*4.1. Sensitivity Analysis of Electricity Profile*

To achieve higher energy and emission reduction benefits, the shares of non-fossil power in the electricity mix should increase to a higher level. China has already introduced relevant policies and measures to develop a low-carbon electricity mix. For example, "The 13th five-year plan for electric power (2016–2020)" has proposed that the installed capacity of non-fossil fuel will be about 770 million kilowatts in 2020, accounting for about 39%, and the installed capacity of natural gas accounts for more than 5% while that of coal will decrease to about 55%. Based on relevant policies and previous studies, this study projects the electricity supply structure of China in the near future (2020) and in the long-term future (2030), as shown in Figure 4.

**Figure 4.** (**a**) The electricity profile of China in 2020; (**b**) The electricity profile of China in 2030.

Power generation technologies in 2020 are assumed to be at the same level as in 2017, but in 2030, advanced technologies and highly efficient equipment are expected to be implemented to improve the energy conversion efficiency and control the pollution emissions. Therefore, this study assumes the energy consumption and GHG emissions would be 10% less than the predicted baseline power generation mix in 2030. The energy and emission intensities of Chinese electricity in 2020 and 2030 are calculated by using the GREET model and the results are shown in Table 9. Therefore, the GHG emissions from power generation are 19.25% and 27.32% lower in 2020 and 2030 than those in 2017. Since the gasoline production technology is relatively mature, this study assumes that the gasoline production-related GHG emissions from in 2020 and 2030 will be the same as those in 2017.


**Table 9.** Energy and emission intensities of power generation in 2020 and 2030.

As shown in Table 10, BEVs will generally achieve 6% and 9% emission reduction in 2020 and 2030, compared with 2017. The reduction benefits for PHEVs are lower, which are about 2.75% in 2020 and 3.80% in 2030. Clearly, the GHG emission differences between the BEVs and PHEVs expand to 31.40% (LFP) and 19.79% (NMC) in 2020. In 2030, the attractiveness of BEVs will be more prominent in that the emission reduction benefits of BEVs are expected to be 25.76% (NMC) – 40.00% (LFP) relative to PHEVs. In fact, if the electricity generation moves to a lower-emission intensity, the advantages of BEVs would be more remarkable.

Since China demonstrates a large amount of diversity in the electricity profiles, the conclusion may not be valid in all cities. Therefore, the break-even point is calculated to present in which cases BEVs outperform PHEVs in terms of the GHG emissions. Results obtained from break-even point analysis show that GHG emission intensity below 973.80 gCO2-eq/kWh and 815.00 gCO2-eq/kWh would make LFP- and NMC-powered BEVs, respectively, favorable options. According to Bauer et al. (2015) [36], where regional electricity profiles in China are analyzed, north, northeast, east, and northwest have about 900.00 gCO2-eq/kWh GHG emission intensities in 2012 and cities like Beijing are estimated to have over 900.00 gCO2-eq/kWh in 2020. Therefore, it is possible that PHEVs are currently preferable in parts of cities in China.


**Table 10.** The life cycle energy consumptions and GHG emissions in the 2020 and 2030 scenarios.

#### *4.2. Sensitivity Analysis of Driving Distance*

The driving distance in this part includes two parts: the lifetime mileage and the all-electric ranges within one charging period. Because EVs have just come onto the market, real world data of lifetime mileage are unavailable. As stated in Table 2, the parameter of lifetime mileage is assumed and thus uncertainty is inevitable. Besides, PHEVs are able to use the battery in electric mode and consume gasoline when the battery charge is depleted [11]. Since the electric mode saves more energy with a lower fuel cost, drivers are often encouraged to use electricity as often as possible within the all-electric range. Therefore, the assumption of the travel distance for single travel and the all-electric range limitation are important parameters for the energy use and GHG emission rate of PHEVs.

#### 4.2.1. Sensitivity Analysis of Lifetime Mileage

In the baseline scenario, the lifetime mileage is assumed to be a certain value and remains the same for BEVs and PHEVs provided that they use the same battery type. To deal with the uncertainty of the lifetime mileage, this parameter is considered as any value within an interval, and the range

of life cycle GHG emissions of each vehicle is calculated accordingly. The equation of life cycle GHG emissions is shown as Equation (2).

$$\text{GHG (life cycle)}\_{i} = \frac{\text{GHG (vehicle cycle)}\_{i}}{r\_{i}} + \text{GHG (fuel cycle)}\_{i} \tag{2}$$

where *i* relates to BEV-LFP, BEV-NMC, PHEV-LFP and PHEV-NMC; *ri* represents the lifetime mileage of each vehicle, with the assumed range of [120,000 km, 160,000 km]; *GHG* (*lif e cycle*)*<sup>i</sup>* , *GHG* (*vehicle cycle*)*<sup>i</sup>* and *GHG* (*f uel cycle*)*<sup>i</sup>* represent GHG emissions of each vehicle for the life cycle, vehicle cycle and fuel cycle, respectively.

The calculated ranges of life cycle GHG emissions for each vehicle are shown in Table 11. Since the minimum emissions of PHEVs exceed the maximum values of their counterpart BEVs, it is highly likely that BEVs outperform PHEVs from the life cycle perspective.

**Table 11.** The calculated range of life cycle GHG emissions for each vehicle.


4.2.2. Sensitivity Analysis of All-Electric Range

The all-electric ranges for one charging period for our representative vehicles are reported as 300 km, 400 km for BEVs and 80 km, 100 km for PHEVs. Under real-world driving conditions, 80% depth discharge is always applied. Therefore, we assume that 80% of the reported ranges are reached within each driven trip. The UF for PHEVs can be assumed as follows:

$$\text{UF}(\mathbf{R}\_{LFP}) = \begin{cases} \mathbf{1}, & \mathbf{0} < \mathbf{R}\_{LFP} \le \mathbf{64}; \\ \mathbf{64}/\mathbf{R}\_{LFP}, & \mathbf{R}\_{LFP} > \mathbf{64} \end{cases} \tag{3}$$

$$\text{UF}(\text{R}\_{\text{NMC}}) = \begin{cases} 1, & 0 < \text{R}\_{\text{NMC}} \le 80; \\ 80/\text{R}\_{\text{NMC}}, & \text{R}\_{\text{NMC}} > 80 \end{cases} \tag{4}$$

where UF(R*LFP*) and UF(R*NMC*) represent the UF of LFP-powered and NMC-powered PHEVs, respectively; R*LFP* and R*NMC* represent the travel distances for each time.

Since BEVs are only able to use electricity to propel the vehicles, the life cycle energy use and GHG emissions on the basis of per km will not change. Our results show that for LFP-powered vehicles, as long as the driven distance is below the range limitation of BEVs (300 km), the BEV has lower life cycle emissions than the PHEV. For NMC-powered vehicles, when the driven distance is below 80 km, the life cycle GHG emissions is 222.76 g CO2-eq/km for the PHEV, 6.65% less than the life cycle emissions of the BEV. When the driven distance is higher than 80 km but less than 96.34 km, the per-km-based emissions result for the PHEV increases with the distance, but is still below that of the BEV. Therefore, the break-even point for UF(NMC) is 0.83, at the point that the travel distance reaches 96.23 km. Additionally, it is highlighted that PHEVs would be the option when the travel distances exceed the range limitation of BEVs.
