode. As depicted in Fig. 4, there is a hot and a cold zone
flow diagram of the test rig including the sensor layout relevant for the inside the packed bed throughout the whole process of partial cycle
investigations carried out in this work. For a detailed description of the operation. Depending on the operating mode (charging or discharging)
sensor layout the authors refer to their previous study [36]. the thermocline moves downwards or upwards within the active zone
The PBTES test rig is provided with hot and cold HTF by an external of the packed bed. Based on the review of existing studies it is expected
air supply unit (ASU). The ASU is capable of supplying up to 400 kg h−1 that the shape of the temperature profile inside the packed bed changes
of air with temperatures between ambient temperature and 400 °C. The (degrades) throughout the partial cycle operation (compare Fig. 4a and
b). The purpose of the experiments carried out in the course of this
HTF mass flow that passes through the system is controlled by the
study is to investigate whether this effect (theromcline degradation) is
ASU and is measured with a hot-wire anemometer. The measurement
more pronounced when the PBTES is operated in REV-mode rather than
uncertainty of this sensor does not exceed 4% with respect to the
in FWD-mode. Therefor partial cycle experiments both in FWD- and
measured value.
REV-mode were conducted using the experimental setup presented in
Section 2.1. Every experiment started with an empty storage (constant
2.2. Experimental procedure temperature of 30 °C) that was then charged with hot air (300 °C)
until a state of charge (SOC) of approximately 50% was reached. In
A PBTES that is operated at short charging/discharging periods and the next step, the storage was operated in partial cycle operation
high power rates, can guarantee maximum exergy efficiency and flex- (between a lower and an upper SOC, 𝑆𝑂𝐶 − and 𝑆𝑂𝐶 + respectively)
ibility when it is cycled around a partially charged(≈50%) state [35]. where it was alternately charged and discharged with a HTF mass
The problem with such an operating strategy is that thermocline degra- flow of 230 kg h−1 and very short cycle periods until a steady cycling
dation may significantly reduce the storage’s efficiency over time. operation was reached. The HTF temperatures during charging and

5
P. Schwarzmayr et al. Applied Energy 360 (2024) 122895

Table 3
Summary of parameters: Test rig geometry, data/properties of storage material,
operating conditions.
Source: Adapted from [36].
Test rig
Storage type Vertical PBTES
Tank material P295
Tank diameter 350 mm (bottom), 500 mm (middle),
650 mm (top)
Tank height 2050 mm
Storage volume 0.405 m3
Thermal insulation 100 mm ceramic wool (𝜆 = 0.055 W m−1 K −1 )
80 mm rock wool (𝜆 = 0.04 W m−1 K −1 )
0.1 mm aluminum sheeting
Storage material
Type of storage material LD-slag (irregular shaped, partly porous)
Composition of CaO (24%–49%), SiO2 (6%–37%),
storage material Fe𝑥 O𝑦 (10%–36%), MgO (0%–13%),
Al2 O3 (0%–7%), Cr2 O3 (0–0.55%)
Grain size 16 mm to 32 mm
Particle density 3800 kg m−3 Fig. 4. Experimental procedure for partial cycle operation of a PBTES: (a) vertical
Fractional void volume 0.42 temperature profile at the end of the first cycle’s charging and discharging period,
Bulk density 2200 kg m−3 (b) vertical temperature profile at the end of the 𝑛th cycle’s (where a steady cycling
Specific heat capacity 900 J kg−1 K −1 operation is established) charging and discharging period.
Operating parameters
Mass of storage material 891 kg
Air mass flow 230 kg h−1 the system boundary due to heat transfer between HTF and storage
Charging temperature 300 °C material can be calculated as
Recovery temperature 30 °C
𝑇h,𝑖
HTF mass flux 0.17 kg m−2 s−1 to 0.58 kg m−2 s−1
𝛿𝑄htf,𝑖 = 𝛿𝑚htf,𝑖 𝑐𝑝 (𝑇 ) d𝑇 . (3)
Cycle duration 40 min to 60 min ∫𝑇c,𝑖

where 𝛿𝑚htf,𝑖 is the amount of HTF that passes through the packed
bed, 𝑐𝑝 (𝑇 ) is the specific heat capacity of the HTF at isobaric condi-
discharging were set to 300 °C and 40 °C respectively. This procedure tions and 𝑇c,𝑖 and 𝑇h,𝑖 are the temperatures of the cold and hot HTF
was repeated for both FWD- and REV-mode operation of the storage entering/exiting the packed bed respectively.
and different charging/discharging periods (40 min to 60 min). The Analogously to the energy balance in Eq. (1) the entropy generation
reason for time-controlled charging/discharging periods rather than equation (second law) applied to the storage volume can be written as
SOC- or cut-of-temperature-controlled is the nature/characteristics of
industrial processes. When integrated into an industrial energy system, ∑
the operation of a TES is determined by the strictly scheduled produc- d𝑆𝑖,𝑗 = 𝛿𝑆Q,𝑖 + 𝛿𝑆gen,𝑖 (4)
𝑗 ∈
tion plan of the upstream industrial process. This operating behavior is
best simulated with time-controlled charging/discharging periods with where the left side is the change in system entropy and the right side is
a 40 min to 60 min cycle duration. the sum of entropy crossing the system boundary due to heat transfer
and the entropy generation due to irreversible effects. The change in
system entropy is defined by the equation of state given in Eq. (5).
3. Theory and calculations
𝑐(𝑇 )
d𝑆𝑖,𝑗 = 𝑚𝑗 d𝑇𝑖,𝑗 (5)
𝑇
In this section all equations that are used for the post-processing of
the measured data are provided and explained. The index 𝑖 represents The entropy introduced into the system due to heat crossing the
system boundary is per definition
the sample number and the index 𝑗 is used to represent the individual
temperature sensors (𝑗 ∈  = [1, 9]) inside the packed bed. Be- ∑ d𝑈𝑖,𝑗
𝛿𝑆Q,𝑖 = (6)
fore calculating energy/exergy efficiencies and evaluating thermocline 𝑗 ∈
𝑇htf,𝑖,𝑗
degradation, the measured data is validated using the first and second
law of thermodynamics applied to the storage material inside the where 𝑇htf,𝑖,𝑗 is the absolute temperature at which the heat that is
equivalent to d𝑈𝑖,𝑗 crosses the system boundary. It is assumed, that the
storage tank (system boundaries are the same as in [36]). The energy
entropy transport due to heat losses to the surrounding is negligible
balance (first law) of this thermodynamic system can be written as
∑ compared to the entropy transport due to heat transfer between HTF
d𝑈𝑖,𝑗 = 𝛿𝑄htf ,𝑖 − 𝛿𝑄loss,𝑖 (1) and storage material. The impact of this assumption on the significance
𝑗 ∈ of the results presented in this study is discussed in Section 4. The
where the left side is the change in the system’s internal energy and absolute temperature 𝑇htf,𝑖,𝑗 is calculated using Eq. (7), where 𝛼 is
the right side is the sum of all heat crossing the system boundary. The the heat transfer coefficient between HTF and storage material and
change in the system’s internal energy is defined by the caloric equation 𝐴 is the volume-specific surface of the packed bed. Both parameters
of state as were determined via model-based parameter optimization using a finite
volume model of the test rig [48].
d𝑈𝑖,𝑗 = 𝑚𝑗 𝑐(𝑇 ) d𝑇𝑖,𝑗 (2) d𝑈𝑖,𝑗
𝑇htf,𝑖,𝑗 = 𝑇𝑖,𝑗 + (7)
where 𝑇𝑖,𝑗 is the measured temperature value of each sensor located in 𝛼𝐴
the packed bed, 𝑚𝑗 is the mass of storage material that is assigned to Considering both Eqs. (1) and (4), the measured data is considered
each temperature sensor and 𝑐(𝑇 ) is the specific heat capacity of the valid if the heat losses to the surrounding 𝛿𝑄loss in Eq. (1) and the
storage material which is assumed to be constant. The heat crossing entropy generation 𝛿𝑆gen,𝑖 in Eq. (4) are positive.

6
P. Schwarzmayr et al. Applied Energy 360 (2024) 122895

The energy and exergy efficiency of the examined PBTES test rig in of energy that is stored in each vertical volume section 𝑈𝑖,𝑗 is weighted
partial cycle operation are defined as with its vertical distance 𝑥𝑗 from the cold end of the storage as
∑ ∑
𝛿𝑄htf ,𝑖 𝑀𝑖 = 𝑥𝑗 𝑈𝑖,𝑗 (16)
𝑖∈ 𝑗 ∈
𝜂E = ∑ ∑ ∑ (8)
𝛿𝑄htf,𝑖 + 𝛿𝑄sat,𝑖 + 𝛿𝑊p,𝑖 With this definition, the 𝑀𝐼𝑋 number is a suitable metric to
𝑖∈ 𝑖∈ 𝑖 ∈ ( ∪ )
describe the thermocline in a PBTES with an uneven distribution of the
and storage mass. Andersen et al. defined the 𝑀𝐼𝑋 number as
∑
𝛿𝐵htf,𝑖 𝑀str,𝑖 − 𝑀exp,𝑖
𝑖∈ 𝑀𝐼𝑋 𝑖 = (17)
𝜂B = ∑ ∑ ∑ . (9) 𝑀str,𝑖 − 𝑀mix,𝑖
𝛿𝐵htf ,𝑖 + 𝛿𝐵sat,𝑖 + 𝛿𝑊p,𝑖
𝑖∈ 𝑖∈
where 𝑀exp,𝑖 is the momentum of energy calculated from the measured
𝑖 ∈ ( ∪ )
data, 𝑀str,𝑖 is the momentum of energy for an entirely stratified storage
The energy efficiency 𝜂E is defined as the ratio of the heat that can and 𝑀mix,𝑖 is the momentum of energy of an entirely mixed storage.
be extracted from the test rig during a discharging period (𝑖 ∈ ) to the Both 𝑀str,𝑖 and 𝑀mix,𝑖 are calculated for a virtual storage with the same
amount of heat that is introduced into the system during the preceding SOC as the real storage. With this definition, the 𝑀𝐼𝑋 number is 0
charging period (𝑖 ∈ ) plus the amount of work that is needed to if the real storage is perfectly stratified and 1 if it is perfectly mixed.
pump the HTF through the packed bed during charging and discharging Details on the calculation of the 𝑀𝐼𝑋 number and the definition of an
periods (𝑖 ∈ ( ∪ )). The exergy efficiency is defined analogously to entirely stratified and mixed storage are provided by Oró et al. [15].
the energy efficiency as given in Eq. (9). The impact of measurement uncertainties on the quality of the
The exergy change that the HTF experiences when it passes through calculated results is estimated using the law of error propagation. The
the packed bed can be calculated as maximum errors of the measuring devices used for the experiments
in this study are ±4% of the measured value for the air mass flow
𝛿𝐵htf,𝑖 = 𝛿𝑄htf,𝑖 − 𝑇ref 𝛿𝑆htf ,𝑖 (10) sensor and ±0.6 K for the temperature sensors. With these two values
where 𝑇ref is the reference temperature (30 °C = 303.15 K) and the maximum errors in the calculated heat and exergy 𝛿𝑄htf ,𝑖 and
𝛿𝐵htf ,𝑖 defined in Eqs. (3) and (10) are well below 4% of the calculated
𝑇h,𝑖 𝑐𝑝 (𝑇 )
𝛿𝑆htf,𝑖 = 𝛿𝑚htf,𝑖 d𝑇 . (11) values. The uncertainties in the energy and exergy efficiency 𝜂E and 𝜂B
∫𝑇c,𝑖 𝑇 are calculated using the repeatability of the utilized mass flow sensor
(±0.5%). This results in a maximum error that is less than 1% of the
It can be seen, that both the energy and exergy efficiency incor-
calculated value for both quantities. Since these errors are nominal,
porate saturation losses (𝛿𝑄sat,𝑖 and 𝛿𝐵sat,𝑖 ) for the charging periods.
they do not have any impact on the quality and significance of the
These losses are caused by increased exhaust gas temperatures and can
results presented in this study.
be calculated as
𝑇c,𝑖
𝛿𝑄sat,𝑖 = 𝛿𝑚htf ,𝑖 𝑐𝑝 (𝑇 ) d𝑇 and 𝛿𝐵sat,𝑖 = 𝛿𝑄sat,𝑖 − 𝑇ref 𝛿𝑆sat,𝑖 (12) 4. Results and discussion
∫𝑇ref

where Before presenting and discussing the core results of this study the
𝑇c,𝑖
validity and significance of the measured data is clarified shortly. The
𝑐𝑝 (𝑇 )
𝛿𝑆sat,𝑖 = 𝛿𝑚htf ,𝑖 d𝑇 . (13) first and second law are used for data validation as already presented in
∫𝑇ref 𝑇 the beginning of Section 3. The experimental data satisfies the energy
𝛿𝑆sat,𝑖 is the entropy difference between the HTF that exits the balance in Eq. (1) with average heat loss rates of 525 W and 287 W
storage and the surrounding air. for charging and discharging respectively. Similarly, the second law
The amount of pumping work 𝛿𝑊p,𝑖 that is needed to overcome the in Eq. (4) is satisfied with an average entropy generation of 0.63 W K −1
for charging and 0.1 W K −1 for discharging. Since these values are all
pressure drop 𝛥𝑝𝑖 caused by the packed bed can be calculated as
positive, the experimental data collected in the course of this study is
ref 𝑝 + 𝛥𝑝𝑖
1 considered valid.
𝛿𝑊p,𝑖 = 𝛿𝑚htf,𝑖 𝜈 d𝑝 with 𝑝 𝜈 𝜅 = const. (14)
𝜂p ∫𝑝ref Multiple experiments for the FWD- and REV-mode with different
charging/discharging periods are carried out in the course of this study.
where 𝑝ref is the ambient pressure, 𝛥𝑝𝑖 is the pressure drop of the
The results reveal, that the length of the charging/discharging periods
packed bed measured during the experiments, 𝜈 is the specific volume
does not have a significant impact on the thermocline degradation.
of the HTF passing through the packed bed, 𝜅 is the isentropic exponent
Therefore, only results from the experiments with charging/discharging
of the HTF and 𝜂p is the efficiency (assumed to be 0.35 [43]) of the
periods of 60 min are discussed in the remainder of this section. Figs. 5
aggregate that is utilized to pump the HTF through the packed bed.
and 6 display the axial temperature distribution inside the test rig
For a more convenient data visualization in Section 4 the SOC is
during the experiments where the TES was operated in FWD- and
used to describe the amount of energy stored in the PBTES at a certain REV-mode respectively. The ordinate represents the dimensionless tem-
time. In this study the SOC is defined as perature where 0 corresponds to the ambient temperature (30 °C) and
∑
𝑗 ∈  𝑈𝑖,𝑗 − 𝑈min
𝑇𝑖,𝑗 1 corresponds to 300 °C. On the abscissa, the dimensionless height of
𝑆𝑂𝐶 𝑖 = where 𝑈𝑖,𝑗 = 𝑚𝑗 𝑐(𝑇 ) d𝑇 . (15) the packed bed is plotted, where 0 corresponds to the bottom (0 m) and
𝑈max − 𝑈min ∫ 𝑇ref
1 to the total height of the packed bed (2.05 m). The blue lines and
𝑈min is the energy that is stored in a fully discharged storage (𝑇𝑗 = areas show the temperature profile inside the packed bed for the first
30 °C ∀ 𝑗 ∈  = [1, 9]) and 𝑈max is the energy for a fully charged storage charging/discharging cycle and the orange lines and areas show the
(𝑇𝑗 = 300 °C ∀ 𝑗 ∈  = [1, 9]). temperature profile of the 𝑛th cycle where steady cycling is reached.
Finally, the 𝑀𝐼𝑋 number [49] can be calculated, to quantify the For the given operating parameters steady cycling is reached after
thermocline degradation of the examined PBTES in partial cycle op- the 10th cycle (corresponding to less than 10 h) for the experiments
eration. The 𝑀𝐼𝑋 number is a dimensionless number that was first carried out in this study. Besides the degradation of the thermocline,
defined by Andersen et al. [50] and is based on the momentum of the most important observation that can be made in these figures are
energy. For the calculation of the momentum of energy, the amount the temperature differences at both ends of the storage. During the first

7
P. Schwarzmayr et al. Applied Energy 360 (2024) 122895

Fig. 5. Thermocline degradation of a PBTES in partial cycle operation (FWD-mode) Fig. 6. Thermocline degradation of a PBTES in partial cycle operation (REV-mode)
with a cycle time of 1 h. with a cycle time of 1 h.

cycle (blue lines and areas) the temperatures at the top and bottom charging/discharging cycle. Efficiencies for the 𝑛th cycle, where steady
end of the packed bed do not change significantly for both operating cycling is reached, are displayed as orange and green bars. Naturally,
the TES in FWD- and REV-mode. For the 𝑛th cycle (red lines and all exergy efficiencies are lower than the energy efficiencies which
areas), however, the temperature differences between the solid and is mainly caused by the temperature difference that is needed as a
dashed lines at the top and bottom ends of the packed bed are clearly driving force to transfer heat from the HTF to the storage material
bigger. This means that during charging, the exhaust temperatures and vice versa. Nevertheless, the exergy efficiencies achieved in the
of the HTF are above the ambient temperature leading to saturation experiments of this study all are greater than 80%, which is higher
losses. Likewise, the exhaust gas temperature during discharging is than other sensible heat TES systems (Ruths steam storage, hot water
significantly lower than the charging temperature. Both effects lead to a tanks) of comparable size can accomplish. These results coincide with
deterioration of the energy and exergy efficiency of the storage, which findings from Knobloch et al. and Soprani et al. [16,17]. Furthermore,
will be confirmed later in this section. the results in Fig. 8 reveal that both the energy and the exergy ef-
As the axial distribution of the storage material mass in the exam- ficiency for the 𝑛th cycle are 5 to 6% points lower compared to the
ined test rig is not even, the comparison of the temperature profiles in efficiencies in the first cycle. This is due to the increased impact of
Figs. 5 and 6 is not reasonable. To make statements on the difference thermocline degradation on the storage’s performance in the 𝑛th cycle.
in thermocline degradation between a PBTES that is operated in FWD- Since thermocline degradation leads to lower discharging temperatures
mode versus a PBTES that is operated in REV-mode, the 𝑀𝐼𝑋 number and higher exhaust losses during charging both the energy and exergy
as it is defined in Eq. (17) comes in handy. As the 𝑀𝐼𝑋 number efficiency are negatively influenced by this effect. An impact of the
is calculated based on the momentum of energy the information of
uneven mass distribution is also taken into account. Fig. 7 presents the
results of evaluating the 𝑀𝐼𝑋 number for the collected data. Again,
the blue lines and area represent the FWD-mode and the orange lines
and area represent the REV-mode. In this figure, the number of charg-
ing/discharging cycles is plotted on the abscissa. It can be observed,
that for both the FWD- and REV-operation the 𝑀𝐼𝑋 number in the first
cycle is similar and starts to diverge with increasing number of cycles.
In FWD-mode the 𝑀𝐼𝑋 number seems to converge to an average
value of approximately 0.18 whereas in REV-mode the 𝑀𝐼𝑋 number is
noticeably higher and still slightly rises after the 11th cycle. A corollary
of these observations is, that when a PBTES is operated in partial cycle
and REV-mode, a much faster and more pronounced degradation of the
thermocline inside the packed bed has to be expected. In the following
paragraphs, the impact of this accelerated thermocline degradation on
the most important thermal key performance indicators is discussed.
In Fig. 8 the energy and exergy efficiency as they are defined
in Eqs. (8) and (9) for both operating modes (FWD and REV) are
illustrated. The evaluation of measurement data revealed, that the
pumping work needed to overcome the pressure drop of the packed
bed reduces the energy and exergy efficiency by less than 0.5% and 3%
points respectively and is therefore negligible compared to the thermal
effects. Notice, that the ordinate in this figure starts at 0.75. The blue Fig. 7. Thermocline degradation of a PBTES in partial cycle operation with a cycle
and violet bars represent the energy and exergy efficiencies for the first time of 1 h: MIX number.

8
P. Schwarzmayr et al. Applied Energy 360 (2024) 122895

values on the ordinate are energy/exergy input rates when the storage
is charged whereas negative values are output rates when the storage is
discharged. All values are normalized to the maximum gross input rate
that is observed during charging (𝑄̂̇ htf = 15.4 kW and 𝐵̂̇ htf = 3.9 kW).
The gross input rate is the sum of the energy/exergy rate transferred to
the packed bed and the saturation losses (𝑄̇ htf + 𝑄̇ sat and 𝐵̇ htf + 𝐵̇ sat
respectively) and is plotted as a black solid line in Figs. 9 and 10.
The dashed lines represent the in-/output rates that are reached in
the first cycle of each experiment whereas the dotted lines are the in-
/output rates achieved in the 𝑛th cycle after steady cycling is reached.
The results from the experiments where the storage is operated in
FWD- and REV-mode are plotted in blue and orange respectively. As
documented in Section 2 the partial cycle experiments start with a
storage that is charged to a state of charge of approximately 50%.
During the partial cycle experiments the storage is cycled between a
lower and an upper SOC (𝑆𝑂𝐶 − and 𝑆𝑂𝐶 + respectively). As 𝑆𝑂𝐶 − and
𝑆𝑂𝐶 + are not exactly the same for every cycle, the x-axes in Figs. 9
and 10 are labeled with the variable names 𝑆𝑂𝐶 − and 𝑆𝑂𝐶 + . For
the experiments presented in Figs. 9 and 10 𝑆𝑂𝐶 − = 0.32 to 0.38 and
Fig. 8. Energy-/exergy roundtrip efficiency of a PBTES in partial cycle operation with 𝑆𝑂𝐶 + = 0.48 to 0.54.
a cycle time of 1 h. Considering the energy in-/output rate for the first cycles (dashed
lines) in Fig. 9, exactly the same behavior is observed for the FWD- and
the REV- mode. In the second half of the charging period both curves
start to diverge from the solid black curve. This difference is caused by
operating mode (FWD or REV) on this degradation of the efficiencies
saturation losses due to an increase in the exhaust gas temperature.
cannot be observed. When comparing the FWD- and REV-operation,
During the discharging period a fairly constant output rate with a
Fig. 8 indicates that the energy efficiency of a PBTES in REV-mode is
maximum of 93% and 89% is reached in the FWD- and the REV-mode
lower by 5% points compared to the FWD-mode. This efficiency loss can
respectively. The difference between the two modes can be explained
be explained by increased heat losses to the surrounding via the metal
by higher heat losses via the metal grating at the bottom of the packed
grating on which the packed bed is sitting and a higher surface-to- bed in the REV-mode. The energy in-/output rate for the 𝑛th cycles,
volume ratio at the bottom half of the storage vessel. The losses of the which are plotted as dotted lines, indicate increased saturation losses
exergy efficiency, when comparing the REV- and the FWD-operation, throughout the whole charging period. Furthermore a degradation of
are between 6 to 7% points, which is mainly caused by the accelerated the maximum output rate for both the FWD- (89%) and REV-mode
thermocline degradation that is also evident in Fig. 7. (84%) can be observed. However, in both modes the energy output rate
Figs. 9 and 10 illustrate the energy and exergy in-/output rates for the 𝑛th cycle (steady cycling) is well above 74% of the maximum
that are accomplished for different experimental conditions. Positive gross input rate 𝑄̂̇ htf .

Fig. 9. Rate of energy in- and output of a PBTES in partial cycle operation with a Fig. 10. Rate of exergy in- and output of a PBTES in partial cycle operation with a
cycle time of 1 h. cycle time of 1 h.

9
P. Schwarzmayr et al. Applied Energy 360 (2024) 122895

Table 4 energy and exergy efficiency are 85% and 80% respectively which are
Summary of the main results: Energy/exergy efficiency are calculated according to
3 to 5% points less than they were in the first cycle.
Eqs. (8) and (9). Energy/exergy output rates are given as relative values with respect
to the maximum gross input rates 𝑄̂̇ htf and 𝐵̂̇ htf .
The energy and exergy in-/output rates were observed to be stable
HTF flow direction FWD-mode REV-mode
for all experiments. Despite minimum saturation losses during charging
and slight reductions of the maximum output rates during discharging
Cycle number 1st nth 1st nth
no substantial deterioration of the thermal performance of the storage
Energy efficiency 𝜂E 95% 90% 90% 85%
could be detected. Energy output rates of 74% to 93% and exergy output
Exergy efficiency 𝜂B 90% 86% 83% 80%
rates of 65% to 89% with respect to the maximum input rates were
min. energy output rate 90% 82% 85% 75%
𝑄̇ htf ∕𝑄̂̇ htf measured.
max. energy output rate 93% 89% 89% 84%
min. exergy output rate 84% 74% 77% 65%
𝐵̇ htf ∕𝐵̂̇ htf CRediT authorship contribution statement
max. exergy output rate 89% 87% 82% 81%

Paul Schwarzmayr: Writing – original draft, Visualization, Valida-

tion, Methodology, Investigation, Formal analysis, Data curation, Con-
Regarding the exergy in-/output rate, similar deductions can be ceptualization. Felix Birkelbach: Writing – review & editing, Method-
drawn from Fig. 10. Compared to the energy input rate, the exergy ology, Formal analysis, Conceptualization. Heimo Walter: Writing – re-
input rate during the charging periods does not diverge from the gross view & editing, Visualization, Supervision, Methodology, Formal anal-
exergy input rate 𝐵̂̇ htf in any experiment. That is because saturation ysis, Conceptualization. René Hofmann: Writing – review & editing,
losses occur at low temperatures near the ambient temperature and are Validation, Supervision, Funding acquisition, Conceptualization,.
therefor mostly composed of anergy. For the discharging process, the
qualitative behavior of all the exergy curves is similar to the energy Declaration of competing interest
curves in Fig. 9. Again, in the first cycles a fairly constant output rate
with a maximum of 89% and 82% can be observed for the FWD- and the The authors declare that they have no known competing finan-
REV-mode respectively. However, the degradation of the exergy output cial interests or personal relationships that could have appeared to
rate in the 𝑛th cycle is slightly more pronounced. For steady cycling the influence the work reported in this paper.
maximum exergy output rates for the FWD-mode and the REV-mode are
87% and 81% respectively. Still noteworthy, the exergy output rate is Data availability
well above 65% of the maximum gross exergy input rate 𝐵̂̇ htf for all
experiments. Data will be made available on request.
The results presented and discussed in this section provide impor-
tant information about the thermal performance of PBTES systems that Acknowledgments
operate under challenging, not yet investigated, conditions. The most
important key performance indicators that can be used to assess the The authors acknowledge funding support of this work through
suitability of PBTES systems for the waste heat recovery in industrial the research project 5DIndustrialTwin as part of the Austrian Climate
processes are summarized in Table 4. These reference values are valid and Energy Fund’s initiative Energieforschung (e!MISSION) 6th call
for a PBTES with a design similar to the test rig used in this study as (KLIEN/FFG project number 881140). Furthermore, the authors ac-
well as similar operating conditions. knowledge TU Wien Bibliothek, Austria for financial support through
its Open Access Funding programme.
5. Conclusion
References
The experiments presented in this work were conducted to assess
[1] Bai Y, Wang L, Lin L, Lin X, Peng L, Chen H. A performance analysis of the
the efficiency of packed bed thermal energy systems when they are spray-type packed bed thermal energy storage for concentrating solar power
operated with industrial exhaust gases. To facilitate the removal of generation. J Energy Storage 2022;51:104187. http://dx.doi.org/10.1016/j.est.
powder particles, a packed bed thermal energy storage is required to be 2022.104187.
charged from the bottom and discharged from the top in this particular [2] Gautam A, Saini RP. A review on sensible heat based packed bed solar
thermal energy storage system for low temperature applications. Sol Energy
application. Since this is contrary to an optimal operation in terms of 2020;207:937–56. http://dx.doi.org/10.1016/j.solener.2020.07.027.
thermal efficiency, this study investigated the thermal performance of [3] Gautam A, Saini RP. A review on technical, applications and economic as-
a packed bed thermal energy storage for different flow directions of pect of packed bed solar thermal energy storage system. J Energy Storage
the heat transfer fluid. To quantify the effect of the flow direction on 2020;27:101046. http://dx.doi.org/10.1016/j.est.2019.101046.
[4] Cui Z, Du Q, Gao J, Bie R, Li D. Development of a direct contact heat
the exergy efficiency and thermocline degradation, a lab-scale test rig
exchanger for energy and water recovery from humid flue gas. Appl Therm Eng
of a packed bed thermal energy storage is used. Multiple partial cycle 2020;173:115214. http://dx.doi.org/10.1016/j.applthermaleng.2020.115214.
experiments in which the storage was periodically charged from the [5] Wang W, Li H, Guo S, He S, Ding J, Yan J, Yang J. Numerical simulation study
top and discharged from the bottom and vice versa were conducted on discharging process of the direct-contact phase change energy storage system.
Appl Energy 2015;150:61–8. http://dx.doi.org/10.1016/j.apenergy.2015.03.108.
and analyzed.
[6] Scharinger-Urschitz G, Schwarzmayr P, Walter H, Haider M. Partial cycle oper-
The results reveal that in a packed bed thermal energy storage ation of latent heat storage with finned tubes. Appl Energy 2020;280:115893.
in partial cycle operation, the thermocline gradually degrades until it http://dx.doi.org/10.1016/j.apenergy.2020.115893.
reaches a steady state after approximately 10 cycles (equivalent to 10 h [7] Zanganeh G, Ambrosetti G, Pedretti A, Zavattoni S, Barbato M, Good P,
in this study). A slightly faster and more pronounced degradation of Haselbacher A, Steinfeld A. A 3 mwth parabolic trough CSP plant operating
with air at up to 650 ◦ c. In: 2014 international renewable and sustainable
the thermocline could be observed when the storage was charged from energy conference (IRSEC). 2014, p. 108–13. http://dx.doi.org/10.1109/IRSEC.
the bottom and discharged from the top. Furthermore, the energy and 2014.7059884, ISSN: 2380-7393.
exergy efficiency of a storage that has reached steady cycling after it [8] Rong Y, Huang SY, Zhou H. Experimental study on storage performance of
was charged from the top and discharged from bottom are 90% and packed bed solar thermal energy storage system using steel slag. J Energy Storage
2024;78:110042. http://dx.doi.org/10.1016/j.est.2023.110042.
86% respectively. These values are 5% points lower than the ones
[9] Mawire A, McPherson M, Heetkamp RRJvd, Mlatho SJP. Simulated performance
achieved in the first cycle. For a storage that has reached steady cycling of storage materials for pebble bed thermal energy storage (TES) systems. Appl
after it was charged from the bottom and discharged from top the Energy 2009;86(7):1246–52. http://dx.doi.org/10.1016/j.apenergy.2008.09.009.

10
P. Schwarzmayr et al. Applied Energy 360 (2024) 122895

[10] Meier A, Winkler C, Wuillemin D. Experiment for modelling high temperature [30] Keplinger T, Haider M, Steinparzer T, Patrejko A, Trunner P, Haselgrübler M.
rock bed storage. Solar Energy Mater 1991;24(1):255–64. http://dx.doi.org/10. Dynamic simulation of an electric arc furnace waste heat recovery system
1016/0165-1633(91)90066-T. for steam production. Appl Therm Eng 2018;135:188–96. http://dx.doi.org/10.
[11] Yang X, Yang X, Qin FG, Jiang R. Experimental investigation of a molten 1016/j.applthermaleng.2018.02.060.
salt thermocline storage tank. Int J Sustain Energy 2014;35(6):606–14. http: [31] Ortega-Fernández I, Rodríguez-Aseguinolaza J. Thermal energy storage for waste
//dx.doi.org/10.1080/14786451.2014.930465, Publisher: Taylor & Francis. heat recovery in the steelworks: The case study of the reslag project. Appl Energy
[12] Anderson R, Shiri S, Bindra H, Morris JF. Experimental results and modeling of 2019;237:708–19. http://dx.doi.org/10.1016/j.apenergy.2019.01.007.
energy storage and recovery in a packed bed of alumina particles. Appl Energy [32] Slimani H, Filali Baba Y, Ait Ousaleh H, Elharrak A, El Hamdani F, Bouzekri H,
2014;119:521–9. http://dx.doi.org/10.1016/j.apenergy.2014.01.030. Al Mers A, Faik A. Horizontal thermal energy storage system for Moroccan steel
[13] Kanojia N, Kaushik S, Singh M, Sah MK. Comprehensive review on packed bed and iron industry waste heat recovery: Numerical and economic study. J Clean
thermal energy storage systems. In: Manik G, Kalia S, Sahoo SK, Sharma TK, Prod 2023;393:136176. http://dx.doi.org/10.1016/j.jclepro.2023.136176.
Verma OP, editors. Advances in mechanical engineering. Lecture notes in [33] Odenthal C, Steinmann W-D, Zunft S. Analysis of a horizontal flow closed loop
mechanical engineering, Singapore: Springer; 2021, p. 165–73. http://dx.doi.org/ thermal energy storage system in pilot scale for high temperature applications –
10.1007/978-981-16-0942-8_15. Part I: Experimental investigation of the plant. Appl Energy 2020;263:114573.
[14] Al-Azawii MMS, Alhamdi SFH, Braun S, Hoffmann J-F, Calvet N, Anderson R. http://dx.doi.org/10.1016/j.apenergy.2020.114573.
Thermocline in packed bed thermal energy storage during charge-discharge cycle [34] Touzo A, Olives R, Dejean G, Pham Minh D, El Hafi M, Hoffmann J-Fc, Py X.
using recycled ceramic materials - Commercial scale designs at high temperature. Experimental and numerical analysis of a packed-bed thermal energy storage
J Energy Storage 2023;64:107209. http://dx.doi.org/10.1016/j.est.2023.107209. system designed to recover high temperature waste heat: an industrial scale up.
[15] Oró E, Castell A, Chiu J, Martin V, Cabeza LF. Stratification analysis in packed J Energy Storage 2020;32:101894. http://dx.doi.org/10.1016/j.est.2020.101894.
bed thermal energy storage systems. Appl Energy 2013;109:476–87. http://dx. [35] Esence T, Bruch A, Molina S, Stutz B, Fourmigué J-F. A review on experience
doi.org/10.1016/j.apenergy.2012.12.082. feedback and numerical modeling of packed-bed thermal energy storage systems.
[16] Knobloch K, Muhammad Y, Costa MS, Moscoso FM, Bahl C, Alm O, Engel- Sol Energy 2017;153:628–54. http://dx.doi.org/10.1016/j.solener.2017.03.032.
brecht K. A partially underground rock bed thermal energy storage with a [36] Schwarzmayr P, Birkelbach F, Walter H, Hofmann R. Standby efficiency and
novel air flow configuration. Appl Energy 2022;315:118931. http://dx.doi.org/ thermocline degradation of a packed bed thermal energy storage: An experimen-
10.1016/j.apenergy.2022.118931. tal study. Appl Energy 2023;337:120917. http://dx.doi.org/10.1016/j.apenergy.
[17] Soprani S, Marongiu F, Christensen L, Alm O, Petersen KD, Ulrich T, Engel- 2023.120917.
brecht K. Design and testing of a horizontal rock bed for high temperature [37] Bayón R, Rivas E, Rojas E. Study of thermocline tank performance in dynamic
thermal energy storage. Appl Energy 2019;251:113345. http://dx.doi.org/10. processes and stand-by periods with an analytical function. Energy Procedia
1016/j.apenergy.2019.113345. 2014;49:725–34. http://dx.doi.org/10.1016/j.egypro.2014.03.078.
[18] Dusek S, Hofmann R, Gruber S. Design analysis of a hybrid storage concept [38] Biencinto M, Bayón R, Rojas E, González L. Simulation and assessment of
combining Ruths steam storage and latent thermal energy storage. Appl Energy operation strategies for solar thermal power plants with a thermocline storage
2019;251:113364. http://dx.doi.org/10.1016/j.apenergy.2019.113364. tank. Sol Energy 2014;103:456–72. http://dx.doi.org/10.1016/j.solener.2014.02.
[19] Xie B, Baudin N, Soto J, Fan Y, Luo L. Chapter 10 - Thermocline packed bed 037.
thermal energy storage system: a review. In: Jeguirim M, editor. Renewable [39] Wang Y, Wang Z, Yuan G. Control strategy effect on storage performance for
energy production and distribution. Advances in renewable energy technologies, packed-bed thermal energy storage. Sol Energy 2023;253:73–84. http://dx.doi.
vol. 1, Academic Press; 2022, p. 325–85. http://dx.doi.org/10.1016/B978-0-323- org/10.1016/j.solener.2023.02.012.
91892-3.24001-6. [40] Zavattoni SA, Barbato MC, Pedretti A, Zanganeh G, Steinfeld A. High temperature
[20] Yang B, Bai F, Wang Y, Wang Z. Study on standby process of an air-based solid rock-bed TES system suitable for industrial-scale CSP plant – CFD analysis
packed bed for flexible high-temperature heat storage: Experimental results and under charge/Discharge cyclic conditions. Energy Procedia 2014;46:124–33. http:
modelling. Appl Energy 2019;238:135–46. http://dx.doi.org/10.1016/j.apenergy. //dx.doi.org/10.1016/j.egypro.2014.01.165.
2019.01.073. [41] Zavattoni SA, Barbato MC, Pedretti A, Zanganeh G. Single-tank TES system
[21] Marti J, Geissbühler L, Becattini V, Haselbacher A, Steinfeld A. Constrained – transient evaluation of thermal stratification according to the second-law of
multi-objective optimization of thermocline packed-bed thermal-energy storage. thermodynamics. Energy Procedia 2015;69:1068–77. http://dx.doi.org/10.1016/
Appl Energy 2018;216:694–708. http://dx.doi.org/10.1016/j.apenergy.2017.12. j.egypro.2015.03.213.
072. [42] Mertens N, Alobaid F, Frigge L, Epple B. Dynamic simulation of inte-
[22] Trevisan S, Jemmal Y, Guedez R, Laumert B. Packed bed thermal energy storage: grated rock-bed thermocline storage for concentrated solar power. Sol Energy
A novel design methodology including quasi-dynamic boundary conditions and 2014;110:830–42. http://dx.doi.org/10.1016/j.solener.2014.10.021.
techno-economic optimization. J Energy Storage 2021;36:102441. http://dx.doi. [43] Zanganeh G, Pedretti A, Zavattoni S, Barbato M, Steinfeld A. Packed-bed
org/10.1016/j.est.2021.102441. thermal storage for concentrated solar power – Pilot-scale demonstration and
[23] Bruch A, Fourmigué JF, Couturier R. Experimental and numerical investigation of industrial-scale design. Sol Energy 2012;86(10):3084–98. http://dx.doi.org/10.
a pilot-scale thermal oil packed bed thermal storage system for CSP power plant. 1016/j.solener.2012.07.019.
Sol Energy 2014;105:116–25. http://dx.doi.org/10.1016/j.solener.2014.03.019. [44] Johnson E, Bates L, Dower A, Bueno PC, Anderson R. Thermal energy storage
[24] Bruch A, Molina S, Esence T, Fourmigué JF, Couturier R. Experimental in- with supercritical carbon dioxide in a packed bed: Modeling charge-discharge
vestigation of cycling behaviour of pilot-scale thermal oil packed-bed thermal cycles. J Supercrit Fluids 2018;137:57–65. http://dx.doi.org/10.1016/j.supflu.
storage system. Renew Energy 2017;103:277–85. http://dx.doi.org/10.1016/j. 2018.03.009.
renene.2016.11.029. [45] Cascetta M, Cau G, Puddu P, Serra F. Experimental investigation of a packed
[25] Miró L, Gasia J, Cabeza LF. Thermal energy storage (TES) for industrial waste bed thermal energy storage system. J Phys Conf Ser 2015;655(1):012018. http:
heat (IWH) recovery: A review. Appl Energy 2016;179:284–301. http://dx.doi. //dx.doi.org/10.1088/1742-6596/655/1/012018, Publisher: IOP Publishing.
org/10.1016/j.apenergy.2016.06.147. [46] McTigue JD, Markides CN, White AJ. Performance response of packed-bed ther-
[26] Manente G, Ding Y, Sciacovelli A. A structured procedure for the selection of mal storage to cycle duration perturbations. J Energy Storage 2018;19:379–92.
thermal energy storage options for utilization and conversion of industrial waste http://dx.doi.org/10.1016/j.est.2018.08.016.
heat. J Energy Storage 2022;51:104411. http://dx.doi.org/10.1016/j.est.2022. [47] Mitterlehner T, Kartnig G, Haider M. Analysis of the thermal ratcheting phe-
104411. nomenon in packed-bed thermal energy storage using Discrete Element Method.
[27] Bianchi G, Panayiotou GP, Aresti L, Kalogirou SA, Florides GA, Tsamos K, FME Trans 2020;48(2):427–31. http://dx.doi.org/10.5937/fme2002427M.
Tassou SA, Christodoulides P. Estimating the waste heat recovery in the European [48] Schwarzmayr P, Birkelbach F, Kasper L, Hofmann R. Development of a digital
Union Industry. Energy Ecol Environ 2019;4(5):211–21. http://dx.doi.org/10. twin platform for industrial energy systems. In: Accelerated energy innovations
1007/s40974-019-00132-7. and emerging technologies, Vol. 25, Cambridge, USA: Energy Proceedings; 2022,
[28] Papapetrou M, Kosmadakis G, Cipollina A, La Commare U, Micale G. Industrial http://dx.doi.org/10.46855/energy-proceedings-9974.
waste heat: Estimation of the technically available resource in the EU per indus- [49] Haller MY, Cruickshank CA, Streicher W, Harrison SJ, Andersen E, Furbo S.
trial sector, temperature level and country. Appl Therm Eng 2018;138:207–16. Methods to determine stratification efficiency of thermal energy storage processes
http://dx.doi.org/10.1016/j.applthermaleng.2018.04.043. – Review and theoretical comparison. Sol Energy 2009;83(10):1847–60. http:
[29] Schwarzmayr P, Birkelbach F, Walter H, Javernik F, Schwaiger M, Hofmann R. //dx.doi.org/10.1016/j.solener.2009.06.019.
Packed bed thermal energy storage for waste heat recovery in the iron and steel [50] Andersen E, Furbo S, Fan J. Multilayer fabric stratification pipes for solar tanks.
industry: A cold model study on powder hold-up and pressure drop. J Energy Sol Energy 2007;81(10):1219–26. http://dx.doi.org/10.1016/j.solener.2007.01.
Storage 2024;75:109735. http://dx.doi.org/10.1016/j.est.2023.109735. 008.

11