Levelized costs of the energy chains of new energy vehicles targeted at carbon neutrality in China

Abstract

The diffusion of new energy vehicles (NEVs), such as battery electric vehicles (BEVs) and fuel cell vehicles (FCVs), is critical to the transportation sector’s deep decarbonization. The cost of energy chains is an important factor in the diffusion of NEVs. Although researchers have addressed the technological learning effect of NEVs and the life cycle emissions associated with the diffusion of NEVs, little work has been conducted to analyze the life cycle costs of different energy chains associated with different NEVs in consideration of technological learning potential. Thus, relevant information on investment remains insufficient to promote the deployment of NEVs. This study proposes a systematic framework that includes various (competing or coordinated) energy chains of NEVs formed with different technologies of power generation and transmission, hydrogen production and transportation, power-to-liquid fuel, and fuel transportation. The levelized costs of three typical carbon-neutral energy chains are investigated using the life cycle cost model and considering the technological learning effect. Results show that the current well-to-pump levelized costs of the energy chains in China for BEVs, FCVs, and internal combustion engine vehicles (ICEVs) are approximately 3.60, 4.31, and 2.21 yuan/GJ, respectively, and the well-to-wheel levelized costs are 4.50, 6.15, and 7.51 yuan/GJ, respectively. These costs primarily include raw material costs, and they vary greatly for BEVs and FCVs from resource and consumer costs. In consideration of the technological learning effect, the energy chains’ well-to-wheel levelized costs are expected to decrease by 24.82% for BEVs, 27.12% for FCVs, and 19.25% for ICEVs by 2060. This work also summarizes policy recommendations on developing energy chains to promote the diffusion of NEVs in China.

Graphical abstract

Keywords

energy chain / new energy vehicle / internal combustion engine vehicle / life cycle cost / technological learning

1 Introduction

The transportation sector is the second largest contributor to global greenhouse gas emissions, and it accounts for 25% of global greenhouse gas emissions and 9% of China’s total greenhouse gas emissions (Yu et al., 2021). It poses a considerable challenge to realizing carbon peaking and neutrality goals. The diffusion of new energy vehicles (NEVs) plays an important role in the deep decarbonization of the transportation sector (Milovanoff et al., 2020; Yang et al., 2020). The diffusion of NEVs is influenced by many factors, such as government subsidies (Zhao et al., 2021a), ownership cost (Lopez et al., 2021), safety (Wang et al., 2021), consumer acceptance (Huang et al., 2021), and the maturity of energy chains (Taalbi and Nielsen, 2021), which are the focus of this study. Energy chains are combinations of technologies that are defined at the level of conversion facilities (e.g., power plants or end-user devices such as cars).

NEVs have several energy chains. For example, different energy chains can be formed for fuel cell vehicles (FCVs) by using different technologies for hydrogen production, storage, transportation, and refueling. The different energy chains of FCVs, battery electric vehicles (BEVs), and internal combustion engine vehicles (ICEVs) compete with each other in the same way that FCVs, BEVs, and ICEVs compete in the market. Existing studies have focused on specific energy chains (e.g., Athanasopoulou et al., 2018; Qian et al., 2020; Abdelkareem et al., 2021; He et al., 2021) or a simple framework diagram composed of these energy chains (Watabe et al., 2020), and little attention has been paid to the overall picture of the energy system consisting of these energy chains. In terms of decarbonization, NEVs compete with ICEVs that use clean liquid fuels synthesized via the Fischer–Tropsch process by hydrogen from electrolysis and carbon dioxide (CO₂) captured from emission sources, and they could contribute to deep decarbonization in the transportation sector. As a result, a systematic framework for these energy chains is required, and research in this area could be beneficial to decision-makers who perform overall planning and industry actors who plan to invest in energy chains related to NEVs or ICEVs using clean liquid fuels.

Regardless of whether NEVs are BEVs or FCVs, their diffusion is inextricably linked to the synchronous development of their supporting energy infrastructure. Taalbi and Nielsen (2021) discovered that the reason for failures of early BEV diffusion is that the energy infrastructure is relatively expensive, resulting in a shortage of BEVs relative to ICEVs and limiting the diffusion of BEVs. The costs of energy chains influence the development of NEVs’ supporting energy infrastructure and thus play a crucial role in the diffusion of NEVs. Hence, investigating the costs of energy chains for the diffusion of NEVs is critical.

The life cycle cost (LCC) method is widely used to analyze the total costs incurred during the entire life cycle (from production to scrapping) of an equipment or system, and it has become a vital tool for project cost analysis. Lagaros et al. (2015) developed an LCC model that includes limit state dependence and LCC for evaluating wind farms. Basing on the concept of LCC, Gan et al. (2017) proposed to coordinate the process of decision-making and reliability assessment for solving a bi-level probabilistic transmission optimization problem. To address the problem of transitioning to clean energy transportation services, Comello et al. (2021) developed a time-driven LCC model for mobility services. However, most existing studies on the LCC of new energies focused only on a single project (i.e., a single technology in an energy chain) rather than the entire energy chain. For example, the entire life cycle of a BEV spans the entire well-to-wheel process, including well-to-pump and pump-to-wheel (Khan et al., 2019). The raw material supply, production, transmission, distribution, sale, use, storage, and recovery of electric power are closely related to the entire life cycle of a BEV. Levelized cost analysis based on LCC could facilitate cost comparisons between different energy chains, and it has already been applied to research on electricity generation (Fan et al., 2019), hydrogen production (Fan et al., 2022), and fuel production (Shirazi et al., 2019).

Meanwhile, the costs of new technologies tend to decrease with the accumulation of people’s experience in and knowledge of the technology, and this phenomenon is known as the technological learning effect (Arrow, 1962; Arthur, 1989). With the technological learning effect, the cost of a new technology is expected to decrease, allowing the technology to be increasingly adopted and permitting further technological learning; as a result, the technology becomes widespread (Rout et al., 2009). Technological learning is regarded as an endogenous driving force for the diffusion of NEVs (e.g., Schwoon, 2007). With regard to the technologies in NEVs’ energy chains, Kavlak et al. (2018) discovered that in the last 40 years, the cost of photovoltaic modules had decreased rapidly due to technological learning, thereby lowering the cost of electricity generation. Lane et al. (2021) reported that the high learning rate of electrolyzers, combined with the long-term trend of declining renewable electricity prices, leads to equal shares for new installations by the mid-term and eventually to electrolyzers having the dominant share of new facilities. Zhou et al. (2018) estimated economic parameters in the research on coal-to-liquids for ICEVs and investigated potential future cost changes. Some researchers studied the learning effect of NEVs and the life cycle emissions associated with the diffusion of NEVs (e.g., Shafiei et al., 2019; Watabe et al., 2020). However, little work has been conducted to analyze the LCCs of different energy chains associated with different NEVs while considering various technological learning rates, particularly for the energy chains of NEVs in China, the world’s largest NEV market.

Compared with previous studies, the current research provides the following primary contributions: (1) development of a systematic framework with various NEV energy chains, thus providing a panoramic picture of the most potential energy chains, including BEVs, FCVs, ICEVs, electricity generation, electricity transmission, charging station, hydrogen production, hydrogen transportation, hydrogen refueling stations, power-to-liquid fuel, fuel transportation, refueling stations, and other technologies; (2) analysis of how technological learning affects the LCCs of three typical energy chains in China on the basis of the systematic framework; and (3) provision of policy recommendations on developing energy chains for promoting the diffusion of NEVs. The systematic framework developed in this study can also serve as a starting point for future work on system optimization and simulation for the systematic diffusion of NEVs, and the method used in this study to estimate the levelized costs of the three focus energy chains can be employed to estimate the cost of other potential energy chains in the future.

The remainder of the paper is structured as follows. The models and data used in this study are introduced in Section 2. Section 3 presents the levelized costs of the three focus energy chains from well-to-pump and well-to-wheel. Section 4 discusses the findings, and Section 5 concludes this study by providing policy recommendations.

2 Model and data

2.1 Systematic framework with various carbon-neutral energy chains of vehicles

Fig.1 depicts the framework of an energy system showing various demands for society’s energy service and various natural resources. Demands and natural resources are linked by various energy chains, which are chains of different energy technologies. This framework is used by many system optimization models in determining the best dynamic technology combinations to meet various constraints, such as demands and emission targets, at the lowest cost (e.g., Gambhir et al., 2013; Victor et al., 2014; Zhang et al., 2017). This framework is also used in the present study.

At present, BEVs are the most common NEVs in use (although plug-in hybrid vehicles have been widely used, they are a transitional form between BEVs and ICEVs; therefore, plug-in hybrid vehicles are not considered in this study). The diffusion of BEVs increases the load on the electricity grid. Although BEVs have zero emission when driven, if the electricity is generated from fossil fuels, the energy chain emissions may not be reduced. Only when the electricity is generated using renewable energy (wind power, hydropower, solar energy, etc.) or carbon capture, utilization, and storage (CCUS) can BEV diffusion truly aid in carbon reduction (Das et al., 2020). FCVs are expected to be another alternative, particularly in logistics and freight transportation (İnci et al., 2021). The ability of FCVs to contribute to deep decarburization in the transportation sector is highly dependent on how hydrogen is produced (Karayel et al., 2021). In short, the energy-saving and emission reduction potentials of NEVs depend on their energy chains.

The integration of various energy chains of NEVs can improve the energy system’s supply-demand balance and increase economic benefits. For example, surplus power generated during peak periods of renewable energy generation can be used to produce hydrogen, which can then be utilized as the energy source for FCVs or to generate electricity during peak periods of electricity consumption. The hydrogen produced can also be combined with CO₂ to produce liquid fuel (Fischer–Tropsch reaction), which can be employed in conventional ICEVs (Atsbha et al., 2021; Schäppi et al., 2022). Thus, integrating different energy chains can aid in the absorption of redundant photovoltaic and wind power, resulting in a balanced load of the electricity grid and additional economic and environmental benefits (Koohi-Fayegh and Rosen, 2020). In addition to BEVs and FCVs, the carbon-neutral energy chains of ICEVs, namely, ICEVs using clean liquid fuel, are also considered in this study.

Fig.2 presents our systematic framework model of energy chains of BEVs, FCVs, and ICEVs using clean liquid fuel. It was developed based on the systematic energy framework in Fig.1.

Our systematic framework model includes two categories of energy chains for BEVs, with the main distinction being where the electricity is generated. In the first category, electricity is generated at the resource side and then transmitted to the consumer side to charge BEVs. The second category generates electricity at the consumer side. The fossil fuel required for power generation (e.g., coal and natural gas) is exploited and then transported to the consumer side to generate electricity for charging BEVs.

The energy chains of FCVs, same as the energy chains of BEVs, are divided into two categories. In the first category, hydrogen is produced at the resource side by electricity (or natural gas) and then transported to the consumer side to refuel FCVs. In the second category, electricity or natural gas is produced or exploited at the resource side before being transmitted to the consumer side to produce hydrogen for refueling FCVs.

The energy chains of ICEVs that use clean liquid fuel are also divided into two categories, with the difference being where the fuel is produced. In the first category, clean liquid fuel is produced at the resource side by combining hydrogen generated by renewable energy and CO₂ captured from the air, fossil fuel combustion, or other sources; the clean liquid fuel is then transported to the consumer side to refuel ICEVs. In the second category, clean liquid fuel is produced at the consumer side by using hydrogen generated by renewable energy and the captured CO₂. The clean liquid fuel is then directly used to refuel ICEVs.

The following shows descriptions of the main technologies considered in our systematic framework model.

● Coal power (supercritical, ultra-supercritical, etc.), hydropower, photovoltaic, wind power, nuclear power, biomass power generation, geothermal power generation, and other technologies are used to generate electricity. Flexible direct current/alternating current (DC/AC), high-voltage DC/AC, ultrahigh-voltage DC/AC, superconductivity, frequency division transmission, and other transmission technologies are among those available. Charging technologies include AC slow charging in the home, DC fast charging piles, power exchange, wireless charging, and mobile charging.

● Hydrogen production technologies include hydrogen production from fossil fuels (natural gas and coal), hydrogen production from industrial byproducts (chlor-alkali and coke oven gas byproducts), hydrogen production from biomass, and hydrogen production from power generation and water electrolysis. Gas-hydrogen trailers, gas-hydrogen pipelines, liquid-hydrogen tank cars, liquid organic matter, liquid ammonia, and other hydrogen transportation technologies are available. Different pressures, such as 35 MPa and 70 MPa, are considered by hydrogen refueling technologies.

● The fossil energy and liquid fuel transportation technologies include road or railway coal transportation, pipeline natural gas transportation, pipeline oil transportation, and tanker transportation.

2.2 Model for estimating levelized costs

An energy chain comprises various projects. A project in this context refers to the implementation of the technology depicted in Fig.2. Wind power generation, for example, is an energy chain technology; the project of this technology is the establishment and use of a wind farm. The costs of various projects are dynamic, based on different learning curves. These complexities and dynamics are reflected in our work estimating an energy chain’s LCC. When calculating the LCC of an energy chain, the costs of all components are discounted (considering the time value of money) and accumulated, and taxes are factored in.

The LCC of an energy chain project is divided into five stages (Heralova, 2017), as shown in Fig.3. These stages are decision-making, design, implementation, operation and maintenance, and scrap recovery.

With the five stages described in Fig.3, the following LCC estimation model is used in our analysis:

(1)LCC=C_{\text{i}}+\sum _{n=1}^N{\displaystyle \frac{C_{\text{o}}^n+R_{\text{v}}^n+R_{\text{s}}^n+R_{\text{i}}^n}{{\left(1+r\right)}^n}}-{\displaystyle \frac{V_{\text{r}}}{{\left(1+r\right)}^N}},

where C_{\text{i}} denotes the initial investment cost of establishing the facilities and C_{\text{o}}^n denotes the operation cost of the project in the nth year. R_{\text{v}}^n, R_{\text{s}}^n, and R_{\text{i}}^n denote the value-added tax, sales tax and surcharges, and income tax in the nth year, respectively. V_{\text{r}} denotes the salvage value of the project. N denotes the total life cycle of the project, and r denotes the discount rate.

The levelized costs of electricity, hydrogen, and fuel (abbreviated as LCOE, LCOH, and LCOF, respectively) are shown in Eq. (2) with LCOE as an example:

(2)LCOE=\frac{LCC}{\sum _{n=1}^N{A}_n{\left(1+r\right)}^{-n}},

where A_n denotes the annual output of electricity. LCOH and LCOF are calculated in the same way as LCOE.

The initial investment cost, C_{\text{i}} in Eq. (1), is the cost at the beginning of the life cycle, and it refers to the total cost of a one-time investment before the project is put into use. C_{\text{i}} mainly consists of the equipment purchase cost ( C_{\text{i1}}), equipment and installation cost ( C_{\text{i2}}), construction cost ( C_{\text{i3}}), and other costs ( C_{\text{i4}}). C_{\text{i1}} refers to the cost of all kinds of equipment used in equipment transportation and the transit cost and incidental expenses of these equipment from the manufacturer to the site. C_{\text{i2}} refers to the related costs incurred during equipment procurement and installation. C_{\text{i3}} consists of related expenses incurred during the construction and production stage and related expenses for project services. C_{\text{i4}} refers to various expenses other than the three above-mentioned expenses and mainly includes requisition of the construction site. Therefore, C_{\text{i}} can be calculated using Eq. (3):

(3)C_{\text{i}}=C_{\text{i}1}+C_{\text{i2}}+C_{\text{i3}}+C_{\text{i4}}.

The operation cost ( C_{\text{o}}^n) in Eq. (1) consists of running cost ( C_{\text{r}}^n), maintenance cost ( C_{\text{m}}^n), failure cost ( C_{\text{f}}^n), and raw material cost ( C_{\text{rm}}^n) in the nth year.

C_{\text{r}}^n refers to the sum of all expenses incurred during the operation of the project, and it mainly includes labor and energy consumption costs. Labor cost mainly includes the salaries, subsidies, and fees for training operators and the labor cost for other operators.

C_{\text{m}}^n refers to the sum of all costs incurred to ensure the normal operation of equipment or systems during project operation, and it mainly includes maintenance labor, equipment troubleshooting, and technical renovation costs. Different equipment has different maintenance costs. The cost structure of C_{\text{m}}^n is based on the comprehensive consideration of the initial investment cost ( C_{\text{i}}), labor cost, material cost, insurance cost, and others. Therefore, C_{\text{m}}^n is calculated for the entire project.

C_{\text{f}}^n refers to the total loss incurred by failure during the operation of the project, and it includes the loss of production stoppage and maintenance cost.

In this study, C_{\text{rm}}^n refers to the cost of production and usage of the main raw materials.

C_{\text{o}}^n can be calculated as

(4)C_{\text{o}}^n=C_{\text{r}}^n+C_{\text{m}}^n+C_{\text{f}}^n+C_{\text{rm}}^n.

V_{\text{r}} in Eq. (1) denotes the salvage value of a project at the retirement stage, and it mainly includes cleaning expenses, environmental restoration cost, and asset reuse cost. The revenue recovered excludes costs at the time of retirement, and the result is generally positive. V_{\text{r}} can be calculated as

(5)V_{\text{r}}=C_{\text{i}}\times r_{\text{v}},

where r_{\text{v}} denotes the salvage value rate.

In this study, different projects with different energy chains are given different technological learning rates from existing literature, and the dynamics of the fixed costs (including C_{\text{i}} and the fixed part of C_{\text{o}}^n) is reflected in the widely used learning curve presented in Eq. (6). The technological learning effect is reflected by the accumulation of the outputs of different projects.

(6)y=a{x}^{-b},

where y denotes the direct cost of constructing the xth project, a denotes the cost of constructing the first project, x denotes the number of construction projects, and b denotes the learning efficiency. Thus, the technological learning rate (LR) is \left(1-2^{-b}\right), indicating that the percentage of direct cost decreases when the amount of construction doubles.

2.3 Data

With the goal of achieving carbon neutrality, this study focused on estimating the levelized costs of three typical (and possibly the most promising) carbon-neutral energy chains in China included in the systematic framework shown in Fig.2. The three carbon-neutral energy chains primarily refer to wind power, hydrogen production, and clean fuel production, and relevant data on the energy chains of BEVs, FCVs, and IECVs were obtained from existing literature and statistical yearbooks.

In our analysis, the US dollar was converted into RMB at a rate of 7.00, r was set to 5% following Newell and Pizer (2004), the tax rate of R_{\text{v}}^n was set to 9% following the Provisional Regulations of the People’s Republic of China on Value-Added Tax, the value of R_{\text{s}}^n was set to 8% of R_{\text{v}}^n following the Provisional Regulations of the People’s Republic of China on Urban Maintenance and Construction Tax, and R_{\text{i}}^n was levied at 25% of the project profits following the Enterprise Income Tax Law of the People’s Republic of China.

2.3.1 Focus energy chain of BEVs

The focus energy chain of BEVs is based on renewable energy power generation, which is as follows: electricity generation with wind power plant (WPP) at the resource side → electricity transmission through the electricity transmission system (ETS) → BEV charging through the charging station (CS) at the consumer side.

The costs of WPP and CS are reduced in this energy chain as a result of the technological learning effect. ETS is assumed to be a mature technology with no technological learning effect. Tab.1 shows the specific parameters of BEV’s energy chain.

2.3.2 Focus energy chain of FCVs

In our techno–economic analysis, the focus energy chain of FCVs is based on renewable energy power generation as follows: electricity generation through WPP at the resource side → hydrogen production through the hydrogen production system (HPS) at the resource side → hydrogen transportation through the hydrogen transportation system (HTS) → FCV refueling hydrogen through the hydrogen refueling station (HRS) at the consumer side.

Wind power generation is used in this energy chain for water electrolysis to produce hydrogen, and alkaline electrolysis is selected as the hydrogen production technology. The tank truck is chosen as the hydrogen transportation technology. The energy consumed in the production of liquid hydrogen is high, and evaporation loss occurs in the storage and transportation of liquid hydrogen, resulting in a high cost. However, the transportation capacity of liquid hydrogen is more than 10 times that of a gas-hydrogen trailer, which greatly improves the transportation efficiency and makes it suitable for large-scale, long-distance transportation. As a result, liquid hydrogen transportation is viewed as the future trend, with the cost primarily derived from hydrogen liquefaction (Wulf and Kaltschmitt, 2012). Tab.2 shows the specific parameters of this energy chain.

2.3.3 Focus energy chain of ICEVs using clean liquid fuel

The focus energy chain of ICEVs using clean liquid fuel is based on renewable energy power generation as follows: electricity generation through WPP at the resource side → hydrogen production through HPS at the resource side → hydrogen and CO₂ are synthesized into liquid fuel through the Fischer–Tropsch synthesis plant (FSP) at the resource side → liquid fuel transportation through the fuel transportation system (FTS) → ICEV refueling through the refueling station (RS) at the consumer side.

The liquid fuel in this energy chain is obtained from the reaction between CO₂ and hydrogen via the Fischer–Tropsch process (Wei et al., 2017). The costs of FSP and RS are assumed to decrease with technological learning effect. FTS is regarded as a mature technology with no technological learning effect. The specific parameters of this energy chain are shown in Tab.3.

2.3.4 Technological learning rates

Tab.4 lists the LR and corresponding b value of the technologies that have learning potential, which were obtained from existing literature (listed in the last column).

3 Results

3.1 LCC of relevant projects and levelized costs of the three energy chains

Tab.5 shows the current LCC, LCOE, LCOH, and LCOF of each project (i.e., application of different technologies), with WPP, HPS, and FSP representing the consumer side costs and CS, HRS, and RS representing the resource side costs. The C_{\text{o}} in Tab.5 denotes the total operation cost, that is, C_{\text{o}}=\sum _{n=1}^N{\displaystyle \frac{C_{\text{o}}^n}{{\left(1+r\right)}^n}}. R_{\text{v}}, R_{\text{s}}, and R_{\text{i}} denote the total value-added tax, sales tax and surcharges, and income tax, respectively.

The dynamics of LCC, LCOE, LCOH, and LCOF are also estimated using the LRs in Tab.4, as shown in Fig.4, where LCOE, LCOH, and LCOF are unified into low heating values (the calorific values of electricity, hydrogen, and clean fuel are 3600 kJ/kWh, 142351 kJ/kg, and 44000 kJ/kg, respectively), C_{\text{rm}} denotes the total raw material cost, that is, C_{\text{rm}}=\sum _{n=1}^N{\displaystyle \frac{C_{\text{rm}}^n}{{\left(1+r\right)}^n}}, and R is the sum of R_{\text{v}}, R_{\text{s}}, and R_{\text{i}} (i.e., R=R_{\text{v}}+R_{\text{s}}+R_{\text{i}}). At the resource side, LCOE in 2020, 2030, and 2060 is 0.83, 0.72, and 0.67 yuan/GJ, respectively, according to Fig.4. With the technological learning effect, LCOE in 2030 and 2060 decreases by 13.25% and 19.28%, respectively, compared with that in 2020. In 2020, 2030, and 2060, LCOH is 2.29, 1.94, and 1.84 yuan/GJ, respectively. With the technological learning effect, LCOH decreases by 15.28% and 19.65% in 2030 and 2060, respectively, compared with that in 2020. In 2020, 2030, and 2060, LCOF is 1.44–1.74, 1.21–1.51, and 1.13–1.44 yuan/GJ, respectively. LCOF decreases by about 0.23 yuan/GJ from 2020 to 2030 and by about 0.08 yuan/GJ from 2030 to 2060 as a result of the technological learning effect.

Given that the source of electricity is renewable energy, LCOE excludes C_{\text{rm}}, and C_{\text{i}} and R account for a sizable proportion. The C_{\text{i}} of WPP gradually decreases with the learning effect (44.50%, 33.13%, and 27.69% in 2020, 2030, and 2060, respectively), whereas R gradually increases (46.66%, 59.19%, and 65.19% in 2020, 2030, and 2060, respectively). With regard to the LCOH of HPS, the proportion of C_{\text{i}} gradually declines (22.43%, 12.23%, and 8.51% in 2020, 2030, and 2060, respectively), but the proportion of C_{\text{rm}} (electricity and water) remains high (56.45%, 57.66%, and 57.89% in 2020, 2030 and 2060, respectively). C_{\text{rm}} (hydrogen, electricity, and CO₂) accounts for a sizable proportion of the LCOF of FSP. When the project captures the source of CO₂ (i.e., S = +240), C_{\text{rm}} accounts for more than 90% (90.69%, 91.20%, and 91.51% in 2020, 2030, and 2060, respectively), and when CO₂ utilization is profitable (i.e., S = −40), the proportion of C_{\text{rm}} decreases noticeably (74.13%, 72.12%, and 71.40% in 2020, 2030, and 2060, respectively).

At the consumer side (well-to-pump), the LCOE in 2020, 2030, and 2060 is 3.60, 3.07, and 2.90 yuan/GJ, respectively. With technological learning effects, the LCOE in 2030 and 2060 decreases by 14.72% and 19.44%, respectively, compared with that in 2020. In 2020, 2030, and 2060, LCOH is 4.31, 3.57, and 3.33 yuan/GJ, respectively. In 2030 and 2060, with the technological learning effect, LCOH decreases by 17.17% and 22.74%, respectively, compared with that in 2020. The LCOF in 2020, 2030, and 2060 is 1.88–2.21, 1.62–1.95, and 1.54–1.87 yuan/GJ, respectively. With the technological learning effect, LCOF decreases by about 0.26 yuan/GJ from 2020 to 2030, and it decreases by about 0.08 yuan/GJ from 2030 to 2060.

The main cost of CS is C_{\text{rm}} (electricity) (55.04%, 56.06%, and 55.79% in 2020, 2030, and 2060, respectively) and R (19.65%, 29.34%, and 33.18% in 2020, 2030, and 2060, respectively). The main cost of HRS is C_{\text{rm}} (hydrogen) (61.85%, 63.76%, and 64.83% in 2020, 2030, and 2060, respectively). For RS, the main cost is also C_{\text{rm}} (clean liquid fuel) when the source of CO₂ is captured by the project itself (i.e., S = +240), and C_{\text{rm}} accounts for more than 90% (91.35%, 91.10%, and 91.06% in 2020, 2030, and 2060, respectively). When CO₂ utilization in the upstream is profitable (i.e., S = −40), similar to the situation at the resource side, the proportion of C_{\text{rm}} decreases (77.11%, 74.98%, and 74.17% in 2020, 2030, and 2060, respectively).

The technological learning effect is assumed to affect only fixed costs in our estimation. Thus, the R and C_{\text{rm}} of renewable energy electricity generation account for a relatively high proportion of the energy chain of BEVs. To reduce R, relevant policies and subsidies should be implemented while reducing the cost of C_{\text{rm}}, and a reasonable price should be set for electricity. Meanwhile, the cost of the energy chain of FCVs is affected by the cost of electricity generation because C_{\text{rm}} accounts for more than 50% of the total cost of the energy chain of FCVs. The cost of C_{\text{rm}} can be effectively reduced by lowering the cost of electricity generation and improving hydrogen production and transportation technology. In the energy chain of ICEVs, C_{\text{rm}} is affected by the cost of electricity generation and hydrogen production. Hydrogen production upgrading, electricity generation, and Fischer–Tropsch synthesis technology can help lower the cost of C_{\text{rm}}. Furthermore, the higher the profit is from using CO₂, the smaller the proportion of C_{\text{rm}} is in the levelized cost.

3.2 LCOE, LCOH, and LCOF

Fig.5(a) and 5(b) show the levelized costs (in low heating value) on the resource and consumer sides, respectively.

On the resource side, LCOE is the lowest among the three, and LCOH is the highest. The cost of clean fuel production is between those of electricity generation and hydrogen production because the Fischer–Tropsch synthesis technology considered in this study has high selectivity for the catalyzer, and a large amount of electricity is used for raw gas circulation to reduce the cost of hydrogen. The cost of hydrogen production shows the steepest downward trend over time due to technological LR (20%) of the fixed cost of hydrogen production. The cost of raw materials belonging to the variable cost also decreases with the learning of power generation technology. The cost structure of clean fuel production is similar to that of hydrogen production, with raw materials accounting for the majority of the cost. LCOF’s downward trend is primarily due to the low electricity and hydrogen prices.

On the consumer side, through the respective transmission/transportation technology, LCOE still increases to about four times that on the resource side over time. This occurrence is due to the fact that the electricity price in China in 2020 was regulated and did not reflect the market value, whereas our estimate considers an electricity price that reflects the market value in 2060. This analysis of electricity prices is not comprehensive, and it may not fully reflect the downward trend of future electricity prices. LCOH always doubles from the resource side to the consumer side. To transport liquid hydrogen via trailers, hydrogen must be liquefied first. Liquid hydrogen evaporates during storage and transportation. This process requires a large amount of electricity, which is the main reason for the cost increase. LCOF does not increase considerably compared with the other energy sources due to the relatively mature technology of liquid fuel storage, transportation, and refueling stations; the technology allows liquid fuel to be transported from the resource side to the consumer side at a low cost.

The cost of each energy source has a greater potential for cost reduction on the consumer side than on the resource side. Notably, technological learning does not occur naturally over time and instead requires financial investment. Thus, improving consumer infrastructure is an efficient way to reduce costs.

Furthermore, the impact of carbon costs on economic performance should be assessed to achieve carbon neutrality. If utilizing CO₂ is profitable, the cost of Fischer–Tropsch synthesis can be considerably reduced (the blue and black lines in Fig.5 are the upper and lower bounds of cost change, respectively). As a result, the advancement of CCUS technology and the formulation of reasonable carbon trading policies can help reduce LCOF.

3.3 Energy efficiency of BEVs, FCVs, and ICEVs

The lowest heating values of the various energy chains were analyzed and compared in the previous section under the assumption that the fuel was completely burned, and the water vapor in its combustion products was in the gaseous state. However, the energy efficiency of various types of vehicles varies. In the following analysis, the energy efficiency of ICEVs, FCVs, and BEVs is assumed to be 30%, 70%, and 80%, respectively (Xiong et al., 2019) on the basis of the current average level of different technologies. According to this assumption (Fig.6(a)), at the well-to-wheel stage, the cost of electricity in 2020, 2030, and 2060 is 4.50, 3.83, and 3.62 yuan/GJ, respectively; the cost of hydrogen in 2020, 2030, and 2060 is 6.15, 5.10, and 4.76 yuan/GJ, respectively; and the cost of clean fuel in 2020, 2030, and 2060 is 6.40–7.51, 5.55–6.66, and 5.26–6.37 yuan/GJ, respectively. In 2060, the energy chains’ well-to-wheel levelized costs are expected to decrease by 19.56% for BEVs, 22.60% for FCVs, and 15.18% (S = +240) and 17.81% (S = −40) for ICEVs.

The energy chain of BEVs is the most economical in this scenario because of the high energy conversion efficiency of BEVs and the LCOE on the consumer side, and the well-to-wheel levelized cost of this energy chain is always low. Although LCOH is the highest on the consumer side, the energy conversion efficiency of FCVs is second only to that of BEVs, resulting in a lower well-to-wheel levelized cost of the energy chain of FCVs compared with that of ICEVs. The consumer-side LCOF is the lowest, but the energy conversion efficiency of ICEVs is only 30%, resulting in the highest well-to-wheel levelized cost of the energy chain of ICEVs. Even with a profit of 40 yuan/t of CO₂, the well-to-wheel levelized cost of ICEVs is slightly higher than that of FCVs.

The energy efficiency of BEVs, FCVs, and ICEVs can be gradually improved. Following Kosai et al. (2018)’s work, we assumed that the energy efficiency of BEVs, FCVs, and ICEVs will increase by 0.17%, 0.15%, and 0.12% per year, respectively. The well-to-wheel levelized costs of the energy chains will be reduced by 24.82% for BEVs, 27.12% for FCVs, and 19.25% (S = +240) and 21.76% (S = −40) for ICEVs in 2060 compared with those in 2020, as shown in Fig.6(b).

According to the preceding analysis, the well-to-wheel levelized costs of the various energy chains can be reduced by the technological learning effect of the energy infrastructure and by improving vehicle energy efficiency. The policy implications of the findings are as follows. To achieve deep decarbonization in the transportation sector, governments need to encourage investments in energy infrastructure through various measures (e.g., subsidies) to induce the technological learning effect. Moreover, governments need to gradually enhance regulations on vehicles’ energy efficiency.

4 Discussion

4.1 Panoramic view of the energy chains of NEVs

Various technology combinations were reviewed in this study, and a panorama of the energy chains of NEVs was formed as a systematic framework (Fig.2). The energy chains primarily included electricity generation, transmission, charging, hydrogen production, transportation, refueling, power-to-liquid fuel, fuel transportation, and refueling technologies.

Although the technologies on the consumer and resource sides are similar, regional climate, natural resources, economic development level, and infrastructure conditions influence the adoption of various technologies. In China, the resource is mostly concentrated in areas with large amounts of land and a small population (especially Northwest China), making them ideal for constructing large-scale power plants with economies of scale. Consumption is concentrated in areas with dense populations and good infrastructure (particularly in Southeast China). Hence, this research focused on renewable energy used for resource-side electricity generation, hydrogen production, and clean fuel production, all of which will eventually be used as energy sources for BEVs, FCVs, and ICEVs on the consumer side. The following shows the representative energy chains.

The energy chain of BEVs is as follows: wind power generation at the resource side → electricity transmission → charging at the consumer side.

The energy chain of FCVs is as follows: wind power generation at the resource side → hydrogen production at the resource side → hydrogen transportation → refueling hydrogen at the consumer side.

The energy chain of ICEVs is as follows: wind power generation at the resource side → hydrogen production at the resource side → hydrogen and CO₂ are synthesized into clean fuel (Fischer–Tropsch reaction) at the resource side → clean fuel transportation → refueling at the consumer side.

Beginning with renewable energy, the aforementioned energy chains provide the transmission process of each energy source from the resource side to the consumer side. This study has focused primarily on the techno–economic analysis of these energy chains. Other factors related to these energy chains, such as water consumption and environmental toxicity, should be investigated in future research.

4.2 Life cycle cost

The LCC model was used to calculate the levelized costs of representative energy chains, namely, LCOE, LCOH, and LCOF. In terms of well-to-pump levelized costs, the order from low to high was determined to be LCOE, LCOF, and LCOH on the resource side, and LCOF, LCOE, and LCOH on the consumer side. These findings are primarily the result of factors, such as the high cost of hydrogen transportation and refueling, regulated electricity prices, and mature liquid fuel transportation technology. Therefore, a market-based electricity price and the development of hydrogen energy technologies may be beneficial to the diffusion of BEVs and FCVs. In terms of well-to-wheel levelized costs, which are also affected by vehicle efficiency, the order from low to high was found to be LCOE, LCOH, and LCOF, indicating that vehicle energy efficiency is important. The technological learning effect is primarily responsible for lowering the fixed costs of energy infrastructure. Vehicle energy efficiency improvements can considerably reduce well-to-wheel levelized costs.

BEVs have the advantages of high energy conversion efficiency and low cost, with the goal of carbon neutrality being based on renewable energy for power generation. The disadvantages of BEVs include the fact that their widespread use will increase grid load, and BEVs do not work well for long-distance and heavy-cargo transportation. FCVs have the advantage of using the hydrogen produced by excess renewable energy generation during peak periods, and they may perform better than BEVs in long-distance and heavy-cargo transportation. The main disadvantage of FCVs is the high costs of hydrogen production, storage, transportation, and hydrogen refueling facilities. ICEVs that use clean liquid fuel have a similar advantage as FCVs in that they can use hydrogen produced by surplus electricity during the peak period of renewable energy generation and liquid fuel produced by the synthesis of hydrogen and CO₂. Another advantage of ICEVs that use clean liquid fuel is that the automobile engine does not need to be transformed, and the fuel supply can rely on existing gasoline stations, resulting in a minimal impact on automobile manufacturers and consumers. However, if ICEVs’ energy efficiency is not improved considerably, BEVs and FCVs will become more cost-effective options in the future, particularly in scenarios with high carbon costs.

4.3 Different scenarios of technological learning

Historical evidence indicates that technological learning is uncertain and that this uncertainty follows a lognormal distribution (McDonald and Schrattenholzer, 2001). If the decision is based on a specific assumed value of LR (e.g., the mean of its lognormal distribution), the future investment may be misled. For example, if the technological LR is overestimated, the cost of technology investment will be underestimated and may potentially result in severe economic risks in the future. Thus, we performed calculations under various technological learning scenarios. The range of LRs was based on existing literature (Kahouli-Brahmi, 2008; Detz et al., 2018; Greene et al., 2020; Thomassen et al., 2020). Tab.6 compares the changes in levelized costs under various scenarios of technological LRs to those in Subsection 3.1, which is considered the baseline scenario.

Tab.6 shows that despite having different LRs, the LCOE of WPP and the LCOH of HPS have similar fluctuation ranges on the resource side. The LCOF of FSP is influenced by the LRs of electricity generation, hydrogen production, and itself. If the learning rates with the lowest values in Tab.6 are compared with our calculation in Subsection 3.1, the LCOF of FSP will increase by about 6% and 7% in 2030 and 2060, respectively. The fixed costs of consumer facilities account for a small portion of the total cost. Hence, the fluctuation of levelized costs on the consumer side (CS, HRS, and RS) is primarily due to the resource side. In summary, the energy chain of ICEVs using clean fuel faces great risks due to the uncertainty of technological learning.

5 Conclusions and policy recommendations

This study developed a systematic framework that provides a panoramic view of the energy chains of NEVs starting from renewable energy. The framework is combined with various technologies to estimate the levelized costs of three representative energy chains of BEVs, FCVs, and ICEVs using clean liquid fuel. This research can be used as a starting point for an in-depth examination of diffusion of NEVs and the development of their supporting energy infrastructure in China. This study’s levelized cost estimation method can also be adopted to estimate the levelized costs of other energy chains. Despite some limitations discussed at the end of this paper, we make the following policy recommendations based on the analysis presented in this paper.

● The energy chain of BEVs is the most cost-effective among the three representative energy chains, and policymakers are advised to encourage the establishment of energy infrastructures supporting BEVs in order to promote the diffusion of BEVs. Furthermore, developing and implementing a market-based electricity price policy could aid in the long-term diffusion of BEVs.

● In terms of the energy efficiency of the energy chains, FCVs are uneconomical compared with BEVs. However, hydrogen can be used to store the energy generated by “light and wind” that would otherwise be lost, and hydrogen production technology based on water electrolysis does not emit carbon. Water resources are the focus of consideration in hydrogen production by water electrolysis, while water resources in China are frequently mismatched with renewable electricity-producing areas, implying that policymakers should consider both water resources and electricity transmission. Low-cost hydrogen production technology, large-scale hydrogen storage and transportation technology, and low-cost and high-efficiency hydrogen utilization technology are the key research and development directions for the future hydrogen industry. Policymakers are advised to pay close attention to energy chains with high technological learning potential and to direct investment toward these energy chains to benefit from the economies of scale and technological learning.

● When Fischer–Tropsch synthesis of liquid fuel is based on green hydrogen and captured CO₂, it is also a carbon-neutral energy chain for ICEVs. This energy chain has the advantage of existing and mature fuel transportation and refueling technology and facilities. The benefit of using CO₂ and the energy efficiency of ICEVs are important factors in this energy chain. In the short and middle terms, policymakers are advised to gradually increase energy efficiency regulations, which will make this energy chain increasingly competitive.

● According to the systematic framework depicted in Fig.2, the energy chains for BEVs, FCVs, and ICEVs using clean fuel compete with each other, but they can also be complementary. Coordination of these energy chains, for example, could balance the load on electricity grids. Given that new technologies and their supporting infrastructure typically take a long time to mature and be adopted, ICEVs could remain dominant vehicles in the short and possibly middle terms. Furthermore, BEVs and FCVs may have different strengths and drawbacks when used for different purposes and in different climates. With these considerations, policymakers are advised to focus on improving all of these energy chains (at least the three representative ones) rather than focusing on just one and ignoring the others.

The first limitation of this study is that it focuses only on three representative energy chains rather than all of the energy chains in the systematic framework depicted in Fig.2. The second limitation is that in the estimations, the interactions of different energy chains are not considered. For example, a wind power project could serve both hydrogenation production for FCVs or ICEVs and electricity production for BEVs; thus, to some extent, the fixed capital investment and the technological learning effect should be considered based on different energy chains. We intend to conduct a systematic analysis of the diffusion of NEVs and the development of their supporting energy infrastructure in China in the future by considering the competition and coordination among different energy chains and heterogeneous climates, development levels, and other variables. The above-mentioned limitations will be addressed in our future work.

References

[1] Abdelkareem, M A Elsaid, K Wilberforce, T Kamil, M Sayed, E T Olabi, A ( 2021). Environmental aspects of fuel cells: A review. The Science of the Total Environment, 752: 141803

[2] Anandarajah, G McDowall, W Ekins, P ( 2013). Decarbonising road transport with hydrogen and electricity: Long term global technology learning scenarios. International Journal of Hydrogen Energy, 38( 8): 3419– 3432

[3] Arrow, K J ( 1962). The economic implications of learning by doing. Review of Economic Studies, 29( 3): 155– 173

[4] Arthur, W B ( 1989). Competing technologies, increasing returns, and lock-in by historical events. Economic Journal, 99( 394): 116

[5] Athanasopoulou, L Bikas, H Stavropoulos, P ( 2018). Comparative well-to-wheel emissions assessment of internal combustion engine and battery electric vehicles. Procedia CIRP, 78: 25– 30

[6] Atsbha, T A Yoon, T Seongho, P Lee, C J ( 2021). A review on the catalytic conversion of CO₂ using H₂ for synthesis of CO, methanol, and hydrocarbons. Journal of CO₂ Utilization , 44: 101413

[7] Borlaug, B Salisbury, S Gerdes, M Muratori, M ( 2020). Levelized cost of charging electric vehicles in the United States. Joule, 4( 7): 1470– 1485

[8] Comello, S Glenk, G Reichelstein, S ( 2021). Transitioning to clean energy transportation services: Life-cycle cost analysis for vehicle fleets. Applied Energy, 285: 116408

[9] Das, R Wang, Y Putrus, G Kotter, R Marzband, M Herteleer, B Warmerdam, J ( 2020). Multi-objective techno–economic–environmental optimisation of electric vehicle for energy services. Applied Energy, 257: 113965

[10] Detz, R J Reek, J N H van der Zwaan, B C C ( 2018). The future of solar fuels: When could they become competitive?. Energy & Environmental Science, 11( 7): 1653– 1669

[11] Fan, J L Wei, S Yang, L Wang, H Zhong, P Zhang, X ( 2019). Comparison of the LCOE between coal-fired power plants with CCS and main low-carbon generation technologies: Evidence from China. Energy, 176: 143– 155

[12] Fan, J L Yu, P Li, K Xu, M Zhang, X ( 2022). A levelized cost of hydrogen (LCOH) comparison of coal-to-hydrogen with CCS and water electrolysis powered by renewable energy in China. Energy, 242: 123003

[13] Gambhir, A Schulz, N Napp, T Tong, D Munuera, L Faist, M Riahi, K ( 2013). A hybrid modelling approach to develop scenarios for China’s carbon dioxide emissions to 2050. Energy Policy, 59: 614– 632

[14] Gan, L Li, G Lin, J Zhou, M ( 2017). A bi-level probabilistic transmission planning with intermittent generations based on life cycle cost. International Journal of Electrical Power & Energy Systems, 90: 306– 314

[15] Greene, D L Ogden, J M Lin, Z ( 2020). Challenges in the designing, planning and deployment of hydrogen refueling infrastructure for fuel cell electric vehicles. eTransportation, 6: 100086

[16] He, X Wang, F Wallington, T J Shen, W Melaina, M W Kim, H C de Kleine, R Lin, T Zhang, S Keoleian, G A Lu, X Wu, Y ( 2021). Well-to-wheels emissions, costs, and feedstock potentials for light-duty hydrogen fuel cell vehicles in China in 2017 and 2030. Renewable & Sustainable Energy Reviews, 137: 110477

[17] Heralova, R S ( 2017). Life cycle costing as an important contribution to feasibility study in construction projects. Procedia Engineering, 196: 565– 570

[18] Hu, H Dong, W Zhou, Q ( 2021). A comparative study on the environmental and economic effects of a resource tax and carbon tax in China: Analysis based on the computable general equilibrium model. Energy Policy, 156: 112460

[19] Huang, X Lin, Y Zhou, F Lim, M K Chen, S ( 2021). Agent-based modelling for market acceptance of electric vehicles: Evidence from China. Sustainable Production and Consumption, 28: 206– 217

[20] İnci, M Büyük, M Demir, M H İlbey, G ( 2021). A review and research on fuel cell electric vehicles: Topologies, power electronic converters, energy management methods, technical challenges, marketing and future aspects. Renewable & Sustainable Energy Reviews, 137: 110648

[21] Jenn, A Highleyman, J ( 2022). Distribution grid impacts of electric vehicles: A California case study. iScience, 25( 1): 103686

[22] Kahouli-Brahmi, S ( 2008). Technological learning in energy–environment–economy modelling: A survey. Energy Policy, 36( 1): 138– 162

[23] Karayel, G K Javani, N Dincer, I ( 2022). Green hydrogen production potential for Turkey with solar energy. International Journal of Hydrogen Energy, 47( 45): 19354– 19364

[24] Kavlak, G McNerney, J Trancik, J E ( 2018). Evaluating the causes of cost reduction in photovoltaic modules. Energy Policy, 123: 700– 710

[25] Khan, M I Shahrestani, M Hayat, T Shakoor, A Vahdati, M ( 2019). Life cycle (well-to-wheel) energy and environmental assessment of natural gas as transportation fuel in Pakistan. Applied Energy, 242: 1738– 1752

[26] Koohi-Fayegh, S Rosen, M A ( 2020). A review of energy storage types, applications and recent developments. Journal of Energy Storage, 27: 101047

[27] Kosai, S Nakanishi, M Yamasue, E ( 2018). Vehicle energy efficiency evaluation from well-to-wheel lifecycle perspective. Transportation Research Part D: Transport and Environment, 65: 355– 367

[28] Lagaros, N D Karlaftis, M G Paida, M K ( 2015). Stochastic life-cycle cost analysis of wind parks. Reliability Engineering & System Safety, 144: 117– 127

[29] Lane, B Reed, J Shaffer, B Samuelsen, S ( 2021). Forecasting renewable hydrogen production technology shares under cost uncertainty. International Journal of Hydrogen Energy, 46( 54): 27293– 27306

[30] Lopez, N S Tria, L A Tayo, L A Cruzate, R J Oppus, C Cabacungan, P Isla Jr, I Ansay, A Garcia, T Cabarrubias-Dela Cruz, K Biona, J B M ( 2021). Societal cost-benefit analysis of electric vehicles in the Philippines with the inclusion of impacts to balance of payments. Renewable & Sustainable Energy Reviews, 150: 111492

[31] McDonald, A Schrattenholzer, L ( 2001). Learning rates for energy technologies. Energy Policy, 29( 4): 255– 261

[32] Milovanoff, A Posen, I D MacLean, H L ( 2020). Electrification of light-duty vehicle fleet alone will not meet mitigation targets. Nature Climate Change, 10( 12): 1102– 1107

[33] Newell, R G Pizer, W A ( 2004). Uncertain discount rates in climate policy analysis. Energy Policy, 32( 4): 519– 529

[34] Qian, F Gao, W Yang, Y Yu, D ( 2020). Economic optimization and potential analysis of fuel cell vehicle-to-grid (FCV2G) system with large-scale buildings. Energy Conversion and Management, 205: 112463

[35] Rout, U K Blesl, M Fahl, U Remme, U Voß, A ( 2009). Uncertainty in the learning rates of energy technologies: An experiment in a global multi-regional energy system model. Energy Policy, 37( 11): 4927– 4942

[36] Rubin, E S Azevedo, I M L Jaramillo, P Yeh, S ( 2015). A review of learning rates for electricity supply technologies. Energy Policy, 86: 198– 218

[37] Sahoo A Mistry K Baker T ( 2019). The costs of revving up the grid for electric vehicles

[38] Schäppi, R Rutz, D Dähler, F Muroyama, A Haueter, P Lilliestam, J Patt, A Furler, P Steinfeld, A ( 2022). Drop-in fuels from sunlight and air. Nature, 601( 7891): 63– 68

[39] Schoots, K Ferioli, F Kramer, G Vanderzwaan, B ( 2008). Learning curves for hydrogen production technology: An assessment of observed cost reductions. International Journal of Hydrogen Energy, 33( 11): 2630– 2645

[40] Schwoon, M ( 2007). A tool to optimize the initial distribution of hydrogen filling stations. Transportation Research Part D: Transport and Environment, 12( 2): 70– 82

[41] Shafiei, E Davidsdottir, B Stefansson, H Asgeirsson, E I Fazeli, R Gestsson, M H Leaver, J ( 2019). Simulation-based appraisal of tax-induced electro-mobility promotion in Iceland and prospects for energy-economic development. Energy Policy, 133: 110894

[42] Shirazi, A Rahbari, A Asselineau, C A Pye, J ( 2019). A solar fuel plant via supercritical water gasification integrated with Fischer–Tropsch synthesis: System-level dynamic simulation and optimisation. Energy Conversion and Management, 192: 71– 87

[43] Sinigaglia, T Lewiski, F Santos Martins, M E Mairesse Siluk, J C ( 2017). Production, storage, fuel stations of hydrogen and its utilization in automotive applications: A review. International Journal of Hydrogen Energy, 42( 39): 24597– 24611

[44] Taalbi, J Nielsen, H ( 2021). The role of energy infrastructure in shaping early adoption of electric and gasoline cars. Nature Energy, 6( 10): 970– 976

[45] Thomassen, G van Passel, S Dewulf, J ( 2020). A review on learning effects in prospective technology assessment. Renewable & Sustainable Energy Reviews, 130: 109937

[46] Victor, N Nichols, C Balash, P ( 2014). The impacts of shale gas supply and climate policies on energy security: The US energy system analysis based on MARKAL model. Energy Strategy Reviews, 5: 26– 41

[47] Wang, H Liu, H Yao, J Ye, D Lang, Z Glowacz, A ( 2021). Mapping the knowledge domains of new energy vehicle safety: Informetrics analysis-based studies. Journal of Energy Storage, 35: 102275

[48] Watabe, A Leaver, J Shafiei, E Ishida, H ( 2020). Life cycle emissions assessment of transition to low-carbon vehicles in Japan: Combined effects of banning fossil-fueled vehicles and enhancing green hydrogen and electricity. Clean Technologies and Environmental Policy, 22( 9): 1775– 1793

[49] Wei, J Ge, Q Yao, R Wen, Z Fang, C Guo, L Xu, H Sun, J ( 2017). Directly converting CO₂ into a gasoline fuel. Nature Communications, 8( 1): 15174

[50] Wulf, C Kaltschmitt, M ( 2012). Life cycle assessment of hydrogen supply chain with special attention on hydrogen refuelling stations. International Journal of Hydrogen Energy, 37( 21): 16711– 16721

[51] Xiong, H Liu, H Zhang, R Yu, L Zong, Z Zhang, M Li, Z ( 2019). An energy matching method for battery electric vehicle and hydrogen fuel cell vehicle based on source energy consumption rate. International Journal of Hydrogen Energy, 44( 56): 29733– 29742

[52] Yang, Z Wang, B Jiao, K ( 2020). Life cycle assessment of fuel cell, electric and internal combustion engine vehicles under different fuel scenarios and driving mileages in China. Energy, 198: 117365

[53] Yu, Y Li, S Sun, H Taghizadeh-Hesary, F ( 2021). Energy carbon emission reduction of China’s transportation sector: An input–output approach. Economic Analysis and Policy, 69: 378– 393

[54] Zhang, G Y Niu, D S Shi, Y M Xu, G L Li, G L Zhang, H Li, J Cao, L Guo, Z Q Wang, Z C Qiu, Y T ( 2017). Nuclear fuel cycle modelling using MESSAGE. Journal of Radioanalytical and Nuclear Chemistry, 311( 2): 1435– 1440

[55] Zhao, D Ji, S F Wang, H P Jiang, L W ( 2021a). How do government subsidies promote new energy vehicle diffusion in the complex network context? A three-stage evolutionary game model. Energy, 230: 120899

[56] Zhao, J Zhou, L Zhou, W Ren, H Yu, Y Wang, F Ma, T ( 2021b). Techno-economic analysis and comparison of coal-based chemical technologies with consideration of water resources scarcity. Energy Strategy Reviews, 38: 100754

[57] Zhou, L Duan, M Yu, Y Zhang, X ( 2018). Learning rates and cost reduction potential of indirect coal-to-liquid technology coupled with CO₂ capture. Energy, 165: 21– 32

RIGHTS & PERMISSIONS

The Author(s). This article is published with open access at link.springer.com and journal.hep.com.cn