Computers and Chemical Engineering 169 (2023) 108087

Contents lists available at ScienceDirect

Computers and Chemical Engineering

journal homepage: www.elsevier.com/locate/compchemeng

Optimization and analysis of different liquid air energy

storage configurations
Zhongxuan Liu a, *, Donghoi Kim b, Truls Gundersen a
a
Department of Energy and Process Engineering, Norwegian University of Science and Technology (NTNU), Kolbjørn Hejes vei 1B, NO-7491, Trondheim, Norway
b
SINTEF Energy Research, Sem Sælands vei 11, NO-7465, Trondheim, Norway

A R T I C L E I N F O A B S T R A C T

Keywords: Energy storage technologies are required to ensure stability of energy systems when the share of renewable
Liquid air energy storage energy forms (wind and solar) is increasing. Liquid air energy storage (LAES) is a promising technology for
Cycle configuration storing electricity with certain advantages, such as high energy density and being geographically unconstrained.
Round-trip efficiency
However, one drawback of a standalone LAES is the relatively low round-trip efficiency (RTE). In this work, the
Compression stages
Expansion stages
performance of a standalone LAES system with different number of compression and expansion stages is studied.
Organic rankine cycle All cases are optimized by using a particle swarm optimization (PSO) algorithm. The optimal results show that
the highest RTE of 66.7% is obtained when there is a 2-stage compressor and a 3-stage expander in the LAES
system. When the number of compression stages is fixed, the highest RTE is obtained when hot and cold streams
have close to parallel temperature profiles in the preheaters of the expansion section.

variations). Thus, energy storage technologies are promising in future

energy markets.
1. Introduction Some energy storage processes have been investigated and proven to
be highly efficient, such as batteries (Zakeri and Syri, 2015), pumped
Due to the concern of climate change, the greenhouse gas emissions hydroelectric energy storage (PHES) (Rehman, Al-Hadhrami, and Alam,
should be reduced by at least 40% by 2030 compared to 1990 (Euro­ 2015), and compressed air energy storage (CAES) (Arabkoohsar et al.,
pean Commission 2014). Therefore, our energy systems tend to adopt 2015). Liquid air energy storage (LAES) (Damak et al., 2020) is a
lower-carbon energy sources, such as renewables. It is predicted that the promising energy storage technology that is limited by its low round-trip
share of renewable energies like solar, wind and geothermal in the total efficiency (RTE). These four energy storage technologies are suitable for
energy consumption will rise to 20-60% by 2050 compared to 5.7% in different scenarios depending on their costs and efficiencies. It is worth
2020 (British Petroleum 2021). In contrast, the share of fossil fuels will noting that batteries are the most efficient alternative for storing energy,
be reduced from 83% in 2020 to somewhere in the range 20-70% by however, the cost of batteries is quite high, so they have mostly been
2050. However, this transformation of energy systems may lead to a used for small-scale energy storage (Zakeri and Syri, 2015). Currently,
critical issue related to instability on the power supply side. Since the larger batteries have been investigated, which expands the range of
availability of renewable energies like solar and wind depends on time applications. PHES, CAES and LAES can be considered for large-scale
and weather, the power generated from these energy sources is inter­ energy storage. Since energy densities for PHES and CAES are rela­
mittent and unpredictable. tively low, large storage space is required for these two technologies.
Thus, additional measures should be taken to maintain the stability Two reservoirs at different levels and considerable sizes are needed for
of energy markets. Energy storage technologies represent an effective PHES, and suitable sites for PHES are at locations with adequate fresh
measure to handle the problem caused by unpredictable renewable water supply. Above ground tanks and underground caverns are used to
energy sources. The principle of energy storage technologies is that store compressed air in CAES systems, and caverns are commonly
available energy is collected and stored, and then provided in different adopted for economic considerations. This means that CAES tends to be
forms when energy is demanded. Energy storage technologies are also located in remote areas far away from other energy conversion pro­
able to store unstable supplied energy and smoothen variations in supply cesses. The geographical constraints for PHES and CAES increase the
(typical for renewable sources) and demand (short-term or mid-term

* Corresponding author.
E-mail address: [email protected] (Z. Liu).

https://doi.org/10.1016/j.compchemeng.2022.108087
Received 7 July 2022; Received in revised form 4 November 2022; Accepted 25 November 2022
Available online 26 November 2022
0098-1354/© 2022 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/).
Z. Liu et al. Computers and Chemical Engineering 169 (2023) 108087

Nomenclature in Inlet
liq Liquid air
Symbols LY Liquid yield
Ė Exergy (kW) M Number of expansion stages
ṁ Mass flowrate (kg/s) N Number of compression stages
p Pressure (bar) net Net power output
T Temperature (◦ C) ORC Organic Rankine Cycle
VF Vapor fraction out Outlet
Ẇ Power (kW) pre preheater
w Specific power (kJ/kg) pump Pump
η Efficiency (%) rec Cold thermal energy recovery part
ΔT Minimum heat transfer approach temperature (◦ C) RT Round-trip efficiency
tur Turbine
Subscripts
air Air Acronyms
comp Compression part or compressor CAES Compressed Air Energy Storage
Con Condensation LAES Liquid Air Energy Storage
cryotur Cryo-turbine LMTD Logarithmic Mean Temperature Difference
dir Direct expansion in the discharging part LNG Liquefied Natural Gas
eva Evaporation/Evaporator MTD Minimum temperature difference
exp Expansion part or expander ORC Organic Rankine Cycle
feed Feed stream PHES Pump Hydroelectrical Energy Storage
h Thermal oil (working fluid in the hot thermal energy PSO Particle Swarm Optimization
recovery cycle) RTE Round-Trip Efficiency

operating complexity and require extra cost for energy transmission. In Morgan et al. (Morgan et al., 2015) used a three-stage turbine to
contrast, the working fluid for the LAES system is liquid air with a high replace the single cold turbine before the air is sent to a phase separator
energy density, thus, the storage space is considerably reduced in their process model of the LAES. In addition, the compression heat
compared to PHES and CAES. Typical energy density values for LAES, was also recovered and used to preheat air in the discharging process.
CAES and PHES are 97 Wh/kg, 45 Wh/kg and 1 Wh/kg, respectively Thus, the liquid yield of air is increased and an RTE of 57% was ob­
(Rahman et al., 2020, Aneke and Wang, 2016). Thus, the location of tained. Liu et al. (Liu et al., 2020) simulated the LAES system with
LAES systems is flexible enough to be placed near other energy con­ different configurations in the charging and discharging processes, see
version processes, which is a significant advantage of LAES over other Fig. 1. The number of compression stages was varied in the range of 2-6,
energy storage technologies. and the number of expansion stages was varied between 3 and 6.
LAES has attracted more focus in recent years due to its character­ Thermal oil is used to transport heat from the compression section to the
istics of compact configurations and flexible locations. It is worth noting expansion section, and the flowrate of thermal oil depends on the
that the round-trip efficiency (RTE), which is defined as the ratio be­ number of compression stages. Depending on the flowrate of thermal oil,
tween the recovered energy in the discharging process and the different situations will occur in the expansion section. For high flow­
consumed energy in the charging process, is commonly used to evaluate rates of thermal oil relative to the flowrate of air in the discharging part,
and compare various energy storage technologies. Highview Power is the compression heat is not fully utilized. For low flowrates of thermal
committed to developing liquid air energy storage, and the first pilot oil, the minimum approach temperature is shifted to the cold end of the
plant of an LAES with a storage capacity of 2.5 MWh was built by the preheaters, and the outlet temperature of air is reduced. The optimal
University of Birmingham for academic research (Highview Power situation is when the flowrate of thermal oil makes the temperature
2022). A commercial-scale plant with a storage capacity of 300 MWh is profiles in the preheaters parallel. The LAES system with 2-stage
underway and will be located at Trafford Energy Park near Manchester. compression and 3-stage expansion had the largest RTE. In addition,
Guizzi et al. (Guizzi et al., 2015) proposed an LAES system based on the Liu et al. (Liu et al., 2022) also tried new working fluids and novel
Claude cycle for air liquefaction. The key feature of their concept is that configurations for the cold thermal energy recovery cycles. The LAES
hot and cold thermal energy are exchanged between the charging and system with a 4-stage compressor and a 4-stage expander was studied.
discharging parts. Compression heat from the charging part can be used Multi-component fluid cycles (MCFCs) and Organic Rankine Cycles
to preheat air before expansion in the discharging part. Likewise, cold (ORCs) were used to transfer the cold regasification energy from the
energy from regasification can be used to partly liquefy air. Such heat discharging process to the charging process. The optimization results of
exchange requires energy storage since charging and discharging oper­ the proposed process models indicate that the LAES system with dual
ate at different times, however, it achieves an RTE for the system of MCFCs has the best performance among 6 cases that were studied, and
54.4%. the RTE can be further improved from 62.4% to 64.7% by reducing
Based on the above arguments and using the concept of Guizzi et al. minimum temperature differences in high-temperature heat exchangers
(Guizzi et al., 2015), the objective of this work is to consider optimi­ from 10◦ C to 5◦ C (Liu et al., 2022).
zation opportunities related to the configuration of LAES systems In addition, the utilization of compression heat that is wasted in the
focusing on the charging part (compression), the discharging part discharge mode has been investigated to further increase the efficiency
(expansion) and the hot thermal energy storage and recycle system. The of a standalone LAES system. Peng et al. (Peng et al., 2018) suggested
cold thermal energy storage and recycle system has been previously that around 2/5 of the compression heat could not be used in the dis­
studied in our group and will not be repeated here. Energy efficiency will charging process and is therefore wasted. Thus, an ORC was adopted to
be optimized and measured by various performance indicators. Cost utilize this part of compression heat to produce extra power and thereby
issues are, however, regarded to be outside the scope of this work. increase the RTE of the system. Two cases related to different heat sink

2
Z. Liu et al. Computers and Chemical Engineering 169 (2023) 108087

Fig. 1. Simplified block diagram of a Liquid Air Energy Storage System with charging, discharging and storage of both liquid air and thermal energy recovery fluids.

temperatures were considered for the ORCs. One at ambient tempera­ compressor stages, however, the economic analysis of the cases indi­
ture and the other at sub-ambient temperature. The sub-ambient tem­ cated that the minimum total cost per liquid air is obtained when there is
perature was achieved by consuming part of the compression heat a 4-stage compressor in the liquefaction process. In addition, liquid air
through an Absorption Refrigeration Cycle. It was observed that the has been framed as an emerging energy vector towards carbon neutrality
LAES system with the ORC operating at ambient temperature had a (Qi et al., 2022).
higher RTE of 62.7% compared to the LAES with the ORC operating at In a standalone LAES system, when the number of compression
sub-ambient temperature. Zhang et al. (Zhang et al., 2020) studied stages is greater than or equal to the number of expansion stages, part of
cascaded storage of compression heat to increase the temperature level the compression heat is wasted rather than utilized (Liu et al., 2020). To
of the hot thermal energy storage. In addition, two LAES systems, one improve the performance of LAES systems, the wasted part of
with an ORC and the other with a Kalina cycle to utilize part of the compression heat can in certain cases be utilized to produce power.
unused enhanced compression heat to generate additional power, were Organic Rankine Cycles (ORCs) (Yu et al., 2016), which represent an
tested and compared. Results indicated that both the ORC and the Kalina effective heat to work technology, are applied in this study to utilize this
cycle can improve the performance of the LAES. By comparing the LAES wasted part of the compression heat. The temperature difference be­
with ORC and the LAES with a Kalina cycle, the system with an addi­ tween thermal oil and ambient temperature is the driving force for the
tional ORC, where n-Pentane is the working fluid for the cycle, had ORC.
better performance (RTE of 56.9%) than the LAES with a Kalina cycle It is well known that power consumption and production will vary
(56.1%). with different number of compressors and expanders when the outlet
In addition to the mentioned studies on the configuration improve­ pressures of compression and expansion are fixed. Most of the existing
ments of a standalone LAES system, various integration strategies be­ literature discuss the expansion section and compression section sepa­
tween the LAES and external hot and cold thermal energy sources have rately for the LAES system. However, the expansion section is strongly
proven to be effective methods for enhancing the performance of the influenced by the compression section when there is a hot thermal en­
LAES. Li et al. (Li et al., 2014) studied the integration with a nuclear ergy storage between these two sections in the LAES. The amount and
power plant as a way to increase the efficiency of LAES system. The high temperature of thermal oil are changing with different number of
temperature steam (287◦ C) from the nuclear power plant was used to compressor stages, which will directly affect the inlet air temperature to
preheat air in the discharging process, and thereby increasing the power expanders and the power output. Thus, there is a lack of studies related
generated from the turbines. The RTE was increased from 34% to 71%. to the complex relations of different number of compression and
Cetin et al. (Cetin et al., 2019) considered to use geothermal heat to expansion stages and how this affects the performance of the LAES
increase the turbine inlet air temperatures in the LAES system. The system. This paper focuses on the potential of system efficiency im­
geothermal heat is supplied at temperatures in the range of 150-250◦ C, provements for different configurations of the LAES system, where the
and the corresponding RTE of the LAES system is reported to be between term configuration relates to having different number of compression
40% and 55%. Qi et al. (Qi et al., 2020) investigated the integration and expansion stages, as well as the potential use of ORC.
between the LAES system and an LNG regasification process. When To have a fair comparison between these different configurations of
energy is demanded, both the discharging process in the LAES system the LAES system, all cases have been optimized to maximize their RTE
and the LNG regasification process are activated and used to generate by using a particle swarm optimization (PSO) method (Eberhart and
power as two separate systems. The LAES discharging process produces Kennedy, 1995). Different configurations of the LAES system are opti­
power by expanding pressurized air, and the LNG regasification process mized and analyzed in Section 5. It is worth noting that the corre­
produces power by operating as heat sink for an ORC system. When sponding pressure ratios of compressors and expanders are changing
excess energy is available and needed to be stored, the LNG is first sent to with different number of compression and expansion stages. The pres­
the cold box and then to aftercoolers in the compression part of the LAES sure ratios will be relatively high when the number of compression (or
system. After delivering part of the cold regasification energy to the expansion) stages is small, which may affect the efficiencies of com­
LAES, LNG is used as a heat sink in an ORC. The novel integrated system pressors (or expanders). However, this work emphasizes purely on a
has an extremely high RTE of 129.2%; larger than 100% because of the thermodynamic analysis, which is why mechanical limitations and cost
“free” cold energy from LNG regasification. Park et al. (Park et al., 2020) issues related to pressure ratios and outlet temperatures of compressors
also considered to use the cold energy from an LNG regasification pro­ and expanders will not be considered.
cess to liquefy air. It was found that the specific energy consumption is
reduced, and the liquid air yield is increased with increasing number of

3
Z. Liu et al. Computers and Chemical Engineering 169 (2023) 108087

2. System description and compared to identify the most suitable cold cycles for the LAES
when the configurations of the compression and expansion parts are
Flowsheets of the liquid air energy storage system are illustrated in fixed. It was found that two multi-component fluid cycles gave the
Figs. 2 and 3. The liquid air energy storage is commonly divided into highest performance (RTE) among six studied cases. Since the focus here
charging, storage and discharging processes based on its operating is on configurations of the compression and expansion parts rather than
mode. However, for a standalone LAES, the overall system can also be working fluids, a slightly simpler set of working fluids with a marginal
decomposed into three parts, which are the compression, the hot and decrease in RTE was used in order to reduce the computational efforts.
cold thermal energy recovery cycles, and the expansion sections, ac­ The cold thermal energy recovery cycles are therefore using a
cording to the function of each part. In the compression section, air is single-component cycle with methanol and a multi-component fluid
compressed in stages by using available electricity. Air is cooled to 30◦ C cycle, which consists of 70 mol% propane, 20 mol% ethane and 10 mol%
after each compressor stage and the compression heat is collected and n-butane. These working fluids are operated at different temperatures.
stored by the hot thermal energy recovery cycle. The high-pressure air is For different LAES processes in this work, the configuration and the
then sent to the cold thermal energy recovery part. It is precooled in the components of cold thermal energy recovery cycles are the same, while
cold box by working fluids in cold thermal energy recovery cycles before the compression and expansion parts are changed by varying the num­
it is expanded in the cryo-turbine to atmospheric pressure. The name of ber of compression and expansion stages.
this turbine refers to the very low (i.e. cryogenic) temperature, and the The process flowsheet of the LAES system with N compression stages
unit is operating with a liquid inlet stream and a two-phase outlet (2≤N≤6) and M expansion stages (3≤M≤5) is shown in Fig. 2. When the
stream. The cryo-turbine can produce additional refrigeration capacity number of compression stages N is greater than or equal to the number
and power, and thereby improve the efficiency of the system. The of expansion stages M, there will always be a part of the compression
partially evaporated stream is then separated into a vapor stream, which heat that cannot be utilized in the expansion part. This is due to the fact
is sent back to the compression part, and a liquid stream that is sent to that the amount of compression heat and the temperature of the air are
storage. When there is a need for power, liquid air is first pumped to a reduced when the number of compression stages is increased. As a
high pressure before it is sent to evaporators, where liquid air is heated result, the temperature of the thermal oil is reduced. Despite the fact that
to be regasified by working fluids in cold thermal energy recovery cy­ the compressor duty is decreased with increasing number of stages, the
cles. After delivering the cold regasification energy, air is sent to the repeated cooling of air will increase the flowrate of thermal oil since it
expansion part, where air is expanded through a series of turbine stages must be split into more branches. In order to improve the performance of
to generate power. There is a preheater before each expander (Heater 1- the system, ORCs are considered to utilize the unused part of compres­
M) to increase the power generation. sion heat. For the additional ORC, R152a (C2H4F2 with critical tem­
The hot thermal energy storage between the compression and perature 113.3◦ C and critical pressure 45.2 bar) is chosen as the working
expansion parts is used to collect the compression heat and release it to fluid. The reason for selecting R152a as the working fluid in the ORC is
increase the inlet air temperature to the turbine stages. The working that it is environment-friendly and non-toxic to humans, and it is also
fluid for the hot energy recovery cycle is Therminol 66 with a wide reported to be more efficient than several other working fluids, such as
operating temperature range from -3 to 350◦ C. This fluid is widely used R134a, R143a, and R32 (Bellos and Tzivanidis, 2019). In addition, the
in the hot storage cycle for LAES systems (Tafone et al., 2018). Two cold critical temperature of R152a is within the temperature range of the
thermal energy recovery cycles are used to transfer the cold regasifica­ compression heat in different LAES configurations. The critical pressure
tion energy of liquid air to the compressed air. In our previous study (Liu of R152a is relatively large and the saturation pressure at ambient
et al., 2022), different cold thermal energy recovery cycles are proposed temperature is only 3.7 bar, therefore, we can take advantage of the

Fig. 2. Process flowsheet of the liquid air energy storage (LAES) system.

4
Z. Liu et al. Computers and Chemical Engineering 169 (2023) 108087

Fig. 3. Process flowsheet of the LAES system with an additional ORC.

large pressure drop during expansion. Fig. 3 illustrates the LAES system temperatures in LAES systems with varying number of compression and
with an additional ORC, where the working fluid of the ORC is assumed expansion stages on the performance of compressors and expanders is
to be cooled to 20◦ C by cooling water. not considered. The isentropic efficiencies of compressors and expanders
are assumed to be 85% and 90%, respectively. Other simulation con­
3. Process simulation ditions for process units are given in Table 1.

In this section, the assumptions and key parameter settings for main 4. Process evaluation, validation and optimization
components during the simulation of the LAES system are provided.
Although the air pre-treatment steps for the removal of CO2, H2O and This section introduces the key performance indicators (KPIs) that
other trace components are important for the operation of LAES systems, are used to evaluate the different process configurations, followed by a
they are not included in our simulation and optimization studies, since validation of the process model that is used against the results presented
the main focus is on the LAES configurations. The LAES system is by Guizzi et al. (Guizzi et al., 2015). Finally, the optimization formu­
modeled and simulated by Aspen HYSYS Version 10.0 (HYSYS, 2017) lation is presented.
with the Peng-Robinson equation of state to calculate thermodynamic
properties of process streams. It is assumed that the air feed consists of 4.1. Process evaluation
78.82 mole% nitrogen, 21.14 mole% oxygen and 0.04 mole% argon. The
mass flowrate of air feed is 61,520 kg/h, which is aimed at a 10 MW Various key performance indicators (KPIs) are selected to evaluate
scale energy storage system, and air is supplied at 20◦ C and 1 bar. In this the thermodynamic performance of different LAES configurations. As
work, the configuration of the cold thermal energy recovery part is mentioned in Section 1, the most widely used parameter for energy
identical for all cases. This also applies to the working fluids of the two
cycles, where methanol and a multi-component fluid are adopted to
Table 1
transfer the cold regasification energy of liquid air. As mentioned in
Simulation conditions for the LAES configurations.
Section 2, each compressor stage is followed by an aftercooler and a
preheater is placed before each expander stage. Pressure drops and heat Design parameters Unit Value

losses of heat exchangers, the phase separator and storage tanks are Ambient temperature ◦
C 20
neglected in the simulation models. To simplify the thermodynamic Ambient pressure bar 1
Isentropic efficiency of compressors % 85
analysis of the LAES system in this work, constant isentropic efficiencies
Isentropic efficiency of expanders % 90
for the compressors and expanders are used in the modeling of the LAES. Isentropic efficiency of cryo-turbine % 75
The effect of different pressure ratios and internal stage pressures and Isentropic efficiency of pumps % 80

5
Z. Liu et al. Computers and Chemical Engineering 169 (2023) 108087

storage technologies is the round-trip efficiency, which is commonly Gundersen, 2015) who only considered the thermo-mechanical exergy.
used for comparison with other technologies in the literature. The spe­ The ETE was further developed to include chemical exergy by Kim and
cific power consumption is a performance parameter to indicate the Gundersen (Kim and Gundersen, 2018). The ETE is defined as the ratio of
efficiency of the liquefaction process in the LAES system. In addition, exergy sinks and exergy sources, and thereby focusing, as the name in­
exergy efficiency (such as the Exergy Transfer Effectiveness - ETE), dicates, on exergy transfer, see Eq. (5).
which evaluates the quality of work and heat in a consistent way, is ∑
ExergySinks
considered as a performance indicator for the charging and discharging ηĖ = ETE = ∑ (5)
ExergySources
parts of the LAES system.
The definition of round-trip efficiency (RTE) is shown in Eq. (1). Any increment in exergy (produced exergy) in the process is regar­
ded as exergy sinks, and exergy decreases (consumed exergy) are exergy
Ẇ out ṁliq wtur wtur
ηRT = = = ηLY ⋅ (1) sources. As mentioned in Section 2, the LAES system is decomposed into
ṁcomp wcomp wcomp
Ẇ in three parts: compression, hot and cold thermal energy recovery, and
Here, ṁliq and ṁcomp denote the mass flowrate of liquid air produced expansion sections. Thus, exergy efficiencies of the three parts are
and air that is sent to compressors, respectively. wcomp and wtur represent evaluated to reveal the performance of each part. The exergy efficiency
the specific work of compressors in the compression part and turbines in of the compression part ηĖcomp is calculated by Eq. (6).
the expansion part, respectively. ηLY is the liquid yield of air. As
Ėout,air,comp + Ėh
mentioned in Section 2, the outlet stream of the cryo-turbine that is ηĖcomp = (6)
downstream of the cold box is split into a vapor and a liquid stream. The Ẇ comp + Ėfeed,air
vapor stream is returned back to the compression part, which leads to a Here, Ẇcomp denotes the work consumed by compressors in the
larger mass flowrate of air entering the compressors than the mass
compression part. Ėout,air,comp , Ėh and Ėfeed,air represent the thermo-
flowrate of liquid air. The definition of liquid yield of air is the ratio
mechanical exergy of outlet air of the compression part, working fluid
between the mass flowrates of liquid air produced and air compressed.
in the hot thermal energy recovery cycle (thermal oil), and the air feed.
Specific power consumption (SPC) is the net work consumption in
The thermo-mechanical exergy of streams, which can be calculated by
the charging part divided by the mass flowrate of liquid air produced,
Eq. (4), is obtained by applying Visual Basic codes in Aspen HYSYS
see Eq. (2).
simulations as proposed in the work of Abdollahi-Demneh et al.
∑
Ẇ net Ẇ comp − Ẇ cryotur (Abdollahi-Demneh et al., 2011). Similar to exergy efficiency for the
SPC = = (2)
ṁliq ṁliq compression section, exergy efficiencies of the cold energy recovery ηĖrec
and expansion section ηĖexp,dir or ηĖexp,dir+ORC are given by (Eqs. 7, 8 and 9).
Here, Ẇcomp is the total work consumed in the compression part and
ηĖexp,dir represents the exergy efficiency of the expansion part that only
Ẇcryotur is the expansion work produced by the cryo-turbine. The above-
considers the multistage turbine, while ηĖexp,dir+ORC is the exergy efficiency
mentioned performance parameters evaluate the system in terms of
energy efficiency. of the expansion section when an additional ORC is part of the LAES
However, changes in temperature, pressure, and composition of system.
process streams also have significant effects on the performance of the Ėout,air,rec + Ẇ cryotur
system. This information cannot be revealed by energy performance ηĖrec = (7)
Ėout,air,comp + Ẇ pump
parameters. Exergy analysis is a comprehensive method to include both
the 1st and 2nd laws of thermodynamics. All variations in process
Ẇ tur,exp
streams, work and thermal energy in the system are considered in the ηĖexp,dir = (8)
Ėout,air,rec + Ėh
concept of exergy. For process streams, exergy represents the maximum
work that can be obtained by varying the stream temperature, pressure
Ẇ tur,exp + Ẇ tur,ORC
and composition in reversible processes to equilibrium with environ­ ηĖexp,dir+ORC = (9)
mental conditions. The exergy of heat is a simple function of tempera­ Ėout,air,rec + Ėh + Ẇ pump,ORC
ture through the Carnot factor, while work is pure (100%) exergy. Not Here, Ėout,air,rec represents the thermo-mechanical exergy of outlet air
including electrical, potential, kinetic, and nuclear exergies, the exergy
from the cold thermal energy recovery part and Ẇtur,exp is the work
of process streams consists of thermo-mechanical (or physical) exergy
TM Ch produced by the multistage turbine in the expansion part. Ẇtur,ORC and
(Ė ) and chemical exergy (Ė ) (Kotas, 2012), see Eq. (3).
Ẇpump,ORC are the expansion work and pump work in the additional ORC.
(3)
TM Ch
Ė = Ė + Ė
4.2. Process validation
Thermo-mechanical exergy is the maximum work produced in ideal
processes when the stream is taken to a situation that has the same
The validation of the simulation model that is used for the analysis
temperature and pressure as the environment, as is given by Eq. (4).
and optimization of the various configurations in this work has been
Ė
TM
= Ḣ(T, p) − Ḣ(T0 , p0 ) − T0 [Ṡ(T, p) − Ṡ(T0 , p0 )] (4) done by comparing with the work by Guizzi et al. (Guizzi et al., 2015).
As described in our previous work (Liu et al., 2022), RTE values for
Chemical exergy is the maximum work obtained in ideal processes different charging pressures in the flowsheet shown in Fig. 2 with N=2
when the stream composition is changed to equilibrium with the compression stages and M=3 expansion stages have been compared
composition of the environment under ambient conditions. The effect of with values presented in (Guizzi et al., 2015). Fig. 4 shows the corre­
chemical exergy in the LAES system is relatively small since there are no sponding RTE values, and the difference between the results is within
chemical reactions; only the separator after the cryo-turbine has 1.4%. This means that our process model has an acceptable accuracy.
compositional changes, and the LAES system experiences minor changes
in chemical exergy.
4.3. Process optimization
Exergy Transfer Effectiveness (ETE) is adopted to calculate the
exergy efficiency of the system, since this is a suitable performance
Due to the thermal energy storage cycles, the charging and dis­
parameter for energy conversion processes. The concept of ETE was first
charging parts are closely connected. The power recovery ratio is crucial
proposed by Marmolejo-Correa and Gundersen (Marmolejo-Correa and
for an energy storage technology, and measures to increase the RTE of

6
Z. Liu et al. Computers and Chemical Engineering 169 (2023) 108087

nitrogen rejection processes from natural gas using PSO. The PSO al­
gorithm is connected to Aspen HYSYS, performing the simulation of
each input generated by the algorithm. The inputs to Aspen HYSYS are
generated randomly by the PSO algorithm, and certain inputs may lead
to infeasible solutions and non-convergence of the Aspen HYSYS simu­
lation. In this case, the objective function is set to a large value.
ṁliq wtur
min − ηRT = − f (x) = −
x ṁcomp wcomp
subject to ΔTa ≥ 10 a ={aftercoolers 1, 2, ..., N, preheaters 1, 2,..., M}
ΔTb ≥ 1 b ={HX − 1, 2, Eva − 1, 2}
VFORC,pump,in = 0
VFORC,tur,in = 1
xLB ≤ x ≤ xUB
(10)
Fig. 4. Model validation by comparing RTE with numbers from Guizzi et al. Here, ΔTa is the minimum temperature difference (MTD) of after­
(Guizzi et al., 2015). coolers and preheaters, while ΔTb is the minimum temperature differ­
ence (MTD) for heat exchangers and evaporators. VFORC,pump,in and
the system are strongly required. This can be solved by using optimi­ VFORC,tur,in denote the vapor fraction of inlet streams of the pump and
zation algorithms in combination with thermodynamic analysis to the turbine in the ORC. x represents decision variables in the LAES
improve the process performance. system, while xLB and xUB are the lower and upper bounds of the
The optimization of the LAES involves highly nonlinear and non­ variables.
convex equations, which is challenging for deterministic optimization The objective function and the constraints of the optimization
algorithms due to the complex thermodynamic property calculations. formulation are provided in Eq. (10). The purpose of optimizing the
Surrogate models or simplified models can be considered as alternatives system is to identify the most promising option among different LAES
for the replacement of rigorous thermodynamic models when perform­ configurations. The decision variables in this work include the pressure
ing the optimization. However, accurate thermodynamic models are ratios of compressors and expanders, the outlet temperature of thermal
prerequisite for reliable results. It is also important to identify the best oil from coolers, the outlet temperature of air and the recycled air stream
decision variables, which form a multi-dimensional space during opti­ from the cold box, and the molar flowrates and temperatures of the
mization. Thus, it is computationally intensive to find global optimal working fluids in the cold thermal energy recovery cycle. In the system
solutions. The built-in optimizer in HYSYS is too weak to get satisfactory with an additional ORC, the evaporation temperature and pressure and
results. To deal with such challenges, stochastic search algorithms such the molar flowrate of the working fluid are also selected as variables.
as Particle Swarm Optimization (PSO) have been suggested used in The degrees of freedom for design are different in different cases, since
energy intensive systems such as liquefaction processes and LNG rega­ the number of compression and expansion stages are varying in the
sification processes (Mofid et al., 2019, Cao et al., 2017, Sun et al., LAES.
2018). Compared to other stochastic search algorithms, the PSO is MTDs for heat exchangers are economic parameters trading off in­
shown to be a technique that is computationally faster and provides vestment cost and operating cost. Transferring heat with large temper­
better optimal solutions (Eghbal et al., 2011, Elbeltagi et al., 2005). ature differences increases irreversibilities in the plant, and these exergy
Another advantage of using the PSO is that few adjusting parameters are losses are paid for by increased compressor work. In sub-ambient pro­
required. A Genetic Algorithm (GA) is another popular evolutionary cesses, these exergy losses depend on both the temperature difference
optimization method to be used in the optimization of chemical engi­ and the absolute temperature level, making it more important to reduce
neering problems (Jin and Lim, 2018). Unlike the GA, the PSO is mainly temperature driving forces at lower temperatures. As a result, MTDs of
applied in continuous problems and does not require coding of variables aftercoolers and preheaters that are operating above ambient tempera­
and operations of crossover and mutation. Thus, the computing time is ture are assumed to be 10◦ C, while MTDs for HX-1,2 and Eva-1,2 that are
reduced, and the implementation is easier for the PSO compared to the operating significantly below ambient temperature are assumed to be
GA. For these reasons, the PSO algorithm is applied in this study to 1◦ C (Higginbotham et al., 2011). This is a simplification, but the focus of
optimize LAES systems. this work is not cost minimization, emphasis is on energy efficiency. It is
The PSO algorithm is implemented in Matlab, where the relative shown in the literature that below ambient, UAmax is a better design
convergence of variables or the fitness function values being unchanged specification than MTD, where UA is the lumped parameter of heat
after 45 iterations are termination conditions. Table 2 lists the param­ transfer coefficient and heat transfer area, also referred to as the heat
eters for the PSO algorithm. Acceleration factors affect the movement of exchanger conductance (Jensen and Skogestad, 2008, Austbø and
particles and the inertia weight decides the convergence behavior of the Gundersen, 2015, Kim and Gundersen, 2017). In the LAES system with
PSO algorithm during optimization. These parameters were suggested in an additional ORC to utilize the unused compression heat, the inlet
a recent work (Hamedi et al., 2018) related to the optimization of stream to the pump should be totally liquid and the inlet stream to the
turbine should be totally vapor. Thus, the vapor fraction of the inlet
stream to the pump and the turbine in the ORC should be 0 and 1,
Table 2 respectively. In reality, the outlet temperature of compressors should be
Parameters for the PSO algorithm. considered as constraints during optimization. However, constraints
Parameters Value related to the pressure ratio and outlet temperature of compressors and
expanders are not considered in this thermodynamic analysis of the
Number of particles 150
Global acceleration factor 1
LAES system. The decision variables and their lower and upper bounds
Personal acceleration factor 1 are listed in Table 3.
Minimum inertia weight 0.5 It is worth noting that the pressure ratios of compressors would be
Maximum inertia weight 1 relatively high when the number of compression (or expansion) stages is

7
Z. Liu et al. Computers and Chemical Engineering 169 (2023) 108087

Table 3 stages is increased with changing number of compression stages from 2

Decision variables with lower and upper bounds. to 4, and then it is reduced with further increasing number of
Variables Lower Upper compression stages from 4 to 6. Optimization results show that there
Bounds Bounds exists an optimal match between the number of compression stages and
Pressure ratio for compressors a, b 1 20 expansion stages. When the number of expansion stages is 3, 4 and 5, the
Pressure ratio for expanders a, b 1 10 highest RTE is obtained with 2, 3, and 4 compression stages in the LAES
Thermal oil temperature (TH21-H2N) (◦ C) a, b 100 230 system, respectively. This is due to the fact that the hot storage cycle
Cold box outlet air temperature (◦ C) a, b -185 -165 connects the compression and expansion parts in a standalone LAES
Cold box outlet recycled air temperature (◦ C) a, b -10 29
Working fluid operating temperature (higher) (◦ C) -90 -20
system. With varying numbers of compression stages, the temperature
a, b and flowrate of thermal oil change, which influences not only the tem­
Working fluid molar flowrate (kmol/h) a, b 0 200 perature of the air entering expanders but also the location of pinch
Working fluid operating temperature (lower) (◦ C)
a, b
-186 -166 points in preheaters before each stage of the expander. The overall best
performance for the LAES system is with a 2-stage compression and a 3-
Working fluid molar flowrate in ORC (kmol/h) b 0.01 800
Working fluid evaporation pressure in ORC (bar) b 1.1 41 stage expansion, and the highest RTE is found to be 66.7%. The
Working fluid evaporation temperature in ORC 10 113.6 temperature-entropy diagram for the energy storage and release mode of
(◦ C) b the three best matches between compression and expansion stages is
a
variable bounds for the LAES when the number of compression stages is less provided in Fig. 6.
than the number of expansion stages (N<M) The properties of the thermal oil transferring heat of compression to
b
variable bounds for the LAES with additional ORC when the number of the expansion section depend only on the configuration of the
compression stages is greater than or equal to the number of expansion stages compression section. With increasing number of compressor stages, both
(N≥M) the total compressor duty (and thereby the compression heat, see
Table 4) and the outlet temperature of air from the compressor stages
small in the LAES. In the ADELE and ADELE-ING projects (RWE Power, are reduced. This obviously also reduces the outlet temperature of
2010, Zunft, 2015), the combination of an axial LP compressor and a thermal oil from the aftercoolers, as shown in Fig. 7. As mentioned in
radial HP compressor has been suggested to elevate the pressure of air Section 2, even though the compressor duty is decreased with increasing
from 1 bar to 100 bar with an outlet temperature of 600◦ C. Axial and number of stages, the repeated cooling of air increases the flowrate of
radial compressors have been used in some aerospace applications with thermal oil since it must be split into more branches. The effect of
a pressure ratio of up to 40 and in some industrial applications with a repeated cooling of air is more important than the reduced compressor
pressure ratio of up to 30 (U.S. Department of Energy 2006). The upper duty. As a result, the flowrate of thermal oil increases with the number of
bound for the pressure ratio of compressors is arbitrarily set to 20 in this compression stages, and this is also shown in Fig. 7.
work, which is achievable according to the literature review. Focusing on the expansion section, the fact that the flowrate and
temperature of thermal oil have opposite trends with respect to the
5. Results and discussion number of compression stages also means that they have opposite effects
on the preheating of air in the expansion section, and therefore also on
In this work, the LAES system with different number of compression the power generation and round-trip efficiency. Thus, there is a trade-off
and expansion stages is optimized and compared. It is worth noting that between mass flowrate and temperature of thermal oil, which is why
an additional ORC is utilized only when the number of compression there exists an optimal match between the number of compression
stages is greater than or equal to the number of expansion stages. stages and the number of expansion stages. This optimal match has
already been presented based on the results in Fig. 5. For 3, 4 and 5
5.1. Performance of different LAES configurations expansion stages, the optimal number of compression stages is 2, 3 and
4, respectively. The best combination with 2 compression stages and 3
Fig. 5 shows the round-trip efficiency of the LAES system with expansion stages has the highest RTE of 66.7%.
different number of compression and expansion stages. It can be seen The easiest way to explain the trends in Fig. 5 is to consider a case
that when there is a 3-stage turbine in the expansion part, the RTE of the with a fixed number of expansion stages, such as 5. Then the thermal oil
system reduces with increasing number of stages in the compression must be split into 5 branches and sent to the air preheaters. With few
part. The trend for an LAES system with a 4-stage turbine is different.
The RTE first increases when the number of compression stages is
changed from 2 to 3, and then, the RTE decreases when the number of
compression stages continues to increase from 3 to 6. For the system
with 5 stages expansion, the RTE that has a maximum at 4 compression

Fig. 6. Energy storage and release mode in a T-S diagram for the three best
Fig. 5. Round-trip efficiency of the LAES for different configurations. matches between compression and expansion stages.

8
Z. Liu et al. Computers and Chemical Engineering 169 (2023) 108087

Table 4
The compression heat recovered by the thermal oil for configurations with 3
expansion stages and 2-6 compression stages.
Compression stages 2 3 4 5 6

Recovered heat duty GJ/h 56.88 47.78 48.11 46.49 41.37

Fig. 7. Molar flowrate and temperature of thermal oil for configurations with 3
expansion stages and 2-6 compression stages.

compression stages, such as 2 or 3, the temperature of thermal oil is

relatively high, but the flowrate of thermal oil is too low, causing a pinch
in the cold end of the preheaters. This means that the expansion part is
not able to take advantage of the high thermal oil temperature, although
the situation is somewhat improved from 2 to 3 compression stages.
With 4 compression stages, the composite curves in the preheaters are
almost parallel, and the RTE reaches its maximum value for the case
with 5 expansion stages. Increasing the number of compression stages
further to 5 or 6 will result in a too high flowrate of thermal oil, and the
pinch will move to the hot end of the preheaters. Air can now be pre­
heated to a temperature that is ΔTmin below the thermal oil temperature,
however, this temperature is reduced due to the relatively large number
of compression stages, and the large flowrate of thermal oil also means
that the compression heat cannot be fully utilized.
Fig. 8 shows the scope for using ORC to produce power from
compression heat that is not fully utilized in the expansion section. As
already explained, such surplus heat will be available when the number
of compression stages is equal to or larger than the number of expansion
stages. The split of thermal oil between the preheaters and the additional
ORC is shown as a function of number of compression stages for 3
expansion stages (Fig. 8(a)), 4 expansion stages (Fig. 8(b)) and 5
expansion stages (Fig. 8(c)). Unfortunately, the power from the ORC is
not enough to compensate for the reduction in power production in the
expansion section in these cases, partly caused by reduced thermal oil
temperature. As a result, for the entire system (charging and discharg­
ing), the RTE will be reduced when the number of compression stages is
increased beyond the optimal number for the given number of ex­
panders (the 2-3, the 3-4, and the 4-5 matches). This explains the trends
in Fig. 5.
The logarithmic mean temperature differences (LMTDs) of pre­
heaters in different LAES configurations are listed in Table 5. When the
LAES system has 3 stages of expansion, a relatively good match between
the temperature profiles of thermal oil and air is obtained with 2 stages
of compression. The flowrate of thermal oil is slightly less than required
to have parallel profiles, and the pinch is in the cold end of the pre­
heaters. As a result, the LMTD is 10.8◦ C in this case. It is observed that
Fig. 8. Optimized distribution of thermal oil between the expansion preheaters
the LMTDs for the cases with surplus thermal oil are 10.3◦ C. Since ΔTmin and the heater in the ORC in different configurations of the LAES: (a) 3-stage
equals 10◦ C, an LMTD of 10.3◦ C indicates a case with close to parallel turbine; (b) 4-stage turbine; (c) 5-stage turbine.
composite curves in preheaters. For the same reason, when the LAES has
4 or 5 stages of expansion, systems with 3 or 4 stages of compression
have better performance compared to other combinations. The com­
posite curves for the cases mentioned above are shown in Fig. 9. In Fig. 9

9
Z. Liu et al. Computers and Chemical Engineering 169 (2023) 108087

Table 5
Optimal values for some variables and key performance indicators in different LAES configurations.
Exp. stages Comp. stages prcomp prexp LMTDpre LMTDheater, ORC ηLY SPC Ẇch Ẇdc ηRT
(◦ C) (◦ C) (%) (kWh/t) (MW) (MW) (%)

3 2 13.11 5.29 10.78 - 86.21 243.95 15.01 10.00 66.65

3 5.24 4.98 10.27 32.85 86.19 204.79 12.60 8.19 64.99
4 3.70 5.43 10.27 20.46 86.62 203.41 12.51 7.90 63.12
5 2.86 5.45 10.27 18.11 86.64 195.93 12.05 7.30 60.60
6 2.25 4.82 10.27 20.17 85.98 176.83 10.88 6.30 57.88
4 2 12.86 3.45 29.11 - 86.34 241.59 14.86 9.55 64.23
3 5.31 3.37 11.62 - 86.26 207.08 12.74 8.32 65.32
4 3.70 3.56 10.27 21.50 86.57 203.05 12.49 8.05 64.41
5 2.86 3.57 10.27 18.24 86.65 195.96 12.06 7.52 62.40
6 2.23 3.21 10.27 19.26 85.93 175.39 10.79 6.30 58.42
5 2 14.19 2.79 39.95 - 86.68 252.99 15.56 9.64 61.93
3 5.34 2.65 21.42 - 86.29 207.85 12.79 8.13 63.61
4 3.72 2.77 10.59 - 86.65 204.07 12.55 8.05 64.10
5 2.89 2.79 10.27 18.26 86.65 197.88 12.17 7.57 62.22
6 2.22 2.54 10.27 19.56 85.91 174.70 10.75 6.40 59.53

Fig. 9. Composite curves of preheaters in the LAES system: (a) 2-stage compressor and a 3-stage turbine; (b) 3-stage compressor and a 4-stage turbine; (c) 4-stage
compressor and a 5-stage turbine; (d) 2-stage compressor and a 4-stage turbine.

10
Z. Liu et al. Computers and Chemical Engineering 169 (2023) 108087

(a)-(c), the composite curves are close to parallel as indicated by the constant temperature at the outlet of the cold box, compression pressure
listed LMTD values, and they have higher RTE values than other cases is the only factor that affects the liquid yield of air. A higher charging
having the same number of expansion stages. Fig. 9(d), in contrast, il­ pressure increases the pressure drop in the cryo-turbine, with atmo­
lustrates an example of composite curves for preheaters when the system spheric pressure at the outlet. This results in a lower outlet temperature
has insufficient amounts of thermal oil in the expansion part. In this and reduced fraction of air in vapor phase, which improves liquid yield
particular case, the LAES system has a 2-stage compressor and a 4-stage of air. For the entire system, the use of the hot thermal recovery cycle
turbine, and the LMTD of the preheaters is 29.1◦ C. As a result, the has decisive effects on the performance (the inlet temperature of air to
relatively large temperature difference between hot and cold streams expanders and the location of pinch points in preheaters), and the
lead to a poor performance and a lower RTE of the system. highest energy recovery ratio of the LAES (RTE) is obtained when the
In addition, when the number of compression stages is greater than composite curves of hot and cold streams in preheaters are close to
or equal to the number of expansion stages, composite curves of pre­ parallel. The major stream data in different optimized LAES configura­
heaters in LAES systems with an additional ORC are similar to and tions are provided in the Supplementary Material.
slightly better than Fig. 9(a)-(c). The reason is that one part of the The detailed optimization results for an additional ORC in different
thermal oil flowrate is sent to the ORC, while the remaining part has a LAES configurations are listed in Table 6. The heat source is stream
cooling curve that is parallel with the air preheating curve. The com­ HORC (thermal oil, see Fig. 3) in the LAES system. The condensation
posite curves of the LAES system with a 4-stage compressor and a 4-stage pressure of the ORC is 5.30 bar for all the cases, which is the saturation
turbine are used as an example and shown in Fig. 10. However, even if pressure of the working fluid at ambient temperature. When the number
there is a good match between the thermal oil and air temperature of compression stages is less than 6 in the LAES system, the evaporation
profiles, the RTE is reduced when the number of compression stages is pressure of the ORC reaches the upper bound, which is set to 41 bar, i.e.
increased from 3 to 4 and the number of expansion stages is 4. This 90% of the critical pressure of the working fluid, resulting in the largest
means that despite the power production in the ORC, there is a larger power output. When the number of compression stages is 6, the heat
reduction in the power production in the expansion part compared to source temperature (thermal oil) is less than the saturation temperature
the case with 3 compression stages. The thermal oil temperature is of the ORC working fluid at critical pressure. Thus, the pressure and
reduced when increasing the number of compression stages from 3 to 4, temperature of the working fluid in the ORC are less than for the other
and the RTE is reduced. cases that have 2-5 compression stages. It is shown that the largest net
Optimal results for compression and expansion pressure ratios and work output of 605.01 kW is obtained when there is a 4-stage
key performance indicators (liquid yield, specific power consumption, compressor and a 3-stage turbine in the LAES system. It is worth
exergy efficiency and round-trip efficiency) for different LAES configu­ noting that the net power output in the ORC is affected by the physical
rations are listed in Table 5. It can be seen that optimal pressure ratios, properties of the working fluid (critical pressure and temperature).
the power consumption of compressors and SPCs are reduced with Thus, the performance of the ORC may be different for other working
increasing number of compression stages for a given number of expan­ fluids.
sion stages in the LAES. This is because a near isothermal operation of
compression requires near minimum power consumption. Moreover, the 5.2. Exergy analysis of different LAES configurations
power production of the discharging part has the same trend as the
power consumption of the charging part. This is because a higher Exergy efficiencies of the compression, cold thermal energy recovery
expansion pressure results in more work produced, but the expansion and expansion parts in the LAES system with different number of
pressure is constrained to be less than the compression pressure due to compression and expansion stages are shown in Fig. 11. From Fig. 11(a),
the cold thermal energy recovery cycles. The cold regasification energy it can be seen that the exergy efficiency of the compression part is
from liquid air is needed to liquefy compressed air in the LAES system. In reduced with increasing number of compression stages. This is mainly
order to have driving forces for cold regasification energy transfer, the caused by the increased exergy losses related to irreversibilities in the
liquid air must be at a lower temperature than the compressed air. After aftercoolers. The same trend is observed in Fig. 11(b) and (c). It is worth
the cold box, the optimal phase separation temperature is around - noting that the exergy efficiencies of the cold thermal energy recovery
176◦ C for all cases, since the amount and temperature of cold thermal part are almost the same in the various cases. The number of compres­
energy in the cold energy recovery part are insufficient. With almost sion and expansion stages has only marginal effects on the cold thermal
energy recovery part. The somewhat obvious reason is that the config­
uration of this part is the same in all cases. The exergy efficiencies of the
expansion part, however, show significant differences. These exergy
efficiencies are affected by both the heat transfer efficiency in preheaters
and the performance of the ORC.
For the direct expansion part, which means the multistage turbine
part, it is observed that when the number of expansion stages is 3, the
highest exergy efficiency is obtained with a 2-stage compressor in the
system. For the system with a 4-stage or 5-stage turbine, the best per­
formance is obtained when there is a 3-stage or 4-stage compressor. This
is in line with the previous discussion and conclusion based on com­
posite curves in the preheaters.
As already established in Section 5.1, the best performance measured
by the RTE is obtained when the flowrate of thermal oil is large enough
to have close to parallel temperature profiles in the preheaters of the
expansion part of the LAES. If the flowrate of thermal oil is too small, the
pinch in the preheaters is in the cold end, and air preheat cannot take
advantage of the thermal oil inlet temperature. If the flowrate of thermal
oil is too large, the pinch in the preheaters is in the hot end, which is an
advantage for air preheat, but then there is a part of the compression
Fig. 10. Composite curves of preheaters in the LAES system with a 4-stage heat transferred by the thermal oil that is not utilized. In such cases, one
compressor and a 4-stage turbine. could envisage that the surplus flowrate of thermal oil could be sent to

11
Z. Liu et al. Computers and Chemical Engineering 169 (2023) 108087

Table 6
Performance of the additional ORC for utilizing surplus compression heat.
Exp. stages Comp. stages Ẇnet porc,con porc,eva Torc,eva Ẇorc,tur Ẇorc,pump
(kW) (bar) (bar) (◦ C) (kW) (kW)

3 2 - - - - - -
3 363.86 5.30 41.00 112.50 394.96 31.10
4 605.01 5.30 41.00 112.00 657.02 52.01
5 576.67 5.30 41.00 110.32 627.33 50.66
6 361.94 5.30 18.64 69.17 379.90 17.96
4 2 - - - - - -
3 - - - - - -
4 281.80 5.30 41.00 113.03 305.74 23.94
5 374.21 5.30 41.00 110.44 407.03 32.82
6 216.20 5.30 21.22 92.48 226.80 10.60
5 2 - - - - - -
3 - - - - - -
4 - - - - - -
5 185.89 5.30 41.00 110.37 202.21 16.32
6 141.81 5.30 20.82 89.56 148.75 6.94

an ORC to produce additional power while making the composite curves

parallel.
The following conclusions can be made based on the results in
Table 5 and Fig. 11(a)-(c) and explained by exergy analysis, as well as
the observations in Section 5.1:

• Increased flowrate of thermal oil is a result of increased number of

compression stages, and despite a reduction in compression work,
irreversibilities in the aftercoolers are increased. Repeated cooling of
air increases with more compression stages, thereby increasing ir­
reversibilities due to heat transfer with temperature differences
larger than zero.
• When the flowrate of thermal oil is larger than required in the
expansion section, the use of ORC both makes the temperature pro­
files in the preheaters parallel (and thereby reduce irreversibilities)
and produces additional power by utilizing otherwise wasted
compression heat.
• Since the work produced by the ORC is quite small, it cannot
compensate for the additional irreversibilities in the compression
part by having more stages.

Considering first the cases without ORC, the maximum exergy effi­
ciency of the expansion part coincides with the maximum RTE (i.e., 2
compression stages for 3 expansion stages, 3 compression stages for 4
expansion stages, and 4 compression stages for 5 expansion stages).
However, while the RTE is reduced for these cases with increased
number of expansion stages (from 66.7% via 65.4% to 64.2%), the
expansion section exergy efficiency is increased with increased number
of expansion stages (from 84.3% via 85.0% to 85.4%). Contributing to
the overall system, exergy efficiency of the compression part is reduced
with increasing number of compression stages (from 89.4% via 88.5% to
87.5%). The larger reduction in exergy efficiency of the charging part
compared to the discharging part explains why there is a reduction in
RTE for the overall system.
As explained before, use of an additional ORC only makes sense
when the number of compression stages is greater than or equal to the
number of expansion stages. As can be seen in Fig. 11, the maximum
exergy efficiency of the expansion section for LAES systems with an
additional ORC is obtained when the number of compression stages is
equal to the number of expansion stages. However, Fig. 11 also indicates
that the exergy efficiency with ORC is only marginally better than for the
optimal case without ORC. For 3 expansion stages, the exergy efficiency
is improved from 84.3% to 84.8%. For 4 expansion stages, a similar
Fig. 11. Exergy efficiencies of the compression, cold thermal energy recovery improvement from 85.0% to 85.7% is observed. Finally, for 5 expansion
and expansion parts in different LAES configurations: (a) 3-stage turbine; (b) 4- stages, the exergy efficiency with the use of an ORC is actually reduced
stage turbine; (c) 5-stage turbine. from 85.4% to 84.5%. In summary then, the power produced by the
additional ORC does not justify the investment cost and added

12
Z. Liu et al. Computers and Chemical Engineering 169 (2023) 108087

complexity. 6. Conclusions
As for a standalone LAES system without external heat sources, the
exergy efficiency of the overall system would be very similar to the The scope of this work has been to investigate opportunities for
definition of the round-trip efficiency, which is why it has not been improving the energy performance of Liquid Air Energy Storage (LAES)
calculated in this work. The major terms in the exergy efficiency for the systems. While previous work in our group considered improving the
overall system are the work produced in the discharging and the work cold thermal energy recovery cycles, this work has focused on the
consumed in the charging. The minor terms are the thermo-mechanical different configurations of the compression and expansion sections,
exergy of the inlet air (which is constant) and the outlet air, and both meaning the number of compressor and expander stages as well as the
terms are negligible. On the other hand, the individual exergy effi­ hot thermal energy recovery cycle.
ciencies of the charging and discharging parts can help explain the The main assumptions in this work are:
change in RTE. However, the economic feasibility of the system should
be analyzed for a comprehensive evaluation of the LAES system before • Constant isentropic efficiencies for varying compressor and expander
any project is implemented, and this is beyond the scope of this work. duties and number of stages.
• Fixed ΔTmin for heat exchangers, however, adjusted values for above
5.3. Effects of additional ORCs (10◦ C) and below (1◦ C) ambient temperature.
• Fixed set and composition of the two cold thermal energy recovery
In addition to the ORC for the unutilized part of compression heat, an cycles.
ORC can also be used to collect the heat carried by the exhaust air from • R152a has been selected as the working fluid for all cases involving
the last stage expander for further improvement of the LAES. In this ORC.
study, we used the same working fluid (R152a, as discussed in Section 2) • Pressure ratios have been allowed to vary between 1 and 20.
as for the ORCs for unutilized compression heat. As mentioned in Sec­ • No heat losses to the surroundings or pressure drops in piping and
tion 5.1, the LAES system with a 2-stage compressor and a 3-stage equipment.
expander has the best performance, and this process configuration is
selected as the design basis. The process flow diagram of this configu­ One important observation from this work is that when the number
ration is shown in Fig. 12. In this process, the temperature of exhaust air of compression stages is greater than or equal to the number of expan­
(stream D11) is 103.6◦ C. Optimization results indicate that only 78.2 kW sion stages, the expansion section is not able to fully utilize the
power is produced by the ORC, and this added power does not justify the compression heat that is carried by the thermal oil in the hot energy
investment in an ORC. The RTE of the LAES system can only be recovery cycle. This can easily be explained by relative slopes of the
improved from 66.7% to 67.2% with this additional ORC. The temperature profiles in the air preheaters. An Organic Rankine Cycle
improvement by producing power from the heat of exhaust air is mar­ (ORC) has been used to turn this unused compression heat into power.
ginal compared to the system where wasted compression heat is utilized The LAES configuration with the best performance has been identified
by ORCs. by a systematic optimization-based comparison of the various cases. The

Fig. 12. Flow diagram of the liquid air energy storage with an additional ORC for exhaust air.

13
Z. Liu et al. Computers and Chemical Engineering 169 (2023) 108087

following conclusions are obtained in this study: Acknowledgments

• There exists an optimal match between the number of compression This publication has been funded by HighEFF - Centre for an Energy-
stages and expansion stages in a standalone LAES system. When the Efficient and Competitive Industry for the Future. The authors gratefully
number of expansion stages is 3, 4 and 5, the highest RTE is obtained acknowledge the financial support from the Research Council of Norway
with 2, 3, and 4 compression stages, respectively. Among these, the and user partners of HighEFF, an 8-years Research Centre under the
LAES system with a 2-stage compression and a 3-stage expansion has FME-scheme (Centre for Environment-friendly Energy Research,
the highest RTE of 66.7%, which compares nicely with the original 257632).
work of Guizzi et al. in 2015 (54.4%).
• ORCs are used to recover compression heat that is not fully utilized Supplementary materials
for preheating air in the expansion section. The largest net work
output obtained for the case with 4-stage compression and 4-stage Supplementary material associated with this article can be found, in
expansion is 605.01 kW, unfortunately this work is less than the the online version, at doi:10.1016/j.compchemeng.2022.108087.
reduced expansion work when increasing the number of compression
stages from 3 to 4, i.e. the optimal match for the case with 4-stage References
expansion. This shows that ORCs can never improve the energy ef­
ficiency of standalone LAES systems. Abdollahi-Demneh, F, Moosavian, MA, Omidkhah, MR, Bahmanyar, H., 2011.
Calculating exergy in flowsheeting simulators: A HYSYS implementation. Energy 36
• For cases with optimal matches between the number of compression (8), 5320–5327.
stages and expansion stages, the exergy efficiency of the compression Aneke, M, Wang, M, 2016. Energy storage technologies and real life applications – a state
part is reduced with increasing number of compression stages (from of the art review. Appl. Energy 179, 350–377.
Arabkoohsar, A, Machado, L, Farzaneh-Gord, M, Koury, RNN., 2015. Thermo-
89.4% via 88.5% to 87.5%), while the expansion section exergy ef­ economicanalysis and sizing of a PV plant equipped with a compressed air energy
ficiency is increased (from 84.3% via 85.0% to 85.4%). The larger storage system. Renew. Energy 83, 491–509.
reduction in exergy efficiency of the charging part compared to the Austbø, B, Gundersen, T., 2015. Optimal distribution of temperature driving forces in
low-temperature heat transfer. AIChE J. 61 (8), 2447–2455.
discharging part explains the reduction in RTE (from 66.7% via Bellos, E, Tzivanidis, C., 2019. Investigation of the environmentally-friendly refrigerant
65.37% to 64.17%). This clearly indicates that exergy efficiency is a R152a for air conditioning purposes. Appl. Sci. 9 (1), 119.
valuable and comprehensive performance indicator for LAES British Petroleum, 2021. BP Statistical Review of World Energy Report. BP, London, UK.
Cao, Y, Flores-Cerrillo, J, Swartz, CL., 2017. Practical optimization for cost reduction of a
systems.
liquefier in an industrial air separation plant. Comput. Chem. Eng. 99, 13–20.
Cetin, TH, Kanoglu, M, Yanikomer, N., 2019. Cryogenic energy storage powered by
The best configuration of the LAES is the system with 2-stage geothermal energy. Geothermics 77, 34–40.
compression and 3-stage expansion (66.7%). Such high RTE makes the Damak, C, Leducq, D, Hoang, HM, Negro, D, Delahaye, A., 2020. Liquid air energy
storage (LAES) as a large-scale storage technology for renewable energy integration
LAES more competitive among various energy storage technologies in –a review of investigation studies and near perspectives of LAES. Int. J. Refrigerat.
terms of energy efficiency. However, the pressure ratios of compressors 110, 208–218.
in the best configuration of the LAES are relatively high, and the cost for Eberhart, R, Kennedy, J., 1995. A new optimizer using particle swarm theory. In:
Proceedings of the Sixth International Symposium on Micro Machine and Human
unconventional compressors with high pressure ratios is obviously Science, pp. 39–43.
higher than the conventional ones. Thus, there are several challenges Eghbal, M, Saha, TK, Hasan, KN., 2011. Transmission expansion planning by meta-
that can be considered in future work: heuristic techniques: a comparison of shuffled frog leaping algorithm, PSO and GA.
In: IEEE Power and Energy Society General Meeting, pp. 1–8.
Elbeltagi, E, Hegazy, T, Grierson, D., 2005. Comparison among five evolutionary-based
• A cost analysis can be conducted and used to indicate the profit­ optimization algorithms. Adv. Eng. Informat. 19 (1), 43–53.
ability and feasibility of the LAES system with high pressure ratio European Commission, 2014. Communication from the Commission to the European
Parliament, the Council, the European Economic and Social Committee and the
compressors. Committee of the Regions - a Policy Framework for Climate and Energy in the Period
• More case studies are needed to analyze the LAES with additional from 2020 to 2030. Brussels.
thermodynamic cycles. Guizzi, GL, Manno, M, Tolomei, LM, Vitali, RM., 2015. Thermodynamic analysis of a
liquid air energy storage system. Energy 93, 1639–1647.
• Going beyond the standalone LAES system, integration with external
Hamedi, H, Karimi, IA, Gundersen, T., 2018. Optimal cryogenic processes for nitrogen
hot and cold thermal energy sources can significantly boost the RTE rejection from natural gas. Comput. Chem. Eng. 112, 101–111.
of the system. Higginbotham, P, White, V, Fogash, K, Guvelioglu, G., 2011. Oxygen supply for oxyfuel
CO2 capture. Int. J. Greenhouse Gas Control 5, S194–S203.
Highview Power, 2022, < www.highviewpower.com/plants/>accessed 23.10.2022.
Author contribution statement HYSYS, A., 2017. Version 10.0. Aspen Technology Inc., Burlington, MA.
Jensen, JB, Skogestad, S., 2008. Problems with Specifying ΔTmin in the Design of
Processes with Heat Exchangers. Ind. Eng. Chem. Res. 47 (9), 3071–3075.
Zhongxuan Liu is the corresponding author and main researcher Jin, C, Lim, Y, 2018. Economic evaluation of NGL recovery process schemes for lean feed
behind the manuscript. She has developed the concept, built the simu­ compositions. Chem. Eng. Res. Des. 129, 297–305.
lation models and carried out the optimizations; Donghoi Kim has been a Kim, D, Gundersen, T., 2017. Constraint formulations for optimisation of dual mixed
refrigerant LNG processes. Chem. Eng. Transcat. 61, 643–648.
discussion partner, in particular on the simulation models and the or­ Kim, D, Gundersen, T., 2018. Development and use of exergy efficiency for complex
ganization of the optimization; Truls Gundersen has been responsible for cryogenic processes. Energy Convers. Manag. 171, 890–902.
the quality of the language and the structure as well as the layout of the Kotas, TJ., 2012. The Exergy Method of Thermal Plant Analysis. Exergon Publishing
Company with Paragon Publishing, London, UK.
manuscript.
Li, Y, Cao, H, Wang, S, Jin, Y, Li, D, Wang, X, et al., 2014. Load shifting of nuclear power
plants using cryogenic energy storage technology. Appl. Energy 113, 1710–1716.
Liu, Z, Kim, D, Gundersen, T., 2022. Optimal recovery of thermal energy in liquid air
Declaration of Competing Interest energy storage. Energy. 240, 122810.
Liu, Z, Yu, H, Gundersen, T., 2020. Optimization of liquid air energy storage (LAES)
using a genetic algorithm (GA). Comput. Aid. Chem. Eng. 48, 967–972.
The authors declare that they have no known competing financial Marmolejo-Correa, D, Gundersen, T., 2015. A new efficiency parameter for exergy
interests or personal relationships that could have appeared to influence analysis in low temperature processes. Int. J. Exergy 17 (2), 135–170.
Mofid, H, Jazayeri-Rad, H, Shahbazian, M, Fetanat, A., 2019. Enhancing the performance
the work reported in this paper.
of a parallel nitrogen expansion liquefaction process (NELP) using the multi-
objective particle swarm optimization (MOPSO) algorithm. Energy 172, 286–303.
Data Availability Morgan, R, Nelmes, S, Gibson, E, Brett, G., 2015. An analysis of a large-scale liquid air
energy storage system. Proceed. Instit. Civil Eng. - Energy 168 (2), 135–144.

Data will be made available on request.

14
Z. Liu et al. Computers and Chemical Engineering 169 (2023) 108087

Park, J, Qi, M, Kim, J, Noh, W, Lee, I, Moon, I, 2020. Exergoeconomic optimization of Sun, Z, Lai, J, Wang, S, Wang, T, 2018. Thermodynamic optimization and comparative
liquid air production by use of liquefied natural gas cold energy. Energy 207, study of different ORC configurations utilizing the exergies of LNG and low grade
118193. heat of different temperatures. Energy 147, 688–700.
Peng, X, She, X, Cong, L, Zhang, T, Li, C, Li, Y, et al., 2018. Thermodynamic study on the Tafone, A, Borri, E, Comodi, G, van den Broek, M, Romagnoli, A., 2018. Liquid air energy
effect of cold and heat recovery on performance of liquid air energy storage. Appl. storage performance enhancement by means of organic Rankine cycle and
Energy 221, 86–99. absorption chiller. Appl. Energy 228, 1810–1821.
Qi, M, Park, J, Kim, J, Lee, I, Moon, I., 2020. Advanced integration of LNG regasification U.S. Department of Energy, 2006. Gas Turbine Handbook. West Virginia, U.S.
power plant with liquid air energy storage: Enhancements inflexibility, safety, and Yu, H, Feng, X, Wang, Y, Biegler, LT, Eason, J., 2016. A systematic method to customize
power generation. Appl. Energy 269, 115049. an efficient organic Rankine cycle (ORC) to recover waste heat in refineries. Appl.
Qi, M, Park, J, Lee, I, Moon, I., 2022. Liquid air as an emerging energy vector towards Energy 179, 302–315.
carbon neutrality: A multi-scale systems perspective. Renew. Sustain. Energy Rev. Zakeri, B, Syri, S., 2015. Electrical energy storage systems: a comparative life cycle cost
159, 112201. analysis. Renew. Sustain. Energy Rev. 42, 569–596.
Rahman, M, Oni, A, Gemechu, E, Kumar, A., 2020. Assessment of energy storage Zhang, T, Zhang, X, He, Y, Xue, X, Mei, S., 2020. Thermodynamic analysis of hybrid
technologies: a review. Energy Convers. Manag. 223, 113295. liquid air energy storage systems based on cascaded storage and effective utilization
Rehman, S, Al-Hadhrami, LM, Alam, MM., 2015. Pumped hydro energy storage system: a of compression heat. Appl. Therm. Eng. 164, 114526.
technological review. Renew. Sustain. Energy Rev. 44, 586–598. Zunft, S., 2015. Adiabatic CAES: The ADELE-ING project. In: SCCER Heat & Electricity
RWE Power, 2010. Adele - adiabatic compressed-air energy storage for electricity for Storage Symposium. Villigen, Switzerland.
electricity supply. RWE power AG. Essen/Köln.

15