Energy Conversion and Management 246 (2021) 114659

Contents lists available at ScienceDirect

Energy Conversion and Management

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

100% saturated liquid hydrogen production: Mixed-refrigerant cascaded

process with two-stage ortho-to-para hydrogen conversion
Muhammad Abdul Qyyum a, 1, Amjad Riaz a, 1, Ahmad Naquash a, 1, Junaid Haider b,
Kinza Qadeer a, Alam Nawaz a, Hyunhee Lee a, Moonyong Lee a, *
a
School of Chemical Engineering, Yeungnam University, Gyeongsan 712-749, Republic of Korea
b
School of Energy and Chemical Engineering, Ulsan National Institute of Science and Technology, 50 UNIST-gil, Eonyang-eup, Ulju-gun, Ulsan 44919, Republic of Korea

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

Keywords: To reduce CO2 emissions and address climate change concerns, most futuristic studies investigating 100%
Liquid hydrogen renewable energy sources and subsequent power-to-gas/fuel/liquid/X technological developments have been
Mixed fluid cascade based on hydrogen (H2). The long-term storage and transportation of H2 over long distances restrict its feasibility
Optimization
as an energy vector, mainly due to its low energy density. Liquefaction is a promising approach for overcoming
Energy consumption
these issues. However, it requires a large amount of energy, and if H2 itself is used to provide this energy, then
Composite curves
Exergy analysis 25% to 35% of the initial quantity of H2 is consumed. The existing H2 liquefaction plants have specific energy
consumption values in the range of 10–12 kWh/kgLH2 and exergy efficiencies in the range of 20%–30% with
complicated configurations. Therefore, a thermodynamically efficient and compact design is required to facilitate
a roadmap to H2 economy. This paper proposes a simple, energy-efficient, and cost-effective process for H2
liquefaction. Three refrigeration cycles with optimal mixed-refrigerant compositions are used, which makes the
proposed process energy-efficient. Additionally, two-stage ortho-to-para conversion makes the process compact.
The proposed process is unique in terms of its configuration and mixed-refrigerant combination. The modified
coordinate descent approach was adopted to identify the optimal design variables for the proposed H2 lique­
faction process. The proposed process consumes an energy of 6.45 kWh/kgLH2 , which is 36.5% and 16.1% lower
than that consumed by the base design of the proposed process and a published base case, respectively. Addi­
tionally, the exergy efficiency of the proposed process is 47.2%. This study will help process engineers achieve a
sustainable green economy by improving the competitiveness of H2 storage and transportation over long
distances.

capacity and transportation in large volumes at high pressures [1,2]. The

1. Introduction low energy density of hydrogen restricts its feasible transportation over
long distances, such as intercontinental export. Similar to liquefied
Hydrogen (H2) is the most abundant element in the universe and natural gas (LNG), pure hydrogen can be liquefied prior to trans­
possesses the highest specific energy among all conventional fuels. It has portation. The resultant cryogenic liquid has an energy density (8.49
long been considered a potential game-changer for global energy sys­ MJ/L) that is three and four times higher compared to compressed
tems and a solution to climate change concerns owing to its green gaseous hydrogen at 250 and 350 bar, respectively [3]. Consequently,
characteristics. From the perspective of sustainable development and bulk hydrogen transportation via bowsers, trains, and/or ships for the
future energy supply chains, H2 may potentially take center stage same energy content is more feasible when the hydrogen is liquefied. A
considering that renewable sources are likely to dominate the global trailer weighing 40 t can transport 3,370 kg of liquid H2 (LH2) compared
energy mix. H2 is not a primary energy source; it is an energy carrier to 530 kg of compressed gaseous H2 at 20 MPa, representing a greater
produced in a manner similar to electricity. than six times increase in capacity [4]. From a H2 economy perspective,
The low energy content per unit volume of hydrogen is considered to the most feasible route in terms of technology, energy, and economics is
be a major challenge for its use as an energy vector in terms of storage the distribution and storage of LH2. Recently, Seo et al. [5] proposed a

* Corresponding author.
E-mail address: [email protected] (M. Lee).
1
These authors contributed equally.

https://doi.org/10.1016/j.enconman.2021.114659
Received 8 February 2021; Accepted 15 August 2021
Available online 3 September 2021
0196-8904/© 2021 Elsevier Ltd. All rights reserved.
M.A. Qyyum et al. Energy Conversion and Management 246 (2021) 114659

commercial H2 liquefaction plants (4.86 kWh/kgLH2 at Ingolstadt). Their

Nomenclature simulation study [14] utilized an MR consisting of 10 components for
subcooling down to − 193 ◦ C. Four H2 Joule–Brayton cascade cycles
LNG Liquid Natural Gas were applied subsequently. The overall energy consumption was re­
SEC Specific Energy Consumption ported to be 5.35 kWh/kgLH2 with efficiency values slightly above 54%.
OPC Ortho-to-Para Conversion Based on the incorporation of ortho-to-para conversion (OPC) through
MFC Mixed Fluid Cascade the installation of numerous reactors integrated with the H2 Joule–­
MR Mixed Refrigerant Brayton cycle, the SEC increased to 5.91 kWh/kgLH2 [15]. Similar results
PMR Precooling Mixed Refrigerant have also been reported in [16]. Sadaghiani and Mehrpooya [17]
CMR Cooling Mixed Refrigerant employed two independent refrigeration cycles with different compo­
LMR Liquefaction Mixed Refrigerant sitions and two conversion reactors. An overall energy consumption of
PR Peng–Robinson 4.41 kWh/kgLH2 and an exergy efficiency of 55.47% were reported.
MITA Minimum Internal Temperature Approach Yuksel [18] reported energy and exergy efficiencies of 70.12% and
MCD Modified Coordinate Descent 57.13%, respectively, for their helium-cooled H2 liquefaction process.
SRE Specific Refrigeration Effect Several integral studies have been conducted attempting to minimize
TDCCs Temperature Difference Composite Curves the energy consumption and cost of the liquefaction process. Re­
THCCs Temperature–Heat flow Composite Curves searchers have used absorption precooling cycles [19], geothermal en­
TH Temperature–Heat flow ergy [20,21], solar energy [22], gasification [23,24], LNG [25,26], and/
SExD Specific Exergy Destruction or a combination of these processes with others [27] to share the energy
COP Coefficient Of Performance load between different sections of the H2 liquefaction cycle. Some
FOM Figure Of Merit advanced tools, such as exergo-economic analysis [28] and optimization
[29,30], have been applied to analyze the performance of H2 liquefac­
tion plants. However, the SEC and exergy efficiency values have not
been significantly improved. A possible reason may be that the lique­
centralized LH2 storage and liquefaction facility as an economically
faction cycle/process is not cost-effective and/or energy-efficient.
beneficial solution considering the economics of scale. H2 is a saturated
Seyam et al. [31] introduced an integral scheme incorporating
liquid at 1 bar and 20 K (approximately − 253 ◦ C), and its liquefaction
geothermal and isobutene power plants to generate the electricity
process requires a large amount of energy. If H2 itself is used to provide
required to power electrolyzers for H2 generation and the subsequent
this energy, then 25% to 35% of the initial quantity of H2 is consumed.
liquefaction process. For the proposed 335 tpd of LH2 production, an
This is much higher than the energy consumed for LNG production,
SEC of 6.41 kWh/kgLH2 was reported and compared to the values re­
which is only approximately 10% of the initial natural gas load [3].
ported in similar studies This SEC value may be higher than that of
Recently developed hydrogen liquefaction plants have specific en­
commercial units, but it is lower than that of some recently proposed
ergy consumption (SEC) values in the range of 10–12 kWh/kgLH2 and
competitive alternative processes. Additionally, the number of equip­
exergy efficiencies in the range of 20%–30%. Theoretically, for a feed at
ment required for this process, including nine multi-stream exchangers,
25 bar, the minimum work required is approximately 2.7 kWh/kgLH2
eight of which serve as catalytic converters, is an indication of the
[6,7]. To allow LH2 to be successful from the perspective of future en­
complexity and high capital cost of the process.
ergy demands, the SEC should be reduced, and the exergy efficiency
Most commercial and/or proposed processes differ significantly from
should be increased. Currently, the estimated liquefaction cost is
one another. Therefore, as pointed out by Berstad et al. [32], such
approximately 40% to 50% of the total investment cost for a new 100
processes cannot be compared directly in terms of energy/exergy effi­
tpd plant [8]. Various conceptual designs have been proposed in recent
ciency or SEC. H2 feed specifications, particularly in the pressure or pre-
years with SEC values as low as 5 to 7 kWh/kgLH2 and exergy efficiencies
compression phases, are a major contributor to the overall exergy and
above 50%. Cardella et al. [9,10] presented an economic analysis of the
energy consumption of any process. The higher the pressure, the higher
scale-up potential of two proposed liquefaction processes, namely a
the efficiency. Wilhelmsen et al. [11] identified the temperature dif­
high-pressure Claude cycle and a dual H2–Neon cycle. It was demon­
ferences between hot and cold layers, and OPC as the two main sources
strated that the SEC could be reduced to slightly above 6 kWh/kgLH2 with
of exergy destruction, contributing up to 69% and 29% to exergy
an increase in plant capacity from 5 to 100 tpd. However, compressors destruction, respectively. Therefore, improving catalyst activity is a
and specially designed heat exchangers with and without embedded plausible option for addressing the issue of exergy destruction. It is
catalysts require higher investment costs. In addition to the capital in­ worth noting that several studies have claimed to achieve exceptional
vestment required, process complexity increases with an increase in the SEC values (≥50% improvement compared to commercial processes).
number of equipment used. However, their process flowsheets reveal extreme complexity. For
Small-to-mid-scale H2 liquefaction plants are often located close to example, Asadnia and Mehrpooya [33] used a combination of an MR-
cryogenic air separation facilities so that liquid nitrogen can be used as a based refrigeration cycle with a Joule–Brayton cycle and six Lin­
refrigerant in the precooling stage. Based on the global requirements for de–Hampson cycles, making an already complex process increasingly
pure oxygen, liquefied nitrogen will not be available as an inexpensive difficult to understand and commercialize.
refrigerant for future large-scale H2 liquefaction plants [11]. The most Therefore, a simple, efficient, and cost-effective process was devel­
energy- and cost-efficient processes proposed in the literature use a oped in this study. A special MR was formulated and employed for H2
mixed refrigerant (MR) instead of liquid nitrogen to precool H2. An MR liquefaction. A strategic cooling methodology not only made the
consists of hydrocarbons, nitrogen, and other components such as neon. developed unit relatively simple but also reduced the SEC. The initial
The composition of an MR can be customized to enable tight thermal design was later optimized and comparatively analyzed thermodynam­
matching in heat exchangers, leading to a higher efficiency in the pre­ ically based on established criteria.
cooling stage [11]. Sadaghiani [12] presented a thermodynamic anal­
ysis of a H2 liquefaction process based on a single MR cycle. An SEC of 2. Proposed hydrogen liquefaction process
7.646 kWh/kgLH2 was reported for a processing capacity of 130 tpd.
In their pilot-scale test rig, Krasae-in et al. [13] did not install or First, a simple design for the H2 liquefaction process was developed.
simulate a conversion reactor; the MR-based process consumed less This design is referred to as base case I. A recently proposed pure H2
energy in the precooling phase (1.76 kWh/kgLH2 ) compared to

2
M.A. Qyyum et al. Energy Conversion and Management 246 (2021) 114659

liquefaction process with outstanding results [33] is referred to as base stream H9 is reduced to the LH2 storage pressure of 1.3 bar by passing it
case II in this paper. The optimized version of base case I is referred to as through the dual-phase expander K-4. At the K-4 outlet, 100% saturated
the proposed process. Categorically, there are two distinct sections of the liquid hydrogen is produced.
proposed H2 liquefaction process: i) the liquefaction of H2 and ii) the The PMR is compressed to 12.5 bar in compressors K-2 and K-3 and
two-stage OPC of H2. cooled to 22 ◦ C by aftercoolers C-1 and C-2. This high-pressure stream is
then cooled to − 153 ◦ C in CHX-1 prior to expansion to 1.5 bar in the
dual-phase expander K-1. The expansion in K-1 further reduces the
2.1. MR cascaded liquefaction process temperature, allowing the PMR to serve as the cold stream P3 in CHX-1
to precool the H2. This cold stream also reduces the temperature of the
The proposed mixed-fluid cascaded (MFC) H2 liquefaction process CMR and LMR hot streams I2 and S2, respectively. The CMR cycle passes
consists of three refrigeration cycles: precooling, cooling, and liquefac­ through a multi-compression system increasing its pressure to 36.2 bar;
tion (saturated liquid), as shown in Fig. 1. it is then cooled to − 153 and − 240 ℃ in CHX-1 and CHX-2, respectively,
The three refrigeration cycles utilize different MRs to perform the prior to expansion to 3.2 bar in K-5. This expanded stream provides the
precooling, cooling, and liquefaction of H2 gas. The refrigerants used in required cold energy for H2 stream cooling and LMR cooling. The cold
each refrigeration cycle are listed below. stream I6 from CHX-2 is returned to the compressor facility to complete
the refrigeration loop. Similar to the other two cycles, the pressure of the
• Precooling MR (PMR) cycle: C1, C2, C3, nC4, iC4, and N2 LMR cycle is increased to 25 bar through two-stage compression facili­
• Cooling MR (CMR) cycle: C1, C2, C3, N2, and H2 tated by aftercoolers. The LMR is then cooled to − 251.8 ◦ C in CHX-1,
• Liquefaction MR (LMR) cycle: H2 and He CHX-2, and CHX-3 prior to expansion to 4.0 bar in K-9. Following the
expansion, the LMR temperature is − 252.8 ℃; thus, it can be utilized to
All the three refrigeration cycles are facilitated by a multi- liquefy the H2 stream and provide cold energy in CHX-1 and CHX-2. The
compression system assisted by phase separators to avoid the presence outlet stream S8 from CHX-1 is then fed into the compression unit to
of any liquid in the compressors and aftercoolers to maintain a complete the LMR refrigeration loop.
compressor outlet temperature of 22 ◦ C. The pressure ratio in each cycle The proposed process was simulated in Aspen HYSYS® v10 using the
is maintained under 3.0. Peng–Robinson (PR) thermodynamic package [34]. PR is a well-known
The hydrogen feed stream n-GH2 is precooled in a cryogenic heat equation of state used to calculate thermodynamic properties. The feed
exchanger CHX-1, where its temperature is reduced to − 153 ◦ C. The conditions and design parameters were selected based on base case II for
precooled stream H2 is then fed into the adiabatic conversion reactor R- a fair comparison, as shown in Table 1. The efficiency of the compressor
1 to convert 25% p-H2 into 33% p-H2. The temperature of the reactor and expander was adjusted to 80% [33]. Base case I is the non-optimized
outlet stream H4 is increased to − 149 ◦ C by the exothermic nature of the version of the proposed process.
OPC reaction. After precooling, stream H4 is cooled to − 240 ◦ C in CHX- The stream conditions of the proposed process are listed in Table S1
2. The CHX-2 outlet stream H5 is then fed into the adiabatic conversion of supplementary material.
reactor R-2 to convert 33% p-H2 into ~92% p-H2. After R-2, stream H7
(at a temperature of − 235 ◦ C) is then passed through CHX-3 for further
temperature reduction. The temperature of the outlet stream H8 is
− 251.8 ◦ C, where H2 is in a liquefied state at 21 bar. The pressure of

Fig. 1. Process flow diagram of the proposed process for H2 liquefaction using MFC two-stage OPC of hydrogen.

3
M.A. Qyyum et al. Energy Conversion and Management 246 (2021) 114659

Table 1
Design parameters for the H2 liquefaction process [33]
Design parameters Proposed Case

Feed temperature 25 ℃
Feed pressure 21 bar
Feed flowrate 1 kg/h
Pressure-drop across aftercoolers 0 bar
Pressure drop across cryogenic heat exchangers 0 bar
Temperature aftercooler 22 ℃
H2 liquefaction rate 100 %
Liquid H2 pressure 1.3 bar

2.2. Two-stage OPC of H2

A H2 molecule exists in two different forms distinguished by the

orientation of its nuclei spin, resulting in slightly dissimilar properties.
At 25 ◦ C, molecular H2 consists of 75% o-H2 with nuclei spin in one
direction and 25% p-H2 with nuclei spin in the opposite direction. This
composition is commonly referred to as normal H2 (Fig. 2). These spin Fig. 3. Ortho-to-para equilibrium concentration and the corresponding heat of
isomers of H2 do not have the same rotational energy and o-H2 is the conversion [39]
excited state with a higher energy level. Therefore, the equilibrium
between the two states is temperature-dependent and shifts toward
100% p-H2 as the temperature decreases to 0 K, as shown in Fig. 3 2M + o − H2 →2MH2 →2M + p − H2 + ΔH
[35,36]. Quantitatively, at 20 K, the enthalpy of OPC is approximately
527 kJ/kg and the heat of vaporization of p-H2 is approximately 447 kJ/ o − H2 →p − H2 + ΔH (1)
kg. This exothermic enthalpy of conversion is an additional duty that where M represents the catalyst in the bed.
enhances the total reversible work by approximately 15% [36,37]. This Commercial simulation software requires reaction kinetics in a spe­
equilibrium shift is of paramount importance from the perspective of cific form that does not exist for OPC in the literature, to the best of our
LH2 storage. knowledge. Therefore, researchers utilize the overall conversion com­
By nature, OPC is a slow reaction in the liquid phase and non-existent bined with composition data to estimate the coefficients of the following
in the gaseous phase. Therefore, rapid cooling and liquefaction will not equation:
affect the compositions significantly. However, if H2 is stored for a long
period without achieving equilibrium, this slow conversion will release Conversion(%) = Co + C1 × T + C2 × T 2 (2)
a sufficient amount of energy to vaporize 50% of the LH2 in 10 days where Co, C1, and C2 are conversion coefficients and T is the H2
[40]. This phenomenon is called boil-off and it begins with an initial rate temperature in Kelvin.
of 1% per hour to 18% per day [41,42]. Boil-off losses can only be The conversion reactor in Aspen HYSYS® is also governed by Eq. (2),
mitigated by converting and storing H2 in the para-form, either before or where the second and third terms are specifically used for temperature-
after liquefaction [13,22]. As OPC is strongly exothermic in the cryo­ sensitive reactions. Table 2 lists the values of the conversion coefficients
genic region (see Fig. 3), it is imperative to promote liquefaction while for each reactor, which were obtained via fitting such that the p-H2
cooling the OPC stream. Commercially, H2 liquefiers operate at a p-H2 percentage and temperature at the reactor outlet coincided with the
fraction above 95% [36]. experimental data reported in [43–45]. A similar approach was adopted
However, the conversion is a second-order reaction (reaction rate = in [17,22,33,46].
0.0114 h− 1), indicating that it takes more than two months to increase
the p-composition from 25% to 95% [40]. Commercially, iron-oxide or
Table 2
nickel-silica are commonly used as catalysts based on their low cost and
Coefficients for percentage conversion to be used in the conversion reactor
good performance in terms of increasing the conversion rate [35]. At the model in Aspen Hysys.
industrial scale, heat exchanger converters are used with a catalyst
Reactor ID Conversion Reactor Coefficients
packed on the H2 side.
Co C1 C2
The following equation describes the chemical reactions inside the
catalytic heat exchanger converters: R-1 66.12 − 0.4125 1.168 × 10-3
R-2 85.35 − 0.3325 1.118 × 10-2

Fig. 2. Schematic of o-H2 and p-H2 [38]

4
M.A. Qyyum et al. Energy Conversion and Management 246 (2021) 114659

3. Process optimization The appropriate selection of an optimization strategy or algorithm

for the rigorous optimization of highly nonlinear and complex processes,
The efficient design and/or operational optimization of plants is such as H2 liquefaction, is critical owing to the uncertainty of identifying
typically achieved by improving energy efficiency, minimizing emis­ global optima. Although several optimization algorithms [48–50] are
sions, and improving the quality of valuable products, while minimizing available in the literature and have shown the potential to optimize
total annualized costs [47]. In this study, the goal of the optimization complicated natural gas liquefaction processes, none of them ensures
process was to identify the optimal design of the proposed H2 lique­ finding a global optimum. The modified coordinate descent (MCD) al­
faction process corresponding to the minimum SEC, which was used as gorithm [51] is the most promising approach for finding the optimal
an objective function. Therefore, it seemed worthwhile to identify the designs of conventional LNG processes. Therefore, several recent studies
key decision (or design) variables and constraints associated with the have investigated the effectiveness of the MCD algorithm for finding
design of the H2 liquefaction process to obtain the desired benefits optimal solutions for various complicated and highly nonlinear LNG
through rigorous optimization. The refrigerant flowrates in the refrig­ processes [52–55], as well as retrofitted natural gas liquid recovery
eration cycles (i.e., precooling, cooling, and liquefaction), refrigerant processes [56]. Therefore, it was worthwhile to evaluate the effective­
condensation pressure, evaporation pressure, precooling temperature, ness of the MCD algorithm for the optimization of the proposed H2
cooling temperature, LMR returning to the precooling exchanger inlet, liquefaction process. To this end, a simulation–optimization framework
and outlet temperature were selected as the key decision variables. The was established by linking Aspen HYSYS® with Microsoft Visual Studio
minimum internal temperature approach (MITA), which represents the using the COM functionality. A detailed flowchart of the MCD algorithm
driving force of the heat transfer between the hot and cold streams inside is presented in Fig. 4. Comprehensive explanations of the MCD approach
the main cryogenic heat exchanger, was adopted as a constraint with a can be found in [47,51].
specified range of 1.0 to 2.0 ◦ C, similar to that used in previous works
[31,33]. Table 3 lists the mathematical formulations of the objective
function, constraints, and decision variables.

Table 3
Objective function, constraints, and decision variables with their lower and upper limits.
)
Objective function: (
∑n W
i
Specific energy consumption (kWh/kgLH2 ) Minimize f(X) = Min.
i=1 mLH2

where,
Wi = WPrecooling + WCooling + WLiquefaction

Constraints: 1.0 < MITA(X)CHX-1 < 2.0

Minimum internal approach temperature (◦ C) 1.0 < MITA(X)CHX-2 < 2.0
1.0 < MITA(X)CHX-3 < 2.0
and,
Xlower < X < Xupper
where X is a vector of the decision variables.
Decision Variables Lower limit Upper limit
PMR cycle
mN2 (kg/h) 0.1 5.0
mC1 (kg/h) 0.1 20.0
mC2 (kg/h) 0.1 25.0
mC3 (kg/h) 0.1 35.0
mnC4 (kg/h) 0.1 55.0
miC4 (kg/h) 0.1 55.0
Suction pressure (stream-P3), PP3 (bar) 1.1 3.0
Discharge pressure (stream-P10-V), PP10-V (bar) 8.0 25.0
Precooling temperature (stream-H2), TH2 (◦ C) –165.0 –135.0
CMR cycle
mH2 (kg/h) 0.1 5.0
mN2 (kg/h) 0.1 10.0
mC1 (kg/h) 0.1 5.0
mC2 (kg/h) 0.1 5.0
mC3 (kg/h) 0.1 45.0
Suction pressure (stream-I4), PI4 (bar) 1.1 4.5
Discharge pressure (stream-I14-V), PI14-V (bar) 20.0 45.0
Cooling temperature (stream-H5), TH5 (◦ C) –245.0 –200.0
LMR cycle
mH2 (kg/h) 0.1 5.0
mHe (kg/h) 0.1 3.0
Suction pressure (stream-S5), PS5 (bar) 1.1 5.5
Discharge pressure (stream-S11), PS11 (bar) 15.0 40.0
LMR temperature-1 (stream-S7), TS7 (◦ C) –170.0 –130.0
LMR temperature-2 (stream-S8), TS8 (◦ C) –50.0 20.0

5
M.A. Qyyum et al. Energy Conversion and Management 246 (2021) 114659

Fig. 4. Flowchart of the MCD algorithm [47,55]

4. Results and discussion: Process analysis [33], respectively. Such a low energy consumption is attributed to the
optimal selection of refrigerants, their compositions, and compressor
4.1. Design variable analysis pressure ratios. In the PMR cycle, the combination of high-boiling re­
frigerants, such as i-C4 and n-C4, and low-boiling refrigerants, such as
The design variable analysis of a liquefaction process is typically N2, provides an optimal mixture for precooling the H2 to − 153.2 ◦ C. To
performed based on the SEC. Table 4 lists the refrigerant-specific mass achieve this temperature, high-boiling refrigerants alone are insuffi­
flowrates for base case I, base case II [33], and the proposed case. The cient, regardless of how their flowrates are adjusted while maintaining
overall refrigerant flowrate in base case II [33] is 60.5% less than that in the pressure above atmospheric pressure. In contrast, a large quantity of
the proposed case. However, base case II [33] utilizes a mixture of low-boiling refrigerants can achieve the specified temperature, but at
conventional and unconventional refrigerants and large quantities of H2 the cost of a high SEC. For example, approximately 200 kg/h of N2 is
and He (78.2% more than the proposed case). Additionally, as shown in required in the PMR cycle to achieve the specified temperature at an
Table 5, the SEC is minimized (6.45 kWh/kgLH2 ) in the proposed case overall SEC of 15 kWh/kg according to the MITA. Therefore, an optimal
with energy savings of 36.5% and 16.1% compared to base cases I and II combination of high- and low-boiling refrigerants is required to enhance

6
M.A. Qyyum et al. Energy Conversion and Management 246 (2021) 114659

Table 4 approximately 77% of the total MR flowrate. In this cycle, C3 is a high-

Comparison of refrigerant mass flowrates for base case II, base case I, and the boiling refrigerant with a critical temperature of 97 ◦ C, whereas H2 is a
proposed case. low-boiling refrigerant with a critical temperature of − 239.9 ◦ C. In
Refrigerants Base case II [33] Base case I Proposed case general, the performance of a refrigeration cycle can also be determined
Precooling (PMR)
through the specific refrigeration effect (SRE). It is defined as the
C1 1.56 25.80 14.00 amount of cooling provided by the system/components; it has higher
C2 3.49 27.36 21.81 values for components with a critical temperature close to the ambient
C3 1.37 24.90 4.50 temperature. Therefore, a large C3 quantity is consumed in this cycle,
i-C4 0.69 11.41 15.25
resulting in a high SRE. Additionally, in the CMR cycle, the discharge
n-C4 0.69 22.00 28.36
n-C5 9.56 – – pressure decreases, and the suction pressure increases after optimiza­
C3H6 3.37 – – tion. A low suction pressure increases the refrigeration load of the PMR
R-14 8.20 – – cycle, thereby increasing the SEC; in contrast, the desired cooling tem­
NH3 0.49 – – perature cannot be achieved using a high suction pressure. If the cooling
N2 2.63 5.75 0.78
H2 1.33 – –
temperature is increased to accommodate a high suction pressure, then
Total 33.42 117.22 84.70 the energy load of the LMR cycle increases. Therefore, the MR flowrates
Cooling (CMR) and pressure ratio are adjusted optimally to achieve a low SEC. Simi­
C1 – 5.70 1.12 larly, the LMR cycle is utilized to achieve the required H2 liquefaction
C2 12.00 3.60
–
temperature. In this cycle, H2 and He are used as refrigerants with a total
C3 – 14.00 25.36
N2 – 8.00 1.46 mass flowrate of 1.90 kg/h in the proposed case. Compared to base case
H2 4.17 1.65 1.48 I, the H2 and He flowrates are reduced by 22.2% and 90.4%, respectively
He 9.98 – – in the proposed case. This reduction in mass flowrate reduces the SEC,
Total 14.16 41.35 33.02 maintaining the MITA within suitable limits.
Liquefaction (LMR)
H2 – 2.25 1.75
He – 1.56 0.15 4.2. Composite curve analysis
Total 3.81 1.90
Total (PMR þ CMR þ LMR) 47.20 162.38 119.62 Composite curve analyses present a holistic overview of fluid
behavior within a multi-stream CHX. The curves that represent the re­
lationships of temperature changes between hot and cold streams with
Table 5 respect to heat flow along the length of the CHX are known as temper­
Design variables and SEC of base case I and the proposed case (optimization ature–heat flow composite curves (THCCs). On the other hand, com­
results). posite curves that relate temperature with the approach temperature
Design Variables Base case I Proposed case MITA are known as temperature difference composite curves (TDCCs).
PMR cycle
Because base case I and the proposed case use three CHXs, Fig. 5, Fig. 6,
Suction pressure (stream-P3), PP3 (bar) 2.50 1.52 and Fig. 7 present the THCCs of all the three CHXs in both base case I and
Discharge pressure (stream-P10-V), PP10-V (bar) 19.50 13.00 the proposed case. Similarly, Fig. 8 presents the TDCCs of base case I and
Pressure ratio 2.793 2.924 the proposed case with respect to their CHXs. Base case II [33] has
Precooling temperature (stream-H2), TH2 (◦ C) − 153.20 − 153.20
different process configurations with different precooling temperatures.
CMR cycle
Suction pressure (stream-I4), PI4 (bar) 2.50 2.98 Therefore, this case [33] cannot be compared to base case I and the
Discharge pressure (stream-I14-V), PI14-V (bar) 41.70 35.60 proposed case from the perspectives of THCCs and TDCCs.
Pressure ratio 2.560 2.29 In Fig. 5 (a), base case I exhibits a wide gap between the hot and cold
Cooling temperature (stream-H5), TH5 (◦ C) − 240.40 − 240.40
composites, particularly above − 140 ◦ C, whereas the proposed case
LMR cycle
Suction pressure (stream-S5), PS5 (bar) 3.80 3.75
exhibits a narrow gap between these curves, particularly in the middle
Discharge pressure (stream-S11), PS11 (bar) 20.00 25.00 and at the lower end, as shown in Fig. 5 (b). The upper end, which is the
Pressure ratio 2.294 2.582 hot end of CHX-1, exhibits a wide gap in both cases (encircled region).
LMR temperature-1 (stream-S7), TS7 (◦ C) − 165.00 − 158.00 The gap between the curves is the pinch point or MITA value, which
LMR temperature-2 (stream-S8), TS8 (◦ C) − 15.00 − 32.00
should conform to the desired value (1 to 2 ℃) along the length of the
Constraints
MITA(X)CHX-1 (◦ C) 3.107 1.085 CHX. The wide gap between these curves signifies further potential for
MITA(X)CHX-2 (◦ C) 3.067 1.063 absorbing heat. This can be accomplished either by choosing re­
MITA(X)CHX-3 (◦ C) 3.329 1.004 frigerants whose boiling and critical points enable them to absorb more
Total compression power required (kW) 10.810 7.027
heat or by selecting an optimal refrigerant composition. For example,
Power generated (kW) 0.660 0.577
Net power generated (kW) 10.15 6.450
Fig. 5 (a) reveals a wide gap at the hot end of CHX-1 with a sharp turning
SEC (kWh/kgLH2 ) 10.15 6.450 point in the cold composite at approximately − 20 ℃. This turning point
is caused by the change of phase at this temperature and is primarily
influenced by n-C4 and i-C4. The cold stream P3 enters CHX-1 with a
the energy efficiency of the process. In Table 4, in the proposed PMR 97% liquid phase and is converted into the vapor phase at approxi­
cycle, 84.70 kg/h of MR is consumed, which primarily consists of n-C4, mately − 20 ◦ C, which is the main cause of this turning point. This
followed by C2 and C1. In addition, the flowrates of all the PMR cycle turning point can be managed by increasing or decreasing the ratios of n-
refrigerants decrease following optimization, excluding C4, whose C4 and i-C4, as shown in the proposed case for CHX-1. The n-C4 ratio is
flowrate increases. This is because low-boiling refrigerants alone are not increased in the proposed PMR cycle, which reverses the divergent
sufficiently thermodynamically efficient to provide refrigeration ca­ behavior of the curves.
pacity in this temperature range. Therefore, a high-boiling refrigerant, Similarly, Fig. 6 (a) and (b) present the THCCs for CHX-2. A wide gap
namely n-C4, contributes the highest flowrate to improve energy between the hot and cold composites is prominent in the middle region
efficiency. between − 180 and − 215 ◦ C (encircled). This gap is significantly
In the CMR cycle, an optimal mixture of C1, C2, C3, H2, and N2 is reduced in the proposed case. This region is mainly influenced by N2 and
utilized to cool H2 to − 240.4 ◦ C. Here, again, the total MR flowrate H2. Increasing the flowrate of N2 and H2 significantly affects this region
decreases by 20% after optimization. C3 alone contributes and reduces the temperature of the hot composite to below that of the

7
M.A. Qyyum et al. Energy Conversion and Management 246 (2021) 114659

Fig. 5. CHX-1 THCCs for the (a) base case I and (b) proposed case.

Fig. 6. CHX-2 THCCs of the (a) base case I and (b) proposed case.

cold composite, leading to thermodynamic infeasibility. Additionally, recovery and reduce entropy generation. A wide gap between these
the hot end of CHX-2 is influenced by C3 such that, by increasing the C3 curves leads to a high entropy generation and high exergy destruction
flowrate, the gap between the curves decreases along with the associated based on large temperature differences.
heat flow. A large heat flow induces greater compressor duties, which In Fig. 8, the TDCCs of all the three CHXs are presented for base case I
increases the SEC. Therefore, an optimal composition of high- and low- and the proposed case. The dark-colored curves represent the TDCCs of
boiling components should be selected to manage the gaps between the the hot streams in the proposed case, whereas the light-colored curves
curves. represent the TDCCs of the hot streams in base case I. The heights of
The THCCs of base case I and the proposed case for CHX-3 are pre­ these curves should be within the limit of 1 to 2 ◦ C. The higher the curve,
sented in Fig. 7 (a) and (b), respectively. The two wide gaps (encircled the higher is the entropy generation. The CHX-1 curves exhibit the
regions) between the hot and cold composites can be reduced by opti­ maximum MITA of 37.6 ◦ C at the hot end of the proposed case, whereas
mizing the design variables of the LMR flowrates and compressor pres­ for base case I, this MITA value is relatively low (approximately 32 ℃).
sure ratios. The hot end of CHX-3 is significantly affected by the H2 This difference in MITA values occurs because the LMR stream I15 en­
flowrate. Increasing the flowrate of H2 eliminates the cold composite ters CHX-1 at − 105.3 ◦ C in base case I and − 108.3 ◦ C in the proposed
turning point and smoothens the curve. However, it widens the gap, case. The higher the temperature difference between hot and cold
which leads to a larger MITA value and higher entropy generation. streams, the higher the MITA value. The temperature differences be­
Similarly, He affects the cold end of the THCCs such that, by decreasing tween streams n-GH2, P12, S1, and I15 induce a positive change in the
the He flowrate, the gap at the cold end decreases. However, a large MITA value at the hot end of CHX-1. For example, by increasing the
decrease in the He flowrate increases the gap at the hot end. Therefore, temperature of the stream I15, the MITA value decreases and vice versa.
this tradeoff should be optimized to not only narrow the gap at both ends Additionally, as the temperature inside CHX-1 decreases, the MITA
but also reduce the SEC. In summary, the gap between the hot and cold value decreases (lower in the proposed case compared to base case I).
composites should be within the desired limits to ensure efficient heat The middle and cold ends of CHX-1 in the proposed case (below − 30 ℃)

8
M.A. Qyyum et al. Energy Conversion and Management 246 (2021) 114659

Fig. 7. CHX-3 THCCs of the (a) base case I and (b) proposed case.

Fig. 8. TDCCs of base case I and the proposed case.

exhibit MITA values between 1 and 10 ℃, demonstrating better heat (− 153.2 ◦ C). This increase in temperature is mainly attributed to the
recovery compared to base case I. The hot end exhibits a potential for H2 stream, whose temperature is increased by the exothermic OPC. A
improvement through the rigorous optimization of refrigerant compo­ similar phenomenon occurs in reactor R-2 between CHX-2 and CHX-3. In
sitions or the introduction of a refrigerant that enhances heat recovery in CHX-3 in the proposed case, the MITA value at the hot end is at its peak
this region. of 12.3 ◦ C, and it sharply decreases along the length of the exchanger.
The CHX-2 TDCCs in the proposed case exhibit a better performance This MITA peak is caused by an abrupt change in temperature based on a
than the CHX-1 TDCCs, where the MITA value lies within the range of phase change in the middle region. This phase change can be adjusted by
1.1–9.9 ◦ C, as shown in the enlarged region. The highest peak lies in the tuning the suction or discharge pressure of the LMR cycle. The lowest
middle region of − 180 to − 215 ℃. Compared to base case I, this MITA MITA values at the cold end in base case I and the proposed case are 3.33
peak is higher, which can be attributed to the C3 flowrate and CMR cycle and 1.00 ◦ C, respectively. These values are attributed to the temperature
suction pressure. Increasing the C3 flowrate and CMR suction pressure gaps between the streams at the cold ends, which can be adjusted by
increases the MITA peak; however, the SEC also increases. The optimal tuning the suction pressure of the LMR cycle.
selection of the pressure and refrigerant flowrates can reduce both the Fig. 9 (a) and (b) present the temperature–heat flow (TH) profiles
SEC and MITA values. Furthermore, the hot-end temperature of CHX-2 (aka grand composite curve) of the H2 feed stream for base case II [33]
(− 149 ◦ C) is higher than the cold-end temperature of CHX-1 and the proposed case. These H2 TH curves represent the temperature

9
M.A. Qyyum et al. Energy Conversion and Management 246 (2021) 114659

Fig. 9. Hydrogen temperature–heat flow composite curves for the (a) base case II [33] and (b) proposed case.

and heat flow changes in the H2 streams. Here, the sloped lines represent
Table 7
the H2 TH profiles within the CHXs and the approximately horizontal
Equipment SExD values for base case I, base case II [33], and the proposed case.
lines represent the profiles in the adiabatic conversion reactors. This
heat flow is influenced by stream conditions, particularly the tempera­ Equipment Base case I Base case II Proposed case
(kWh/kg) (kWh/kg) (kWh/kg)
ture across the CHXs. For example, under the same feed conditions, base
case II [33] yields a heat flow of 0.28 kW, whereas the proposed case Compressors 2.06 1.24 1.73
Pumps 0.05 0.001 0.03
yields a heat flow of 0.70 kW. This difference in heat flow occurs because
After coolers 0.90 0.75 0.90
of the temperature differences between the inlet and outlet H2 (71.43 ℃ Phase separators 2.42 0.28 2.05
in CHX-I of base case II [33] and 178.2 ℃ in the proposed case). In the Reactors 0.42 – 0.42
proposed case, by changing the temperature of the precooling H2 from Expander 1.71 1.54 1.56
− 153.2 to − 130 ◦ C, the heat flow in CHX-1 changes from 0.703 to 0.611 Cryogenic 1.71 1.26 − 0.56
exchangers
kW while in CHX-2, the heat flow increases from 0.47 kW to 0.57 kW. Total 9.27 5.11 6.45
The heat flow of the stream increases based on the exothermic OPC in
the adiabatic reactor, which also leads to an increase in temperature
from − 240 ◦ C to − 235 ◦ C resulting in p-H2 concentration of ≃92%. The
Table 8
commercial processes, however, operate at around − 243 ◦ C with a p-H2
SExDs of CHXs for base case I and the proposed case.
concentration of ≥95%. In the present study, the inlet temperature of R-
2 has been maintained at − 240 ◦ C to avoid potential freezing of hy­ Equipment Base case I (kWh/kg) Proposed case (kWh/kg)

drocarbons (C1, C2, C3) in the CMR cycle. In order to increase the p-H2 CHX-1 2.93 1.83
concentration beyond this point requires a new optimal composition of CHX-2 − 2.19 − 2.79
CHX-3 0.96 0.40
MR, majorly comprised of low boiling refrigerants such as H2, He and
Total 1.71 − 0.56
Ne. Another probable solution is to consider an isothermal reactor
placed after CHX-3 instead of the adiabatic reactor (R-2). In base case II
[33], the increase in heat flow is most prominent after CHX-3 because of 4.3. Exergy analysis and thermodynamic evaluation
the large temperature difference of 90.89 ℃ in CHX-4. Additionally, in
base case II [33], reactor R-2 provides no prominent contributions to the This section compares base case I, base case II [33], and the proposed
heat flow and temperature because the OPC is low (2.903%). A high OPC case based on exergy analysis and thermodynamic efficiency. Thermo­
is predominant below − 143.15 ℃, as shown in Fig. 3. Therefore, the dynamic efficiency is evaluated based on the coefficients of performance
reactor should be placed based on the ortho-to-para equilibrium profile (COPs) and the figures of merit (FOMs).
of hydrogen. Fig. 9 also illustrates the impact of OPC on the heat flow.
4.3.1. Exergy analysis
Exergy analysis is an important thermodynamic tool for assessing
thermal systems. Unlike energy analysis, exergy analysis can help us
identify and locate the sources of thermodynamic inefficiency. Exergy
Table 6
Exergy destruction equations for the equipment.
Equipment Exergy destruction (kW) Table 9
Overall process exergy efficiency for base case I, base case II, and the proposed
Compressor Ex = ṁ(Exin − Exout ) − Ẇ
case.
Aftercoolers Ex = ṁ(Exin − Exout )
Base Case- Base Case-II Proposed
LNG exchanger ∑
Ex = (ṁ)Exin −
∑
˙
(ṁ)Ex
˙ out
I [33] case
Turbine Ex = ṁ(Exin − Exout ) + Ẇ
Overall process exergy efficiency 29.9 39.5 47.0
Phase separators / Reactors Ex = ṁ(Exin − Exout ) (%)

10
M.A. Qyyum et al. Energy Conversion and Management 246 (2021) 114659

can be physical, chemical, kinetic, or potential. Considering a system at configuration and optimal refrigerant composition, which generate
rest relative to the environment, the kinetic and potential exergies are lower entropy. Additionally, the process of base case II [33] consumes
neglected. The physical and chemical exergies of a stream are calculated 16% more energy than that of the proposed case, which is the main
as reason for its low exergy efficiency. Such a high energy consumption
mainly occurs in the low-temperature section, where high percentages
exph,i = (hi − ho ) − To (si − so ) (3)
of H2 and He are used as refrigerants. Consequently, the low-
∑ ∑ temperature section induces a high exergy destruction, particularly in
exch,i = xi eCH + RTo xi lnxi (4)
i the expanders (1.47 kWh/kg) and aftercoolers (0.64 kWh/kg) [33].
where exph,i and exch,i denote the physical and chemical exergies, Additional evaluations of the base cases and proposed case in terms of
respectively, xi denotes the mole fraction of the ith component in the refrigeration cycles are presented in Fig. 10. The proposed case exhibits
stream, and eCH represents the standard chemical exergy of the ith a higher SED in the precooling section, but a lower SED in the cooling
i
component taken from the literature [57]. In this study, the physical section compared to base case II [33]. The exergy destruction in the
exergy values were derived from Aspen HYSYS® stream properties and precooling section of the proposed case is primarily induced by CHX-1
the chemical exergy values were calculated using Eq. (5). Exergy (1.83 kW), followed by the compressors (0.80 kW). The main reason
destruction for the equipment involved is calculated using equations for such a high exergy destruction is the large temperature difference
from [58], which are shown in Table 6. between the inlet and outlet streams. Additionally, CHX-1 facilitates all
The overall process exergy efficiency is calculated as the three refrigeration cycles, unlike CHX-2 and CHX-3, which
( ) ( ) contribute to the overall exergy destruction. Furthermore, the precool­
Exproduct − ExFeed ing temperature of base case II [33] is − 107.3 ℃, which is higher than
μex = (5)
Wtotal the precooling temperature of base case I and the proposed case (− 153.2
℃). A lower precooling temperature compared to that used in base case
The specific exergy destruction (SExD) values of each equipment
II [33] requires more refrigeration, which ultimately increases the
were calculated and compared between base case I, base case II [33],
exergy destruction. In base case II [33], the high SED values in the
and the proposed case, as shown in Table 7.
cooling section are mainly attributed to the high energy loads and large
As presented in Table 7, the overall SExD in the CHXs of the proposed
amounts of low-boiling refrigerants (i.e., H2 and He consumption). For
case is calculated as − 0.56 kWh/kg. The SExDs of the CHXs for base case
example, base case II [33] consumes 68% more energy in the cooling
I and the proposed case are listed in Table 8.
section and requires 73.6% higher H2 and He refrigerant flowrate
The exergy destruction of a piece of equipment cannot be negative,
compared to the CMR and LMR cycles of the proposed case.
according to its definition. However, as per the authors’ understanding,
this negative value for CHX-2 might be because of the impact of
cascaded configuration of refrigeration cycles. Furthermore, Table 7 4.3.1.1. Thermodynamic efficiency. The thermodynamic efficiency of
indicates that the proposed case exhibits a lower exergy destruction than the proposed process was evaluated based on the COP and FOM. The
the base case I, but a higher exergy destruction than the base case II [33]. COP of a liquefaction process is defined as the ratio of the heat removed
This is because base case II [33] does not consider the impact of from the system to the work required to remove that energy. It describes
chemical exergy in its calculations. Furthermore, base case II [33] does the efficiency of a thermal system and is calculated as
not include the exergy destruction of the conversion reactors. Therefore, ṁ × (hGH2 − hLH 2 )
fair comparisons between the base cases and the proposed case can only COP = (6)
Wtotal
be drawn with respect to the overall exergy efficiency, as shown in
Table 9. where hGH2 and hLH2 are the enthalpies of H2 in the gaseous and
Table 9 indicates that the proposed case is significantly more effi­ liquid states, respectively, ṁ is the mass flowrate of H2, and Wtotal is the
cient than base case I and base case II [33] in terms of exergy. Such a amount of work required to convert gaseous H2 into LH2.
high exergy efficiency is mainly attributed to the simple process The FOM is defined as the COP of the actual liquefaction cycle
compared to that of the ideal case (Carnot cycle). Therefore, the FOM
uses the actual available energy (exergy) to evaluate the process per­

Fig. 10. Exergy destruction in the refrigeration cycles of base case I, base case Fig. 11. Thermodynamic evaluation of base case I, base case II [33], and the
II [33], and the proposed case. proposed case.

11
M.A. Qyyum et al. Energy Conversion and Management 246 (2021) 114659

formance. The FOM can also be defined as the ratio of the minimum Acknowledgment
required work to the actual work provided. It is calculated as
This work was supported by Priority Research Centers Program
ṁ × (ExLH 2 − ExGH2 )
FOM = (7) through the National Research Foundation of Korea (NRF) funded by the
Wtotal
Ministry of Education (2014R1A6A1031189) and by “Human Resources
where ExLH2 and ExGH2 are the exergy values of gaseous and liquid Program in Energy Technology” of the Korea Institute of Energy Tech­
H2, respectively, and ṁ is the mass flowrate of H2. Fig. 11 compares base nology Evaluation and Planning (KETEP), granted financial resource
case I and the proposed case to base case II [33] in terms of their COP from the Ministry of Trade, Industry & Energy, Republic of Korea. (No.
and FOM. 20204010600100)
The improved COP and FOM values of the proposed case demon­
strate that the proposed process is more energy-efficient than base case I Appendix A. Supplementary data
and base case II [33], mainly because of its low energy consumption and
high exergy efficiency. Supplementary data to this article can be found online at https://doi.
org/10.1016/j.enconman.2021.114659.
5. Conclusions
References
Considering the process complexity and energy intensiveness of H2
liquefaction, an efficient, simple, and compact liquefaction process was [1] Zheng J, Chen L, Wang J, Xi X, Zhu H, Zhou Y, et al. Thermodynamic analysis and
proposed in this paper. The proposed process utilized three MFC comparison of four insulation schemes for liquid hydrogen storage tank. Energy
Convers Manag 2019;186:526–34. https://doi.org/10.1016/j.
refrigeration cycles to precool, cool, and liquefy H2 via two-stage OPC. enconman.2019.02.073.
This unique process configuration was optimized using the well-known [2] Abdalla AM, Hossain S, Nisfindy OB, Azad AT, Dawood M, Azad AK. Hydrogen
MCD methodology to achieve the maximum thermodynamic potential. production, storage, transportation and key challenges with applications: A review.
Energy Convers Manag 2018;165:602–27. https://doi.org/10.1016/j.
The following conclusions can be drawn from this study.
enconman.2018.03.088.
[3] International Energy Agency (IEA). The Future of Hydrogen - Seizing today’s
• The significant energy savings of the proposed process (36.5%) opportunities. 2019. 10.1787/1e0514c4-en.
[4] Wolf J. Liquid hydrogen technology for vehicles. In: Vielstich W, Gasteiger HA,
indicate that it exhibits a superior performance compared to base
Lamm A, Yokokawa H, editors. Handb. Fuel Cells, John Wiley & Sons, Ltd; 2010.
case I. Such a high energy efficiency is mainly attributed to the se­ 10.1002/9780470974001.f302008.
lection of optimal refrigerant compositions. [5] Seo SK, Yun DY, Lee CJ. Design and optimization of a hydrogen supply chain using
• Additionally, the proposed process exhibits a lower energy con­ a centralized storage model. Appl Energy 2020;262:114452. https://doi.org/
10.1016/j.apenergy.2019.114452.
sumption (6.45 kWh/kgLH2 ) compared to base case II, representing an [6] Stolzenburg K, Berstad D, Decker L, Elliott A, Haberstroh C, Hatto C, et al. Efficient
energy saving of 16.12%. Liquefaction of Hydrogen: Results of the IDEALHY Project 2013:7–9.
• Further thermodynamic evaluations of the proposed process [7] Aasadnia M, Mehrpooya M. Large-scale liquid hydrogen production methods and
approaches : A review. Appl Energy 2018;212:57–83. https://doi.org/10.1016/j.
revealed a higher exergy efficiency (47.2%) compared to base cases I apenergy.2017.12.033.
and II. The liquefaction cycle exhibited major improvements in terms [8] Andersson J, Grönkvist S. Large-scale storage of hydrogen. Int J Hydrogen Energy
of exergy, mainly based on the optimal refrigerant flowrates in the 2019. https://doi.org/10.1016/j.ijhydene.2019.03.063.
[9] Cardella U, Decker L, Klein H. ScienceDirect Roadmap to economically viable
precooling and cooling cycles. hydrogen liquefaction. Int J Hydrogen Energy 2017;42:13329–38. https://doi.org/
• The high COP and FOM values demonstrate the superior thermody­ 10.1016/j.ijhydene.2017.01.068.
namic performance of the proposed case compared to base cases I [10] Cardella U, Decker L, Sundberg J, Klein H. ScienceDirect Process optimization for
large-scale hydrogen liquefaction. Int J Hydrogen Energy 2017;42:12339–54.
and II.
https://doi.org/10.1016/j.ijhydene.2017.03.167.
[11] Wilhelmsen Ø, Berstad D, Aasen A, Nekså P, Skaugen G. Reducing the exergy
The only limitation of the process is its slightly lower p-H2 concen­ destruction in the cryogenic heat exchangers of hydrogen liquefaction processes.
Int J Hydrogen Energy 2018;43:5033–47. https://doi.org/10.1016/J.
tration (i.e., 92%) compared to commercial processes (i.e., 95%). which
IJHYDENE.2018.01.094.
may be overcome by employing low-boiling refrigerants (H2, He and Ne) [12] Sadaghiani MS, Mehrpooya M, Ansarinasab H. ScienceDirect Process development
in the CMR cycle. The proposed process can be further enhanced by and exergy cost sensitivity analysis of a novel hydrogen liquefaction process. Int J
finding a new optimal design (decision) variables using any evolu­ Hydrogen Energy 2017;42:29797–819. https://doi.org/10.1016/j.
ijhydene.2017.10.124.
tionary or metaheuristic algorithm with or without changes in the pro­ [13] Krasae-in S, Bredesen AM, Stang JH, Neksa P. Simulation and experiment of a
cess configuration. hydrogen liquefaction test rig using a multi-component refrigerant refrigeration
system. Int J Hydrogen Energy 2010;36:907–19. https://doi.org/10.1016/j.
ijhydene.2010.09.005.
CRediT authorship contribution statement [14] Krasae-in S, Stang JH, Neksa P. Simulation on a proposed large-scale liquid
hydrogen plant using a multi-component refrigerant refrigeration system. Int J
Muhammad Abdul Qyyum: Conceptualization, Investigation, Hydrogen Energy 2010;35:12531–44. https://doi.org/10.1016/j.
ijhydene.2010.08.062.
Methodology, Software, Writing - original draft. Amjad Riaz: Investi­ [15] Krasae-in S. ScienceDirect Optimal operation of a large-scale liquid hydrogen plant
gation, Writing - original draft, Writing - review & editing. Ahmad utilizing mixed fluid refrigeration system. Int J Hydrogen Energy 2014;39:
Naquash: Formal analysis, Resources, Writing - original draft, Writing - 7015–29. https://doi.org/10.1016/j.ijhydene.2014.02.046.
[16] Aasadnia M, Mehrpooya M. Conceptual design and analysis of a novel process for
review & editing. Junaid Haider: Writing - review & editing. Kinza
hydrogen liquefaction assisted by absorption precooling system. J Clean Prod
Qadeer: Formal analysis, Software. Alam Nawaz: Formal analysis, 2018;205:565–88. https://doi.org/10.1016/j.jclepro.2018.09.001.
Visualization. Hyunhee Lee: Writing - review & editing. Moonyong [17] Sadaghiani MS, Mehrpooya M. ScienceDirect Introducing and energy analysis of a
novel cryogenic hydrogen liquefaction process configuration. Int J Hydrogen
Lee: Supervision, Funding acquisition, Writing - review & editing.
Energy 2017;42:6033–50. https://doi.org/10.1016/j.ijhydene.2017.01.136.
[18] Emre Y, Ozturk M, Dincer I. ScienceDirect Analysis and assessment of a novel
Declaration of Competing Interest hydrogen liquefaction process. Int J Hydrogen Energy 2017;42:11429–38. https://
doi.org/10.1016/j.ijhydene.2017.03.064.
[19] Yilmaz C. A case study : Exergoeconomic analysis and genetic algorithm
The authors declare that they have no known competing financial optimization of performance of a hydrogen liquefaction cycle assisted by
interests or personal relationships that could have appeared to influence geothermal absorption precooling cycle. Renew Energy 2018;128:68–80. https://
the work reported in this paper. doi.org/10.1016/j.renene.2018.05.063.
[20] Kanoglu M, Yilmaz C, Abusoglu A. ScienceDirect Geothermal energy use in
absorption precooling for Claude hydrogen liquefaction cycle. Int J Hydrogen
Energy 2016;41:11185–200. https://doi.org/10.1016/j.ijhydene.2016.04.068.

12
M.A. Qyyum et al. Energy Conversion and Management 246 (2021) 114659

[21] Yilmaz C. ScienceDirect Optimum energy evaluation and life cycle cost assessment [40] Sherif SA, Goswami DY, Stefanakos EK, Steinfeld A. Handbook of Hydrogen
of a hydrogen liquefaction system assisted by geothermal energy. Int J Hydrogen Energy. Taylor & Francis; 2014.
Energy 2019. https://doi.org/10.1016/j.ijhydene.2019.03.105. [41] Weitzel DH, Loebenstein WV, Draper JW, Park OE. Ortho-para catalysis in liquid-
[22] Ghorbani B, Mehrpooya M, Aasadnia M, Shariati M. Hydrogen liquefaction process hydrogen production. J Res Natl Bur Stand 1934;1958(60):221–7. https://doi.org/
using solar energy and organic Rankine cycle power system. J Clean Prod 2019; 10.6028/jres.060.026.
235:1465–82. https://doi.org/10.1016/j.jclepro.2019.06.227. [42] Baker CR, Shaner RL. A study of the efficiency of hydrogen liquefaction 1978;3:
[23] Yilmaz F, Ozturk M, Selbas R. Design and thermodynamic analysis of coal- 321–34.
gasification assisted multigeneration system with hydrogen production and [43] Woolley HW, Scott RB, Brickwedde FG. Compilation of thermal properties of
liquefaction. Energy Convers Manag 2019;186:229–40. https://doi.org/10.1016/j. hydrogen in its various isotopic and ortho-para modifications. J Res Natl Bur Stand
enconman.2019.02.053. 1934;1948(41):379–475. https://doi.org/10.6028/jres.041.037.
[24] Emre Y, Ozturk M, Dincer I. Energy and exergy analyses of an integrated system [44] Scott RB, Denton WH, Nichollls CM, editors. Technology and Uses of Liquid
using waste material gasi fi cation for hydrogen production and liquefaction. Hydrogen. London: Pergamon Press; 1964.
Energy Convers Manag 2019;185:718–29. https://doi.org/10.1016/j. [45] Harkness RW, Deming WE. The Equilibrium of Para and Ortho Hydrogen. J Am
enconman.2019.02.033. Chem Soc 1932;54:2850–2. https://doi.org/10.1021/ja01346a503.
[25] Yang JH, Yoon Y, Ryu M, An SK, Shin J, Lee CJ. Integrated hydrogen liquefaction [46] Hammad A, Dincer I. Analysis and assessment of an advanced hydrogen
process with steam methane reforming by using liquefied natural gas cooling liquefaction system. Int J Hydrogen Energy 2018. https://doi.org/10.1016/j.
system. Appl Energy 2019;255:113840. https://doi.org/10.1016/j. ijhydene.2017.10.158.
apenergy.2019.113840. [47] Qyyum MA, He T, Qadeer K, Mao N, Lee S, Lee M. Dual-effect single-mixed
[26] Riaz A, Abdul Qyyum M, Min S, Lee S, Lee M. Performance improvement potential refrigeration cycle: An innovative alternative process for energy-efficient and cost-
of harnessing LNG regasification for hydrogen liquefaction process: Energy and effective natural gas liquefaction. Appl Energy 2020;268:115022. https://doi.org/
exergy perspectives. Appl. Energy 2021;301. https://doi.org/10.1016/j. 10.1016/j.apenergy.2020.115022.
apenergy.2021.117471. [48] Qyyum MA, Qadeer K, Lee M. Comprehensive Review of the Design Optimization
[27] Seyam S, Dincer I, Agelin-Chaab M. Development of a clean power plant integrated of Natural Gas Liquefaction Processes: Current Status and Perspectives. Ind Eng
with a solar farm for a sustainable community. Energy Convers Manag 2020;225: Chem Res 2018;57:5819–44. https://doi.org/10.1021/acs.iecr.7b03630.
113434. https://doi.org/10.1016/j.enconman.2020.113434. [49] Qyyum MA, Chaniago YD, Ali W, Saulat H, Lee M. Membrane-assisted removal of
[28] Ansarinasab H, Mehrpooya M, Mohammadi A. Advanced exergy and hydrogen and nitrogen from synthetic natural gas for energy-efficient liquefaction.
exergoeconomic analyses of a hydrogen liquefaction plant equipped with mixed Energies 2020;13. https://doi.org/10.3390/en13195023.
refrigerant system 2017;144:248–59. [50] Nawaz A, Qyyum MA, Qadeer K, Khan MS, Ahmad A, Lee S, et al. Optimization of
[29] Yilmaz C, Kaska O. ScienceDirect Performance analysis and optimization of a mixed fluid cascade LNG process using a multivariate Coggins step-up approach:
hydrogen liquefaction system assisted by geothermal absorption precooling Overall compression power reduction and exergy loss analysis. Int J Refrig 2019;
refrigeration cycle. Int J Hydrogen Energy 2018;43:20203–13. https://doi.org/ 104:189–200. https://doi.org/10.1016/j.ijrefrig.2019.04.002.
10.1016/j.ijhydene.2018.08.019. [51] Park JH, Khan MS, Lee M. Modified coordinate descent methodology for solving
[30] Bae J-E, Wilailak S, Yang J-H, Yun D-Y, Zahid U, Lee C-J. Multi-objective process design optimization problems: Application to natural gas plant. J Nat Gas
optimization of hydrogen liquefaction process integrated with liquefied natural gas Sci Eng 2015;27:32–41. https://doi.org/10.1016/j.jngse.2014.10.014.
system. Energy Convers Manag 2021;231:113835. https://doi.org/10.1016/j. [52] Rehman A, Qyyum MA, Qadeer K, Zakir F, Ding Y, Lee M, et al. Integrated
enconman.2021.113835. biomethane liquefaction using exergy from the discharging end of a liquid air
[31] Seyam S, Dincer I, Agelin-chaab M. A clean hydrogen liquefaction plant integrated energy storage system. Appl Energy 2020;260:114260. https://doi.org/10.1016/j.
with a geothermal system. J Clean Prod 2019. https://doi.org/10.1016/j. apenergy.2019.114260.
jclepro.2019.118562. [53] Abdul Qyyum M, Qadeer K, Lee M. Closed-Loop Self-Cooling Recuperative N2
[32] Berstad DO, Stang JH, Neksa P. Comparison criteria for large-scale hydrogen Expander Cycle for the Energy Efficient and Ecological Natural Gas Liquefaction
liquefaction processes 2009;34:1560–8. 10.1016/j.ijhydene.2008.11.058. Process. ACS Sustain Chem Eng 2018;6:5021–33. https://doi.org/10.1021/
[33] Asadnia M, Mehrpooya M. A novel hydrogen liquefaction process configuration acssuschemeng.7b04679.
with combined mixed refrigerant systems. Int J Hydrogen Energy 2017;42: [54] Qyyum MA, Ali W, Long NVD, Khan MS, Lee M. Energy efficiency enhancement of
15564–85. https://doi.org/10.1016/j.ijhydene.2017.04.260. a single mixed refrigerant LNG process using a novel hydraulic turbine. Energy
[34] Peng D, Robinson DB. A new equation of state. Nature 1929;123:507. 2018;144. https://doi.org/10.1016/j.energy.2017.12.084.
[35] Jacob H. Stang, Nekså P, E. Brendeng. On the design of an effient hydrogen [55] Qyyum MA, Qadeer K, Ahmad A, Ahmed F, Lee M. Two-phase expander
liquefaction process. 16th World Hydrog. Energy Conf. 2006, WHEC 2006, 2006. refrigeration cycles with ethane–nitrogen: A cost-efficient alternative LNG
[36] D. Stolten V. Scherer Transition to Renewable Energy Systems. 2013 10.1002/ processes for offshore applications. J Clean Prod 2020;248:119189. https://doi.
9783527673872. org/10.1016/j.jclepro.2019.119189.
[37] Stetson NT, McWhorter S, Ahn CC. Introduction to hydrogen storage. Compend. [56] Long NVD, Minh LQ, Pham TN, Bahadori A, Lee M. Novel retrofit designs using a
Hydrog. Energy, Woodhead Publishing 2016:3–25. https://doi.org/10.1016/b978- modified coordinate descent methodology for improving energy efficiency of
1-78242-362-1.00001-8. natural gas liquid fractionation process. J Nat Gas Sci Eng 2016;33:458–68.
[38] Valenti G. Hydrogen liquefaction and liquid hydrogen storage. Elsevier Ltd. 2016; https://doi.org/10.1016/j.jngse.2016.05.038.
23. https://doi.org/10.1016/B978-1-78242-362-1.00002-X. [57] Szargut J. Chemical exergies of the elements. Appl Energy 1989;32:269–86.
[39] McCarty RD, Hord J, Roder HM. Selected Properties of Hydrogen (Engineering https://doi.org/10.1016/0306-2619(89)90016-0.
Design Data), National Bureau of Standards Monograph 168. 1981. Library of [58] Venkatarathnam G. Cryogenic Mixed Refrigerant Processes. 2008. https://doi.org/
Congress Catalog Card Number: 80-600195. 10.1007/978-0-387-78514-1.

13