Review on the design and optimization of hydrogen liquefaction processes

Abstract

The key technologies of liquefied hydrogen have been developing rapidly due to its prospective energy exchange effectiveness, zero emissions, and long distance and economic transportation. However, hydrogen liquefaction is one of the most energy-intensive industrial processes. A small reduction in energy consumption and an improvement in efficiency may decrease the operating cost of the entire process. In this paper, the detailed progress of design and optimization for hydrogen liquefaction in recent years are summarized. Then, based on the refrigeration cycles, the hydrogen liquefaction processes are divided into two parts, namely precooled liquefaction process and cascade liquefaction process. Among the existing technologies, the SEC of most hydrogen liquefaction processes is limited in the range of 5–8 kWh/ {\mathrm{k}\mathrm{g}}_{{\mathrm{L}\mathrm{H}}_2} : liquid hydrogen). The exergy efficiencies of processes are around 40% to 60%. Finally, several future improvements for hydrogen liquefaction process design and optimization are proposed. The mixed refrigerants (MRs) as the working fluids of the process and the combination of the traditional hydrogen liquefaction process with the renewable energy technology will be the great prospects for development in near future.

Key words： hydrogen liquefaction; energy consumption; efficiency; optimization

1 Introduction

As an energy carrier and the most plentiful element on Earth, hydrogen is primarily derived from water and can address issues of sustainability, environmental emissions, and energy security [1]. The demand for liquid hydrogen (LH₂), particularly driven by clean fuel cell applications, is expected to have a rapid growth [2], and the number of hydrogen refueling stations in the world will also rise sharply in the near future, as shown in Table 1 [3]. In view of a mobility based on hydrogen, the distribution and storage of hydrogen as a liquid is one of the most feasible options in terms of energy density, technical, and economic perspectives [4]. In addition, the density of LH₂ is much higher than gaseous hydrogen, resulting in higher energy content. In the coming decades, innovative energy supplies, advanced energy systems, and upgraded infrastructure will be needed to sustainably meet the increasing energy demands [5]. Hydrogen liquefaction processes will play indispensable roles in clean energy chain.

Tab.1 Number of hydrogen refueling stations in the world in the near future

Area	2017	2020	2025	2030
Japan	100	160	320	900
South Korea	14	80	210	520
China	15	100	350	1000
America	69 (35 in California)	—	320	—
Germany	56	—	400	—
France	3	—	—	400–1000
UK	3	65	300	1150
Denmark	11	15	185	—
Spain	6	20	—	—
Sweden	4	14	—	—
Belgium	—	25	75	—

Liquefaction of hydrogen is a cost-efficient way to store and transport large quantities of hydrogen over extended distances and can offer a low-pressure, high energy density fuel to be used in a variety of applications [6–8]. Hydrogen gas is liquefied when it is cooled down to a temperature below ‒253°C at 101.325 kPa. Hydrogen was first liquefied by Sir James Dewar at 0.24 L/h in 1898 [9]. Hydrogen gas was pressured to 18000 kPa and precooled to ‒250°C with carbonic acid by Dewar. There are five hydrogen liquefaction plants in the Praxair of United States with the production rates between 6 and 35 tons per day (TPD) LH₂ [10]. The specific energy consumptions (SECs) are between 12.5 and 15 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2} [11]. Bracha et al. [12] illustrated the largest German hydrogen liquefier, Linde large-scale N₂ pre-cooled Claude plant in Ingolstadt. The liquefaction capacity of Ingolstadt was 4.4 TPD and the SEC was 13.58 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2}. Shimko and Gardiner [13] developed and modeled a large capacity (50 TPD) hydrogen liquefaction cycle precooled by helium with four ortho-para catalyst beds with 8.73 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2}. Tang [14] proposed a hydrogen liquefaction process based on liquid nitrogen (LN₂) precooling and helium refrigeration with a capacity of 50 TPD. Yuksel et al. [15] analyzed and assessed a novel supercritical hydrogen liquefaction process thermodynamically based on helium cooled hydrogen liquefaction cycles to produce LH₂. Krasae-in et al. [16] explained an LH₂ plant using a mixed refrigerants (MRs) precooling cycle and a four H₂ Joule-Brayton (J-B) cascade refrigeration system. The overall power consumption of the proposed plant was 5.35 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2} and the exergy efficiency (EXE) was 0.54. Ansarinasab et al. [17] analyzed a hydrogen liquefaction plant equipped with the MR system for the liquefaction capacities of 100 TPD.

With the development of hydrogen energy, liquefied hydrogen will play an important role in long-distance and large-scale transportation. However, hydrogen liquefaction is usually a high energy consumption and relatively low efficiency process. Therefore, a review on the current development status of the hydrogen liquefaction process will lay a solid foundation for future development. Specific energy consumption (SEC) and EXE are investigated to confirm whether the performance of hydrogen liquefaction process is improved. Additionally, in order to facilitate the analysis and comparison, the hydrogen liquefaction processes are classified into two main kinds, namely precooled liquefaction process and cascade liquefaction process. Moreover, the process design and optimization approaches of hydrogen liquefaction are summarized and the development prospects for future improvements are proposed as well.

2 Ortho-para hydrogen conversion

Ortho hydrogen and para hydrogen are two spin isomers of molecular hydrogen, which is caused by two possible couplings of the proton nuclear spins of the atoms. As can be seen in Fig. 1, for ortho hydrogen and para hydrogen, two proton nuclear spins are in the same direction and in the opposite direction, respectively. Hydrogen is typically an equilibrium mixture of ortho and para hydrogens. At room temperature in a thermally equilibrium state, the percent of hydrogen composition is approximately 75% ortho hydrogen and 25% para hydrogen. Furthermore, at a temperature of ‒253°C, the composition is almost 100% para hydrogen.

The energy level of ortho hydrogen is higher than that of para hydrogen. In the process of liquefaction hydrogen, as the temperature decreases, the ortho hydrogen will gradually convert into para hydrogen to reach the equilibrium hydrogen. The released heat of 670 kJ/kg is greater than the latent heat of hydrogen evaporation of 452 kJ/kg, which will lead to the evaporation of a large amount of LH₂. This deal heat of transformation can evaporate as much as 50% of the LH₂ over 10 days without catalyst [18]. Therefore, it is necessary to add ortho-para hydrogen converters in the process to accelerate the conversion rate using catalysts. Different conversion catalyst materials can be selected, such as Fe(OH)₃, CrO₂ on Al₂O₃ [19], hydrous ferric-oxide (HFO) [20], Ni on silica catalyst [21], and CrO₃ on silica catalyst [22].

Fig.1 Spin isomers of molecular hydrogen: ortho hydrogen and para hydrogen.

Full size|PPT slide

3 Basic cycles of hydrogen liquefaction

Basic cycles of hydrogen liquefaction include simple Claude, Kapitza, dual pressure Claude, precooled Linde-Hampson, precooled dual-pressure Linde-Hampson, precooled simple Claude, precooled dual pressure Claude, helium-precooled Claude, and precooled mixed refrigerant (MR) cycles [23]. Figure 2 depicts the design of a simple Claude cycle, in which the throttle valve is replaced by the expander to recover a portion of the expansion work and the liquefying fluid itself is used as the refrigerant. The performance of the system will increase if the cycle is pre-cooled with LN₂. Compared to the simple Claude cycle, one heat exchanger (HX) is missed and the cycle pressure is lower in the Kapitza cycle, as shown in Fig. 3. Since the maximum conversion temperature of hydrogen (205 K) is lower than the ambient temperature, pre-cooling must be added to liquefy hydrogen for the simple Linde-Hampson cycle, unlike the simple Claude cycle. Figure 4 is the process flow diagram of an LN₂ pre-cooled Linde-Hampson cycle. In fact, the LH₂ can be obtained at a desirable liquefaction rate only when the pressure is as high as 10–15 MPa and the temperature is reduced to 50–70 K for throttling [14]. A comparison of the basic hydrogen liquefaction cycles is listed in Table 2.

Fig.2 Design of a simple Claude cycle.

Full size|PPT slide

Fig.3 Design of a Kapitza cycle.

Full size|PPT slide

Fig.4 Design of an LN₂ pre-cooled Linde-Hampson cycle.

Full size|PPT slide

Tab.2 Comparison of basic hydrogen liquefaction cycles

Basic hydrogen liquefaction cycle	Liquid yield/%	SEC/( \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2} ‒1)	EXE/%
Simple Claude [24]	8	22.1	18.1
Precooled Linde-Hampson [25]	12–17	72.8–79.8	4.5–5.0
Precooled dual-pressure Linde-Hampson [26]	41	12.14	27
Precooled simple Claude [25]	16–20	28–39.2	9.2–13
Precooled dual pressure Claude [24]	—	12.26	—
Helium-precooled Claude [25]	100	33.6–56	6.5–11

In the liquefaction systems, the SEC, coefficient of performance (COP), and EXE are often used to evaluate the system performance. According to the first law of thermodynamics, the SEC is defined as the rate of total energy consumption in the process to the mass flowrate of the productions [27]. That is

(1) \mathrm{SEC}={\displaystyle \frac{\dot{W}}{{\dot{m}}_{{\mathrm{LH}}_{\mathrm{2}}}}},

where \dot{W} is the net power of the system and {\dot{m}}_{{\mathrm{LH}}_{\mathrm{2}}} is the mass flow rate of the LH₂.

The COP is defined as the ratio of cooling capacity provided by the working fluid to the net input power of the cycle [24], given as

(2) \mathrm{COP}={\displaystyle \frac{{\dot{m}}_{{\mathrm{LH}}_{\mathrm{2}}}\times (h_{\mathrm{feed}}-h_{{\mathrm{LH}}_{\mathrm{2}}})}{\dot{W}}},

where h_{\mathrm{L}{\mathrm{H}}_2} is the enthalpy of the streams and h _feed is the enthalpy of the inlet hydrogen gas.

The EXE for a liquefaction process is the ratio of the ideal liquefaction work to the actual liquefaction consumption [27]. It can be expressed as

(3) \mathrm{EXE}={\displaystyle \frac{W_{\mathrm{rev}}}{W_{\mathrm{act}}}}={\displaystyle \frac{{\dot{m}}_{{\mathrm{LH}}_{\mathrm{2}}}[(h_{{\mathrm{LH}}_{\mathrm{2}}}-h_{\mathrm{0}})-t_{\mathrm{0}}(s_{{\mathrm{LH}}_{\mathrm{2}}}-s_{\mathrm{0}})]}{\dot{W}}},

where t ₀ is the ambient temperature, t ₀ = 25°C, W _rev is the ideal reversible liquefaction work, W _act is the actual liquefaction work, h ₀ and s ₀ are the hydrogen enthalpy and entropy at ambient temperature, and s_{\mathrm{L}{\mathrm{H}}_2} is the entropy of the streams LH₂.

4 Precooled cycles of hydrogen liquefaction

4.1 Precooled liquefaction process for hydrogen

For hydrogen liquefaction process with a wide temperature range, single or multi-stage precooling can increase thermal efficiency and decrease energy consumption. Moreover, by lowering the temperature of the working fluid prior to throttling, the high pressure required in an ideal liquefaction process can be reduced [24].

4.1.1 Nitrogen precooled cycles

Baker and Shaner [28] presented a flowsheet of hydrogen liquefaction with LN₂ precooled dual pressure Claude cycle. The process could produce 250 TPD of LH₂ and the SEC and EXE are 10.85 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2} and 0.36, respectively. Approximately ten of these modules would be needed for a major airport in the 1990s. Bracha et al. [12] illustrated the largest German hydrogen liquefier, Linde large-scale N₂ pre-cooled Claude plant in Ingolstadt. The liquefaction capacity of Ingolstadt was 4.4 TPD and the SEC was 13.58 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2}. Apart from the liquefier, the entire plant includes an air separation plant purification unit, a high pressure hydrogen compressor station, an LH₂ storage tank, an LN₂ tank, and filling stations for GH₂ and LH₂ trailers. The World Energy NET work (WE-NET) project [29] suggested building large-scale hydrogen liquefaction plants based on a Claude cycle with nitrogen pre-cooling. This cycle was similar to the plant in Ingolstadt, while the liquefaction capacities of it were greater (300 TPD). Kuzmenk et al. [30] designed, calculated, and compared four versions of hydrogen liquefiers with similar capacities (5.4 TPD). In version 4, the delivered LN₂ with a SEC of 0.5 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2} was used as the precooling cycle and the helium gas was the refrigerant of the liquefaction cycle. The SEC and the EXE of version 4 were 12.7 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2} and 0.346, respectively.

Tang [14] proposed a hydrogen liquefaction process based on LN₂ precooling and helium refrigeration with a capacity of 50 TPD. He used MATLAB programming to simulate the process and simulated two cases of Joule-Thomson (J-T) throttle valve and expander for final expansion. The result showed that the exergy efficiencies were 38.52% and 40.17% for the final stage of the J-T valve and expander process. Garceau et al. [31] built and investigated a small scale hydrogen liquefaction plant with a single stage Gifford-McMahon cryocooler and LN₂ precooler. The process flow diagram of the completed system is shown in Fig. 5. Two ortho-para hydrogen converters were placed in an LN₂ precooler where H₂ was precooled and converted, and then went into the liquefier. After liquefaction, LH₂ can be transferred to a 5 L storage vessel designed to limit the evaporation of LH₂, using two 1 m long G10-CR necks to reduce conduction, a radiation shield, and multi-layer insulated (MLI) with high vacuum to reduce radiation and convective heat leak from the outer shell to the inner tank. Hammad and Dincer [32] proposed and analyzed an advanced hydrogen liquefaction system with catalyst infused heat exchangers (HXs). The hydrogen gas was precooled by the running LN₂, refrigerated through expansion of high pressure hydrogen gas (2 MPa) in three expanders placed in series and liquefied by the J-T cycle. The simulation of process was developed in Aspen Plus and the EXE was 0.1158.

Fig.5 Process flow diagram of completed system (reprinted with permission from Ref. [31].)

Full size|PPT slide

4.1.2 Helium precooled cycles

Shimko and Gardiner [13] (2008) developed and modeled a large capacity (50 TPD) hydrogen liquefaction cycle precooled by helium with four ortho-para catalyst beds. The study showed that the SEC of this process was 8.73 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2} and the EXE was 0.446. Staats et al. [33] proposed and analyzed a supercritical hydrogen liquefaction cycle numerically with a helium pre-cooling Claude cycle. The liquefaction capacity was still 50 TPD and three ortho-para catalyst beds were needed. A simulation program was written in MATLAB to investigate the effects of altering component efficiencies and various system parameters on the cycle efficiency. The result showed that the four-stage helium expansion cycle was efficient, at 0.356, among the two-, three- and four-stage cycles and it was expected that adding another stage to a cycle would not be worth the considerable associated increase in capital cost. In addition, in order to determine an effective way to improve performance, they changed the parameters by manual optimization and re-conducted the simulation at these new operating points. Finally, the result indicated that the efficiency could possibly be made as high as 0.440 with increases in HX area.

Yuksel et al. [15] analyzed and assessed a novel supercritical hydrogen liquefaction process thermodynamically based on helium cooled hydrogen liquefaction cycles to produce LH₂, which is modified from the process of Staats et al. [33], as demonstrated in Fig. 6. In this process, the four helium expanders produce electricity to power the hydrogen compressor and helium compressor to pressurize the hydrogen and helium flows to the required pressure. Hydrogen was liquefied by four-stage helium expansion and refrigeration, passing through eight HXs. The liquefaction capacity, energy and exergy efficiencies of the liquefaction process were found to be 50 TPD, 70.12% and 57.13%, respectively. They studied some parameters to test the effects of different design variables on the EXE and exergy destruction rates of the hydrogen liquefaction process. The results illustrated that any increase in hydrogen mass flow rate and the inlet pressure of helium expanders resulted in increasing the process EXE. On the other hand, a 14°C increase in the pinch temperature of the catalyst bed led to an increase in the exergy destruction rates of the liquefaction process and a decrease in the EXE.

Fig.6 Schematic diagram of super-critical hydrogen liquefaction process (reprinted with permission from Ref. [15].)

Full size|PPT slide

4.1.3 J-B precooled cycles

To meet WE-NET project objectives, Matsuda and Nagamei [34] compared four J-B precooling Claude cycles with different refrigerants, namely hydrogen, helium, base neon, and neon with cold pump J-B precooling Claude. The liquefaction capacity of the four cases is 300 TPD. They summarized that the refrigerant and HXs cost of the hydrogen case was minimum and its compressors and safety cost were maximum. The HXs cost of the helium case was minimum. The refrigerant, maintenance, and plant cost of the neon with cold pump case was maximum and the operation, management of refrigerant and expanders cost was minimum. Accordingly, the last one for the neon with cold pump case had the best SEC of 8.49 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2} and best EXE of 0.471, compared with 8.60 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2} and 0.464 for the hydrogen case, 8.76 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2} and 0.456 for the helium case, and 8.65 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2} and 0.462 for the base neon case. Quack [35] proposed an ethane-propane J-B precooling Claude cycle with helium-neon refrigerant in the cryogenic section, as exhibited in Fig. 7. This result in SEC was 5 to 7 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2}, depending on the pressure of the feed and the temperature of the product. Besides, the EXE of the total plant in the order of 0.60 was feasible. Valenti [36] selected a cascade of four helium reversed, closed, and recuperative J-B cycles to liquefy hydrogen gas, with a capacity of 864 TPD LH₂. It was simulated on a computer with the aid of the commercial software Aspen Plus, developed by Aspen Tech, which included a wide databank and robust routines for fluid property calculations. The computed SEC was 5.04 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2} and EXE was 0.477.

Fig.7 Overall flow diagram (reprinted with permission from Ref. [35].)

Full size|PPT slide

4.1.4 MR precooled cycles

Stang et al. [37] conducted a conceptual plant based on the MR cycle process for precooling hydrogen and a liquid helium bath was used in the cryogenic section. The proposed plant showed that an EXE of 0.60 and SEC of 7.0 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2} was feasible. Krasae-in et al. [16] explained a LH₂ plant using a MR precooling cycle and a four H₂ J-B cascade refrigeration system. The cycle that was capable of producing 100 TPD of LH₂ was simulated in Aspen HYSYS. The MR system with ten components was used to precool feed normal hydrogen gas from 25°C to the equilibrium temperature of - 193°C with a high efficiency. In addition, for the transition from the equilibrium temperature of the hydrogen gas of - 193°C to - 253°C, the proposed four H₂ J-B cascade refrigeration system was recommended. The overall power consumption of the proposed plant was 5.35 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2} and the EXE was 0.54. They indicated that the sizes of HXs, compressor motors, and crankcase compressors were smaller. Krasae-in [38] gave an improvement for more realistic large-scale LH₂ cycles based on a previously published work. A simplified five-component mixture of refrigerant (4% hydrogen, 18% nitrogen, 24% methane, 28% ethane, and 26% butane) and a four H₂ J-B cycle refrigeration system with optimization were proposed, as displayed in Fig. 8. The most important of this paper was the inclusion of a preliminary study of the variables and constraints, as well as methods to adjust for optimal steady-state operation. He programmed a C-language source code to simulate and optimize the hydrogen liquefaction process by trial and error. The optimization process contains 23 variables and 26 constraints. The result of the optimization was that the SEC and EXE of this plant were 5.91 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2} and 0.489.

Sadaghiani and Mehrpooya [39] introduced and analyzed a novel cryogenic hydrogen liquefaction process configuration, as shown in Fig. 9. They used the MR with nine components for the precooling cycle and MR with three components (10% neon, 6.5% hydrogen and 83.5% helium) which could produce 300 TPD of LH₂ for the cryogenic J-B cycle in the process. They simulated the process by using the Aspen HYSYS software. Besides, they adjusted the operating condition of the equipment to the optimum values for the process optimization, and then studied the sensitivity of the process outputs versus effective parameters. The SEC and EXE of the process were 4.410 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2} and 0.5547. The energy analysis revealed that the COP of the proposed process was 0.1797. Asadnia and Mehrpooya [40] proposed and analyzed a novel hydrogen liquefaction process configuration with a MR of 11 components for the precooling cycle and an MR of two components (hydrogen and helium) for cryogenic J-B cycles. The Aspen HYSYS simulator was also utilized to partial validating of thermodynamic properties of the process in this paper and the feed flow rate was 100 TPD. The result of the simulation specified that the total SEC of the plant was 7.69 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2}, the EXE was 0.395, and the COP was 0.1710.

Cardella et al. [41] outlined a novel approach to process development for large-scale hydrogen liquefaction which could produce 100 TPD of LH₂. In this paper, a new process simulation model was adopted and implemented in the modeling software UniSim Design for the process calculation of the entire hydrogen liquefaction process. Moreover, the process simulation coupled to MATLAB was optimized by the gradient-based fmincon solver with the sequence quadratic programming algorithm. The SEC was selected as the objective function, 18 variables were to be optimized, and 22 nonlinear inequality constraints were chosen to represent reasonable boundary conditions. Finally, two hydrogen liquefaction processes precooled with MR were proposed. One was designed with a high-pressure hydrogen Claude cycle for refrigeration and liquefaction, while the other was devised with a dual H₂-Ne cycle. The result demonstrated that the SEC of the two processes was around 6.0 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2}.

Fig.8 Flowsheet for large-scale 100 TPD LH₂ plant utilizing MR and four H₂ J-B refrigeration cycles (reprinted with permission from Ref. [38].)

Full size|PPT slide

Fig.9 Flowsheet of proposed process for liquefaction of hydrogen (reprinted with permission from Ref. [39].)

Full size|PPT slide

4.1.5 Liquefied natural gas (LNG) precooled cycles

Kuendig et al. [42] analyzed a large-scale hydrogen liquefaction plant integrated with an LNG precooling plant for 50 TPD. The LNG unloaded from the tanker ships as saturated liquid at ambient pressure was compressed by the pumps and the temperature of the LNG was increased to -148°C. Then, the hydrogen was precooled to -138°C by the LNG in the HX. Further cooling of the LH₂ was conducted by the nitrogen and hydrogen refrigeration loops step by step. Kramer [43] implied that the SEC of this process could be reduced to 4 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2}. However, Krasae-in et al. [10] held that the process could only be used for hydrogen gas made from LNG, and the plant should be located near a seaport.

4.2 Cascade liquefaction process for hydrogen

The cascade process is actually an extension of the precooling system where each individual refrigeration cycle precools the other refrigeration cycle or uses multiple auxiliary refrigeration cycles to precool the primary gas stream [23].

Ansarinasab et al. [17] analyzed a hydrogen liquefaction plant equipped with an MR system for the liquefaction capacities of 100 TPD. In the process, five refrigeration cycles were adopted to provide the required coldness. Four of five refrigeration cycles were single stage expansion refrigeration system which used pure hydrogen as refrigerant, while the fifth one was an MR cycle composed of five components, which was cooled by a throttle valve to provide cooling energy to the main hydrogen stream. To perform the energy, exergy, and economic analyses, the system was simulated in Aspen HYSYS.

Ansarinasab et al. [44] analyzed a conventional hydrogen liquefaction process of 300 TPD based on exergy, economy, and environment. As illustrated in Fig. 10, the liquefaction process was composed by three sections, one main section containing pure hydrogen stream and two auxiliary designed sections in order to provide the coldness to the main stream and work as a refrigeration process. The first stage MR refrigeration cycle consisting of nine components cooled the temperature of hydrogen to -195°C. In the second refrigeration stage, the temperature of hydrogen was decreased to - 254.55°C by the hydrogen of 6.19%, the helium of 83.61%, and the neon of 10.20%. They adopted the Aspen HYSYS software to simulate the hydrogen liquefaction process. Besides the traditional energy and exergy analysis, they conducted an economic analysis and applied the life cycle assessment (LCA) approach to evaluate the environmental impact of the studied hydrogen liquefaction. Figure 11 showed a proposed framework for evaluating the economic and environmental impacts of a hydrogen liquefaction process. The amount of SEC, EXE and COP of this process was 1.102 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2}, 0.5547, and 0.1797, respectively.

Fig.10 Process flow diagram of proposed liquefaction cycle (reprinted with permission from Ref. [44].)

Full size|PPT slide

Fig.11 Proposed framework for evaluating economic and environmental impacts of a selected hydrogen liquefaction process (reprinted with permission from Ref. [44].)

Full size|PPT slide

Aasadnia and Mehrpooya [45] proposed a novel process which involved an absorption refrigeration system that cooled a part of hydrogen streams in the precooling and liquefied sections of the process. The schematic diagram of the proposed plant was illustrated in Fig. 12. The process included an MR refrigeration cycle with 11 components and an absorption refrigeration cycle that precooled the temperature of feed hydrogen from 25°C to -199.9°C. A new MR with 6.5% of hydrogen, 83.5% of helium, and 10% of neon was used in a cascade J-B cycle that deep-cooled the low-temperature gaseous hydrogen from -199.9°C to -252.2°C in the liquefied section of the plant. The process was simulated in the Aspen HYSYS simulator and the production rate of the LH₂ was 90 TPD. Furthermore, the process was optimized by trial and error, which was a functional and simple method of complex system analysis. The optimization result of SEC was as high as 6.47 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2}. The EXE of the process was evaluated to be 0.455 and the COP of the overall system was 0.2034.

Fig.12 A hybrid of MR hydrogen liquefaction system, MR cascade J-B refrigeration cycle, and an absorption refrigeration cycle (Reprinted with permission from Ref. [45].)

Full size|PPT slide

5 Results and discussion

Numerous types of researches for hydrogen liquefaction have been discussed. Different structures, operating fluids, and working conditions lead to different SEC, EXE, COP and so on. Table 3 lists the differences between those proposed systems for hydrogen liquefaction based on key parameters. It can be found that in the previous hydrogen liquefaction processes, the nitrogen precooled and MR precooled cycles are the most adopted, and the precooling MR cycle has more applications in the last decade. At present, the SEC of most hydrogen liquefaction cycles is mainly in the range of 5 to 8 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2} and the EXE of the cycles is around 40% to 60%. For long-distance and large-scale transportation, non-flammable, non-toxic refrigerants such as nitrogen and helium should be first considered. The J-B cycles and MRs should be selected for onshore large-scale hydrogen liquefaction systems. The cascade cycles are suitable for hybrid liquefaction systems combining renewable energy technologies with traditional technologies.

Tab.3 Differences between proposed hydrogen liquefaction cycles

Proposed hydrogen liquefaction cycle		Capacity/TPD	Liquid yield/%	SEC/( \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2} ‒1)	EXE/%	Remarks
Nitrogen precooled cycles	Baker and Shaner [28]	250	100	10.85	36.0	H2 expander and compressor isentropic efficiency (eff.) = 0.79, liquefier max p = 4137 kPa
	Bracha et al. [12]	4.4	100	13.58		Lngolstadt by Linde
	WE-NET project [29]	300	100			Process efficiency>0.40, liquid pressure= 105 kPa
	Kuzmenk et al. [30]	5.4	100	12.7	34.6	H2 compressor isothermal eff. = 0.6, He compressor isothermal eff. = 0.53, N2 compressor isothermal eff. = 0.53, N2 expanders isothermal eff. = 0.73–0.75, He expanders isothermal eff. = 0.85
	Tang [14]	50	100		40.17	H2 compressor isothermal eff. = 0.6, He compressor isothermal eff. = 0.8, He expanders isothermal eff. = 0.85, wet expanders isothermal eff. = 0.80, liquefier max p = 2100 kPa
	Hammad and Dincer [32]				11.58	liquefier max p = 2000 kPa
Helium precooled cycles	Shimko and Gardiner [13]	50	100	8.73	44.6	expander isentropic eff. = 0.83–0.86, wet expander isentropic eff. = 0.90, He and H2 compressor eff. = 0.8
	Staats et al. [33]	50	100		35.6	expander isentropic eff. = 0.83–0.86, wet expander isentropic eff. = 0.90, He compressor isothermal eff. = 0.80, H2 compressor isothermal eff. = 0.60
	Yuksel et al. [15]	50	100		57.13	He and H2 compressors isentropic eff. = 0.80, He expanders eff. = 0.85
J-B precooled cycles	Matsuda and Nagamei [34]	300	100	8.49	47.1	expanders isentropic eff. = 0.85 compressor isentropic eff. = 0.80, liquefier max p ≈ 5000 kPa
	Quack [35]	173	100	5–7	60.7	expanders isentropic eff. = 0.85–0.90, compressor isentropic eff. = 0.85, liquefier max p ≈ 8000 kPa
	Valenti [36]	864	100	5.04	47.7	He expanders isothermal eff. = 0.88–0.93, He compressors polytrophic eff. = 0.92, H2 expander polytrophic eff. = 0.85, J-B cycle max p = 4000 kPa
MR precooled cycles	Stang et al. [37]		100	7.0	60.0	—
	Krasae-in et al. [16]	100	100	5.35	54.0	Ten-component mixture, compressors and expanders isentropic eff. = 0.80
	Krasae-in [38]	100	100	5.91	48.9	Five-component mixture, J-B cycles max p = 4000 kPa, MR cycle max p = 1800 kPa
	Sadaghiani and Mehrpooya [39]	300	100	4.41	55.47	Expanders adiabatic eff. = 0.85, compressor isentropic eff. = 0.90, J-B cycles max p = 1000 kPa, MR cycle max p = 1600 kPa
	Asadnia and Mehrpooya [40]	100	100	7.69	39.5	Expander and compressor isentropic eff. = 0.80, J-B cycles max p = 3000 kPa, MR cycle max p = 1805 kPa
	Cardella et al. [41]	100	100	6	—	Expanders isentropic eff. = 0.78–0.88, compressor isentropic eff. = 0.76–0.86
LNG precooled cycles	Kuendig et al. [42]	50	100	4	—	—
Cascade cycles	Ansarinasab et al. [17]	100	—	—	—	Expanders isentropic eff. = 0.80, compressor isentropic eff. = 0.90, MR cycle max p = 1800 kPa
	Ansarinasab et al. [44] (2019)	300	100	1.102	55.47	H2 cycle max p = 2100 kPa, MR cycle max p = 1600 kPa
	Aasadnia and Mehrpooya [45]	90	100	6.47	45.5	Expander and compressor isentropic eff. = 0.80, H2 cycle max p = 2100 kPa, MR cycle max p = 1805 kPa

Multigenerational systems with high efficiency are a potential route for the green supply energy chain and the studies of using various renewable energies in hydrogen production and liquefaction systems were proposed by Muradov and Veziroğlu [46]. Yilmaz and Kaska [47] modeled and analyzed a hydrogen liquefaction system with an absorption precooling cycle assisted by geothermal water. High-temperature geothermal water in the absorption refrigeration cycle was used to reduce electricity power consumption. Yilmaz [48] analyzed the thermodynamic performance and life cycle cost of a geothermal energy-assisted hydrogen liquefaction system based on his previous research. Yuksel [49] presented the definition of multigenerational sub-systems and working conditions to analyze a thermodynamic investigation of the solar concentrating system based integrated system for electricity, hydrogen, heating-cooling, dry product, fresh water, and hot water production. Zhang et al. [50] reviewed the typical magneto caloric materials and prototypes in the temperature range of nitrogen and hydrogen liquefaction.

From the review of the above hydrogen liquefaction processes, it can be found that most researchers used the chemical software such as Aspen HYSYS, Aspen Plus, Pro/II, and UniSim Design even C language programs written by themselves for simulation calculation. Meanwhile, new hydrogen liquefaction processes for the choice of different refrigerants and the modification of the processes were proposed. In addition, most studies focused on energy and exergy analyses, some of which also conducted economic and environmental analyses. From an optimization perspective, the total energy consumption and SEC were selected as objective functions in most optimization studies. The lack of optimization studies of the total cost resulted from the complexity of estimation of the equipment costs, like that for the LNG process [51]. Some optimization methods, such as trial and error, the sequence quadratic programming algorithm, and the genetic algorithm could provide better optimization results.

Reducing energy consumption and increasing efficiency still play a significant role in the hydrogen liquefaction processes. Several promising developments in the design and optimization of future hydrogen liquefaction processes may include:

1) MRs usually use hydrocarbons and nitrogen as refrigerants. In the hydrogen liquefaction process, the MR liquefaction process provides cooling to hydrogen from ambient temperature to a lower temperature all the time. By optimizing the component ratio of the MR, the cold and hot composite curves of hydrogen and MR can be better matched to reduce energy consumption.

2) With the increasing demand for hydrogen and the diversity of hydrogen utilization, small and medium scale hydrogen liquefaction processes have attracted increasing attentions. On the other hand, different design features should be considered in comparison to traditional large-scale hydrogen liquefaction processes to accommodate the difficulties in selecting the appropriate equipment, like that for the LNG process [52].

3) Most previous studies choose energy consumption as the optimization objective function. However, operational cost is an important parameter in the hydrogen liquefaction process. In one research, capital expenditures, energy expenditures, and operational and maintenance expenditures are reported as 63%, 29%, and 8% of the total cost of liquefaction respectively [11]. Economic analysis has been discussed in a part of studies, and the capital expenditures should be further optimized as an objective function. It is best to consider both the capital expenditures and the SEC of the liquefied system.

4) In the review of the hydrogen liquefaction process, dynamic simulations are not mentioned. As the parameters of the hydrogen liquefaction process are different with time, static simulations do not provide a deeper understanding of the process. Additionally, dynamic simulation can be used for robust control structures in the design process for hydrogen liquefaction, and the dynamic simulation has been adopted in the LNG process [51]. Unlike static simulation, dynamic simulation requires the knowledge of the detailed design parameters of the device.

5) Considering the highly nonlinear processes of hydrogen liquefaction, better optimization methods should be developed. In previous studies, many researchers optimized important parameters manually, which was time-consuming and inevitably inaccurate. Developing a new and better optimization algorithm to obtain global optimum parameter will be beneficial to process design and optimization.

6) Renewable energy such as solar energy and geothermal energy can be coupled with the hydrogen liquefaction process to decrease energy consumption to some extent.

6 Conclusions

The detailed progresses on the design and optimization of hydrogen liquefaction in recent years are summarized and comprehensively reviewed in this paper. The basic cycles of hydrogen liquefaction are introduced and compared. The cycles can be classified into two sections based on the theoretical cycles of hydrogen, namely precooled liquefaction process and cascade liquefaction process. In the previous hydrogen liquefaction processes, the nitrogen precooled and MR precooled cycles are the most adopted and the precooling MR cycles have been widely applied in the last decade. In addition, the SEC, EXE, and COP for different hydrogen liquefaction processes are discussed. Considering the existing technologies, the SEC of most hydrogen liquefaction processes are limited in the range of 5–8 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2} and only a small part of SEC are achieved around 1 \mathrm{k}\mathrm{W}\mathrm{h}/\mathrm{k}{\mathrm{g}}_{\mathrm{L}{\mathrm{H}}_2}, while the EXE of most processes are around 40% to 60%. Moreover, the latest technologies in the literature are also addressed and the challenges and future improvements are identified. The MRs as the working fluids of the liquefaction process and the combination of the traditional hydrogen liquefaction process with the renewable energy technology will be the great prospects for development in near future.

References

1 Turner J A. Sustainable hydrogen production. Science, 2004, 305(5686): 972–974

2 Cardella U, Decker L, Klein H. Roadmap to economically viable hydrogen liquefaction. International Journal of Hydrogen Energy, 2017, 42(19): 13329–13338

3 LBST, Hydrogen Refueling Station Worldwide. 2019–06–18

4 Valenti G. Hydrogen liquefaction and liquid hydrogen storage. In: Gupta R B, Basile A, Nejat T. Compendium of Hydrogen Energy. Woodhead Publishing, 2016, 27–51

5 Acar C, Dincer I. Hydrogen energy. In: Dincer I, eds. Comprehensive Energy Systems, Elsevier, 2018, 1: 568–605

6 Ball M, Wietschel M. The Hydrogen Economy: Opportunities and Challenges. Cambridge: Cambridge University Press, 2009

7 Li M, Bai Y F, Zhang C Z, Song Y, Jiang S, Grouset D, Zhang M. Review on the research of hydrogen storage system fast refueling in fuel cell vehicle. International Journal of Hydrogen Energy, 2019, 44(21): 10677–10693

8 Nash D, Aklil D, Johnson E, Gazey R, Ortisi V. Hydrogen storage: compressed gas. Comprehensive Renewable Energy, 2012, 4: 131–155

9 Dewar J. Liquid hydrogen. Science, 1900, 11(278): 641–651

10 Krasae-in S, Stang J H, Neksa P. Development of large-scale hydrogen liquefaction processes from 1898 to 2009. International Journal of Hydrogen Energy, 2010, 35(10): 4524–4533

11 Drnevich R. Hydrogen delivery: liquefaction and compression. In: Strategic initiatives for Hydrogen Delivery Workshop, Tonawanda, NY, USA, 2003

12 Bracha M, Lorenz G, Patzelt A, Wanner M. Large-scale hydrogen liquefaction in Germany. International Journal of Hydrogen Energy, 1994, 19(1): 53–59

13 Shimko M, Gardiner M. Innovative hydrogen liquefaction cycle. FY 2008 Annual Progress Report, DOE Hydrogen Program, 2008

14 Tang L. Hydrogen liquefaction process design and system simulation based on liquid nitrogen precooling. Dissertation for the Master Degree. Hangzhou: Zhejiang University, 2011, 23–32

15 Yuksel Y E, Ozturk M, Dincer I. Analysis and assessment of a novel hydrogen liquefaction process. International Journal of Hydrogen Energy, 2017, 42(16): 11429–11438

16 Krasae-in S, Stang J H, Neksa P. Simulation on a proposed large-scale liquid hydrogen plant using a multi-component refrigerant refrigeration system. International Journal of Hydrogen Energy, 2010, 35(22): 12531–12544

17 Ansarinasab H, Mehrpooya M, Mohammadi A. Advanced exergy and exergoeconomic analyses of a hydrogen liquefaction plant equipped with mixed refrigerant system. Journal of Cleaner Production, 2017, 144: 248–259

18 Stetson N T, Olson G L, Bowman R C Jr. Overview of hydrogen storage, transportation, handling and distribution. In: Sherif S A, Goswami D Y, Stefanakos E K. et al. Handbook of Hydrogen Energy. CRC Press, 2014: 567–592

19 Vander Arend P. Large-scale liquid hydrogen production. Chemical Engineering Progress, 1961, 57(10): 62–67

20 Hutchinson H L. A kinetics study of the para-ortho shift of hydrogen. Dissertation for the Master Degree. Colorado: University of Colorado, 1964

21 Silvera I F. The solid molecular hydrogens in the condensed phase: fundamentals and static properties. Reviews of Modern Physics, 1980, 52(2): 393–452

22 Sullivan N, Zhou D, Edwards C. Precise and efficient in situ ortho-para-hydrogen converter. Cryogenics, 1990, 30(8): 734–735

23 Barron R F. Liquefaction cycles for cryogens. In: Timmerhaus K D. Advances in Cryogenic Engineering. Boston: Springer, 1972, 20–36

24 Aasadnia M, Mehrpooya M. Large-scale liquid hydrogen production methods and approaches: a review. Applied Energy, 2018, 212: 57–83

25 Nandi T K, Sarangi S. Performance and optimization of hydrogen liquefaction cycles. International Journal of Hydrogen Energy, 1993, 18(2): 131–139

26 Peschka W. Liquid Hydrogen: Fuel of the Future. Springer Science & Business Media, 2012

27 Yin L, Ju Y L. Comparison and analysis of two nitrogen expansion cycles for BOG Re-liquefaction systems for small LNG ships. Energy, 2019, 172: 769–776

28 Baker C R, Shaner R L. A study of the efficiency of hydrogen liquefaction. International Journal of Hydrogen Energy, 1978, 3(3): 321–334

29 Mitsugi Ch, Harumi A, Kenzo F W E N E T. Japanese hydrogen program. International Journal of Hydrogen Energy, 1998, 23(3): 159–165

30 Kuz’menko I F, Morkovkin I M, Gurov E I. Concept of building medium-capacity hydrogen liquefiers with helium refrigeration cycle. Chemical and Petroleum Engineering, 2004, 40(1/2): 94–98

31 Garceau N M, Baik J H, Lim C M, Kim S Y, Oh I H, Karng S W. Development of a small-scale hydrogen liquefaction system. International Journal of Hydrogen Energy, 2015, 40(35): 11872–11878

32 Hammad A, Dincer I. Analysis and assessment of an advanced hydrogen liquefaction system. International Journal of Hydrogen Energy, 2018, 43(2): 1139–1151

33 Staats W L. Analysis of a supercritical hydrogen liquefaction cycle. Dissertation for the Master Degree. Massachusetts: Massachusetts Institute of Technology, 2008

34 Matsuda H, Nagami M. Study of large hydrogen liquefaction process. Hydrogen Energy, 1997, 8: 175–175

35 Quack H. Conceptual design of a high efficiency large capacity hydrogen liquefier. AIP Conference Proceedings, 2002, 613: 255–263

36 Valenti G, Macchi E. Proposal of an innovative, high-efficiency, large-scale hydrogen liquefier. International Journal of Hydrogen Energy, 2008, 33(12): 3116–3121

37 Stang J, Neksa P, Brendeng E. On the design of an efficient hydrogen liquefaction process. In: 16th World Hydrogen Energy Conference 2006 (WHEC 2006), Lyon, France, 2006: 1–6

38 Krasae-in S. Optimal operation of a large-scale liquid hydrogen plant utilizing mixed fluid refrigeration system. International Journal of Hydrogen Energy, 2014, 39(13): 7015–7029

39 Sadaghiani M S, Mehrpooya M. Introducing and energy analysis of a novel cryogenic hydrogen liquefaction process configuration. International Journal of Hydrogen Energy, 2017, 42(9): 6033–6050

40 Asadnia M, Mehrpooya M. A novel hydrogen liquefaction process configuration with combined mixed refrigerant systems. International Journal of Hydrogen Energy, 2017, 42(23): 15564–15585

41 Cardella U, Decker L, Sundberg J, Klein H. Process optimization for large-scale hydrogen liquefaction. International Journal of Hydrogen Energy, 2017, 42(17): 12339–12354

42 Kuendig A, Loehlein K, Kramer G, Huijsmans J. Large scale hydrogen liquefaction in combination with LNG re-gasification. In: 16th World Hydrogen Energy Conference 2006 (WHEC 2006), Lyon, France, 2006: 3326–3333

43 Kramer G J, Huijsmans J, Austgen D. Clean and green hydrogen. In: 16th World Hydrogen Energy Conference 2006 (WHEC 2006), Lyon, France, 2006: 3317–3325

44 Ansarinasab H, Mehrpooya M, Sadeghzadeh M. An exergy-based investigation on hydrogen liquefaction plant-exergy, exergoeconomic, and exergoenvironmental analyses. Journal of Cleaner Production, 2019, 210: 530–541

45 Aasadnia M, Mehrpooya M. Conceptual design and analysis of a novel process for hydrogen liquefaction assisted by absorption precooling system. Journal of Cleaner Production, 2018, 205: 565–588

46 Muradov N Z, Veziroğlu T N. “Green” path from fossil-based to hydrogen economy: an overview of carbon-neutral technologies. International Journal of Hydrogen Energy, 2008, 33(23): 6804–6839

47 Yilmaz C, Kaska O. Performance analysis and optimization of a hydrogen liquefaction system assisted by geothermal absorption precooling refrigeration cycle. International Journal of Hydrogen Energy, 2018, 43(44): 20203–20213

48 Yilmaz C. Optimum energy evaluation and life cycle cost assessment of a hydrogen liquefaction system assisted by geothermal energy. International Journal of Hydrogen Energy, 2019

49 Yuksel Y E, Ozturk M, Dincer I. Energetic and exergetic assessments of a novel solar power tower based multigeneration system with hydrogen production and liquefaction. International Journal of Hydrogen Energy, 2019, 44(26): 13071–13084

50 Zhang H, Gimaev R, Kovalev B, Kamilov K, Zverev V, Tishin A. Review on the materials and devices for magnetic refrigeration in the temperature range of nitrogen and hydrogen liquefaction. Physica B, Condensed Matter, 2019, 558: 65–73

51 He T B, Karimi I A, Ju Y L. Review on the design and optimization of natural gas liquefaction processes for onshore and offshore applications. Chemical Engineering Research & Design, 2018, 132: 89–114

52 Kwak D H, Heo J H, Park S H, Seo S J, Kim J K. Energy-efficient design and optimization of boil-off gas (BOG) re-liquefaction process for liquefied natural gas (LNG)-fuelled ship. Energy, 2018, 148: 915–929