Effective and Realistic Strategies for Large-Scale Liquid Hydrogen Production

Abstract

The excessive use of fossil fuels could bring about a global environmental crisis. Transitioning from a carbon-based to a hydrogen-based economy is an important way to realize the low-carbon energy transition. The key to this economy transformation lies in the efficient utilization of hydrogen. Hydrogen liquefaction is an efficient technology for the transportation and storage of hydrogen, and the liquid hydrogen produced is also a direct feedstock for many important fields. Large-scale liquefaction of hydrogen has not been commercialized due to its high energy consumption (>10 kWh/kg_LH2) and low efficiency (<30%). However, conceptual designs for hydrogen liquefaction with a low energy consumption (about 6.4 kWh/kg_LH2) and high efficiency (>40%) are frequently reported in the existing literature. Therefore, in this paper, the production process of liquid hydrogen is reviewed from three aspects, which are hydrogen pre-cooling, hydrogen cryo-cooling, and ortho-para hydrogen (OPH) conversion. The focus is to summarize effective and realistic hydrogen liquefaction schemes in the existing studies to provide process guidance for the subsequent practical production of liquid hydrogen. The development of open and closed refrigeration cycles for hydrogen pre-cooling is reviewed following the lead of pre-coolant types. The implementation methods of structural optimization of different hydrogen cryo-cooling cycles are clarified and the performance improvements achieved are compared. Different modes of OPH conversion are presented and their realization in simulation and practical applications is summarized. Finally, subjective recommendations are given regarding the content of the review.

1. Introduction

Over the past centuries, the global economy has exploded in a phenomenal growth rate, in a carbon-based economy development model dominated by the use of fossil fuels [1,2,3]. The world’s industry has also seen many revolutions. However, due to excessive carbon emissions, many environmental problems such as global warming, polar ice melting, and the lowering of seawater pH are occurring worldwide [4,5]. Failure to curb the environmental crisis brought about by a carbon-based economy will certainly threaten the survival of humankind. Consequently, carbon-based economies that use predominantly fossil fuels must transform to renewable and low-carbon energy models [6].

As a clean energy source, hydrogen is multifunctional. It can be used as a fuel for electricity transportation and generation and as an energy storage medium [7]. In particular, the replacement of fossil fuels with hydrogen-based energy in the field of transportation is an important way to limit carbon emissions and achieve a low-carbon transformation [8]. Hydrogen energy has been used in a variety of industries, including long-distance road and rail transportation, aviation, and maritime. Hydrogen, as an important industrial raw material, has been widely used in industrial fields such as ammonia synthesis, methanol synthesis, petrochemicals and metallurgy. The scale of hydrogen-based energy applications in the industrial field is expected to grow rapidly under the constraints of low-carbon targets. In addition, the synergies between hydrogen and renewable energy are remarkable [9]. The combination of renewable power generation and hydrogen production allows a flexible conversion between electrical and chemical energy, thus enabling energy peaking to smooth the power characteristics. Renewable energy sources can be efficiently reused through hydrogen production, thereby increasing the efficiency of utilization during power curtailment.

The hydrogen-based economy is now supported by unprecedented political, commercial, and governmental policies [10,11]. By the end of 2023, more than 50 countries and regions around the world have published their hydrogen energy development strategies. Japan planned to achieve domestic hydrogen production of 3, 12, and 20 million tons in 2030, 2040, and 2050, respectively. Korean government planned to universalize 30,000 hydrogen-powered commercial vehicles by 2030 in their hydrogen economy development strategy announced in 2022. Russia was set to launch the first commercial-scale hydrogen energy project, with the plan to export 2 million tons of hydrogen per year from 2025 to 2035. The U.S. Department of Energy released the U.S. National Clean Hydrogen Strategic Roadmap in 2023. The roadmap proposed that clean hydrogen demand will reach 10, 20, and 50 million tons per year by 2030, 2040 and 2050, respectively. As an important part of national energy, the development of hydrogen energy industry will be strongly promoted and guaranteed.

The transition from a carbon-based to a hydrogen-based economy depends on the efficient utilization of hydrogen. Due to its low volumetric energy density, large-scale transportation and storage of hydrogen is the biggest challenge for its effective utilization [12,13]. Therefore, densification of hydrogen is required. The three methods for densification of hydrogen are physically based (compressed gas/liquid), chemically based (liquid organic carriers and reformed fuels) and materials based (adsorption) [14,15,16,17].

Table 1. The main characteristics of different forms of hydrogen storage and transportation ¹ [18].

Ref.	Compressed Gas (35 MPa)	Compressed Gas (70 MPa)	Liquid Hydrogen	Methanol	NH3
Volumetric density (kg/m3)	23.3	39.2	70.9	786	682
Volumetric hydrogen density (kgH2/m3)	23.3	39.2	70.9	99	121
Mass hydrogen density (wt%)	100	100	100	12.5	17.8
Specific energy consumption (kWh/kg)	2.8	5.6	6.4–15.0	17.6–30.4	19.1–30.0
Levelized cost of end product ($/kg)	0.22–0.28	2.83	0.5–3	2–4.17	1–2.17

¹: Data from Ref [18].

summarizes the main characteristics of different forms of hydrogen storage and transportation. Among them, liquid hydrogen (LH₂) has higher volumetric hydrogen density, which is the biggest advantage compared to the gaseous hydrogen storage method. For chemical-based methods, the need for synthesis and decomposition of chemical carriers not only sacrifices the mass hydrogen density of stored hydrogen, but also entails higher energy consumption and costs. Therefore, LH₂, with both high volumetric and mass hydrogen densities, is the most promising solution for transportation and storage of hydrogen [18,19,20,21,22,23].

Hydrogen liquefaction is an energy-intensive process, and inefficiencies in equipment and processes are obstacles to commercial production. Hydrogen has an extremely low boiling point (approximately 20 K) and inversion temperature (approximately 200 K), and is difficult to liquefy without being sufficiently pre-cooled. Therefore, it was once considered a permanent gas [24]. It was not until 1898 that Jamaz [25] first prepared LH₂ with the help of liquid air. A few years later, the pre-cooled Linde–Hampson (L-H) cycle for laboratory preparation of LH₂ was developed [26]. In the twentieth century, several more efficient hydrogen liquefaction systems were invented, including the Claude cycle, the Kapitza cycle, and the helium refrigeration cycle [27]. The theoretical minimum energy consumption for hydrogen liquefaction is approximately 2.7 kWh/kg_LH2. The current average energy consumption level of hydrogen liquefaction plants is around 10 kWh/kg_LH2, with an exergy efficiency (EXE) of about 30%. The energy consumption for liquefying hydrogen is equivalent to 30% of the calorific value of hydrogen. And this value is only 5–10% for LNG. Therefore, it is necessary to reduce the energy consumption of hydrogen liquefaction. Since 1950, a number of large hydrogen liquefaction plants have been built in the United States [28]. They were used primarily in the growing aerospace and petroleum industries. These plants were built based on the pre-cooled Claude process, in which a liquid nitrogen (LN₂) pre-cooling system cooled hydrogen to −193 °C and a refrigeration system in the Claude cycle cooled further hydrogen to −253 °C [28]. Air Products and Praxair were the major suppliers of LH₂ in North America. After 1970, hydrogen liquefaction plants were built in a number of countries in Europe and Asia [28]. The liquefaction process used was similar to that of the first plant built in the United States. The plants for which process descriptions were currently available in the literature included the plants built by Praxair in the United States and the two plants built by Linde in Germany.

Praxair built five hydrogen liquefaction plants in the United States. The capacities of these plants were within 30 t/d and the EXE was between 20% and 30% [29]. The process used for hydrogen liquefaction was an improved Claude cycle with LN₂ pre-cooling. The modified Claude cycle was equipped with two stages of expansion refrigeration and one stage of throttling refrigeration. The specific energy consumption (SEC) of the plant was in the range of 12.5–15 kWh/kg_LH2 [30]. Linde built two liquefaction plants, in Ingolstadt in 1991 and in Leuna, Germany, in 2008, with capacities of 4.4 and 5 t/d [31], respectively. The liquefaction plants were all based on the modified Claude cycles where LN₂ was used as a pre-coolant for hydrogen. The Leuna plant has been designed with a number of process improvements compared to the Ingolstadt plant, which have resulted in a higher efficiency [32]. The first difference was the treatment of the flash hydrogen. In contrast to the flash hydrogen recovery process adopted in the Ingolstadt plant, the Leuna plant employed an injector to accomplish both the introduction of the flash hydrogen and the depressurization of the mainstream hydrogen [33]. The refrigeration cycle in the Leuna plant allowed for a more flexible arrangement of expanders due to the absence of flash hydrogen recovery. The second was the ortho- and para-hydrogen (OPH) conversion mode. Different from the isothermal conversion adopted at the Ingolstadt plant, the Leuna plant realized the continuous conversion of OPH by filling the heat exchanger with catalysts. These improvements not only made the Leuna plant more efficient, but also provided valuable experience in the design and construction of the subsequent liquefaction plants.

With the increasing demand for LH₂, hydrogen liquefaction plants are becoming larger in size. The earlier throttling refrigeration cycle has been eliminated and more efficient liquefaction cycles with expanders are more widely available. On this basis, the emergence of pre-cooled hydrogen liquefaction cycles has greatly advanced the development of hydrogen liquefaction technology [34,35,36,37,38] because it can offer at least a 30% increase in liquefaction efficiency compared to direct hydrogen liquefaction. In addition, the integration of ortho–para hydrogen (OPH) conversion and hydrogen liquefaction processes has become a new breakthrough point to drive LH₂ production. In the future, reducing the energy consumption of hydrogen liquefaction requires two types of efforts [39,40]. One is to design more efficient hydrogen liquefaction configurations, and the other is to improve the efficiency of major components including compressors, expanders, and heat exchangers [41]. Different from the fact that the energy consumption of existing hydrogen liquefaction plants is still at a high level, a number of new conceptual designs for hydrogen liquefaction with energy consumption below 6.4 kWh/kg_LH2 have been proposed in recent years [42,43,44]. This value is widely considered to be the benchmark for commercial large-scale hydrogen liquefaction plants [45]. However, many of the options are unrealistic, either using extremely complex systems or efficient equipment that exceeds reality. There are also liquefaction solutions that do not demonstrate efficient configurations to match their low energy consumption.

Therefore, in this paper, the production process of LH₂ is reviewed from three aspects, which are hydrogen pre-cooling, hydrogen cryo-cooling, and OPH conversion [46]. The focus is to summarize effective and realistic hydrogen liquefaction schemes in the existing literature, especially more efficient process designs, to provide process guidance for subsequent practical production of LH₂.

2. Hydrogen Pre-Cooling Process

Pre-cooling processes have become an indispensable stage in large hydrogen liquefaction systems, both in existing commercial plants and in conceptualized designs in the literature. Existing studies have proved that the introduction of hydrogen pre-cooling is an essential way to reduce the SEC [47]. A number of pre-coolants are available for hydrogen pre-cooling, with pre-cooling temperatures ranging from approximately −30 °C to −194 °C. Pre-cooling cycles can be categorized into open and closed cycles based on the reusability of pre-coolants.

2.1. Open Pre-Cooling Cycle

Open pre-cooling cycles are most frequently used in existing hydrogen liquefaction plants. The most typical working medium is LN₂. In addition, LNG and liquid air are also frequently introduced into the conceptual design of hydrogen liquefaction as efficient pre-cooling agents.

2.1.1. LN₂ Cycle

LN₂, as an air separation product, has an extremely mature production process and is quite inexpensive. In addition, it is an environmentally friendly refrigerant that causes no additional pollution. It is, therefore, the most frequently used working medium in the industry and laboratories. LN₂ has been the favored pre-coolant for the hydrogen liquefaction since it was first achieved. LN₂ has a temperature of about −195 °C and can cool hydrogen to approximately −193 °C [28]. Currently, almost all commercial large-scale hydrogen liquefaction plants in service have adopted LN₂ as a pre-coolant [28]. These hydrogen liquefaction plants were generally located immediately adjacent to air separation units for access to LN₂. Therefore, in some studies, LN₂ used for hydrogen pre-cooling was not considered for energy consumption. However, the supply of oxygen required globally is limited, while the demand for LH₂ will increase dramatically in the future [48]. As a result, this hinders the application of LN₂ in hydrogen liquefaction systems. In most studies that considered LN₂ production, the energy consumption for the production of LN₂ for hydrogen pre-cooling was considered to be 0.5 kWh/kg [49]. Considering the poor accuracy of the above energy consumption values, Khodaparast et al. [50] discussed the energy consumption of supplying LN₂. It has been demonstrated that the pre-cooling process at the expense of LN₂ production did not meet the low energy requirement for large-scale liquefaction of hydrogen. Therefore, it is already a reality that LN₂ pre-cooling cycle will slowly be withdrawn from large-scale hydrogen liquefaction processes.

2.1.2. LNG Cycle

High-quality cold energy is contained in the LNG, with a temperature of about −162 °C and stored at LNG-receiving terminals. Generally, the cold energy (830 kJ/kg) of LNG is released in seawater during its gasification process [51]. The concept of combining the cold energy of LNG with the hydrogen pre-cooling was first proposed in 2006 [32]. LNG is ideally suited as a pre-coolant for hydrogen liquefaction processes due to the realization of the dual benefits of natural gasification of LNG and hydrogen pre-cooling [52]. However, natural gas needs to be protected from leakage because of its high global warming potential and its flammability and explosiveness.

Depending on the pressure of the gasified natural gas, different pre-cooling schemes can be implemented [53]. In the LNG pre-cooling process where the product is low-pressure natural gas, the gasified high-pressure natural gas needs to be expanded and depressurized. The temperature of the expanded natural gas is reduced, which is referred to as secondary cold energy in this paper. A few studies directly adopted the LNG pre-cooling process, as shown in Figure 1a, and did not recover the secondary cold energy. This resulted in low natural gas temperatures at the expander outlet, which not only wasted cold energy, but also required an additional heat exchanger to bring it up to temperature. In a system combining hydrogen liquefaction, steam methane reforming (SMR), and waste heat recovery [54] to meet the requirement of natural gas into SMR, part of the recovered heat was used to heat the natural gas. The wasted secondary cold energy of LNG was also encountered in the designed integrated cryogenic system by Zarsazi et al. [55]. A specialized heat exchanger was applied to heat the expanded natural gas. An effective cold energy utilization strategy is shown in Figure 1b. The expanded natural gas was used again to pre-cool the hydrogen, which reduced the amount of LNG required by approximately 12% [39] compared to the pre-cooling configuration shown in Figure 1a. A two-phase expander is generally required in this scenario. The above LNG pre-cooling processes are mostly present in small-scale hydrogen liquefaction systems.

A more common product of the LNG gasification process is high pressure natural gas that is easy to transport. In general, LNG is gasified in extremely large quantities, and the hydrogen liquefaction process that accompanies it is of a large scale. LNG releases more cold energy within its slip temperature compared to that released at other temperatures due to continuous gasification. As a result, the cold energy of the LNG is in excess during the hydrogen pre-cooling process. When the pressure of gasified natural gas is lower than 5 MPa, it is important to further utilize the remaining cold energy in the gasification stage [56,57,58].

LNG is on the warm side and is slightly less effective at pre-cooling hydrogen. Therefore, LNG can be coupled with a refrigeration cycle with lower refrigeration temperature. In this scenario, LNG can pre-cool hydrogen and other refrigeration cycles at the same time. Bi et al. [59] included a nitrogen Brayton cycle as an intermediate cycle for hydrogen pre-cooling using LNG. The integration of LNG resulted in a 26.3% reduction in the SEC of the system compared to the case where only LN₂ was used for pre-cooling. In order to minimize the SEC, Cho et al. [60] proposed a pre-cooling method with a combination of LNG cycle and a mixed refrigerant cycle. The proposed system reduced the SEC by more than 26.4% and the costs of producing LH₂ by approximately 7.7% compared to the base system without LNG. In order to better match the heat transfer characteristics of LNG, Li et al. [61] selected a binary fluid consisting of nitrogen and propane as the refrigerants of an intermediate refrigeration cycle instead of LN₂. They derived the thermodynamic relationship between enthalpy and temperature of refrigerant based on heat transfer temperature difference and composite curves, and solved for the optimal binary refrigerants includes 39.79% propane and 60.21% nitrogen.

2.1.3. Liquid Air Cycle

Liquid air energy storage (LAES) is an important energy storage technology to solve the problem of renewable energy storage [62]. In recent years, emerging LAES have been integrated into hydrogen liquefaction systems, making liquid air a new option for hydrogen pre-cooling [63]. The potential of liquid air in hydrogen liquefaction was evaluated in their studies by Taghavi et al. [64] and Naquash et al. [65]. In addition to being used to pre-cool hydrogen, the warmed air can be burned in the fuel chamber to output power through the powertrain. Two integrated systems demonstrated liquefaction energy consumption of 5.955 kWh/kg_LH2 and 6.71 kWh/kg_LH2, respectively, which showed better low-energy hydrogen liquefaction capability. However, the instability of the supply of liquid air is the biggest obstacle to its development, making it difficult to play a role in the large-scale preparation of LH₂. With the further development of LAES in the future, it may serve as an intermediate storage link to facilitate small-scale distributed LH₂ preparation. For example, considering the volatility of the gasification loads at LNG-receiving stations, LAES technology was used as an intermediate stage by Chen et al. [66]. The technology could stably store LNG gasification cold energy, thus enabling the hydrogen liquefaction process to be protected from fluctuations in the LNG gasification load.

2.2. Closed Pre-Cooling Cycle

The reverse Brayton refrigeration cycle (BRC) is a typical recuperative cycle [67]. The mixed refrigerant cycle (MRC) is also a refrigeration cycle developed on the basis of the recuperators. The absorption refrigeration cycle (ARC) is representative of the regenerative cycle, whose main equipment includes the compressor, evaporator, regenerator and condenser [68,69]. They are commonly used for hydrogen pre-cooling.

2.2.1. Reverse Brayton Refrigeration Cycle

The Brayton cycle is a common power cycle. Its reverse variant is a refrigeration cycle known as the reverse BRC. The effective refrigeration temperature of the BRC is generally the boiling point of the working fluid employed. The refrigeration temperature range depends on the pressure ratio between the compressor and the expander outlet and the number of stages of the refrigeration circuit.

For a single-stage BRC as illustrated in Figure 2a, it can provide small cooling range and cannot accomplish pre-cooling of hydrogen on its own. Therefore, it is often used in combination with LNG or MRC for hydrogen pre-cooling, thus maximizing the refrigeration effect. Yang et al. [70] evaluated the ability combining the use of LNG and the nitrogen BRC in reducing costs and energy of hydrogen liquefaction process. Hydrogen pre-cooling using a combination of nitrogen BRC and LNG was used by Bae et al. [71]. Multi-objective optimization and cost analysis were performed to integrate energy and environmental aspects. Furthermore, Asadnia et al. [72] proposed a combined form of MRC with nitrogen BRC in a hydrogen liquefaction process. This pre-cooling solution solved the problem of insufficient cooling temperatures due to the lack of low boiling point refrigerants in the mixed refrigerant.

A schematic diagram of a two-stage Brayton cascade refrigeration cycle is shown in Figure 2b. The two refrigeration circuits allow for the pre-cooling of hydrogen over a larger temperature range. The increase in the number of refrigeration circuits contributes to the reduction in the energy consumption of the refrigeration system. After replacing the single-stage nitrogen Brayton refrigeration cycle with a two-stage refrigeration cycle, the SEC of the hydrogen liquefaction system was reduced from 7.06 to 6.70 kWh/kg_LH2 [49].

2.2.2. Mixed Refrigerant Cycle

To intensify the heat transfer capability between the refrigerant and hydrogen, MRC has been developed [73], as depicted in Figure 3. The mixed refrigerant used in the MRC contains a variety of refrigerants with different saturation temperatures. Therefore, the heavier liquid-phase refrigerant can be separated from the lighter gas-phase refrigerant by means of a gas–liquid separator. Among them, the liquid-phase refrigerant after throttling may be used to pre-cool hydrogen. The gas-phase refrigerant needs to be further cooled in the heat exchanger to reduce the temperature. Different numbers of gas–liquid separators can be set in the MRC as required.

MRC can provide significant advantages in reducing the energy demand for hydrogen liquefaction. As shown in

Table 2. Components of mixed refrigerant in different hydrogen liquefaction studies.

Ref.	H2	N2	CH4	C2H6	C3H8	n-C4H10	i-C4H10	C5H12	R14	Acetone	C2H4	NH3	SEC 1	EXE 2
Bian et al. [75]	0.013	0.136	0.136	0.162	0.064	0.017	0.017	0.189	0.152	0.114	- 3	-	5.633	53.15
Azizabadi et al. [43]	0.1	0.15	0.18	0.1	0.16	-	-	0.08	0.08	-	0.15	-	4.54	-
Taghavi et al. [76]	0.01	0.16	0.17	0.07	0.18	0.02	-	0.15	0.08	-	0.16	-	4.32	53.35
Chorbani et al. [77]	0.04	0.18	0.24	0.28	-	0.26	-	-	-	-	-	-	5.462	58.73
Xu et al. [78]	-	0.11	0.222	0.115	0.166	-	-	0.238	-	-	0.149	-	6.422	52.66
	-	0.065	0.16	0.028	0.225	-	-	0.297	-	-	0.225	-	6.872	49.24
Noh et al. [79]	0.01	0.16	0.17	0.07	0.18	0.02	-	0.15	0.08	-	0.16	-	5.613	-
Asadnia et al. [72]	0.0002	0.0642	0.1021	0.1925	0.0532	0.0235	0.0243	0.2982	0.0986	0.1273	-	0.0158	7.69	-
Asadnia et al. [80]	0.0002	0.0642	0.1021	0.1925	0.0532	0.0235	0.0243	0.2982	0.0986	0.1273	-	0.0158	6.47	-
Krasae-in et al. [81]	0.012	0.136	0.136	0.162	0.064	0.017	0.017	0.189	0.152	0.114	-	-	5.35	54.02
Luo et al. [82]	0.0143	0.0712	0.1772	0.149	0.0479	0.0209	0.0202	0.2736	0.1208	0.0913	-	0.1208	6.15	-
Sleiti et al. [83]	0.01	0.14	0.17	0.07	0.21	0.02	-	0.16	0.07	-	0.17	-	-	-
Ansarinasab et al. [84]	0.04	0.18	0.24	0.28	-	0.26	-	-	-	-	-	-	-	-
Krasae-in et al. [85]	0.04	0.18	0.24	0.28	-	0.26	-	-	-	-	-	-	5.91	-
Ebrahimi et al. [86]	0.01	0.16	0.17	0.07	0.18	0.02	-	0.15	0.08	-	0.16	-	-	-
Ghorbani et al. [44]	0.01	0.16	0.17	0.07	0.18	0.02	-	0.15	0.08	-	0.16	-	6.642	-
Sadaghiani et al. [45]	0.01	0.16	0.17	0.07	0.18	0.02	-	0.15	0.08	-	0.16	-	4.36	55.47
Faramarzi et al. [87]	-	0.132	0.211	-	0.169	-	-	0.143	-	-	0.345	-	5.31	-
Qyyum et al. [46]	-	0.0112	0.3523	0.2927	0.0412	0.1058	0.1969	-	-	-	-	-	6.45	47.2
Jackson et al. [88]	-	0.101	0.324	0.274	0.031	0.27	-	-	-	-	-	-	7.1	-

^1,2 Due to the lack of data from operating hydrogen liquefaction plants, this paper cites data from a large number of references, which are based on simulations. SEC and EXE are the principal performance parameters for evaluating hydrogen liquefaction processes. SEC is the net energy consumption per kg of hydrogen liquefied. EXE is defined as the ratio between the minimum energy consumption for hydrogen liquefaction and the actual energy consumption. These two parameters are influenced by the equation of state and assumptions used in the simulation. The commonly used equations of state are the P-R equation and the MBWR equation, the latter being considered more appropriate for hydrogen. The assumptions to be considered for the simulation include the initial conditions of the feed hydrogen (pressure and temperature) and the performance parameters of the key parameters (isentropic efficiency and pressure ratio of the compressor, isentropic efficiency of the expander, minimum heat transfer temperature difference in the heat exchanger and natural cooling temperature of the cooler, etc.). Different assumptions are adopted in different studies. ³ Since the data is not provided in the literature, it is marked with “-”.

, which lists the performance of hydrogen liquefaction systems using MRC, the SEC of liquefaction systems generally decreases to less than 7.0 kWh/kg_LH2, with EXE generally exceeding 50%. To test the performance of MRC, Krasae-in et al. [74] conducted experiments on the developed MRC. The mixed refrigerant used contained 28% butane, 30% ethane, 26% methane, 12% nitrogen, and 4% neon. In the experiment. The hydrogen of 0.6 kg/h was cooled from 25 °C to −158 °C by the MRC with an energy consumption of 1.76 kWh/kg_LH2. Two conclusions were drawn from the experimental study. (1): The power consumption of the refrigeration system obtained by simulation was basically equal to that measured experimentally. This verified the validity of the simulation model commonly used. (2): The energy consumption of the MRC was lower compared to that of regular refrigeration systems.

The composition of the mixed refrigerant has a significant impact on the performance of the MRC. Therefore, many researchers have optimized the components and ratios used in mixed refrigerant. These studies are investigated, and the mixed refrigerant they adopted are summarized in Table 2. Nitrogen, methane, ethane, and propane were selected as refrigerants for hydrogen pre-cooling in almost all studies. Among them, methane accounted for the most, about 20%, while nitrogen, ethane, and propane mostly accounted for between 10 and 15%. In addition, n-Butane, pentane, R14, ethylene, and hydrogen were commonly available refrigerants. Of these, ethylene and pentane made up a high percentage, ranging from about 15% to 20%, and hydrogen was typically within 4%. Other refrigerants were used less frequently and included neon, i-Butane, acetone, and ammonia. Notably, mixed refrigerant including 16% nitrogen, 17% methane, 7% ethane, 18% propane, 2% isobutane, 15% n-pentane, 8% R14, 16% ethylene, and 1% hydrogen have been adopted in many studies. Determining the optimum for each component in the mixed refrigerant requires the aid of an optimization algorithm. Genetic algorithms, particle swarm optimization, and Bayesian optimization have been applied to the optimization of mixed refrigerant composition. Sleiti et al. [83] facilitated the optimization process by using a knowledge-based optimization procedure in combination with the above optimization algorithms. Equbal et al. [89] evaluated the efficacy of various artificial intelligence and machine learning techniques being used to optimize the composition of mixed refrigerant. It was demonstrated that Support Vector Machines were the optimal model followed by Genetic Algorithms.

2.2.3. Absorption Refrigeration Cycle

ARC is a heat-driven refrigeration system that is typically used to recover low-quality waste heat from renewable energy sources and industrial processes. Figure 4 displays the structural configuration of an ARC. The working fluid of an ARC is a binary solution with saturated solubility related to temperature and pressure. Currently, the use of ARC is receiving more and more attention and is being used in hydrogen liquefaction processes due to the importance of renewable energy development and waste heat recovery [90,91,92,93,94,95,96,97,98,99].

Table 3. Details of several hydrogen liquefaction systems integrated with ARCs.

Ref.	Work Fluid	Hot Source	Function	Pre-Cooling Temperature (°C)	SEC (kWh/kgLH2)
Li et al. [68]	ammonia-water	waste heat	pow output	- 1	6.61
Taghavi et al. [76]	ammonia-water	waste heat	hydrogen pre-cooling	−31.7	4.32
Azizabadi et al. [43]	ammonia-water	waste heat	hydrogen pre-cooling	−30	4.54
Ghorbani et al. [91]	ammonia-water	waste heat	hydrogen pre-cooling	−55	7.208
Yilmaz et al. [92]	ammonia-water	waste heat	refrigerant cooling	−57	5.413
Cao et al. [93]	-	geothermal energy	pow output	-	-
Faramarzi et al. [69]	ammonia-water	geothermal energy	hydrogen pre-cooling	−26.9	8.81
Faramarzi et al. [94]	ammonia-water	geothermal power	hydrogen pre-cooling	−27	8.69
Yilmaz et al. [95]	ammonia-water	geothermal power	hydrogen pre-cooling	−30	11.88
Yilmaz et al. [96]	ammonia-water	geothermal power	hydrogen pre-cooling	−30	10.06
Zhang et al. [97]	ammonia-water	solar energy	hydrogen pre-cooling	−28.65	-
Yan et al. [98]	ammonia-water	solar energy	refrigerant cooling	-	5.2201
Ghorbani et al. [44]	-	solar energy	refrigerant cooling	-	4.02
Aasadnia et al. [80]	ammonia-water	solar energy	refrigerant cooling	−23.5	6.47
Aasadnia et al. [99]	ammonia-water	solar energy	refrigerant cooling	−23.17	12.7

¹ Since the data is not provided in the literature, it is marked with “-”.

summarizes the details of several hydrogen liquefaction systems integrated with ARCs. In terms of the type of heat source, the plant waste heat and renewable energy sources, including solar and geothermal energy, were the main options. From the perspective of working fluids, ammonia–water refrigerant was almost a preferred option. This refrigerant can provide a cooling temperature of about −30 °C. When two cycles were cascaded, the refrigeration temperature can be reduced to approximately −60 °C [91,92]. Since the energy-consuming component in an ARC is the circulation pump, its energy demand is extremely low. However, the high refrigeration temperature of the ARC leads to high energy consumption for hydrogen liquefaction when it is used independently for hydrogen pre-cooling. It can be used for hydrogen pre-cooling in combination with an inverse BRC or a MRC to achieve lower liquefaction energy consumption, down to less than 5 kWh/kg_LH2 [43,44]. In addition to pre-cooling hydrogen, ARC can also be used to pre-cool refrigerants or to output power. The effect of the temperature at the compressor inlet on the SEC of liquefaction system was investigated by Jackson et al. [88]. It showed that energy consumption increased by approximately 20% when refrigerant temperatures were increased from 5 to 50 °C. Therefore, in several studies, ARC has been applied to pre-cool the refrigerant entering compressors, thus reducing the compression energy consumption. Li et al. [68] compared the ability of ORC and ARC to recover waste heat in a hydrogen liquefaction system. It was found that the SEC of the system using ARC was 2.2% lower than that using ORC.

In the large-scale hydrogen liquefaction conceptual design, ARC can be useful in many ways to reduce the energy demand of the system. However, there are no reports of ARC being included in formal programs among the currently publicized projects for hydrogen liquefaction plants. In the future, as renewable energy sources play a more important role in hydrogen production, ARC is likely to be applied as a solution to improve efficiency and reduce energy consumption.

3. Hydrogen Cryo-Cooling Process

The hydrogen liquefaction cycle could be classified according to hydrogen cryo-cooling cycle. The basic liquefaction cycles are classified into three categories, namely the Linde-Hampson cycle, the Claude cycle, and the liquefaction cycle with an external refrigeration cycle [100].

3.1. Linde-Hampson Cycle

The L-H cycle is the simplest liquefaction cycle in which the most central refrigeration element is the J-T valve. The isenthalpic expansion process of a gas has a refrigeration effect only if its temperature is lower than the inversion temperature. The inversion temperature of hydrogen is approximately 200 K, implying that the hydrogen must be pre-cooled in the L-H cycle to liquefy hydrogen [101]. Figure 5 shows an L-H liquefaction cycle with pre-cooling. LN₂ is the most commonly used pre-coolant. To obtain LH₂ during the throttling process, the hydrogen in the L-H cycle needs to be compressed to a high pressure and cooled to a lower temperature before entering a J-T valve [14]. As a result, a significant amount of compression power is required to compress hydrogen in the L-H cycle. Furthermore, only a small portion of hydrogen is liquefied and the rest of the gas needs to be passed through a recuperator for cooling hydrogen [55,102,103,104,105]. Therefore, hydrogen liquefaction systems with L-H cycles are extremely energy intensive and are no longer suitable for hydrogen liquefaction.

3.2. Claude Cycle

In contrast to the L-H cycle, the Claude liquefaction cycle enables the utilization of an expander to provide hydrogen cooling. Figure 6 illustrates the flow diagram of a simple Claude cycle. A portion of hydrogen split from the mainstream hydrogen is applied as the refrigerant. After passing through the expander and lowering its temperature, it can be used to provide cryo-cooling for the mainstream hydrogen. The SEC of the Claude liquefaction system depends mainly on the hydrogen liquefaction rate. The hydrogen liquefaction rate is related to the pre-cooling temperature and the structural configuration of the Claude cycle.

Table 4. Details of several hydrogen liquefaction systems with different pre-cooling cycles.

Ref.	Pre-Cooling Method	Pre-Cooling Temperature (°C)	Cycle Type	Production (t/d)	Liquefaction Rate	Feed Pressure (Bar)	SEC/EXE(kWh/kg)
Bae et al. [71]	N2 BRC + LNG	−173.15	simple	300	0.12	20	10.76/-
Kwon et al. [106]	N2 BRC + LNG	−163	simple	- 1	0.1783	20	11.02/-
Yilmaz et al. [96]	ARC	−26.90	simple	507.86	0.2856	32	10.06/-
Aasadnia et al. [99]	ARC	−23.17	simple	261	0.2337	1.13	12.7/31.6
Yamin et al. [92]	ARC	−28.65	simple	85	0.4873	30	-/50.22
Yang et al. [107]	LN2	−194	two-stage	12	0.8375	35	5.02/-
Im et al. [108]	CO2 BRC	−50.63	two-stage	100	0.2047	20	7.3/33
Yang et al. [107]	LN2	−194	two-stage	120	-	21	5.62/-
Seyam et al. [109]	N2 Claude	−193.15	three-stage	355	1	20	5.24/-
Kim et al. [110]	N2 BRC + LNG	−193	two-stage	300	1	20	7.78/52.4
Berstad et al. [111]	MRC	−159.15	four-stage	132.10	0.9463	20	6.57/-
Cardella et al. [49]	Dual-N2	−193.15	three-stage	25	1	-	6.7/-
	MRC	−193.15	three-stage	100	1	-	6.2/-

¹ Since the data is not provided in the literature, it is marked with “-”.

summarizes the details of several Claude hydrogen liquefaction process studies. In the studies by Bae et al. [71] and Kwon et al. [106], the refrigeration portion of the Claude cycle was used to pre-cool hydrogen, resulting in no cold energy for hydrogen cryo-cooling. As a result, the performance of the Claude liquefaction system they constructed was similar to that of the L-H system, exhibiting low liquefaction rates and high energy consumption. When a simple Claude cycle is combined with a lower temperature ARC, the hydrogen liquefaction rate is typically less than 50% and the SEC is higher than 10 kWh/kg_LH2. The reduction in pre-cooling temperature is beneficial to increase the liquefaction rate of hydrogen. As the pre-cooling temperature decreases, the SEC of hydrogen liquefaction can be reduced to less than 10 kWh/kg_LH2.

Due to the increasing capacity of hydrogen liquefaction plants, reducing operating costs has become a priority issue. In order to reduce energy consumption, several improved Claude hydrogen liquefaction cycles have been developed.

First, the form of expansion refrigeration in the Claude cycle can be improved. The initial single expander is replaced by two expanders in a series and a heat exchanger is added between them, as shown in Figure 7a. It can be called series expansion refrigeration. The advantages of such a setup are twofold. First, two expanders in the series have a smaller expansion ratio and better refrigeration performance than a single expander. Second, the refrigerant between the two expanders can be further cooled in a heat exchanger, allowing it to function over a wider temperature range. Therefore, the series expansion refrigeration allows for a more flexible distribution of the refrigeration capacity between the two temperature intervals. Cammarata et al. [112] first constructed a form of series expansion refrigeration based on the simple Claude cycle and showed that it was suitable for liquefaction of gases such as hydrogen and helium. Series expansion refrigeration maximizes the performance of a single-stage refrigeration cycle. This Claude cycle is suitable for the liquefaction of high-pressure hydrogen, preferably in combination with a pre-cooling process with a low pre-cooling temperature.

Secondly, multiple expansion refrigeration cycles can be combined to further reduce the energy consumption, as shown in Figure 7b. In recent years, Claude liquefaction cycles with multi-stage expansion refrigeration technology have been frequently developed. The multi-stage expansion refrigeration system in the Claude cycle provides sufficient cooling capacity for hydrogen at different temperature intervals, thus making the hydrogen cryo-cooling process more efficient. In addition, hydrogen can be liquefied at a rate of more than 90% due to the higher quality of cold energy and larger refrigeration capacity that can be provided by a multi-stage refrigeration system. Current research has revealed that once the number of stages of the expansion refrigeration cycles exceeds four, there is little benefit to be gained from increasing the number of stages. Therefore, to balance operating and fixed costs, the number of stages is generally limited to four. The performance of the Claude liquefaction cycles using multiple expansion refrigeration technology is given in Table 4. These Claude liquefaction cycles are highly efficient with SEC reduced to 6–8 kWh/kg_LH2. In addition, due to the lesser amount of flash hydrogen, 100% liquefaction of hydrogen can be achieved in some studies by either early depressurization of hydrogen or the introduction of an injector. This is also an important improved form of the Claude cycle.

3.3. Joule-Brayton Refrigeration Cycle

In addition to the Claude liquefaction process, cryo-cooling and liquefaction of hydrogen using an external refrigeration cycle is also a common option. A 100% liquefaction of hydrogen can be achieved easily using an external refrigeration cycle, but not using the Claude cycle. The cascaded Joule–Brayton (J-B) refrigeration cycles are the most commonly used external refrigeration cycles for hydrogen cryo-cooling. Figure 8 presents three-stage cascaded J-B cycles that share a compression system. Each J-B cycle is equipped with a recuperator to adjust the refrigeration temperature. An expander is connected after the recuperator. The expanded refrigerant enables cryo-cooling of hydrogen. In recent years, numerous scholars have carried out research on the cascaded J-B cycles to improve its performance. These improvement studies can be categorized into two types, including adjusting the number of cascaded cycles and improving the configuration of cascaded cycles.

During the process of hydrogen being cryo-cooled, hydrogen at different temperatures requires different cooling capacities. The cooling capacity provided by the refrigeration cycle depends on the mass flow rate of the refrigerant. Each additional J-B cycle allows an increased degree of freedom to regulate the refrigerant flow. Therefore, the more J-B cycles are cascaded, the more efficient the cooling of hydrogen. However, as the number of cascaded cycles increases, the fixed costs increase almost linearly.

Table 5. Details of several hydrogen liquefaction systems with cascade J-B refrigeration cycles.

Ref.	Pre-Cooling Method	Pre-Cooling Temperature (°C)	Stages	SEC (kWh/kg)	EXE (%)	Production (t/d)
Geng et al. [114]	MRC	−193	2	5.963	52.61	302.4
Yu et al. [115]	N2 BRC	−193.15	2	11.41	26.1	10
Geng et al. [116]	MRC	−185.3	3	6.3476	49.26	302.4
Bian et al. [75]	MRC	−195.9	3	5.633	53.15	100
Aasadnia et al. [80]	MRC	−199.9	3	6.47	45.5	90
Luo et al. [82]	MRC	−200.9	3	6.15	- 1	100
Sadaghiani et al. [45]	MRC	−195	3	4.36	55.47	300
Valenti et al. [117]	- 1	-	4	5	48	864
Krasae-in et al. [85]	MRC	−193	4	5.91	-	100
Krasae-in et al. [81]	MRC	−193	4	5.35	54.02	100
Faramarzi et al. [118]	LNG	−142.15	4	8.85	47	369
Faramarzi et al. [87]	MRC	−159.4	4	5.31	-	1.512
Ghorbani et al. [119]	MRC	−194.6	6	6.642	-	100
Asadnia et al. [80]	MRC	−198.2	6	7.69	39.5	100
Bian et al. [39]	LNG	−156	4	6.88	-	-
	LNG	−156	Improved	6.6	47	120
Yang et al. [113]	LNG	−156	Improved	6.61	46.9	120
Yang et al. [56]	LNG	−156	Improved	6.59	47	12
Yang et al. [120]	LNG + N2 BRC	−194	Improved	6.29	48.7	120

¹ Since the data is not provided in the literature, it is marked with “-”.

counts the details of several hydrogen liquefaction systems with cascaded J-B refrigeration cycles in recent years. More researchers were choosing to adopt three- or four-stage cascaded J-B refrigeration cycles to cool hydrogen in the temperature range from about −190 to −252 °C. This suggests that three- or four-stage cascaded J-B cycles can balance the operating and fixed costs of hydrogen liquefaction systems. To further cut fixed costs, Yang et al. [56,113] creatively proposed a dual-pressure J-B cycle and a parallel J-B cycle based on the discovery that the most cooling capacity is required for hydrogen near its critical temperature, as depicted in Figure 9b,c. The improved J-B refrigeration cycles allow the amount of refrigerant to be adjusted in two temperature ranges, ensuring that more refrigerant is used to cool hydrogen near the critical temperature. The utilization efficiency of refrigerant in the improved J-B cycles has been significantly improved, allowing the number of cascaded cycles and refrigerant usage to be reduced while ensuring low energy consumption.

In addition to increasing the number of cascaded cycles, the heat exchange efficiency between hydrogen and refrigerant can be improved by adjusting the refrigeration temperatures of J-B cycles. The refrigeration temperature range of the J-B cycle depends on the pre-cooling temperature of the recuperator and the ratio of the refrigerant pressurization pressure to the expansion pressure (referred to in this paper as the refrigeration pressure ratio of the J-B cycle). The pre-cooling temperature of the refrigerant in the recuperator can be freely adjusted. Therefore, the strategy to improve the configuration is to construct cascaded J-B cycles with different refrigeration pressure ratios. First, it is possible to do this based on a multi-stage compression system. A multi-stage compression system is arranged with several compressors in the series. At the outlet of each stage, a portion of refrigerant can be diverted. Similarly, the inlet of each stage can be fed with a portion of return refrigerant. Therefore, the J-B cycle can be made to have a specific refrigeration pressure ratio by adjusting the refrigerant outflow and return positions in the compression system. In this way, multi-stage cascaded J-B cycles can be combined to yield different refrigeration pressure ratios. Second, multiple J-B cycles with different refrigeration pressure ratios can be used in parallel for hydrogen cryo-cooling. These cycles are independent of each other. Each cycle possesses an exclusive compression system and the refrigeration pressure ratio can be adjusted at will. The second scheme is more flexible than the first, but has the disadvantage that the number of stand-alone equipment is particularly high, which is not conducive to cost control and operational management.

Differently from the pre-cooling process, where there are more refrigerant options available, the refrigerants used for hydrogen cryo-cooling generally include only neon, hydrogen, and helium [48]. The thermodynamic properties of the three refrigerants are different. Hydrogen and helium are more capable of heat transfer, while neon is more compressible [45]. Therefore, better combined performance can be obtained by mixing different refrigerants in certain proportions. It has been shown that the use of mixed refrigerant containing neon, hydrogen and helium was more conducive to reducing energy consumption, by up to 28.4% [121].

3.4. Emerging Refrigeration Technologies

Existing refrigeration technology relies on compressors and expanders. In recent years, several researchers have proposed emerging refrigeration technologies for hydrogen liquefaction by replacing compressors or expanders with other equipment. Expanders used for regular hydrogen liquefaction face problems with high-speed rotation and bearing sealing. Zou et al. [122] designed a supersonic two-phase expander that enabled the direct liquefaction of hydrogen without moving parts. This refrigeration technology with a supersonic nozzle as a core component has been frequently used in natural gas liquefaction research and was expected to be an emerging hydrogen liquefaction technology. Conventional compressor refrigeration technology is less energy efficient, therefore an uncompressed magnetic refrigeration technology (MRT) is starting to be introduced for hydrogen liquefaction. MRT ideally offers higher performance than conventional gas compression refrigeration technology. Tang et al. [123] discovered a material suitable for operation in the full temperature range (20–77 K) required for hydrogen liquefaction. This discovery contributed to the implementation of MRT in hydrogen liquefaction. Ansarinasab et al. [47] compared two hydrogen liquefaction processes using compressor refrigeration technology and MRT, respectively. The results showed that the SEC and the coefficient of performance of the hydrogen liquefaction process using MRT were 6.3% and 18.2% higher than those of the process using compressor refrigeration technology, respectively.

4. Ortho- and Para-Hydrogen Conversion

The greatest difference in the liquefaction of hydrogen from that of other gases lies in the existence of conversion reaction. Due to the fact that the spin directions of the two atoms that make up the hydrogen molecule are both in the opposite direction and in the same direction, there are two forms of hydrogen [39]. The two forms of hydrogen have different energy states and can be converted into each other until equilibrium is reached. The conversion reaction is of secondary order and is extremely slow without catalytic. The equilibrium state of hydrogen is temperature dependent, as shown in Figure 10. The heat generated by the conversion reaction is approximately 527 kJ/kg, which is greater than the latent heat of LH₂, 445 kJ/kg [124]. Therefore, during non-equilibrium hydrogen storage, the heat released from the conversion reaction can result in about 18% boiling after 1 day and about 50% boiling after 1 week [125].

To avoid excessive loss of LH₂ in storage, it is necessary to complete the OPH conversion while the hydrogen is being liquefied [125]. As shown in Figure 11, there are three modes of conversion depending on where the catalyst is inserted in the liquefaction process, namely isothermal, adiabatic, and continuous conversion. The effects of different conversion modes on the energy consumption of hydrogen liquefaction were quantitatively analyzed and discussed by Teng et al. [126]. The SEC of the systems using three different conversion modes was 11.38, 14.56 and 15.97 kWh/kg_LH2, respectively.

4.1. Isothermal Conversion

Isothermal conversion of OPH occurs in a reactor where the catalyst bed is kept at a constant temperature. Therefore, the reactor should be placed in a pool containing boiling refrigerant, as shown in Figure 11a. LN₂ and LH₂ are commonly used as fluid media to maintain isothermal operation of the converter. Due to simpler operation and higher conversion efficiency, two OPH isothermal converters with LN₂ and LH₂ were used in some of the early commercial plants, such as the hydrogen liquefaction plant in Ingolstadt [31]. This approach was also adopted in the medium-scale hydrogen liquefaction concept proposed by Kuz’menko et al. [127]. However, isothermal conversion increases the fixed costs significantly due to the need to equip a boiling pool. Moreover, it is considered to be the least energy efficient method, which is not conducive to the low-energy operation of hydrogen liquefaction plants. Therefore, the isothermal conversion of OPH has almost been replaced by adiabatic and continuous conversion.

4.2. Adiabatic Conversion

The adiabatic conversion of OPH is generally arranged in the converter with a catalyst bed, as shown in Figure 11b. Since there is no heat output, the temperature of hydrogen will rise significantly after the reaction is completed. The equilibrium hydrogen can either enter into the next heat exchanger to be cooled, or can return to the previous heat exchanger to be cooled. Both methods are common in research. There are two reactor models in Aspen HYSYS that can be utilized for adiabatic conversion simulations of OPH. The most widely used is a conversion reactor. For the conversion reactor, the conversion rate can be determined by changing reaction coefficients. The relationship between them is shown in Equation (1) [128].

$$Conversion=C_0+C_1\times T+C_2\times T^2$$

(1)

where $C_0$, $C_1$ and $C_2$ are equilibrium conversion coefficients. T is the temperature of the hydrogen stream.

Conversion reactors are applied in most hydrogen liquefaction studies that employ adiabatic conversion. It is first necessary to determine the temperature of hydrogen at the inlet of each conversion reactor, so that the conversion coefficients can be calculated iteratively to ensure that the outlet hydrogen is in equilibrium. In the subsequent simulation calculations, there is no need to change the reaction coefficients as long as it is ensured that the temperatures of hydrogen at the inlet of all conversion reactors remain constant. However, this imposes severe limitations on the parameter optimization. Since the conversion coefficients cannot be recalculated during the optimization process, the inlet temperatures of conversion reactors cannot be used as variables to be optimized, otherwise the hydrogen at outlet will deviate from equilibrium. This is the reason why some studies did not optimize the inlet temperatures of conversion reactors. Riaz et al. [48] solved this problem by replacing the conversion reactor with an equilibrium reactor.

The equilibrium reactor can recalculate the equilibrium concentration as the hydrogen temperature changes, which in turn can update the heat duty required for the reaction [48]. The temperature-dependent equilibrium hydrogen concentration relationship cannot be directly implanted in the equilibrium reactor. The K_eq, defined as Equation (2), needs to be entered in the equilibrium reactor. K_eq can be calculated from the relationship between the concentrations of OPH in hydrogen. The introduction of equilibrium reactors in the hydrogen liquefaction process makes it possible to optimize the reactor inlet temperatures, thus providing greater flexibility. Currently, this method was only adopted in studies by Naquash et al. [65] and Riaz et al. [48]. Therefore, it needs to be further validated to prove its worthiness to be adopted in more studies.

$$K_{eq}={\displaystyle \frac{n_{p-H_2}}{n_{o-H_2}}}$$

(2)

The number of stages in the converters has a significant impact on energy consumption when adiabatic conversion of OPH is adopted. It has been demonstrated that the SEC of a liquefaction system with five-stage adiabatic converters was 16.39% lower than that of a liquefaction system with a one-stage converter [129]. As the number of adiabatic converters increases, energy consumption decreases more slowly. The energy consumption of a liquefier with five-stage adiabatic converters was only 4.94% higher than that of continuous conversion [130]. Therefore, the number of adiabatic converters should be rationalized. In addition, the conversion temperatures are often chosen to be the temperatures of some of the important streams in the process, and the temperatures of these streams have a significant impact on the liquefaction system. Therefore, the conversion temperatures need to be included within the parameters being optimized.

4.3. Continuous Conversion

Continuous catalytic OPH conversion technology can be traced back to the chemical plant design concept of coupling a strong reaction process and a strong heat transfer process [131]. Based on this concept, the heat exchanger that can realize the continuous conversion of OPH has been proposed. The flow channel of the heat exchanger is filled with a conversion catalyst, as shown in Figure 11c, which enables a continuous conversion of OPH during the flow and heat transfer process [132]. The adoption of continuous conversion in the hydrogen liquefaction process reduces SEC and improves the compactness of the system [133]. The design and development of a continuous conversion heat exchanger relies on the catalytic properties of the catalyst and the flow and heat transfer characteristics of heat exchangers. These are based on a large amount of experimental data and complex numerical simulations. However, there is a lack of data on cryogenic hydrogen conversion, flow, and heat transfer. For example, the hydrogen liquefaction plant built in Germany in 2008 employed OPH continuous conversion technology, but it has not been disclosed in much detail in the literature. This limits the design and development of high-performance continuous conversion heat exchangers.

More research is currently focused on exploring the changes in the flow and heat transfer characteristics of heat exchangers after the addition of catalysts, adopting a combination of experimental and numerical simulations. Commonly used low-temperature heat exchangers include plate-fin heat exchangers and tube-wound heat exchangers. Both of them can be filled with catalysts for the continuous conversion of OPH, as shown in Figure 12. As regular heat exchangers are filled with catalysts, the flow and heat transfer characteristics change dramatically. Therefore, many scholars have conducted extensive research on continuous conversion heat exchangers. First, the mass, momentum and energy equations are combined with the dynamic model of OPH conversion and EOS of hydrogen to construct a mathematical model for solving heat exchangers. A one-dimensional computational model applicable to plate-fin heat exchangers was proposed by Donaubauer et al. [134] and validated with experimental test cases. Park et al. [135] developed a mathematical computational model for the plate-fin heat exchanger channel, but did not couple the conversion of OPH. A combination of numerical simulations and experimental tests was applied by Fan et al. [136] to investigate the performance of a tube-wound heat exchanger at the temperature of LH₂. Xu et al. [137,138] revealed the coupling mechanism of the catalyst-fin-filled channel by comparing existing dynamic models of OPH conversion. Moreover, they utilized the novel thermal-hydraulic correlations instead of Ergun’s and Peter’s equation in the improved numerical model of the plate-fin heat exchanger.

Secondly, continuous conversion heat exchanger performance studies are carried out based on the new model. The tube-wound and plate-fin heat exchangers were analyzed by Skaugen et al. [139], respectively. The results revealed that the viscous resistance, heat transfer temperature difference, and OPH conversion accounted for 84%, 14%, and 2% of the exergy loss generated in the tube-wound heat exchanger and 64%, 30% and 6% in the plate-fin heat exchanger, respectively. The performance of the heat exchanger was evaluated with conversion efficiency and thermal-hydraulic coefficient as performance indicators. After filling with catalyst, the flow properties of the channel deteriorated, leading to an increase in the heat transfer performance and a decrease in the thermal–hydraulic coefficient of heat exchangers [130]. Considering the heat transfer performance, equipment volume, and conversion efficiency, many researchers have concluded that plate–fin heat exchangers were more suitable for the continuous conversion of ortho–para–hydrogen [140,141].

5. Summary and Conclusions

This paper reviews the technological advances in hydrogen liquefaction systems from various aspects of LH₂ production, including hydrogen pre-cooling, hydrogen cryo-cooling, and OPH conversion. The existing hydrogen pre-cooling cycles and their development history are summarized. The performance of different cryo-cooling processes and their improved forms are compared. The conversion modes of OPH are introduced and the current research status of continuous conversion heat exchangers is described. The following is the summary and conclusions:

• The choice of a future hydrogen pre-cooling process depends on the availability of a stable cold source. LN₂ from air separation units and LNG from LNG-receiving stations are the best pre-coolants for hydrogen pre-cooling. Among them, the supply of LN₂ is limited, while the application of LNG is worth anticipating. The co-production of natural gas and LH₂ is an important future direction for large-scale hydrogen liquefaction plants. Strategies for efficient utilization of LNG cold energy when used for hydrogen pre-cooling can be further investigated. For the absence of a cold source, the MRC is almost the only option for hydrogen pre-cooling. The components of mixed refrigerant must be simplified to ensure stable operation of the MRC. New refrigerants can be explored as replacements for existing mixed refrigeration options.
• The Claude cycle and the cascaded J-B refrigeration cycles are the dominant schemes used for hydrogen cryo-cooling. In large-scale hydrogen liquefaction systems, researchers prefer the Claude cycle. The active hydrogen liquefaction plant at Leuna provides good experience. An appropriate flash hydrogen processing method is adopted to achieve 100% hydrogen liquefaction. Thus, the injector may become a central component in future Claude cycles. For small and medium scale hydrogen liquefaction systems, the cascaded J-B refrigeration cycles are more accepted. Whether for the Claude cycle or the cascaded J-B refrigeration cycles, the heat exchange efficiency between hydrogen and refrigerant can be improved by adjusting the number of cascade cycles and the refrigeration temperatures of the individual cycles. Considerable research remains to be performed in this area.
• The continuous conversion of OPH is the only option in the commercial production of LH₂. The other two conversion modes do not meet the requirement for low energy consumption. From the available literature, the continuous conversion of OPH is in the stage of attack of key equipment. Recently, few studies have been reported for experimental testing of OPH conversion. This leads to low predictability and accuracy of kinetic modeling of the continuous conversion of OPH based on the available experimental data. Therefore, the center of future research should be placed on simulations and experiments related to OPH conversion.

In addition to the main scheme described above, BRC, ARC, jet refrigeration cycle, MRT, and other thermodynamic cycles can be involved in hydrogen liquefaction under specific conditions.