Review of Recent Advances in Transcritical CO 2 Heat Pump and Refrigeration Cycles and Their Development in the Vehicle Field

: Refrigerant substitution is an urgent need in the context of reducing carbon emissions and slowing global warming. CO 2 is now being proposed as a promising solution based on its excellent properties and system performance, especially in low-temperature environments. This paper presents an overview of recent advances in system conﬁguration and operation characteristics to improve the performance of transcritical CO 2 heat pump and refrigeration systems. The paper ﬁrst introduces the basic research background, system cycle, and thermodynamic characteristics. Secondly, CO 2 cycle improvements with single modiﬁcations and modiﬁcation combinations are reviewed. Then, some important operation characteristics and control methods are discussed. Additionally, the paper provides a detailed description of the development of transcritical CO 2 heat pump and refrigeration systems in the vehicle ﬁeld. At the end of this review, conclusions and opportunities for future work in this ﬁeld are presented.


Introduction
Though the substitution of hydrofluorocarbons (HFCs) for hydrochlorofluorocarbons (HCFCs) has been beneficial for ozone layer protection, HFCs contribute considerably to global warming due to their high global warming potential (GWP) [1].The Kigali Amendment to the Montreal Protocol and the Paris Agreement further restrict the use of HFCs (such as HFC-32 or HFC-134a) since 2019 to achieve carbon neutrality and slow the rate of global warming [2].Thus, refrigerants with substantially lower GWP, such as HFC-1234yf, HFC-1234ze, or CO 2 , have recently received increased attention.Among these low-GWP refrigerants, CO 2 shows such potential that some people regard it as an ultimate solution.CO 2 is safe and clean because of its properties of being non-toxic and non-flammable, with 0 ozone depletion potential (ODP) and 1 GWP [3].CO 2 also exhibits competitive cooling performance and outstanding heating performance compared with other refrigerants [4].As a common byproduct of many industrial processes, CO 2 is cheap and easy to obtain.These unique properties make CO 2 a promising refrigerant in refrigeration and heat pump systems.
As a natural working fluid, carbon dioxide has been used as a refrigerant since the 19th century.However, it was replaced with synthetic refrigerants such as chlorofluorocarbons (CFCs) and HCFCs due to its poor subcritical cycle performance.The CO 2 refrigerant did not draw any further attention from researchers until a new transcritical CO 2 cycle was proposed by Pettersen and Lorentzen in the 1990s [5].They validated it in a prototype and pointed out that the CO 2 transcritical system has several advantages in terms of cost, performance, and dimensions.Since then, research into the use of the CO 2 transcritical cycle in various applications has continued.
There are several review papers that focus on CO 2 refrigerant or heat pump systems, including cycle design and improvements, heat transfer characteristics, components, or applications.Ma et al. submitted a comprehensive review including the novel CO 2 transcritical cycle, component improvements, and CO 2 heat transfer characteristics in the supercritical or two-phase region [6].They suggested that CO 2 should be a more common refrigerant for the further restriction of harmful emissions.Their article mainly focused on cycle modifications and components.Ehsan et al. reviewed the supercritical CO 2 heat transfer and pressure drop characteristics in various applications [7].The tube diameter, heat flux, and operating mass velocity were considered to be important factors that influence heat transfer effectiveness.Yu et al. presented a review that focused on the cycle modifications made in recent years and demonstrated the merits or drawbacks of different improvements [8], but they did not consider the combination of different modifications.Dilshad et al. also introduced the changing process of refrigerants and different CO 2 cycle modification measures in their review.However, only applications for commercial use worldwide were highlighted in their review [9].Barta et al. summarized that CO 2 vapor compression cycles are highly effective for stationary applications in nearly all climates.In transportation applications, the transcritical CO 2 cycle is competitive but without general recognition [10].Soylemez et al. provided an overview of the application of the CO 2 refrigeration system on fishing vessels.Three major operating conditions-low seawater temperatures, high freezing capacity, and additional freezing capacity in high seawater temperatures-are considered suitable for CO 2 refrigeration system application [11].Gullo et al. concluded that the multi-ejector solution is the main driving force for the spread of CO 2 supermarket refrigeration systems, but stores in areas of high ambient temperatures lack confidence in CO 2 refrigeration systems [12].Bruno et al. presented a comprehensive review that included cycle modifications, exergy analysis, modeling tools, and operation considerations [13].Although that article was more comprehensive than other reviews, its focus was not on modification combinations or operation optimization.Some other reviews related to CO 2 refrigeration and heat pump systems include refs.[14,15].
The abovementioned review papers mainly focused on cycle modifications and performance analysis but contain little content about CO 2 cycle operation characteristics, control strategies, and configuration modification combinations.Thus, this paper aims to provide a comprehensive, state-of-the-art review of these topics for the CO 2 cycle.This review will discuss refrigerant charges, operation process characteristics, and multi-modification configurations.In addition, we present considerable information about the development of CO 2 heat pump systems in the vehicle field, considering the enhanced attention.Guidance for research on CO 2 cycle operation optimization and configuration modification combinations are provided.
This paper starts with an introduction to the basic thermodynamic characteristics and description of the transcritical CO 2 cycle in Section 2. This is followed by a comprehensive presentation in Section 3 on system configuration improvements, especially modification combinations.Then, the operation characteristics and control methods of the transcritical CO 2 cycle including modification configurations are discussed in Section 4. In Section 5, the applications of CO 2 refrigerant or heat pump systems in the vehicle field are discussed.Finally, we present our opinions on the outlook of CO 2 refrigeration and heat pump fields.

Thermodynamic Characteristics and Cycle Introduction
A basic transcritical CO 2 cycle employs the same devices as a conventional cycle, namely, a compressor, condenser, expansion valve, and evaporator.Thus, the thermodynamic diagrams are similar.The p-h diagrams of the transcritical CO 2 cycle and conventional subcritical cycle are shown in Figure 1.
Processes 1-2 and 1 -2 are isentropic compression processes, processes 3-4 and 3 -4 are adiabatic expansion processes, and processes 4-1 and 4 -1 are evaporation processes under constant pressure.The main difference between the two cycles is the heat rejection process.In the conventional subcritical cycle, the heat rejection process (process 2-3) occurs   C, 7.38 MPa) in the subcritical region.This process is usually referred to as the condensation process since it involves a phase transition in the twophase region.In contrast, in the transcritical CO 2 cycle, the heat rejection process (process 2 -3 ), which is called the gas cooling process, takes place above the critical point in the supercritical region [16].During this process, the distinction between liquid and gas phases vanishes, and the temperature of the refrigerant continues to decrease due to the absence of the two-phase region.This temperature glide feature of CO 2 is advantageous as it fits the temperature profile of the coolant side, contributing to the transcritical CO 2 cycle's higher efficiency [17].Another noteworthy advantage of the CO 2 cycle is its high compression efficiency resulting from its lower compression ratio [18].Additionally, the large-volume refrigeration capacity of CO 2 makes for more compact heat exchanger [19].Processes 1-2 and 1′-2′ are isentropic compression processes, processes 3-4 and 3′-4′ are adiabatic expansion processes, and processes 4-1 and 4′-1′ are evaporation processes under constant pressure.The main difference between the two cycles is the heat rejection process.In the conventional subcritical cycle, the heat rejection process (process 2-3) occurs below the critical point (31.2 °C, 7.38 MPa) in the subcritical region.This process is usually referred to as the condensation process since it involves a phase transition in the two-phase region.In contrast, in the transcritical CO2 cycle, the heat rejection process (process 2′-3′), which is called the gas cooling process, takes place above the critical point in the supercritical region [16].During this process, the distinction between liquid and gas phases vanishes, and the temperature of the refrigerant continues to decrease due to the absence of the two-phase region.This temperature glide feature of CO2 is advantageous as it fits the temperature profile of the coolant side, contributing to the transcritical CO2 cycle's higher efficiency [17].Another noteworthy advantage of the CO2 cycle is its high compression efficiency resulting from its lower compression ratio [18].Additionally, the large-volume refrigeration capacity of CO2 makes for more compact heat exchanger [19].
In contrast to the subcritical region, the pressure and temperature of the refrigerant in the supercritical region are two independent variables [20].Thus, the optimum high pressure and refrigerant temperature at the outlet of the gas cooler are the two most important cycle characteristics in the transcritical CO2 cycle for improving system performance.
With a drastic change in pressure along the isothermal curve as it approaches the critical point, an optimum coefficient of performance (COP) exists at a constant refrigerant outlet temperature of the gas cooler, a phenomenon reported in several papers.Sarkar and Agrawals optimized a parallel-compression transcritical CO2 cycle and developed a formula related to the temperature at the evaporator and gas cooler exit [21].Chen and Gus reported that the compressor discharge pressure has a significant influence on the system performance of a CO2 refrigerant system, whereas the evaporating temperature has little effect on the optimum high pressure [22].Zhang et al. validated the existence of In contrast to the subcritical region, the pressure and temperature of the refrigerant in the supercritical region are two independent variables [20].Thus, the optimum high pressure and refrigerant temperature at the outlet of the gas cooler are the two most important cycle characteristics in the transcritical CO 2 cycle for improving system performance.
With a drastic change in pressure along the isothermal curve as it approaches the critical point, an optimum coefficient of performance (COP) exists at a constant refrigerant outlet temperature of the gas cooler, a phenomenon reported in several papers.Sarkar and Agrawals optimized a parallel-compression transcritical CO 2 cycle and developed a formula related to the temperature at the evaporator and gas cooler exit [21].Chen and Gus reported that the compressor discharge pressure has a significant influence on the system performance of a CO 2 refrigerant system, whereas the evaporating temperature has little effect on the optimum high pressure [22].Zhang et al. validated the existence of an optimum high pressure in a CO 2 heat pump system with two throttle valves [23].Optimizing high pressure is an important and common parameter for operational optimization.Aprea and Maiorinos claimed that optimizing high pressure is an easy way to improve CO 2 air conditioning system performance [24].
Furthermore, a literature review of the optimum high pressure reveals that the temperature of the gas cooler exit is an important influence factor on the optimum high pressure and system performance.Figure 2 shows variations in the optimum high pressure and COP with changing gas cooler exit temperature at a 15 • C evaporating temperature.
The system COP increases with the decrease in the gas cooler exit temperature.Conversely, the optimum high pressure increases as the gas cooler exit temperature increases.This phenomenon is particularly well-documented in high ambient temperature conditions.Both Laipradit et al. [25] and Sawalha [26] have previously reported on this phenomenon in their works.Liang et al. presented that the gas cooler exit temperature accounts for 96.38% of the variation in the optimum high pressure for the transcritical CO 2 cycle using response surface methodology [27].Consequently, the ability of the transcritical CO 2 cycle to provide effective cooling is a persistent and widespread challenge.Furthermore, a literature review of the optimum high pressure reveals that the temperature of the gas cooler exit is an important influence factor on the optimum high pressure and system performance.Figure 2 shows variations in the optimum high pressure and COP with changing gas cooler exit temperature at a 15 °C evaporating temperature.The system COP increases with the decrease in the gas cooler exit temperature.Conversely, the optimum high pressure increases as the gas cooler exit temperature increases.This phenomenon is particularly well-documented in high ambient temperature conditions.Both Laipradit et al. [25] and Sawalha [26] have previously reported on this phenomenon in their works.Liang et al. presented that the gas cooler exit temperature accounts for 96.38% of the variation in the optimum high pressure for the transcritical CO2 cycle using response surface methodology [27].Consequently, the ability of the transcritical CO2 cycle to provide effective cooling is a persistent and widespread challenge.

Transcritical CO2 Cycle Improvements
The unique characteristics of the transcritical CO2 cycle necessitate modifications to CO2 heat pump or refrigeration systems in order to enhance system efficiency.These modifications can be broadly classified into three categories.
The first category involves decreasing the gas cooler outlet temperature.As explained in the thermodynamic analysis of the transcritical CO2 cycle in Section 2, the gas cooler outlet temperature has a significant impact on the system COP.Decreasing this temperature can increase system COP significantly, particularly in high ambient temperatures.To achieve this, common modifications include the use of an internal heat exchanger (IHX) [28,29], a cascade cycle [30], thermoelectric subcooling [31], and magnetic subcooling [32].
The second category involves improving the throttling process.In the transcritical cycle, the pressure difference on either side of the throttling process is considerable, resulting in higher exergy loss compared to other common refrigerant cycles.To reduce the exergy loss, measures such as the use of an ejector [33] or an expander [34] are considered effective.
The third category involves improving the compression process.When the evaporative temperature decreases, the density of CO2 vapor also decreases, resulting in a small mass flow rate in a transcritical CO2 cycle system operating in a low-temperature environment.Flash gas bypass [35] and parallel compression [36] are commonly used to address this issue.Alternatively, a two-stage compression process [37] can provide a more efficient

Transcritical CO 2 Cycle Improvements
The unique characteristics of the transcritical CO 2 cycle necessitate modifications to CO 2 heat pump or refrigeration systems in order to enhance system efficiency.These modifications can be broadly classified into three categories.
The first category involves decreasing the gas cooler outlet temperature.As explained in the thermodynamic analysis of the transcritical CO 2 cycle in Section 2, the gas cooler outlet temperature has a significant impact on the system COP.Decreasing this temperature can increase system COP significantly, particularly in high ambient temperatures.To achieve this, common modifications include the use of an internal heat exchanger (IHX) [28,29], a cascade cycle [30], thermoelectric subcooling [31], and magnetic subcooling [32].
The second category involves improving the throttling process.In the transcritical cycle, the pressure difference on either side of the throttling process is considerable, resulting in higher exergy loss compared to other common refrigerant cycles.To reduce the exergy loss, measures such as the use of an ejector [33] or an expander [34] are considered effective.
The third category involves improving the compression process.When the evaporative temperature decreases, the density of CO 2 vapor also decreases, resulting in a small mass flow rate in a transcritical CO 2 cycle system operating in a low-temperature environment.Flash gas bypass [35] and parallel compression [36] are commonly used to address this issue.Alternatively, a two-stage compression process [37] can provide a more efficient compression process for lower pressure ratios and can serve as another improvement.These above improvements have been proved to be useful and are commonly used.Bellos and Tzivanidiss compared the performance of transcritical CO 2 refrigeration systems with internal heat exchangers, mechanical subcooling, parallel compressors, and two-stage compression.Their simulation results showed that all of the modifications make system performance better than the base system and that the mechanical subcooling performs the best [38].Liu et al. replaced the auxiliary compressor and high pressure control valve with an ejector.This replacement reduces the maximum power consumption by 22.8% in the transcritical refrigeration cycle [39].A detailed description of these improvements can be found in the comprehensive reviews of Yu et al. [8] or Bruno et al. [13].It should be noted that some new modifications based on these existing systems have been proposed and proved to be effective, such as the integrated mechanical subcooling system [40] and multi-Energies 2023, 16, 4011 5 of 21 ejector expansion system [41].This section highlights the effectiveness of the combined use of these individual modification technologies.
Although every single improvement has been proved to be effective in promoting performance, the specific process influenced by each modification varies.Furthermore, the effects of combining different modifications are still unclear.Consequently, the system performance influenced by combinations of two or more modifications has been studied by many researchers.For example, as an easy and useful measurement, IHXs are taken as a familiar improvement in all sorts of transcritical CO 2 cycles; thus, the simultaneous use of an IHX and other modifications is one of the most common combinations.The combination of an IHX and an ejector contributes more to the system COP than using either of them alone [42].However, not all combinations of IHXs and other modifications result in improved system performance.Adding an IHX can actually decrease the COP of a basic cycle with an expander, as shown in several studies [43].
Single and combination modifications are categorized and visualized in Figure 3. Every circle represents a category of modification and their intersection signifies a combination of modifications.
performance better than the base system and that the mechanical subcooling performs the best [38].Liu et al. replaced the auxiliary compressor and high pressure control valve with an ejector.This replacement reduces the maximum power consumption by 22.8% in the transcritical refrigeration cycle [39].A detailed description of these improvements can be found in the comprehensive reviews of Yu et al. [8] or Bruno et al. [13].It should be noted that some new modifications based on these existing systems have been proposed and proved to be effective, such as the integrated mechanical subcooling system [40] and multi-ejector expansion system [41].This section highlights the effectiveness of the combined use of these individual modification technologies.
Although every single improvement has been proved to be effective in promoting performance, the specific process influenced by each modification varies.Furthermore, the effects of combining different modifications are still unclear.Consequently, the system performance influenced by combinations of two or more modifications has been studied by many researchers.For example, as an easy and useful measurement, IHXs are taken as a familiar improvement in all sorts of transcritical CO2 cycles; thus, the simultaneous use of an IHX and other modifications is one of the most common combinations.The combination of an IHX and an ejector contributes more to the system COP than using either of them alone [42].However, not all combinations of IHXs and other modifications result in improved system performance.Adding an IHX can actually decrease the COP of a basic cycle with an expander, as shown in several studies [43].
Single and combination modifications are categorized and visualized in Figure 3. Every circle represents a category of modification and their intersection signifies a combination of modifications.Table 1 summarizes some of the combination modifications applied to the transcritical CO 2 cycle and their corresponding major findings, particularly in energy efficiency.The modifications were implemented in a cycle consisting of a compressor, a gas cooler, an expansion valve, and an evaporator.The "+" means a combination of two modifications.
It should be noted that ejectors (EJEs) and IHXs are the two commonly used modifications to improve the transcritical CO 2 cycle performance because of their simplicity and low cost.The vortex tube (VT) has not been extensively studied with other modifications due to the lack of research.Combinations of modifications in the same category are relatively uncommon compared to those consisting of modifications from different categories.These modifications are mainly employed in refrigeration scenarios where cooling performance under high-temperature environments is poor.Therefore, thorough evaluation is necessary before implementing these modification systems in other settings.• PC has a similar performance to PCR, better than TS + FGB.

•
PC is more beneficial for lower temperature applications compared with PCR.

IHX + EJE
[42] • The additional IHX is beneficial to both the basic cycle and ejector system's COP.

•
The maximum COP is about 27% higher than the basic cycle.
Experimental Refrigeration EXP + IHX [43][44][45][46] • The use of an IHX increases the COP of the throttling valve cycle while decreasing the COP of the expander cycle.

•
The new cycle presents 9% less COP improvement than the cycle with an expander only.

•
The modification increases the COP by a maximum of 31.49% in the warm period and the heat capacity by a maximum of 12% in the cold period.

•
The application of the expander produces a larger COP due to work recovery; the expander with an IHX has a lower COP compared with the expander system but has a higher COP compared with the IHX system.

Simulation Refrigeration
Dual EJE + IHX [48] • The compressor pressure ratio is reduced by up to 19.1% and the COP is 15.9-27.1% higher than the IHX system.• The new cycle improves the COP slightly more than the cycle with an IHX only, but it can decrease the discharge temperature and achieve a higher cooling capacity.
Experimental Refrigeration BS + PC + DMS [52,53] • The CC behaves better than other configurations at high ambient temperatures at the cost of worse performance below the ambient temperature of 14 • C.

•
The system with DMS presents better performance than the system with PC at high temperatures.

•
High subcooling can also make the performance of DMS in warm climates better.

Simulation Refrigeration TSMI + EJE + IHX [54]
• The new cycle behaves better under low-pressure working conditions.Simulation Refrigeration TS + PC + SAPCC [55] • The new system has the greatest COP and the energy-saving share is improved with the increase of solar radiation.

•
The improvements are influenced by location and climate.

Simulation Refrigeration TSFI + EXP
[56] • The TSFI system has a larger COP than the TSI system.

•
The complete substitution of the expander for the throttling valve leads to a smaller improvement in the TSFI system compared to the improvement in the TSI system.• Compared with the TSI system with or without an IHX, the COP of these new modifications is significantly improved.However, after multi-parameter optimization, the addition of an IHX is harmful to the new system's COP in low-temperature refrigeration, regardless of the IHX's position Simulation Refrigeration PC +EJE + BS [58] • PC and ejector are profitable to decrease exergy destruction and can be used for a better exergy efficiency simultaneously.
Enhancing the parallel compressor efficiency can make system performance better.An overfed evaporator has little effect on exergy destruction decrease.
Simulation Refrigeration PC + BS + multi EJE + IHX + EC [59] • With an IHX and EC, the COP is improved by up to 40% and the power input ratio is reduced by up to 26%.

TES + IHX [60]
• The COP improvement is less when the thermoelectric works as a generator compared with the cycle modified with an IHX only.

•
The system COP is reduced when the thermoelectric works as a cooler because the thermoelectric consumes additional power.
Simulation Refrigeration FGB + CC + BS [61] • Considering annual energy consumption, the PC + BS system is suitable for moderate conditions while the FGB + PC + BS system is suitable for warm conditions.

•
The PC + BS system is suitable for convenience stores due to lower energy consumption under both moderate and warm climates.

Simulation Refrigeration EJE + DMS
[62] • The system with an ejector and DMS presents better performance than the control group (TES + EJE), especially in warm and hot regions.

•
The performance is influenced by the mass ratio of the mixed refrigerant in DMS.

simulation TS + two EJE [63]
• The COP is improved by up to 80% compared with the basic system when the new system works under high ambient temperatures.

•
The COP ratio of the new system to the conventional cycle rises above 2 in specific situations Simulation Refrigeration TS + IHX + two EJE [64] • Without an IHX, the COP is 10.4% on average and the volumetric heating capacity is 3.6-8.2%higher than the TS system under different evaporation temperatures.

•
With an IHX, the COP is 10.5-30.6% and the volumetric heating capacity is 6.6-25.9%higher than the TS system.

Optimal-Pressure-Seeking Methods
Due to the great significance of optimum high pressure in system performance, the search for optimum high pressure has been a heated research area in the transcritical CO 2 cycle.Inokuty utilized a graphical method to transform the search for the optimum pressure into a curve comparison in the p-h diagram of CO 2 [65].The assumption underlying the research is that the pressure and temperature at the end of refrigeration and the temperature at the beginning of throttling expansion are known values.Kauf considered the graphical method as time-consuming; thus, he proposed a correlation to determine the optimum pressure based on the ambient temperature or the gas cooler exit refrigerant temperature [66].Liao et al. found that the value of optimum high pressure is influenced by several key factors, including compressor performance, evaporation temperature, and gas cooler outlet temperature [67].They established a model involving these parameters and validated it with a standard deviation of less than 1%.Several similar correlations have also been developed separately by different researchers [68][69][70].Cabello et al. compared the correlations proposed by Liao, Chen, Kauf, and Sarkar and the results showed that Sarkar's correlation provides the most precise results [71].Yin et al. selected four characteristic variables-the evaporative temperature, the gas cooler exit temperature, the ambient temperature, and the water temperature of the outlet-as important parameters using the group method of data handling (GMDH).The discharge pressure is obtained through particle swarm optimization-back propagation (PSO-BP) [72].This method was later validated by Song et al. experimentally [73].Model predictive control (MPC) was also applied in optimum high pressure search [74].These offline methods, which rely on prior data, are precise enough in the corresponding systems and operation conditions, but only in particular conditions.The predefined parameters in such empirical correlations are influenced strongly by the system and operating conditions, and therefore, when applied to other systems or conditions, these models may fail to satisfy the requirement.Figure 4 shows a comparison between the optimal value and some correlation results.The maximum COP deviation is −5.4%.Cecchinato et al. proposed that the correlations obtained from a specific system must be assessed before they are used in another system or condition [75].Shao et al. considered the influence of high pressure on COP under different working conditions and suggested that a constrained optimal high pressure equation can result in a minor COP decrease despite significant pressure reduction [76].Yang et al. optimized the optimum pressure correlation through minimizing the difference between the actual COP and the maximum COP [77].selected four characteristic variables-the evaporative temperature, the gas cooler exit temperature, the ambient temperature, and the water temperature of the outlet-as important parameters using the group method of data handling (GMDH).The discharge pressure is obtained through particle swarm optimization-back propagation (PSO-BP) [72].This method was later validated by Song et al. experimentally [73].Model predictive control (MPC) was also applied in optimum high pressure search [74].These offline methods, which rely on prior data, are precise enough in the corresponding systems and operation conditions, but only in particular conditions.The predefined parameters in such empirical correlations are influenced strongly by the system and operating conditions, and therefore, when applied to other systems or conditions, these models may fail to satisfy the requirement.
Figure 4 shows a comparison between the optimal value and some correlation results.The maximum COP deviation is −5.4%.Cecchinato et al. proposed that the correlations obtained from a specific system must be assessed before they are used in another system or condition [75].Shao et al. considered the influence of high pressure on COP under different working conditions and suggested that a constrained optimal high pressure equation can result in a minor COP decrease despite significant pressure reduction [76].Yang et al. optimized the optimum pressure correlation through minimizing the difference between the actual COP and the maximum COP [77].In recent years, many new online methods have been applied to the optimum high pressure search under a wide operation range.In contrast to offline methods, online methods generally do not rely on models or correlations.Online methods obtain the optimal high pressure through real-time optimization techniques, such as based on the system transient response or by comparing the results prior to and following optimization.A correlation-free control method based on the steepest descent method was validated by Zhang and Zhang [78].According to the relationship between the COP and compressor energy consumption, Penarrocha et al. transformed the target of maximizing the COP to minimizing energy consumption with the 'perturb and observe' algorithm [79].Cecchinato et al. obtained a system model using an artificial neural network (ANN) and determined the optimal pressure in real time via PSO [80].Extremum seeking control (ESC) has In recent years, many new online methods have been applied to the optimum high pressure search under a wide operation range.In contrast to offline methods, online methods generally do not rely on models or correlations.Online methods obtain the optimal high pressure through real-time optimization techniques, such as based on the system transient response or by comparing the results prior to and following optimization.A correlation-free control method based on the steepest descent method was validated by Zhang and Zhang [78].According to the relationship between the COP and compressor energy consumption, Penarrocha et al. transformed the target of maximizing the COP to minimizing energy consumption with the 'perturb and observe' algorithm [79].Cecchinato et al. obtained a system model using an artificial neural network (ANN) and determined the optimal pressure in real time via PSO [80].Extremum seeking control (ESC) has been utilized for optimal pressure seeking in several applications [81][82][83][84].The searching process of ESC is shown in Figure 5.The settling time in the start phase and sudden change Energies 2023, 16, 4011 9 of 21 phase are about 5000 s and 1000 s, respectively.Kim et al. suggested that ESC could be time-consuming because it depends on the system's transient response [85].These online methods present excellent performance but require a lengthy searching process.Thus, their practical use remains unrealistic.
Energies 2023, 16, x FOR PEER REVIEW 9 of 21 been utilized for optimal pressure seeking in several applications [81][82][83][84].The searching process of ESC is shown in Figure 5.The settling time in the start phase and sudden change phase are about 5000 s and 1000 s, respectively.Kim et al. suggested that ESC could be time-consuming because it depends on the system's transient response [85].These online methods present excellent performance but require a lengthy searching process.Thus, their practical use remains unrealistic.To combine the technical merits of offline methods and online methods, hybrid methods have become the focus of attention.Okasha et al. employed a non-dominated sorting genetic algorithm-II (NSGA-II) technique to find the best high pressure and then optimized it with an online optimizer [86].Ji et al. compared a hybrid method with traditional offline and online methods, and the new method presented high efficiency and fast speed under various operating conditions [87].The hybrid method has potential to be a feasible solution for optimum high pressure search with fast speed and high precision.The variation of optimum-high-pressure-seeking methods is shown in Table 2. Little attention has been given to the transition between subcritical and transcritical cycles in the literature.Jin et al. activated the optimal gas cooler pressure control only when the high pressure exceeded the minimum operating pressure in winter and fell below the maximum operating pressure in summer for efficiency and safety reasons [88].Ge and Tassou proposed a floating discharge pressure control strategy in which the setting value is a function of ambient temperature [89].Shao and Zhang found that the optimal To combine the technical merits of offline methods and online methods, hybrid methods have become the focus of attention.Okasha et al. employed a non-dominated sorting genetic algorithm-II (NSGA-II) technique to find the best high pressure and then optimized it with an online optimizer [86].Ji et al. compared a hybrid method with traditional offline and online methods, and the new method presented high efficiency and fast speed under various operating conditions [87].The hybrid method has potential to be a feasible solution for optimum high pressure search with fast speed and high precision.The variation of optimum-high-pressure-seeking methods is shown in Table 2. been utilized for optimal pressure seeking in several applications [81][82][83][84].The s process of ESC is shown in Figure 5.The settling time in the start phase and change phase are about 5000 s and 1000 s, respectively.Kim et al. suggested could be time-consuming because it depends on the system's transient respo These online methods present excellent performance but require a lengthy search cess.Thus, their practical use remains unrealistic.To combine the technical merits of offline methods and online methods, hybr ods have become the focus of attention.Okasha et al. employed a non-dominate genetic algorithm-II (NSGA-II) technique to find the best high pressure and th mized it with an online optimizer [86].Ji et al. compared a hybrid method with tr offline and online methods, and the new method presented high efficiency and fa under various operating conditions [87].The hybrid method has potential to be a solution for optimum high pressure search with fast speed and high precision.T tion of optimum-high-pressure-seeking methods is shown in Table 2. Little attention has been given to the transition between subcritical and tra cycles in the literature.Jin et al. activated the optimal gas cooler pressure con when the high pressure exceeded the minimum operating pressure in winter an low the maximum operating pressure in summer for efficiency and safety reasons and Tassou proposed a floating discharge pressure control strategy in which th value is a function of ambient temperature [89].Shao and Zhang found that the Little attention has been given to the transition between subcritical and transcritical cycles in the literature.Jin et al. activated the optimal gas cooler pressure control only when the high pressure exceeded the minimum operating pressure in winter and fell below the maximum operating pressure in summer for efficiency and safety reasons [88].Ge and Tassou proposed a floating discharge pressure control strategy in which the setting value is a function of ambient temperature [89].Shao and Zhang found that the optimal high pressure appears in the subcritical region adjacent to the critical point [90].Considering Energies 2023, 16, 4011 10 of 21 pressure control devices, the TXV (thermostatic expansion valve) is not suitable in the transcritical CO 2 cycle for controlling varied objects.Presently, the EXV (electric expansion valve) has emerged as the main pressure-adjustable component due to its precision and large regulating range [91].The adjustable needle [92] and vortex control [93] are receiving great interest for their high work recovery efficiency.

Refrigerant Charge Characteristics
Irrespective of system types and operating conditions, the charge amount of the refrigerant is a crucial performance-determining factor [94,95].In the case of CO 2 , system performance is more sensitive to refrigerant charge amount [96].Figure 6 shows the COP ratio variation caused by charge deviation in R22, R410A, R407C, and CO 2 systems.
Energies 2023, 16, x FOR PEER REVIEW 10 of 21 high pressure appears in the subcritical region adjacent to the critical point [90].Considering pressure control devices, the TXV (thermostatic expansion valve) is not suitable in the transcritical CO2 cycle for controlling varied objects.Presently, the EXV (electric expansion valve) has emerged as the main pressure-adjustable component due to its precision and large regulating range [91].The adjustable needle [92] and vortex control [93] are receiving great interest for their high work recovery efficiency.

Refrigerant Charge Characteristics
Irrespective of system types and operating conditions, the charge amount of the refrigerant is a crucial performance-determining factor [94,95].In the case of CO2, system performance is more sensitive to refrigerant charge amount [96].Figure 6 shows the COP ratio variation caused by charge deviation in R22, R410A, R407C, and CO2 systems.[100,101].Wang et al. studied the effect of refrigerant charge in an electric bus CO2 air conditioning system and supposed that overcharge has a more substantial impact than insufficient charge on system exergy efficiency [102].Yin et al. discussed the refrigerant charge amount under different working conditions and proposed that the optimum charge is not a definite value, but a range of charge amounts [103].Kim mentioned that varying the refrigerant charge or inside volume of the high side are proper methods to control the high side pressure [20].A system with a low-pressure buffer or medium-pressure buffer was introduced to control high side charge.Changing the high side volume is also an effective way to regulate the high side pressure.He et al. analyzed the refrigerant charge distribution in a transcritical CO2 heat pump water heater [104].The simulation results showed that the optimum COP reaches at an optimal high pressure while the high pressure is influenced by the refrigerant charge amount.A system charge management method based on heat exchanger resizing or additional gas cooler volume has been proposed to improve system performance without controlling the high pressure.Refrigerant migration and distribution characteristics are not only important to this method, but also the main reason for cycling loss [105].Wang et al. studied the CO2 transient migration behavior under the period of start-up and shutdown experimentally with the help of a visualized accumulator [106].The results show that the system reveals rapid refrigerant migration and transient response because CO2 exists in a superheated vapor state after the  [100,101].Wang et al. studied the effect of refrigerant charge in an electric bus CO 2 air conditioning system and supposed that overcharge has a more substantial impact than insufficient charge on system exergy efficiency [102].Yin et al. discussed the refrigerant charge amount under different working conditions and proposed that the optimum charge is not a definite value, but a range of charge amounts [103].Kim mentioned that varying the refrigerant charge or inside volume of the high side are proper methods to control the high side pressure [20].A system with a low-pressure buffer or medium-pressure buffer was introduced to control high side charge.Changing the high side volume is also an effective way to regulate the high side pressure.He et al. analyzed the refrigerant charge distribution in a transcritical CO 2 heat pump water heater [104].The simulation results showed that the optimum COP reaches at an optimal high pressure while the high pressure is influenced by the refrigerant charge amount.A system charge management method based on heat exchanger resizing or additional gas cooler volume has been proposed to improve system performance without controlling the high pressure.Refrigerant migration and distribution characteristics are not only important to this method, but also the main reason for cycling loss [105].Wang et al. studied the CO 2 transient migration behavior under the period of start-up and shutdown experimentally with the help of a visualized accumulator [106].The results show that the system reveals rapid refrigerant migration and transient response because CO 2 exists in a superheated vapor state after the shutdown of the system for CO 2 mobile air conditioning (MAC).Thus, better transient performance and a higher energy-saving effect are achieved.A simulation study of refrigerant distribution and migration characteristics was executed by Wang et al. [107].The ambient temperature and air volume flow rate on both the gas cooler side and evaporator side affect the CO 2 distribution, evidently, and should be treated properly during the system design and operation.Refrigerant leakage, responsible for over 20% of the lost refrigerant charge escaping into the atmosphere annually, results in significant system performance degradation [108].To the best of the authors' knowledge, the refrigerant charge fault detection of the CO 2 system is rarely reported in papers.Uren et al. applied a method based on the energy graph on CO 2 transcritical heat pump fault diagnosis [109].Yin recommended that the discharge temperature, suction temperature, and the temperature at the inlet of EXV are the best parameters for appropriate charge judgment [103].

Operation Control in Modified Cycles
Optimal pressure and refrigerant charge amount are two common characteristics in the operation process of all transcritical CO 2 cycles.In the modified transcritical CO 2 cycle, some extra parameters are needed to control for better performance.
The system using the ejector has limitations on high pressure control and expansion work recovery due to the constant ejector geometry.Liu et al. investigated the performance enhancement with the substitution of a controllable ejector in a transcritical CO 2 cycle [110,111].The ejector has two adjustable parameters, suction nozzle throat area and motive nozzle throat area.They found that the ejection ratio and COP are dependent on the motive nozzle throat diameter.The motive nozzle efficiency and the suction nozzle efficiency are affected by their respective diameter to mixing section diameter ratios.To improve the adaptability of the ejector, a multi-ejector system with several parallel working ejectors was proposed.Every ejector can be activated independently, offering several possible combinations under different operation conditions [112].
For parallel compression or multi-stage compression systems, the interstage pressure provides an additional degree of freedom to optimize system performance.Cecchinato et al. demonstrated that the inter-stage cooling during compression is more beneficial than staged throttling in transcritical CO 2 cycles.Moreover, they noted that the optimal intermediate pressure is sensitive to cooling agent temperature [113].Singh et al. showed that the optimum interstage pressure is not the geometric mean interstage pressure in a multi-stage refrigeration system, no matter with flash gas intercooler (FGI), flash gas bypass (FGB), or intercooler (IC).Additionally, they found that the percentage deviation is influenced by gas cooler outlet temperature [47].
Nebot et al. described an integrated mechanical subcooling CO 2 refrigeration cycle that replaces the dedicated mechanical cycle with an internal cycle.They pointed out that not only the discharge pressure but also the subcooling degree are needed for optimization.The subcooling degree is influenced by both the evaporation temperature and the environment temperature.Nonetheless, high pressure has a greater impact on the COP than the subcooling degree [114].A similar system was presented by Song et al. and the system performance was analyzed via simulation.The medium temperature, defined as the EXV inlet temperature, affects the main and sub-cycles' performance in opposite ways [115].Figure 7 shows the COP variation of different cycles under different medium temperatures.
They further presented that the COP of the main cycle does not follow a particular pattern with the intermediate temperature under different operating conditions.Conversely, the sub-cycle COP increases with the increase of intermediate temperature.Less than 5% variation of the whole system COP was observed under different operating conditions [116].They further presented that the COP of the main cycle does not follow a particular pattern with the intermediate temperature under different operating conditions.Conversely, the sub-cycle COP increases with the increase of intermediate temperature.Less than 5% variation of the whole system COP was observed under different operating conditions [116].

Application of CO2 Cycle in the Vehicle Field
The use of transcritical CO2 cycles in automotive AC systems dates back to thirty years ago [5].From then on, several simulations or experiments were executed by different researchers [117][118][119].Mathur compared the thermodynamic performance of automotive air conditioning (AC) systems using R134a and CO2 as working fluids [120].The study revealed that the COP of the CO2 system is approximately 12% lower than that of the R134a system.Liu et al. found that the lubricant type and refrigerant mass charge affect the CO2 system performance significantly [121].Niu et al. validated a wet-compression absorption CO2 refrigeration system with a sustainable cooling capacity for vehicle AC.The high pressure reduces to less than 35 bar [122].These studies demonstrate that CO2 is an acceptable but not very excellent refrigerant for vehicle air conditioning.Hence R134a plays the primary role in vehicle refrigerants.
However, with the development of electric vehicles, the heat shortage challenge stands out in extremely cold environments.The R134a heat pump or positive temperature coefficient (PTC) consumes enormous amounts of energy; thus, the driving range of electric vehicles decreases rapidly [123,124].CO2 draws researchers' attention again for its excellent heating performance.
Tamura et al. tested the heating performance of a CO2 heat pump system in a highly efficient automobile with an engine.In comparison with the R134a/PTC system, the relative COP of the CO2 system was 1.31 [125].A comparative study between R134a and CO2 heat pump systems with equal size components was presented by Dong et al. [126].The CO2 system was modified with an IHX; thus, the cooling performance improved.Their results showed that the CO2 system provides a competitive cooling capacity.Under 40 °C operating conditions, the CO2 system reached 3.2 kW with a COP of 1.6, while the R134a system reached 3.2 kW with a COP of 1.41, as depicted in Figure 8.Additionally, the CO2 system achieved a COP of 2.16 with 7502 W heating capacity under −20 °C.

Application of CO 2 Cycle in the Vehicle Field
The use of transcritical CO 2 cycles in automotive AC systems dates back to thirty years ago [5].From then on, several simulations or experiments were executed by different researchers [117][118][119].Mathur compared the thermodynamic performance of automotive air conditioning (AC) systems using R134a and CO 2 as working fluids [120].The study revealed that the COP of the CO 2 system is approximately 12% lower than that of the R134a system.Liu et al. found that the lubricant type and refrigerant mass charge affect the CO 2 system performance significantly [121].Niu et al. validated a wet-compression absorption CO 2 refrigeration system with a sustainable cooling capacity for vehicle AC.The high pressure reduces to less than 35 bar [122].These studies demonstrate that CO 2 is an acceptable but not very excellent refrigerant for vehicle air conditioning.Hence R134a plays the primary role in vehicle refrigerants.
However, with the development of electric vehicles, the heat shortage challenge stands out in extremely cold environments.The R134a heat pump or positive temperature coefficient (PTC) consumes enormous amounts of energy; thus, the driving range of electric vehicles decreases rapidly [123,124].CO 2 draws researchers' attention again for its excellent heating performance.
Tamura et al. tested the heating performance of a CO 2 heat pump system in a highly efficient automobile with an engine.In comparison with the R134a/PTC system, the relative COP of the CO 2 system was 1.31 [125].A comparative study between R134a and CO 2 heat pump systems with equal size components was presented by Dong et al. [126].The CO 2 system was modified with an IHX; thus, the cooling performance improved.Their results showed that the CO 2 system provides a competitive cooling capacity.Under 40 • C operating conditions, the CO 2 system reached 3.2 kW with a COP of 1.6, while the R134a system reached 3.2 kW with a COP of 1.41, as depicted in Figure 8.Additionally, the CO 2 system achieved a COP of 2.16 with 7502 W heating capacity under −20 • C.
Wang et al. considered the impact of battery heating/cooling demand on the performance of the CO 2 heat pump system further [127].Ko et al. analyzed the start-up and the pull-down processes of a CO 2 AC system for an electric vehicle [128].The pull-down time and energy consumption increase with the decrease of target temperature or the increase of initial temperature.Increasing the initial RH leads to more energy consumption and almost unchanged pull-down time.Apart from the research about system performance [106,129], several studies focused on the design and optimization of components.For example, flow unsteadiness and tangential leakage were proposed for compressor performance improvement [130,131].Additionally, the heat transfer performance of a micro-channel gas cooler was analyzed by simulation and experiments [132,133].A configuration called series gas cooler was introduced to improve the heat exchange capacity on the high-pressure side in heating mode [134].Mahvi et al. delayed the frosting process with a superhydrophobic heat exchanger, but the results showed that the defrosting process is unavoidable [135].Due to the performance degradation caused by frosting, some necessary defrosting methods and control strategies are applied to vehicle CO 2 heat pump systems.Wang et al. proposed a frost-free control strategy and demonstrated the application scope.However, the study does not consider the influence of defrosting on performance [136].Steiner and Rieberers simulated the reverse cycle and hot gas defrosting process and found that there exists an optimal valve-opening and ideal point of time in the two methods to conduct the defrost operation for electric vehicles, respectively [137,138].Wang et al. considered the impact of battery heating/cooling demand on the performance of the CO2 heat pump system further [127].Ko et al. analyzed the start-up and the pull-down processes of a CO2 AC system for an electric vehicle [128].The pull-down time and energy consumption increase with the decrease of target temperature or the increase of initial temperature.Increasing the initial RH leads to more energy consumption and almost unchanged pull-down time.Apart from the research about system performance [106,129], several studies focused on the design and optimization of components.For example, flow unsteadiness and tangential leakage were proposed for compressor performance improvement [130,131].Additionally, the heat transfer performance of a microchannel gas cooler was analyzed by simulation and experiments [132,133].A configuration called series gas cooler was introduced to improve the heat exchange capacity on the high-pressure side in heating mode [134].Mahvi et al. delayed the frosting process with a superhydrophobic heat exchanger, but the results showed that the defrosting process is unavoidable [135].Due to the performance degradation caused by frosting, some necessary defrosting methods and control strategies are applied to vehicle CO2 heat pump systems.Wang et al. proposed a frost-free control strategy and demonstrated the application scope.However, the study does not consider the influence of defrosting on performance [136].Steiner and Rieberers simulated the reverse cycle and hot gas defrosting process and found that there exists an optimal valve-opening and ideal point of time in the two methods to conduct the defrost operation for electric vehicles, respectively [137,138].
To improve system performance or reduce energy consumption, several modified transcritical CO2 cycles were applied for electric vehicles.The IHX has become a routine alternative for its significant improvement in the cooling mode under high-temperature environments.Fang used two needle valves to adjust IHX effectiveness.The experiment results showed that the IHX enhances the cooling COP all the time, but the cooling capacity reveals in a parabola form as the IHX effectiveness increases.It should be noted that the discharge temperature grows rapidly with the increment of IHX effectiveness [139].The ejector was also considered an appropriate and easy-to-use option for performance improvement [140].However, the non-adjustable feature of the ejector is not suitable for vehicles, which operate in a wide range of temperature, no matter heating or cooling.Thus Zou et al. designed a CO2 heat pump system with dual ejectors.The two ejectors were optimized based on the weight of operating conditions and used in cooling mode and heating mode separately [141].The system demonstrated better adaptability to different environmental conditions and presented large improvements under both heating and To improve system performance or reduce energy consumption, several modified transcritical CO 2 cycles were applied for electric vehicles.The IHX has become a routine alternative for its significant improvement in the cooling mode under high-temperature environments.Fang used two needle valves to adjust IHX effectiveness.The experiment results showed that the IHX enhances the cooling COP all the time, but the cooling capacity reveals in a parabola form as the IHX effectiveness increases.It should be noted that the discharge temperature grows rapidly with the increment of IHX effectiveness [139].The ejector was also considered an appropriate and easy-to-use option for performance improvement [140].However, the non-adjustable feature of the ejector is not suitable for vehicles, which operate in a wide range of temperature, no matter heating or cooling.Thus Zou et al. designed a CO 2 heat pump system with dual ejectors.The two ejectors were optimized based on the weight of operating conditions and used in cooling mode and heating mode separately [141].The system demonstrated better adaptability to different environmental conditions and presented large improvements under both heating and cooling modes.In another study, intermediate cooling enhanced the heating capacity and COP significantly in the temperature range of −20 • C to 0 • C and contributed a significant increase in cooling capacity and maximum COP at 45 • C, compared to the basic CO 2 system.A lower discharge temperature was reached when intermediate cooling was applied and this characteristic extended the application scope of the CO 2 heat pump system effectively [142,143].The application of a vapor-injection compressor could make the heating capacity and COP of the transcritical CO 2 cycle greater than the basic cycle, but the cooling performance was not reported [144].Some mixture refrigerants, such as CO 2 -propane and CO 2 -R41, were also utilized to improve the performance of automotive heat pump systems, and the experimental results showed positive changes in performance and parameters [145,146].Yu et al. combined the CO 2 -propane mixture refrigerant with an auto-cascade heat pump system, and the modification improved the system performance in heating mode compared with the basic pure CO 2 single-stage system [147].Due to the increased cooling demand taken by the fast charge and discharge of the battery, more modifications need to be employed in the vehicle transcritical CO 2 cycle to improve the cooling performance.To the best of our knowledge, no one has studied the direct refrigerant cooling of CO 2 for batteries in electric vehicles.
Besides pure electric vehicles, CO 2 heat pump systems were also applied in fuel cell electric vehicles, buses, freight transport vehicles, and railways.The cooling performance and heating performance in fuel cell vehicles were evaluated [148][149][150].Kim et al. considered the influence of heat exchanger arrangements on system performance, as shown in Figure 9.The new radiator-front arrangement increased the heating performance with a decrement in the cooling performance because of preheated air [151].
CO2-R41, were also utilized to improve the performance of automotive heat pump systems, and the experimental results showed positive changes in performance and parameters [145,146].Yu et al. combined the CO2-propane mixture refrigerant with an auto-cascade heat pump system, and the modification improved the system performance in heating mode compared with the basic pure CO2 single-stage system [147].Due to the increased cooling demand taken by the fast charge and discharge of the battery, more modifications need to be employed in the vehicle transcritical CO2 cycle to improve the cooling performance.To the best of our knowledge, no one has studied the direct refrigerant cooling of CO2 for batteries in electric vehicles.
Besides pure electric vehicles, CO2 heat pump systems were also applied in fuel cell electric vehicles, buses, freight transport vehicles, and railways.The cooling performance and heating performance in fuel cell vehicles were evaluated [148][149][150].Kim et al. considered the influence of heat exchanger arrangements on system performance, as shown in Figure 9.The new radiator-front arrangement increased the heating performance with a decrement in the cooling performance because of preheated air [151].Wang et al. conducted a refrigerant charge experimental study in an electric bus and pointed out that the lack of refrigerant charge enhances the irreversible losses in both exchangers significantly [102].Song et al. compared the CO2 system with the R407C system from the standpoint of energy, economy, and environment and the results revealed that the CO2 system has an environmental advantage in most areas of the world.The CO2 system and R407C system have their respective advantages in their respective suitable regions considering energy consumption and economy [152].Shi et al. designed a combined Wang et al. conducted a refrigerant charge experimental study in an electric bus and pointed out that the lack of refrigerant charge enhances the irreversible losses in both exchangers significantly [102].Song et al. compared the CO 2 system with the R407C system from the standpoint of energy, economy, and environment and the results revealed that the CO 2 system has an environmental advantage in most areas of the world.The CO 2 system and R407C system have their respective advantages in their respective suitable regions considering energy consumption and economy [152].Shi et al. designed a combined CO 2 system for refrigerated trucks that can achieve refrigeration and power generation, leading to a 7.4% and 4.9% fuel economy increase under refrigeration and freezing conditions, respectively [153].Fabris et al. found that ejector efficiency increases with the increase of ambient temperature in a CO 2 system with an ejector [154].

Conclusions and Suggestions for Future Work
After a long time of stillness, CO 2 comes into focus again for its clean, low-carbon, and economical characteristics.This review summarizes recent advances in modification combination, operating characteristics, control methods, and development in the vehicle field relevant to transcritical CO 2 heat pump and refrigeration cycles.CO 2 heat pump or refrigeration cycle performance can be improved through optimizing the system configuration and operating characteristics.This study is expected to facilitate the development and further investigation of this field.With a deeper understanding of the transcritical cycle and better improvement of performance, CO 2 would become a proper choice in many fields.The main conclusions and suggestions for the future are as follows: 1.
Previous studies about combination modifications in CO 2 systems have mainly focused on simulating for commercial refrigeration; more validation experiments and applications in other scenes are needed.

2.
Different combinations yield improvements in several aspects of system performance; the selection of combinations depends on the comprehensive assessment of operating conditions, major improvement objectives, cost, climate, and control complexity.

3.
Online methods of optimal pressure seeking require a long time with good performance, while offline methods acquire inaccurate results with a fast speed.A hybrid method combining the online and offline methods may be a feasible solution that provides both good performance and fast speed.4.
Although the effect of charge amount on transcritical CO 2 system performance has been investigated extensively, more research on refrigerant migration and distribution characteristics is needed.Additionally, refrigerant leakage detection methods for transcritical CO 2 systems are imperative due to the exceptionally high pressure.5.
The transcritical CO 2 cycle can provide enough heat with a high COP in low-temperature environments.However, the cooling performance in high-temperature environments is doubtful, especially with the increasing demand for cooling in electric vehicles.More evaluation and employment of vehicle operating conditions are necessary to find beneficial modification combinations.

Figure 2 .
Figure 2. Variation of optimum high pressure and COP with the change of gas cooler exit temperature [23].

Figure 2 .
Figure 2. Variation of optimum high pressure and COP with the change of gas cooler exit temperature [23].

Figure 3 .
Figure 3. Three categories of mainstream improvement measures and their combinations for transcritical CO2 cycle.Figure 3. Three categories of mainstream improvement measures and their combinations for transcritical CO 2 cycle.

Figure 3 .
Figure 3. Three categories of mainstream improvement measures and their combinations for transcritical CO2 cycle.Figure 3. Three categories of mainstream improvement measures and their combinations for transcritical CO 2 cycle.

Energies 2023 ,
16, x FOR PEER REVIEW 8 of 21 by several key factors, including compressor performance, evaporation temperature, and gas cooler outlet temperature [67].They established a model involving these parameters and validated it with a standard deviation of less than 1%.Several similar correlations have also been developed separately by different researchers [68-70].Cabello et al. compared the correlations proposed by Liao, Chen, Kauf, and Sarkar and the results showed that Sarkar's correlation provides the most precise results [71].Yin et al.

Figure 4 .
Figure 4. COP as a function of the cycle high pressure and water inlet temperature for the baseline coaxial heat exchanger [75].

Figure 4 .
Figure 4. COP as a function of the cycle high pressure and water inlet temperature for the baseline coaxial heat exchanger [75].

Figure 6 .
Figure 6.COP ratios of R22, R410A, R407C, and CO 2 systems with deviation from optimal charge[96].Aprea et al.  presented that the COP and cooling capacity reach a maximum value at the optimum charge.The exergy destruction in the compressor contributes the most to total exergy loss[97].Wang et al. comprehensively analyzed the influence of refrigerant charge on energy, exergy, economy, and the environment[98].Hazarika et al. quantified the influence of refrigerant charge amount on system performance and showed the effects of gas cooler face velocity[99].Zhang et al. and Wang et al. investigated the refrigerant charge characteristic in transcritical CO 2 heat pumps with EXVs and capillary tubes, respectively[100,101].Wang et al. studied the effect of refrigerant charge in an electric bus CO 2 air conditioning system and supposed that overcharge has a more substantial impact than insufficient charge on system exergy efficiency[102].Yin et al. discussed the refrigerant charge amount under different working conditions and proposed that the optimum charge is not a definite value, but a range of charge amounts[103].Kim mentioned that varying the refrigerant charge or inside volume of the high side are proper methods to control the high side pressure[20].A system with a low-pressure buffer or medium-pressure buffer was introduced to control high side charge.Changing the high side volume is also an effective way to regulate the high side pressure.He et al. analyzed the refrigerant charge distribution in a transcritical CO 2 heat pump water heater[104].The simulation results showed that the optimum COP reaches at an optimal high pressure while the high pressure is influenced by the refrigerant charge amount.A system charge management method based on heat exchanger resizing or additional gas cooler volume has been proposed to improve system performance without controlling the high pressure.Refrigerant migration and distribution characteristics are not only important to this method, but also the main reason for cycling loss[105].Wang et al. studied the CO 2 transient migration behavior under the period of start-up and shutdown experimentally with the help of a visualized accumulator[106].The results show that the system reveals rapid refrigerant migration and transient response because CO 2 exists in a superheated vapor state after the shutdown of the system for CO 2 mobile air conditioning (MAC).Thus, better transient performance and

Figure 7 .
Figure7.The COP variation of different cycles with the increase of medium temperature[115].

Figure 7 .
Figure7.The COP variation of different cycles with the increase of medium temperature[115].

Figure 8 .
Figure 8. Cooling performance comparison of R134a and R744 system [126].(a) the compressor speed is equal to 5600 RPM, (b) the cooling capacity is equal to 3.2 kw.

Figure 8 .
Figure 8. Cooling performance comparison of R134a and R744 system [126].(a) the compressor speed is equal to 5600 RPM, (b) the cooling capacity is equal to 3.2 kw.

Figure 9 .
Figure 9.The CO2 heat pump system configuration considering the arrangement of heat exchanger [151].

Figure 9 .
Figure 9.The CO 2 heat pump system configuration considering the arrangement of heat exchanger [151].

Table 1 .
Review of combination modifications applied to the transcritical CO 2 cycle.

Table 2 .
optimal pressure seeking methods analysis.

Table 2 .
Optimal pressure seeking methods analysis.

Table 2 .
optimal pressure seeking methods analysis.