J O U RN A L OF E N V I RO N ME N TA L SC IE N CE S 3 3 (2 0 1 5) 88 – 9 6
Available online at www.sciencedirect.com
ScienceDirect www.journals.elsevier.com/journal-of-environmental-sciences
Vehicular volatile organic compounds losses due to refueling and diurnal process in China: 2010–2050 Xiaofan Yang1 , Huan Liu1,2,3,⁎, Hongyang Cui1 , Hanyang Man1 , Mingliang Fu1 , Jiming Hao1,2,3 , Kebin He1,2,3 1. School of Environment, State Key Joint Laboratory of Environment Simulation and Pollution Control, Tsinghua University, 100084, China. E-mail: [email protected] 2. State Environmental Protection Key Laboratory of Sources and Control of Air Pollution Complex, Beijing 100084, China 3. Collaborative Innovation Center for Regional Environmental Quality, China
AR TIC LE I N FO
ABS TR ACT
Article history:
Volatile organic compounds (VOCs) are crucial to control air pollution in major Chinese
Received 25 September 2014
cities since VOCs are the dominant factor influencing ambient ozone level, and also an
Revised 5 January 2015
important precursor of secondary organic aerosols. Vehicular evaporative emissions have
Accepted 19 January 2015
become a major and growing source of VOC emissions in China. This study consists of lab
Available online 29 April 2015
tests, technology evaluation, emissions modeling, policy projections and cost-benefit analysis to draw a roadmap for China for controlling vehicular evaporative emissions. The
Keywords:
analysis suggests that evaporative VOC emissions from China's light-duty gasoline vehicles
Evaporative VOC emissions
were approximately 185,000 ton in 2010 and would peak at 1,200,000 ton in 2040 without
China roadmap
control. The current control strategy implemented in China, as shown in business as usual
Vehicle
(BAU) scenario, will barely reduce the long-term growth in emissions. Even if Stage II
Refueling emission control
gasoline station vapor control policies were extended national wide (BAU + extended Stage II), there would still be over 400,000 ton fuel loss in 2050. In contrast, the implementation of on-board refueling vapor recovery (ORVR) on new cars could reduce 97.5% of evaporative VOCs by 2050 (BAU + ORVR/BAU + delayed ORVR). According to the results, a combined Stage II and ORVR program is a comprehensive solution that provides both short-term and long-term benefits. The net cost to achieve the optimal total evaporative VOC control is approximately 62 billion CNY in 2025 and 149 billion CNY in 2050. © 2015 The Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences. Published by Elsevier B.V.
Introduction Understanding volatile organic compound (VOC) emission is crucial to control air pollution in major Chinese cities since VOCs are the dominant factor influencing ambient ozone level as well as secondary organic aerosols (SOAs) (Guo et al., 2011a; Lei et al., 2011; Duan et al., 2008; Chan and Yao, 2008; Geng et al., 2007; Liu et al., 2010). In major city clusters, vehicular emissions have become the most important source in VOC emission
inventory. For example, vehicles were responsible for 26%–50% of VOC emissions in inland Pearl River Delta (PRD) cities in 2004– 2007, and 48% ± 4% in Hong Kong from 2001 to 2007 (Srivastava et al., 2005; Guo et al., 2011b; Gentner et al., 2009; Lee et al., 2002; Liu et al., 2008; Walsh, 2014). Vehicular VOC emissions include tailpipe emissions and evaporative losses, e.g., release of gasoline vapors resulting from diurnal temperature variations, refueling, hot soak, running losses and permeation. In developed countries, vehicular evaporative VOC emission
⁎ Corresponding author. E-mail: [email protected] (Huan Liu).
http://dx.doi.org/10.1016/j.jes.2015.01.012 1001-0742/© 2015 The Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences. Published by Elsevier B.V.
J O U RN A L OF E N V I RO N ME N TA L SC IE N CE S 3 3 (2 0 1 5) 8 8– 9 6
without accounting losses during refueling process usually account for 30% of gasoline-related VOCs. A 2007 research study notes that evaporative emissions contribute to 12% of total VOC emissions in Beijing with an ozone formation potential that is higher than for most other VOC sources (Song et al., 2007). Although vehicular evaporative VOC losses are important, the regulating and controlling process is just getting started in China and the roadmap of evaporative emission control is still a controversy. In this article, we evaluated different VOC control scenarios based on two existing technologies: the Stage II vapor recovery system (Stage II) and on-board refueling vapor recovery (ORVR). Stage II collects gasoline vapors from vehicle fuel tanks while customers dispense gasoline at gasoline dispensing facilities (US EPA, 1991). ORVR systems are carbon canisters installed in automobiles to capture gasoline vapors evacuated from the gasoline tank. In China, Stage II has already been required in some major urban areas since 2007, whereas numerous service stations, scattered across medium and small cities, suburbs and freeways, remain uncontrolled. Stage II is widely used in Europe, while both Stage II and ORVR were used in the US until 2012. The US EPA has started to phase-out Stage II because of the widespread penetration of ORVR. The US EPA identified ORVR as the best available control technology (BACT) because of its high efficiency, ability to function without specific monitoring or maintenance, and low cost. The purposes of this article include: (1) understanding current controls and evaporative emissions from vehicles' refueling and diurnal processes in China; (2) evaluating a possible future control roadmap and estimating the effects on emissions reductions; (3) generating cost-benefit analyses for the scenarios evaluated. In this paper, we established emission models to evaluate total evaporative emissions. Five policy scenarios were designed for future vehicular VOC control based on two available technologies. For each technology, the basic emission rates were generated by theoretical calculation, while control efficiency of different technologies was measured by laboratory tests. In addition, cost and benefits of each scenario were further evaluated to provide a comprehensive understanding of future policy options.
1. Methodology 1.1. Real-world emission rates calculation (non-control) The basic emission rate of refueling VOCs without any control can by calculated by the Reddy–Wade equation. This equation was developed in 2010 and now is widely used in all over the world (Reddy, 2010; US EPA, 2008). The real-world refueling vapor emission is influenced by local temperature and the Reid vapor pressure (RVP) of the gasoline. According to a fuel survey in China (Huo et al., 2012a), the RVP of market fuel ranges from 42.5 to 89.0 kPa, and the average is 61.3 kPa. Thus 60 kPa was selected as a representative value in this analysis. For the emission inventory, temperatures in each province from 6:00 am to 21:00 pm, during which most refueling occurs, were used.
89
where, Q(g/L) is the concentration of VOCs in the air that is displaced to the environment, R(gal·psi)/(g·mol·K) is the universal gas constant 0.3187, f(g/gal) is the air entrainment 0.2, Td(K) is the temperature of the dispensed fuel, Tv(K) is the temperature of the vehicle's fuel tank, and Tv–Td is approximately 2 K. Patm(psi) is the atmosphere pressure, Ptank(psi) is the tank fuel vapor pressure, and Pdisp(psi) is the dispensed fuel vapor pressure. Ptank and Pdisp could be further calculated by the RVP and temperature: P ¼ A T RVP expð−B=T Þ
ð2Þ
where, A and B are the constants 25.61 and 2789.78, respectively, RVP is the fuel RVP in psi, and T(K) is the temperature. Evaporative VOC emissions also include hot soak and diurnal emissions. Evaporative emissions from a vehicle not related to refueling are very complex. They range from breathing losses from the vehicle tank to leaks around gaskets and hoses. Additionally, VOCs can escape around engine cylinders and then leak from the oil pan. The US EPA developed an empirical equation for a 24-hr evaporative emission factor (US EPA, 2001, 2012a). The International Vehicle Emission (IVE) model uses adjustment factors for evaporative emission factors to address the difference between real fuel vapor pressure and the EPA experience value (CE-CERT (Center for Environmental Research) et al., 2008). We use both the US EPA equation and the IVE adjustment factors to generate diurnal emission rates for Chinese fuel vapor pressure (8.7 psi). To convert the gasoline consumption-based emission factor into a distance-specific emission factor, the average mileage of Chinese passenger vehicles was introduced using data from previous research (Huo et al., 2012a).
1.2. Lab tests of ORVR control efficiency A series of refueling emission tests were finished to determine efficiency of ORVR technology using local vehicles and gasoline. Two in-use vehicles equipped with original ORVR systems were recruited: (1) a 2009 Chrysler–Dodge Journey with a 2.7 L ORVR canister and 77.6 L fuel tank and (2) a 2012 Chrysler–Jeep Compass with a 2.0 L ORVR canister and a 51.1 L fuel tank. The whole test was conducted in Gas-tight Imtech Variable Temperature Sealed Housing for Evaporative Determination chambers (VT-SHED, VT-SHED software, Imtech Automotive Testing Solutions Beijing Co., Ltd. Beijing, China). This SHED is the most advanced one in China. This system is equipped with a canister loading device, fuel temperature control system, as well as total evaporative emission test system. China Automotive Technology & Research Center Beijing laboratory (CATARC) provided all the facilities and technical support. Each vehicle was tested five times to ensure the reliability. For each test, the processes include canister purge, canister adsorption to breakthrough, type VII driving defined by China vehicle emission control standard, fuel drain, soak, refueling test and canister weighing in quadruplicate. A detailed test procedure was provided in supporting materials, Fig. S1.
1.3. Current control policy and future scenario design Q ¼ 18:2
P disp f P atm P atm −P tank þ R Td R Tv P atm −P disp
ð1Þ
The Chinese government launched regulation of emissions from service stations in 2007, requiring Stage II to be installed
90
J O U RN A L OF E N V I RO N ME N TA L SC IE N CE S 3 3 (2 0 1 5) 88 – 9 6
in selected locations before 2015. The Ministry of Environmental Protection (MEP) conducted investigations to track the implementation of Stage II, and the following levels of implementation were found. (1) Beijing: completed upgrading all of the service stations by 30 June, 2008; (2) Tianjin and cities in Hebei Province: 60% completed by 30 June, 2008; (3) Pearl River Delta: 90% completed before October 2010; (4) Shanghai: 90% completed before the World Expo in May 2010; and (5) Other regions: no policies or plans yet. According to the results from the investigation, we readjusted the percentage of service stations equipped with Stage II on the basis of the governmental control policy. Considering the poor supervision of the Stage II system, the real operational percentage is even lower. Thus, this study has a positive projection of Stage II efficiency. There has been discussion in China on the possibility of requiring ORVR on new cars starting with the China 5 standard (equivalent to Euro 5 standard). Hence, potential control scenarios for the next few decades are shown in Table 1.
1.4. Mathematics of emissions projection under different scenarios Both refueling emissions and diurnal/hot soak emissions are taken into consideration in these projections. The emissions projection equations under different scenarios were listed in Table 1. Emiref and Emidh are VOC emissions through refueling and diurnal/hot soak processes without any control (g/year). They could be calculated by using following equations:
ηORVR and ηCO (%) refer to the removal efficiency of Stage II, ORVR, and when ORVR and Stage II systems work simultaneously. GNum and GNumII are the numbers of total service stations and service stations that have applied Stage II. VP and VPORVR refer to the vehicle population and population with ORVR systems. The calculations for scenarios b and c are based on the proportion of non-controlled and Stage II-controlled sites. The calculations for scenarios d and e are more complicated. A matrix, including two vehicle types (ORVR and non-ORVR) and two service station types (Stage II and non-Stage II), is used to calculate the refueling emissions under four refueling possibilities. The hypothesis is that vehicles with or without ORVR have the same chance to be refueled at a Stage II station, and the possibility only depends on the percentage of stations with Stage II in the province. All of the input data and their sources are summarized in Table 2. In total, 93,173 service stations are included in this study as announced by MEP. Gasoline vehicle population was projected by the Gompertz equation, a widely accepted approach to forecast vehicle population based on gross domestic production (GDP) per person (Jong et al., 2004; Dargay and Gately, 1999). The historical data of vehicle population and gasoline consumption comes from the Year Book of National Bureau of Statistics of China (till 2010). The economic and population growth outlook is a compilation from several different researchers in China (Hu, 2007; Li et al., 2003; Yao and He, 1994).
2. Results and discussion
Emire f ¼ EFre f FC
ð3Þ
2.1. Emission factors of evaporative VOCs
Emidh ¼ EFdh VKT VP
ð4Þ
The measured average refueling emission factor was 0.961 (+0.078, −0.123) g/L. The difference between the measured value and the Reddy equation result (based on test conditions) was very small (5.67%) and could be ignored. Taking realworld temperatures and RVP variation into consideration, the real-world refueling emission factor for China was 1.17 g/L as an average. To understand the importance of vehicular evaporative emissions in China, the evaporative refueling and diurnal emission factors were compared with tailpipe emission factors
where, EFref is the VOC emission factor through the refueling process (g/ton) and EFdh is the diurnal and hot soak emission factor based on mileage (g/km). Method to calculate these two emission factors were defined in Section 1.1. FC is the fuel consumption (ton/year); VKT is the average vehicle distance traveled (km/year). For scenarios other than non-control, the efficiency of control technologies (ORVR and Stage II) was taken into consideration. ηII,
Table 1 – Potential control scenarios designed for future vehicular evaporative volatile organic compound (VOC) control in China. Scenario
Stage II policies
Non-control BAU BAU + extended Stage II BAU + ORVR scenario
None Current policies Stage II is expanded to all service stations in China before 2025. Current policies
BAU + delayed ORVR
Current policies
ORVR policies None None None ORVR is required on new gasoline vehicle from 2016 ORVR is required on new gasoline vehicle from 2019
VOC emissions (Emi) calculation Emi = Emiref × Emidh Emi ¼ Emire f 1−ηII GGNumII þ Emidh Num h ORVR Emi ¼ Emire f 1−ηII GGNumII 1− VPVP −ηORVR Num i ORVR ORVR 1− GGNumII −ηCO GGNumII VPVP VPVP Num Num ORVR þEmidh ηORVR VPVP
BAU: business as usual; ORVR: on-board refueling vapor recovery; Emiref: VOC emissions through refueling; Emidh: VOC emissions through diurnal/hot soak processes without any control; GNum: the number of total service stations; GNumII: the number of service stations that have applied Stage II; VP: the vehicle population; VPORVR: the vehicle population with ORVR systems; ηII: the removal efficiency of Stage II; ηORVR: the removal efficiency of ORVR; ηCO: the removal efficiency when ORVR and Stage II system work simultaneously.
91
J O U RN A L OF E N V I RO N ME N TA L SC IE N CE S 3 3 (2 0 1 5) 8 8– 9 6
Table 2 – Input data for evaporative emission estimation. Input
Value
Source
Emission factors through the refueling process (non-control) Diurnal and hot soak emission factor (non-control) Temperature
1.621 kg/ton
Calculated as introduced in Section 1.1
0.6 kg/ton Average for each province
Information for service stations Vehicle mileage traveled Fuel economy
Location of each service station 20,000 km/year 11.03 km/L
EPA MOVES model Average measured temperature (6:00–22:00) from April 2011 to March 2012 in each province website MEP investigation and digital map Previous research (Wang, 2008) Previous research (Huo et al., 2012a; He et al., 2005)
MEP: the Ministry of Environmental Protection.
that are based on per-distance units. Fig. 1a includes tailpipe HC emission factors from the EMBEV model (a vehicle emission model developed by Tsinghua University) with a deterioration rate based on mileage traveled; Fig. 1b contains tailpipe HC emissions based on Portable Emissions Measurement System (PEMS) tests (Huo et al., 2012b). It can be easily concluded that, total evaporative VOC emissions (0.12 g/km), including emissions from refueling, diurnal and hot soak processes, exceed both the tailpipe standards and in-use tailpipe levels of China 3 (equivalent to Euro 3 standard) and China 4 (equivalent to Euro 4 standard) vehicles (0.09 and 0.02 g/km). From another perspective, the overall evaporative VOC emissions from any gasolinefueled car in China are approximately equivalent to exhaust emissions from a 5-year-old China 3 car or a 7-year-old China 4 car. Starting to control VOC emission from non-tailpipe sources is important.
2.2. Control efficiencies of the two technologies ORVR systems show good measurement repeatability and efficiency in the tests, which were described in Section 1.2. The range of efficiencies was 98.92% to 99.74% (average 99.40%). The efficiency in the model was set as 99%. Compared with other literatures, this efficiency is reasonable. EPA's documentation associated with its 1994 ORVR final rule making package projected an average efficiency for ORVR of 95% based on 1160 in-use vehicle tests. However, more recently, EPA's extensive testing of in-use vehicles has demonstrated a fleet efficiency of approximately 98% (US EPA, 2012b). Besides, the 99% reduction
efficiency was also adopted for recovery of hot soak and diurnal emissions since an ORVR system works exactly the same as for refueling vapor. And according to US Manufacturers of Emission Controls Association (MECA), the overall reduction of efficiency of ORVR is above 98%, equalling to reduction efficiency through refueling process (MECA, 2011). Furthermore, no significant deterioration of ORVR system was spotted during life time of a car. Therefore, lifetime of vehicles was used in the model instead of lifetime of ORVR system. The efficiency of Stage II system was generated by literature review. EPA state implementation plan guidance estimates that 92% efficiency can be achieved with semi-annual inspection and strong follow-up enforcement, while an efficiency of 86% has been associated with annual inspection. A lower end of 62% has been associated with minimal inspection (US EPA, 1991). In 2009, MEP investigated the service stations equipped with Stage II systems in Beijing, Tianjin, Hebei and the Pearl River Delta. The results reveal that, after running for one year, the passing rate of Stage II systems was very low, especially in seal reliability. An overall efficiency of approximately 70% has been estimated. When Stage II and ORVR are used simultaneously, the recovery efficiency will be reduced because of the lack of compatibility between the two controls. The ORVR canister captures the gasoline emissions from the motor vehicle fuel tank. Instead of drawing vapor-laden air from the vehicle fuel tank into the underground storage tank, the vacuum pump of the Stage II system draws fresh air into the underground storage tank. The fresh air causes gasoline in the underground
Fig. 1 – Emission rate comparisons between evaporative and tailpipe volatile organic compounds (VOCs). (a) Tailpipe emission factors derived from new vehicle emission standards considering vehicle deterioration; (b) tailpipe emission factors derived from in use vehicle Portable Emissions Measurement System (PEMS) tests. Chinese government adopted vehicle emission standards following emission standards in Europe since 1999. The emission levels 1 to 4 in China (China 1–4) are equivalent to Euro 1 to 4 standards, respectively.
92
J O U RN A L OF E N V I RO N ME N TA L SC IE N CE S 3 3 (2 0 1 5) 88 – 9 6
tank to evaporate inside the underground tank and thus creates an increase in pressure in the underground storage tank. As a result, gasoline vapors may be forced out of the underground storage tank vent pipe into the ambient air. Generally speaking, the recovery rate decreases by 1%–10% when Stage II and ORVR work at the same time (94% in this research) (US EPA, 2012a).
2.3. Vehicle population and gasoline consumption projections The projections for future vehicle population and gasoline consumption are shown in Fig. 2. According to the projection, gasoline vehicle in China will reach 584 million by 2050. Compared with the projection of Huo and Wang (2012), which is 558 million by 2050, projection in this research is 4.7% higher. This is because the historical data in this study was updated to 2011 and which was higher than previous forecast. To get an outlook of future gasoline consumption in China, fuel economy and its future development is also required. Basic fuel economy used here is from He et al. (2005) and Huo et al. (2012a). For future fuel economy improvement, we adopted the moderate scenario from the three scenarios designed by National Academy of Science (NAS) (US NRC, 2002; Wang, 2008).
2.4. Evaporative emission inventory of China in 2010 The total vehicular evaporative emissions (refueling, diurnal and hot-soak) were calculated to reach approximately 185,000 ton in 2010, of which 124,000 ton were from refueling and the remaining 61,000 ton were from diurnal and hot soak processes. The average refueling loss is approximately 0.26% for total gasoline consumption in China (68,862,000 ton by China energy statistic). This estimation is very reasonable compared with US experience. In 1999, U.S. gasoline loss through refueling was approximately 0.1% of total gasoline consumption (475,000 ton emission versus 455,000,000 ton gasoline) (US EPA, 2004). By 1999, there were over 31,000 gas stations in the US with Stage II vapor recovery, and there had already been two years of new ORVR vehicles introduced into the fleet. Thus, the emission factor was reduced from 1.32 g/L in AP-42 to 1.1 g/L in the US.
With provincial statistical data from National Bureau of Statistics (2012a, 2012b), emission rates of each province with available data were calculated following the equations in methodology (Table 3). As shown in Table 3, VOC emissions per unit gasoline consumed in regions under Stage-II control are significantly lower than other regions. However, the control effect is still not satisfying. The emission distribution shows that VOC emission intensity is higher in well-developed regions – especially Beijing, Tianjin, the Yangzi River Delta and the Pearl River Delta – where ozone and PM2.5 pollution are also serious. Service stations in these regions are under Stage-II control, but vehicle population or gasoline consumption density is higher, which lead to high vapor losses in total.
2.5. Evaporative emissions in 2010–2050 The projected outlook for vehicular evaporative VOC emission is shown in Fig. 3. Without control, the peak would reach almost 1,200,000 ton/year approximately in 2040, which would be ten times the current emissions rate. With Stage II control under the BAU scenario, emissions would reach 809,000 ton in 2050, about 21.5% reduction compared with non-control scenario. The emission trend of BAU scenario shows that refueling losses would be controlled during the initial few years. However, the current control of Stage II only in urban districts is not able to restrain the increasing trend of VOC emissions due to the fast-growing vehicle population. Even with an extended Stage II strategy (scenario c), which means all of the service stations after 2025 are equipped with Stage II, the future VOC emission growth would only be slowed down rather than being reduced. On one hand, there is a significant contribution of new emissions from new vehicles each year, which is represented by the gray belt in Fig. 3. On the other hand, vapor losses from diurnal and hot soak are non-control, approximately 300,000 ton by 2050. The implementation of ORVR on new cars (scenario d) could reduce 97.5% of total evaporative VOCs by 2030 and maintain at a low emission level. The reason for this huge reduction is the high new vehicle penetration rate, high sales and phaseout rate, in China. According the vehicle population growth
Fig. 2 – Gasoline passenger vehicle population, fuel consumption projection and control policy. BAU: business as usual; ORVR: on-board refueling vapor recovery.
93
J O U RN A L OF E N V I RO N ME N TA L SC IE N CE S 3 3 (2 0 1 5) 8 8– 9 6
Table 3 – Provincial evaporative VOC emission through refueling process. Province
Vehicle population (million)
Gasoline consumption (million ton)
Emission (ton)
Emission rate (kg per ton gasoline consumed)
Beijing Tianjin Hebei Shanxi Inner Mongolia Liaoning Jilin Heilongjiang Shanghai Jiangsu Zhejiang Anhui Fujian Jiangxi Shandong Henan Hubei Hunan Guangdong Guangxi Hainan Chongqing Sichuan Guizhou Yunnan Shaanxi Gansu Qinghai Ningxia Xinjiang
4.44 1.68 4.63 2.32 1.76 2.76 1.45 1.73 1.64 5.87 5.56 1.81 1.86 1.23 6.96 3.90 1.87 2.01 7.45 1.43 0.37 0.99 3.41 1.01 2.07 1.92 0.73 0.29 0.36 1.15
3.72 2.05 2.39 2.28 3.26 5.93 1.67 3.64 4.15 7.50 5.87 1.57 3.33 1.55 8.02 2.97 4.58 2.62 10.86 2.48 0.53 1.03 5.42 1.43 2.32 2.55 0.57 0.26 0.23 1.31
1219 657 2148 2203 2679 4808 1199 2375 1646 9340 7723 1983 4888 2137 9433 3575 6135 3539 17299 3875 873 1417 6746 1640 2950 3027 474 189 218 1005
0.33 0.32 0.90 0.97 0.82 0.81 0.72 0.65 0.40 1.25 1.32 1.26 1.47 1.38 1.18 1.20 1.34 1.35 1.59 1.56 1.65 1.38 1.24 1.15 1.27 1.19 0.83 0.73 0.95 0.77
Data of passenger vehicles from available provinces are included in this table.
projection, average annual growth rate of gasoline vehicle will be 8.8% during 2015–2020, and the average annual phase-out ratio of old vehicles is about 5.3%. The emission difference between new and old vehicles is about 99%. Thus, the ORVR could reduce emission by the second year following implementation. Therefore, if ORVR implemented earlier, it's easier to bring more significant reduction in a short time. If the ORVR strategy was delayed to 2019, when the China 6 standard (equivalent to Euro 6 standard) is expected to be implemented
(scenario e), a cumulative total of 1,570,000 ton more VOCs will be emitted to the atmosphere, which is approximately 8 times higher than the total 2010 emissions. Considering the potentially high diurnal and hot soak emissions in the future in China, ORVR seems to be more advantageous than Stage II. If no ORVR is to be adopted, the amount of diurnal and hot soak emissions will rise beyond the total vehicular VOCs of 2010 by the year 2030 and become the most significant source of vehicle emissions.
Fig. 3 – Projection of vehicular evaporative volatile organic compound (VOC) emissions in China to 2050. (a) Total emission trend; (b) portion of diurnal and hot-soak vs. refueling emissions.
94
J O U RN A L OF E N V I RO N ME N TA L SC IE N CE S 3 3 (2 0 1 5) 88 – 9 6
2.6. Cost and benefit analysis Fig. 4 shows the cost and benefit results of the two single technologies, as well as different scenarios. The costs for installing Stage II system in a single gas station is 300,000 CNY. And annual running cost, including system test, nozzle replacement and other maintenance, is 90,500 CNY per year on average. For ORVR (Fig. 4b), average cost is approximately 200 CNY per vehicle and no maintenance cost is required. Assuming that all the VOCs recovered are turned to gasoline, a gas station equipped with Stage II system never recovers its costs during lifetime. On the other hand, 0.145 L gasoline was saved per 1000 km traveled for an ORVR car, which means an ORVR vehicle would achieve cost-benefit balance point in 9–10 years. The BAU + extended stage II scenario has the highest net cost but the second lowest benefit (Fig. 4c). The BAU + ORVR scenario has lower net cost but the highest recovery profit. If ORVR was delayed, compared with the BAU + ORVR scenario, the net cost is about the same but fuel recovery profit is 16 billion CNY lower in 2025. To achieve the optimal total VOC control (BAU + ORVR scenario), the total investment would be 109 billion CNY in 2025 and 466 billion CNY in 2050 (the net cost is approximately 62 billion CNY in 2025 and 149 billion CNY in 2050). The investment for Stage II systems is the majority of the total investment (80%). Based on US experience, Stage II systems can be phased out when ORVR penetration reaches widespread use (or approximately 75% of vehicles). If ORVR was implemented by 2015, the widespread use of ORVR would be accomplished by about 2026 (as illustrated in Fig. 2). In this situation, 170 billion CNY of net cost would be saved through 2050. The benefits in this study only account for the fuel cost; neither an environmental nor a human health benefit is included.
3. Conclusions Vehicular evaporative emissions are becoming a more significant source of VOCs in China. In 2010, total VOC emissions from vehicular refueling, diurnal and hot soak processes were 185,000 ton. Single vehicle evaporative losses (1.17 g/L) have exceeded tailpipe VOC emissions of China 3 and China 4 vehicles. Beijing, Yangzi River Delta and Pearl River Delta (the most economic developed regions in China) bear the highest intensity of vehicular evaporative VOC emissions, even though Stage II systems have already been applied in these regions. The current control of refueling losses by implementing Stage II in the urban built-up area cannot turn over emission growth trend in next 30 years. With current control regulations, vehicular VOC emission will exceed 800,000 ton in 2040, over 4 times of 2010 emission. The best choice to control future evaporative losses is to require ORVR system with China 5 vehicle emission standards combined with current Stage II regulation. This measure would reduce VOC emissions in the coming five years and gradually reduce the emissions to a very low level (less than 5% compared with 2010) by 2030. The cost of installing ORVR on a car would be balanced by fuel saving in 9-10 years.
Acknowledgments This work was supported by the National Natural Science Foundation of China (No. 71101078), the National High Technology Research and Development Program of China (No. 2013AA065303D), and the National Environmental Protection Public Welfare Research Fund (No. 201209003 and No.
Fig. 4 – Cost-benefit of two control technologies and four control scenarios. (a) A single Stage II system; (b) a single ORVR system; (c) national wide costs and benefits.
J O U RN A L OF E N V I RO N ME N TA L SC IE N CE S 3 3 (2 0 1 5) 8 8– 9 6
201409021). The authors would like to gratefully acknowledge Mr. David Vance Wagner from ICCT for polishing English of this paper.
Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.jes.2015.01.012.
REFERENCES CE-CERT (Center for Environmental Research and Technology, College of Engineering, University of California at Riverside), GSSR (Global Sustainable Systems Research), ISSRC (the International Sustainable Systems Research Center), 2008. IVE model user manual: version 2.0. Available at:, Http.www.issrc. org/ive (last access: 1 May 2014). Chan, C.K., Yao, X.H., 2008. Air pollution in mega cities in China. Atmos. Environ. 42 (1), 1–42. Dargay, J., Gately, D., 1999. Income's effect on car and vehicle ownership, worldwide: 1960–2015. Transp. Res. A Policy Pract. 33 (2), 101–138. http://dx.doi.org/10.1016/S0965-8564(98)00026-3. Duan, J.C., Tan, J.H., Yang, L., Wu, S., Hao, J.M., 2008. Concentration, sources and ozone formation potential of volatile organic compounds (VOCs) during ozone episode in Beijing. Atmos. Res. 88 (1), 25–35. http://dx.doi.org/10.1016/j. atmosres.2007.09.004. Geng, F.H., Zhao, C.S., Tang, X., Lu, G.L., Tie, X.X., 2007. Analysis of ozone and VOCs measured in Shanghai: a case study. Atmos. Environ. 41 (5), 989–1001. http://dx.doi.org/10.1016/j.atmosenv. 2006.09.023. Gentner, D.R., Harley, R.A., Miller, A.M., Goldstein, A.H., 2009. Diurnal and seasonal variability of gasoline-related volatile organic compound emissions in Riverside, California. Environ. Sci. Technol. 43 (12), 4247–4252. http://dx.doi.org/10.1021/ es9006228. Guo, H., Cheng, H., Ling, Z., Louie, P.K.K., Ayoko, G.A., 2011a. Which emission sources are responsible for the volatile organic compounds in the atmosphere of Pearl River Delta? J. Hazard. Mater. 188 (1-3), 116–124. http://dx.doi.org/10.1016/j. jhazmat.2011.01.081. Guo, H., Zou, S., Tsai, W., Chan, L., Blake, D.R., 2011b. Emission characteristics of nonmethane hydrocarbons from private cars and taxis at different driving speeds in Hong Kong. Atmos. Environ. 45 (16), 2711–2721. http://dx.doi.org/10.1016/j. atmosenv.2011.02.053. He, K., Huo, H., Zhang, Q., He, D., An, F., Wang, M., et al., 2005. Oil consumption and CO2 emissions in China's road transport: current status, future trends and policy implications. Energ Policy 33, 1499–1507. Hu, A.G., 2007. China's medium-and long-term integrated demographic development strategy (2000–2050). J. Tsinghua Univ. 22 (5), 84–91. Huo, H., Wang, M., 2012. Modeling future vehicle sales and stock in China. Energ Policy 43, 17–29. http://dx.doi.org/10.1016/j.enpol. 2011.09.063. Huo, H., He, K.B., Wang, M., Yao, Z.L., 2012a. Vehicle technologies, fuel-economy policies, and fuel-consumption rates of Chinese vehicles. Energ Policy 43, 30–36. http://dx.doi.org/10.1016/j. enpol.2011.09.064. Huo, H., Yao, Z.L., Zhang, Y.Z., Shen, X.B., Zhang, Q., Ding, Y., et al., 2012b. On-board measurements of emissions from light-duty gasoline vehicles in three mega-cities of China. Atmos.
95
Environ. 49, 371–377. http://dx.doi.org/10.1016/j.atmosenv. 2011.11.005. Jong, G.D., Fox, J., Daly, A., Pieters, M., Smitb, R., 2004. Comparison of car ownership models. Transp. Rev. 24 (4), 379–408. http://dx.doi.org/10.1080/0144164032000138733 . Lee, S.C., Chiu, M.Y., Ho, K.F., Zou, S.C., Wang, X.M., 2002. Volatile organic compounds (VOCs) in urban atmosphere of Hong Kong. Chemosphere 48 (3), 375–382. Lei, Y., Zhang, Q., He, K.B., Streets, D.G., 2011. Primary anthropogenic aerosol emission trends for China, 1990–2005. Atmos. Chem. Phys. 11 (3), 931–954. http://dx.doi.org/10.5194/ acp-11-931-2011. Li, S.T., Hou, Y.Z., Zhai, F., 2003. Potential and projection of China's economic growth in the next fifty years. Rev. Econ. Res. (2), 51–60. Liu, Y., Shao, M., Fu, L.L., Lu, S.H., Zeng, L.M., Tang, D.G., 2008. Source profiles of volatile organic compounds (VOCs) measured in China: part I. Atmos. Environ. 42 (25), 6247–6260. http://dx.doi.org/10.1016/j.atmosenv.2008.01.070. Liu, X.H., Zhang, Y., Xing, J., Zhang, Q., Wang, K., Streets, D.G., et al., 2010. Understanding of regional air pollution over China using CMAQ, part II. Process analysis and sensitivity of ozone and particulate matter to precursor emissions. Atmos. Environ. 44 (30), 3719–3727. http://dx.doi.org/10.1016/j. atmosenv.2010.03.036. MECA, 2011. Written comments of the Manufacturers of Emission Controls Association on the U.S. Environmental Protection Agency's proposal concerning widespread use for onboard refueling vapor recovery and stage II waiver docket ID No. EPA-HQ-OAR-2010-1076. Available at:, http://www.meca.org/ galleries/files/EPA_ORVR_091311.pdf (Date accessed: 01 October 2014). National Bureau of Statistics, 2012a. China Statistical Yearbook. China Statistic Press, Beijing, China. National Bureau of Statistics, 2012b. China Energy Statistical Yearbook. China Statistic Press, Beijing, China. Reddy, S.R., 2010. Mathematical models for predicting vehicle refueling vapor generation. SAE Technical Paper http://dx.doi.org/10.4271/2010-01-1279. Song, Y., Shao, M., Liu, Y., Lu, S., Kuster, W., Goldan, P., Xie, S.D., 2007. Source apportionment of ambient volatile organic compounds in Beijing. Environ. Sci. Technol. 41 (12), 4348–4353. http://dx.doi.org/10.1021/es0625982. Srivastava, A., Sengupta, B., Dutta, S.A., 2005. Source apportionment of ambient VOCs in Delhi City. Sci. Total Environ. 343 (1), 207–220. http://dx.doi.org/10.1016/j.scitotenv. 2004.10.008. US EPA, 1991. Technical guidance — stage II vapor recovery system for control of vehicle refueling emissions at gasoline dispensing facilities. Available at:, http://www.epa.gov/ttn/ atw/gasdist/technica.pdf (Date accessed: 01 October 2014). US EPA, 2001. Modeling hourly diurnal emissions and interrupted diurnal emissions based on real-time diurnal data. Available at:, http://www.epa.gov/otaq/models/mobile6/r01019. pdf (Date accessed: 01 October 2014). US EPA, 2004. Stage II vapor recovery systems issues paper. In: Driscol (Ed.), U. S. EPA Office of Air Quality Planning and Standards, Emissions Monitoring and Analysis Division, Emissions Factors and Policy Applications Group (D243 02). Research Triangle Park (Available at: http://www.epa.gov/ttn/ naaqs/ozone/ozonetech/stage2/isspaper081204.pdf. Date accessed: 01 October 2014). US EPA, 2008. Transportation and marketing of petroleum fuels. Available at:, http://www.epa.gov/ttnchie1/ap42/ch05/final/ c05s02.pdf (Date accessed: 01 October 2014). US EPA, 2012a. Development of evaporative emissions calculations for the motor vehicle emissions simulator MOVES 2010. Available at:, http://www.epa.gov/otaq/models/moves/ documents/420r12027.pdf (Date accessed: 01 October 2014).
96
J O U RN A L OF E N V I RO N ME N TA L SC IE N CE S 3 3 (2 0 1 5) 88 – 9 6
US EPA, 2012b. Air quality: widespread use for onboard refueling vapor recovery and stage II waiver. Available at:, http://www. federalregister.gov/articles/2012/05/16/2012-11846/air-qualitywidespread-use-for-onboard-refueling-vapor-recovery-andstage-ii-waiver (Date accessed: 01 October 2014). US NRC, 2002. Effectiveness and impact of Corporate Average Fuel Economy (CAFE) standards. Available at:, http://www.nhtsa. gov/cars/rules/cafe/docs/162944_web.pdf (Date accessed: 01 October 2014).
Walsh, M.P., 2014. PM2.5: global progress in controlling the motor vehicle contribution. Front. Environ. Sci. Eng. 8 (1), 1–17. Wang, T., 2008. Study on Oil Consumption and Pollutants Emission Modeling of China On-road Vehicles. Master thesis. Tsinghua, Beijing, China. Yao, Y.F., He, J.H., 1994. Outlook of economic and social development in China (1990–2050). Rev. Econ. Res. (36), 24–35.
Available online at www.sciencedirect.com
ScienceDirect www.journals.elsevier.com/journal-of-environmental-sciences
Vehicular volatile organic compounds losses due to refueling and diurnal process in China: 2010–2050 Xiaofan Yang1 , Huan Liu1,2,3,⁎, Hongyang Cui1 , Hanyang Man1 , Mingliang Fu1 , Jiming Hao1,2,3 , Kebin He1,2,3 1. School of Environment, State Key Joint Laboratory of Environment Simulation and Pollution Control, Tsinghua University, 100084, China. E-mail: [email protected] 2. State Environmental Protection Key Laboratory of Sources and Control of Air Pollution Complex, Beijing 100084, China 3. Collaborative Innovation Center for Regional Environmental Quality, China
AR TIC LE I N FO
ABS TR ACT
Article history:
Volatile organic compounds (VOCs) are crucial to control air pollution in major Chinese
Received 25 September 2014
cities since VOCs are the dominant factor influencing ambient ozone level, and also an
Revised 5 January 2015
important precursor of secondary organic aerosols. Vehicular evaporative emissions have
Accepted 19 January 2015
become a major and growing source of VOC emissions in China. This study consists of lab
Available online 29 April 2015
tests, technology evaluation, emissions modeling, policy projections and cost-benefit analysis to draw a roadmap for China for controlling vehicular evaporative emissions. The
Keywords:
analysis suggests that evaporative VOC emissions from China's light-duty gasoline vehicles
Evaporative VOC emissions
were approximately 185,000 ton in 2010 and would peak at 1,200,000 ton in 2040 without
China roadmap
control. The current control strategy implemented in China, as shown in business as usual
Vehicle
(BAU) scenario, will barely reduce the long-term growth in emissions. Even if Stage II
Refueling emission control
gasoline station vapor control policies were extended national wide (BAU + extended Stage II), there would still be over 400,000 ton fuel loss in 2050. In contrast, the implementation of on-board refueling vapor recovery (ORVR) on new cars could reduce 97.5% of evaporative VOCs by 2050 (BAU + ORVR/BAU + delayed ORVR). According to the results, a combined Stage II and ORVR program is a comprehensive solution that provides both short-term and long-term benefits. The net cost to achieve the optimal total evaporative VOC control is approximately 62 billion CNY in 2025 and 149 billion CNY in 2050. © 2015 The Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences. Published by Elsevier B.V.
Introduction Understanding volatile organic compound (VOC) emission is crucial to control air pollution in major Chinese cities since VOCs are the dominant factor influencing ambient ozone level as well as secondary organic aerosols (SOAs) (Guo et al., 2011a; Lei et al., 2011; Duan et al., 2008; Chan and Yao, 2008; Geng et al., 2007; Liu et al., 2010). In major city clusters, vehicular emissions have become the most important source in VOC emission
inventory. For example, vehicles were responsible for 26%–50% of VOC emissions in inland Pearl River Delta (PRD) cities in 2004– 2007, and 48% ± 4% in Hong Kong from 2001 to 2007 (Srivastava et al., 2005; Guo et al., 2011b; Gentner et al., 2009; Lee et al., 2002; Liu et al., 2008; Walsh, 2014). Vehicular VOC emissions include tailpipe emissions and evaporative losses, e.g., release of gasoline vapors resulting from diurnal temperature variations, refueling, hot soak, running losses and permeation. In developed countries, vehicular evaporative VOC emission
⁎ Corresponding author. E-mail: [email protected] (Huan Liu).
http://dx.doi.org/10.1016/j.jes.2015.01.012 1001-0742/© 2015 The Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences. Published by Elsevier B.V.
J O U RN A L OF E N V I RO N ME N TA L SC IE N CE S 3 3 (2 0 1 5) 8 8– 9 6
without accounting losses during refueling process usually account for 30% of gasoline-related VOCs. A 2007 research study notes that evaporative emissions contribute to 12% of total VOC emissions in Beijing with an ozone formation potential that is higher than for most other VOC sources (Song et al., 2007). Although vehicular evaporative VOC losses are important, the regulating and controlling process is just getting started in China and the roadmap of evaporative emission control is still a controversy. In this article, we evaluated different VOC control scenarios based on two existing technologies: the Stage II vapor recovery system (Stage II) and on-board refueling vapor recovery (ORVR). Stage II collects gasoline vapors from vehicle fuel tanks while customers dispense gasoline at gasoline dispensing facilities (US EPA, 1991). ORVR systems are carbon canisters installed in automobiles to capture gasoline vapors evacuated from the gasoline tank. In China, Stage II has already been required in some major urban areas since 2007, whereas numerous service stations, scattered across medium and small cities, suburbs and freeways, remain uncontrolled. Stage II is widely used in Europe, while both Stage II and ORVR were used in the US until 2012. The US EPA has started to phase-out Stage II because of the widespread penetration of ORVR. The US EPA identified ORVR as the best available control technology (BACT) because of its high efficiency, ability to function without specific monitoring or maintenance, and low cost. The purposes of this article include: (1) understanding current controls and evaporative emissions from vehicles' refueling and diurnal processes in China; (2) evaluating a possible future control roadmap and estimating the effects on emissions reductions; (3) generating cost-benefit analyses for the scenarios evaluated. In this paper, we established emission models to evaluate total evaporative emissions. Five policy scenarios were designed for future vehicular VOC control based on two available technologies. For each technology, the basic emission rates were generated by theoretical calculation, while control efficiency of different technologies was measured by laboratory tests. In addition, cost and benefits of each scenario were further evaluated to provide a comprehensive understanding of future policy options.
1. Methodology 1.1. Real-world emission rates calculation (non-control) The basic emission rate of refueling VOCs without any control can by calculated by the Reddy–Wade equation. This equation was developed in 2010 and now is widely used in all over the world (Reddy, 2010; US EPA, 2008). The real-world refueling vapor emission is influenced by local temperature and the Reid vapor pressure (RVP) of the gasoline. According to a fuel survey in China (Huo et al., 2012a), the RVP of market fuel ranges from 42.5 to 89.0 kPa, and the average is 61.3 kPa. Thus 60 kPa was selected as a representative value in this analysis. For the emission inventory, temperatures in each province from 6:00 am to 21:00 pm, during which most refueling occurs, were used.
89
where, Q(g/L) is the concentration of VOCs in the air that is displaced to the environment, R(gal·psi)/(g·mol·K) is the universal gas constant 0.3187, f(g/gal) is the air entrainment 0.2, Td(K) is the temperature of the dispensed fuel, Tv(K) is the temperature of the vehicle's fuel tank, and Tv–Td is approximately 2 K. Patm(psi) is the atmosphere pressure, Ptank(psi) is the tank fuel vapor pressure, and Pdisp(psi) is the dispensed fuel vapor pressure. Ptank and Pdisp could be further calculated by the RVP and temperature: P ¼ A T RVP expð−B=T Þ
ð2Þ
where, A and B are the constants 25.61 and 2789.78, respectively, RVP is the fuel RVP in psi, and T(K) is the temperature. Evaporative VOC emissions also include hot soak and diurnal emissions. Evaporative emissions from a vehicle not related to refueling are very complex. They range from breathing losses from the vehicle tank to leaks around gaskets and hoses. Additionally, VOCs can escape around engine cylinders and then leak from the oil pan. The US EPA developed an empirical equation for a 24-hr evaporative emission factor (US EPA, 2001, 2012a). The International Vehicle Emission (IVE) model uses adjustment factors for evaporative emission factors to address the difference between real fuel vapor pressure and the EPA experience value (CE-CERT (Center for Environmental Research) et al., 2008). We use both the US EPA equation and the IVE adjustment factors to generate diurnal emission rates for Chinese fuel vapor pressure (8.7 psi). To convert the gasoline consumption-based emission factor into a distance-specific emission factor, the average mileage of Chinese passenger vehicles was introduced using data from previous research (Huo et al., 2012a).
1.2. Lab tests of ORVR control efficiency A series of refueling emission tests were finished to determine efficiency of ORVR technology using local vehicles and gasoline. Two in-use vehicles equipped with original ORVR systems were recruited: (1) a 2009 Chrysler–Dodge Journey with a 2.7 L ORVR canister and 77.6 L fuel tank and (2) a 2012 Chrysler–Jeep Compass with a 2.0 L ORVR canister and a 51.1 L fuel tank. The whole test was conducted in Gas-tight Imtech Variable Temperature Sealed Housing for Evaporative Determination chambers (VT-SHED, VT-SHED software, Imtech Automotive Testing Solutions Beijing Co., Ltd. Beijing, China). This SHED is the most advanced one in China. This system is equipped with a canister loading device, fuel temperature control system, as well as total evaporative emission test system. China Automotive Technology & Research Center Beijing laboratory (CATARC) provided all the facilities and technical support. Each vehicle was tested five times to ensure the reliability. For each test, the processes include canister purge, canister adsorption to breakthrough, type VII driving defined by China vehicle emission control standard, fuel drain, soak, refueling test and canister weighing in quadruplicate. A detailed test procedure was provided in supporting materials, Fig. S1.
1.3. Current control policy and future scenario design Q ¼ 18:2
P disp f P atm P atm −P tank þ R Td R Tv P atm −P disp
ð1Þ
The Chinese government launched regulation of emissions from service stations in 2007, requiring Stage II to be installed
90
J O U RN A L OF E N V I RO N ME N TA L SC IE N CE S 3 3 (2 0 1 5) 88 – 9 6
in selected locations before 2015. The Ministry of Environmental Protection (MEP) conducted investigations to track the implementation of Stage II, and the following levels of implementation were found. (1) Beijing: completed upgrading all of the service stations by 30 June, 2008; (2) Tianjin and cities in Hebei Province: 60% completed by 30 June, 2008; (3) Pearl River Delta: 90% completed before October 2010; (4) Shanghai: 90% completed before the World Expo in May 2010; and (5) Other regions: no policies or plans yet. According to the results from the investigation, we readjusted the percentage of service stations equipped with Stage II on the basis of the governmental control policy. Considering the poor supervision of the Stage II system, the real operational percentage is even lower. Thus, this study has a positive projection of Stage II efficiency. There has been discussion in China on the possibility of requiring ORVR on new cars starting with the China 5 standard (equivalent to Euro 5 standard). Hence, potential control scenarios for the next few decades are shown in Table 1.
1.4. Mathematics of emissions projection under different scenarios Both refueling emissions and diurnal/hot soak emissions are taken into consideration in these projections. The emissions projection equations under different scenarios were listed in Table 1. Emiref and Emidh are VOC emissions through refueling and diurnal/hot soak processes without any control (g/year). They could be calculated by using following equations:
ηORVR and ηCO (%) refer to the removal efficiency of Stage II, ORVR, and when ORVR and Stage II systems work simultaneously. GNum and GNumII are the numbers of total service stations and service stations that have applied Stage II. VP and VPORVR refer to the vehicle population and population with ORVR systems. The calculations for scenarios b and c are based on the proportion of non-controlled and Stage II-controlled sites. The calculations for scenarios d and e are more complicated. A matrix, including two vehicle types (ORVR and non-ORVR) and two service station types (Stage II and non-Stage II), is used to calculate the refueling emissions under four refueling possibilities. The hypothesis is that vehicles with or without ORVR have the same chance to be refueled at a Stage II station, and the possibility only depends on the percentage of stations with Stage II in the province. All of the input data and their sources are summarized in Table 2. In total, 93,173 service stations are included in this study as announced by MEP. Gasoline vehicle population was projected by the Gompertz equation, a widely accepted approach to forecast vehicle population based on gross domestic production (GDP) per person (Jong et al., 2004; Dargay and Gately, 1999). The historical data of vehicle population and gasoline consumption comes from the Year Book of National Bureau of Statistics of China (till 2010). The economic and population growth outlook is a compilation from several different researchers in China (Hu, 2007; Li et al., 2003; Yao and He, 1994).
2. Results and discussion
Emire f ¼ EFre f FC
ð3Þ
2.1. Emission factors of evaporative VOCs
Emidh ¼ EFdh VKT VP
ð4Þ
The measured average refueling emission factor was 0.961 (+0.078, −0.123) g/L. The difference between the measured value and the Reddy equation result (based on test conditions) was very small (5.67%) and could be ignored. Taking realworld temperatures and RVP variation into consideration, the real-world refueling emission factor for China was 1.17 g/L as an average. To understand the importance of vehicular evaporative emissions in China, the evaporative refueling and diurnal emission factors were compared with tailpipe emission factors
where, EFref is the VOC emission factor through the refueling process (g/ton) and EFdh is the diurnal and hot soak emission factor based on mileage (g/km). Method to calculate these two emission factors were defined in Section 1.1. FC is the fuel consumption (ton/year); VKT is the average vehicle distance traveled (km/year). For scenarios other than non-control, the efficiency of control technologies (ORVR and Stage II) was taken into consideration. ηII,
Table 1 – Potential control scenarios designed for future vehicular evaporative volatile organic compound (VOC) control in China. Scenario
Stage II policies
Non-control BAU BAU + extended Stage II BAU + ORVR scenario
None Current policies Stage II is expanded to all service stations in China before 2025. Current policies
BAU + delayed ORVR
Current policies
ORVR policies None None None ORVR is required on new gasoline vehicle from 2016 ORVR is required on new gasoline vehicle from 2019
VOC emissions (Emi) calculation Emi = Emiref × Emidh Emi ¼ Emire f 1−ηII GGNumII þ Emidh Num h ORVR Emi ¼ Emire f 1−ηII GGNumII 1− VPVP −ηORVR Num i ORVR ORVR 1− GGNumII −ηCO GGNumII VPVP VPVP Num Num ORVR þEmidh ηORVR VPVP
BAU: business as usual; ORVR: on-board refueling vapor recovery; Emiref: VOC emissions through refueling; Emidh: VOC emissions through diurnal/hot soak processes without any control; GNum: the number of total service stations; GNumII: the number of service stations that have applied Stage II; VP: the vehicle population; VPORVR: the vehicle population with ORVR systems; ηII: the removal efficiency of Stage II; ηORVR: the removal efficiency of ORVR; ηCO: the removal efficiency when ORVR and Stage II system work simultaneously.
91
J O U RN A L OF E N V I RO N ME N TA L SC IE N CE S 3 3 (2 0 1 5) 8 8– 9 6
Table 2 – Input data for evaporative emission estimation. Input
Value
Source
Emission factors through the refueling process (non-control) Diurnal and hot soak emission factor (non-control) Temperature
1.621 kg/ton
Calculated as introduced in Section 1.1
0.6 kg/ton Average for each province
Information for service stations Vehicle mileage traveled Fuel economy
Location of each service station 20,000 km/year 11.03 km/L
EPA MOVES model Average measured temperature (6:00–22:00) from April 2011 to March 2012 in each province website MEP investigation and digital map Previous research (Wang, 2008) Previous research (Huo et al., 2012a; He et al., 2005)
MEP: the Ministry of Environmental Protection.
that are based on per-distance units. Fig. 1a includes tailpipe HC emission factors from the EMBEV model (a vehicle emission model developed by Tsinghua University) with a deterioration rate based on mileage traveled; Fig. 1b contains tailpipe HC emissions based on Portable Emissions Measurement System (PEMS) tests (Huo et al., 2012b). It can be easily concluded that, total evaporative VOC emissions (0.12 g/km), including emissions from refueling, diurnal and hot soak processes, exceed both the tailpipe standards and in-use tailpipe levels of China 3 (equivalent to Euro 3 standard) and China 4 (equivalent to Euro 4 standard) vehicles (0.09 and 0.02 g/km). From another perspective, the overall evaporative VOC emissions from any gasolinefueled car in China are approximately equivalent to exhaust emissions from a 5-year-old China 3 car or a 7-year-old China 4 car. Starting to control VOC emission from non-tailpipe sources is important.
2.2. Control efficiencies of the two technologies ORVR systems show good measurement repeatability and efficiency in the tests, which were described in Section 1.2. The range of efficiencies was 98.92% to 99.74% (average 99.40%). The efficiency in the model was set as 99%. Compared with other literatures, this efficiency is reasonable. EPA's documentation associated with its 1994 ORVR final rule making package projected an average efficiency for ORVR of 95% based on 1160 in-use vehicle tests. However, more recently, EPA's extensive testing of in-use vehicles has demonstrated a fleet efficiency of approximately 98% (US EPA, 2012b). Besides, the 99% reduction
efficiency was also adopted for recovery of hot soak and diurnal emissions since an ORVR system works exactly the same as for refueling vapor. And according to US Manufacturers of Emission Controls Association (MECA), the overall reduction of efficiency of ORVR is above 98%, equalling to reduction efficiency through refueling process (MECA, 2011). Furthermore, no significant deterioration of ORVR system was spotted during life time of a car. Therefore, lifetime of vehicles was used in the model instead of lifetime of ORVR system. The efficiency of Stage II system was generated by literature review. EPA state implementation plan guidance estimates that 92% efficiency can be achieved with semi-annual inspection and strong follow-up enforcement, while an efficiency of 86% has been associated with annual inspection. A lower end of 62% has been associated with minimal inspection (US EPA, 1991). In 2009, MEP investigated the service stations equipped with Stage II systems in Beijing, Tianjin, Hebei and the Pearl River Delta. The results reveal that, after running for one year, the passing rate of Stage II systems was very low, especially in seal reliability. An overall efficiency of approximately 70% has been estimated. When Stage II and ORVR are used simultaneously, the recovery efficiency will be reduced because of the lack of compatibility between the two controls. The ORVR canister captures the gasoline emissions from the motor vehicle fuel tank. Instead of drawing vapor-laden air from the vehicle fuel tank into the underground storage tank, the vacuum pump of the Stage II system draws fresh air into the underground storage tank. The fresh air causes gasoline in the underground
Fig. 1 – Emission rate comparisons between evaporative and tailpipe volatile organic compounds (VOCs). (a) Tailpipe emission factors derived from new vehicle emission standards considering vehicle deterioration; (b) tailpipe emission factors derived from in use vehicle Portable Emissions Measurement System (PEMS) tests. Chinese government adopted vehicle emission standards following emission standards in Europe since 1999. The emission levels 1 to 4 in China (China 1–4) are equivalent to Euro 1 to 4 standards, respectively.
92
J O U RN A L OF E N V I RO N ME N TA L SC IE N CE S 3 3 (2 0 1 5) 88 – 9 6
tank to evaporate inside the underground tank and thus creates an increase in pressure in the underground storage tank. As a result, gasoline vapors may be forced out of the underground storage tank vent pipe into the ambient air. Generally speaking, the recovery rate decreases by 1%–10% when Stage II and ORVR work at the same time (94% in this research) (US EPA, 2012a).
2.3. Vehicle population and gasoline consumption projections The projections for future vehicle population and gasoline consumption are shown in Fig. 2. According to the projection, gasoline vehicle in China will reach 584 million by 2050. Compared with the projection of Huo and Wang (2012), which is 558 million by 2050, projection in this research is 4.7% higher. This is because the historical data in this study was updated to 2011 and which was higher than previous forecast. To get an outlook of future gasoline consumption in China, fuel economy and its future development is also required. Basic fuel economy used here is from He et al. (2005) and Huo et al. (2012a). For future fuel economy improvement, we adopted the moderate scenario from the three scenarios designed by National Academy of Science (NAS) (US NRC, 2002; Wang, 2008).
2.4. Evaporative emission inventory of China in 2010 The total vehicular evaporative emissions (refueling, diurnal and hot-soak) were calculated to reach approximately 185,000 ton in 2010, of which 124,000 ton were from refueling and the remaining 61,000 ton were from diurnal and hot soak processes. The average refueling loss is approximately 0.26% for total gasoline consumption in China (68,862,000 ton by China energy statistic). This estimation is very reasonable compared with US experience. In 1999, U.S. gasoline loss through refueling was approximately 0.1% of total gasoline consumption (475,000 ton emission versus 455,000,000 ton gasoline) (US EPA, 2004). By 1999, there were over 31,000 gas stations in the US with Stage II vapor recovery, and there had already been two years of new ORVR vehicles introduced into the fleet. Thus, the emission factor was reduced from 1.32 g/L in AP-42 to 1.1 g/L in the US.
With provincial statistical data from National Bureau of Statistics (2012a, 2012b), emission rates of each province with available data were calculated following the equations in methodology (Table 3). As shown in Table 3, VOC emissions per unit gasoline consumed in regions under Stage-II control are significantly lower than other regions. However, the control effect is still not satisfying. The emission distribution shows that VOC emission intensity is higher in well-developed regions – especially Beijing, Tianjin, the Yangzi River Delta and the Pearl River Delta – where ozone and PM2.5 pollution are also serious. Service stations in these regions are under Stage-II control, but vehicle population or gasoline consumption density is higher, which lead to high vapor losses in total.
2.5. Evaporative emissions in 2010–2050 The projected outlook for vehicular evaporative VOC emission is shown in Fig. 3. Without control, the peak would reach almost 1,200,000 ton/year approximately in 2040, which would be ten times the current emissions rate. With Stage II control under the BAU scenario, emissions would reach 809,000 ton in 2050, about 21.5% reduction compared with non-control scenario. The emission trend of BAU scenario shows that refueling losses would be controlled during the initial few years. However, the current control of Stage II only in urban districts is not able to restrain the increasing trend of VOC emissions due to the fast-growing vehicle population. Even with an extended Stage II strategy (scenario c), which means all of the service stations after 2025 are equipped with Stage II, the future VOC emission growth would only be slowed down rather than being reduced. On one hand, there is a significant contribution of new emissions from new vehicles each year, which is represented by the gray belt in Fig. 3. On the other hand, vapor losses from diurnal and hot soak are non-control, approximately 300,000 ton by 2050. The implementation of ORVR on new cars (scenario d) could reduce 97.5% of total evaporative VOCs by 2030 and maintain at a low emission level. The reason for this huge reduction is the high new vehicle penetration rate, high sales and phaseout rate, in China. According the vehicle population growth
Fig. 2 – Gasoline passenger vehicle population, fuel consumption projection and control policy. BAU: business as usual; ORVR: on-board refueling vapor recovery.
93
J O U RN A L OF E N V I RO N ME N TA L SC IE N CE S 3 3 (2 0 1 5) 8 8– 9 6
Table 3 – Provincial evaporative VOC emission through refueling process. Province
Vehicle population (million)
Gasoline consumption (million ton)
Emission (ton)
Emission rate (kg per ton gasoline consumed)
Beijing Tianjin Hebei Shanxi Inner Mongolia Liaoning Jilin Heilongjiang Shanghai Jiangsu Zhejiang Anhui Fujian Jiangxi Shandong Henan Hubei Hunan Guangdong Guangxi Hainan Chongqing Sichuan Guizhou Yunnan Shaanxi Gansu Qinghai Ningxia Xinjiang
4.44 1.68 4.63 2.32 1.76 2.76 1.45 1.73 1.64 5.87 5.56 1.81 1.86 1.23 6.96 3.90 1.87 2.01 7.45 1.43 0.37 0.99 3.41 1.01 2.07 1.92 0.73 0.29 0.36 1.15
3.72 2.05 2.39 2.28 3.26 5.93 1.67 3.64 4.15 7.50 5.87 1.57 3.33 1.55 8.02 2.97 4.58 2.62 10.86 2.48 0.53 1.03 5.42 1.43 2.32 2.55 0.57 0.26 0.23 1.31
1219 657 2148 2203 2679 4808 1199 2375 1646 9340 7723 1983 4888 2137 9433 3575 6135 3539 17299 3875 873 1417 6746 1640 2950 3027 474 189 218 1005
0.33 0.32 0.90 0.97 0.82 0.81 0.72 0.65 0.40 1.25 1.32 1.26 1.47 1.38 1.18 1.20 1.34 1.35 1.59 1.56 1.65 1.38 1.24 1.15 1.27 1.19 0.83 0.73 0.95 0.77
Data of passenger vehicles from available provinces are included in this table.
projection, average annual growth rate of gasoline vehicle will be 8.8% during 2015–2020, and the average annual phase-out ratio of old vehicles is about 5.3%. The emission difference between new and old vehicles is about 99%. Thus, the ORVR could reduce emission by the second year following implementation. Therefore, if ORVR implemented earlier, it's easier to bring more significant reduction in a short time. If the ORVR strategy was delayed to 2019, when the China 6 standard (equivalent to Euro 6 standard) is expected to be implemented
(scenario e), a cumulative total of 1,570,000 ton more VOCs will be emitted to the atmosphere, which is approximately 8 times higher than the total 2010 emissions. Considering the potentially high diurnal and hot soak emissions in the future in China, ORVR seems to be more advantageous than Stage II. If no ORVR is to be adopted, the amount of diurnal and hot soak emissions will rise beyond the total vehicular VOCs of 2010 by the year 2030 and become the most significant source of vehicle emissions.
Fig. 3 – Projection of vehicular evaporative volatile organic compound (VOC) emissions in China to 2050. (a) Total emission trend; (b) portion of diurnal and hot-soak vs. refueling emissions.
94
J O U RN A L OF E N V I RO N ME N TA L SC IE N CE S 3 3 (2 0 1 5) 88 – 9 6
2.6. Cost and benefit analysis Fig. 4 shows the cost and benefit results of the two single technologies, as well as different scenarios. The costs for installing Stage II system in a single gas station is 300,000 CNY. And annual running cost, including system test, nozzle replacement and other maintenance, is 90,500 CNY per year on average. For ORVR (Fig. 4b), average cost is approximately 200 CNY per vehicle and no maintenance cost is required. Assuming that all the VOCs recovered are turned to gasoline, a gas station equipped with Stage II system never recovers its costs during lifetime. On the other hand, 0.145 L gasoline was saved per 1000 km traveled for an ORVR car, which means an ORVR vehicle would achieve cost-benefit balance point in 9–10 years. The BAU + extended stage II scenario has the highest net cost but the second lowest benefit (Fig. 4c). The BAU + ORVR scenario has lower net cost but the highest recovery profit. If ORVR was delayed, compared with the BAU + ORVR scenario, the net cost is about the same but fuel recovery profit is 16 billion CNY lower in 2025. To achieve the optimal total VOC control (BAU + ORVR scenario), the total investment would be 109 billion CNY in 2025 and 466 billion CNY in 2050 (the net cost is approximately 62 billion CNY in 2025 and 149 billion CNY in 2050). The investment for Stage II systems is the majority of the total investment (80%). Based on US experience, Stage II systems can be phased out when ORVR penetration reaches widespread use (or approximately 75% of vehicles). If ORVR was implemented by 2015, the widespread use of ORVR would be accomplished by about 2026 (as illustrated in Fig. 2). In this situation, 170 billion CNY of net cost would be saved through 2050. The benefits in this study only account for the fuel cost; neither an environmental nor a human health benefit is included.
3. Conclusions Vehicular evaporative emissions are becoming a more significant source of VOCs in China. In 2010, total VOC emissions from vehicular refueling, diurnal and hot soak processes were 185,000 ton. Single vehicle evaporative losses (1.17 g/L) have exceeded tailpipe VOC emissions of China 3 and China 4 vehicles. Beijing, Yangzi River Delta and Pearl River Delta (the most economic developed regions in China) bear the highest intensity of vehicular evaporative VOC emissions, even though Stage II systems have already been applied in these regions. The current control of refueling losses by implementing Stage II in the urban built-up area cannot turn over emission growth trend in next 30 years. With current control regulations, vehicular VOC emission will exceed 800,000 ton in 2040, over 4 times of 2010 emission. The best choice to control future evaporative losses is to require ORVR system with China 5 vehicle emission standards combined with current Stage II regulation. This measure would reduce VOC emissions in the coming five years and gradually reduce the emissions to a very low level (less than 5% compared with 2010) by 2030. The cost of installing ORVR on a car would be balanced by fuel saving in 9-10 years.
Acknowledgments This work was supported by the National Natural Science Foundation of China (No. 71101078), the National High Technology Research and Development Program of China (No. 2013AA065303D), and the National Environmental Protection Public Welfare Research Fund (No. 201209003 and No.
Fig. 4 – Cost-benefit of two control technologies and four control scenarios. (a) A single Stage II system; (b) a single ORVR system; (c) national wide costs and benefits.
J O U RN A L OF E N V I RO N ME N TA L SC IE N CE S 3 3 (2 0 1 5) 8 8– 9 6
201409021). The authors would like to gratefully acknowledge Mr. David Vance Wagner from ICCT for polishing English of this paper.
Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.jes.2015.01.012.
REFERENCES CE-CERT (Center for Environmental Research and Technology, College of Engineering, University of California at Riverside), GSSR (Global Sustainable Systems Research), ISSRC (the International Sustainable Systems Research Center), 2008. IVE model user manual: version 2.0. Available at:, Http.www.issrc. org/ive (last access: 1 May 2014). Chan, C.K., Yao, X.H., 2008. Air pollution in mega cities in China. Atmos. Environ. 42 (1), 1–42. Dargay, J., Gately, D., 1999. Income's effect on car and vehicle ownership, worldwide: 1960–2015. Transp. Res. A Policy Pract. 33 (2), 101–138. http://dx.doi.org/10.1016/S0965-8564(98)00026-3. Duan, J.C., Tan, J.H., Yang, L., Wu, S., Hao, J.M., 2008. Concentration, sources and ozone formation potential of volatile organic compounds (VOCs) during ozone episode in Beijing. Atmos. Res. 88 (1), 25–35. http://dx.doi.org/10.1016/j. atmosres.2007.09.004. Geng, F.H., Zhao, C.S., Tang, X., Lu, G.L., Tie, X.X., 2007. Analysis of ozone and VOCs measured in Shanghai: a case study. Atmos. Environ. 41 (5), 989–1001. http://dx.doi.org/10.1016/j.atmosenv. 2006.09.023. Gentner, D.R., Harley, R.A., Miller, A.M., Goldstein, A.H., 2009. Diurnal and seasonal variability of gasoline-related volatile organic compound emissions in Riverside, California. Environ. Sci. Technol. 43 (12), 4247–4252. http://dx.doi.org/10.1021/ es9006228. Guo, H., Cheng, H., Ling, Z., Louie, P.K.K., Ayoko, G.A., 2011a. Which emission sources are responsible for the volatile organic compounds in the atmosphere of Pearl River Delta? J. Hazard. Mater. 188 (1-3), 116–124. http://dx.doi.org/10.1016/j. jhazmat.2011.01.081. Guo, H., Zou, S., Tsai, W., Chan, L., Blake, D.R., 2011b. Emission characteristics of nonmethane hydrocarbons from private cars and taxis at different driving speeds in Hong Kong. Atmos. Environ. 45 (16), 2711–2721. http://dx.doi.org/10.1016/j. atmosenv.2011.02.053. He, K., Huo, H., Zhang, Q., He, D., An, F., Wang, M., et al., 2005. Oil consumption and CO2 emissions in China's road transport: current status, future trends and policy implications. Energ Policy 33, 1499–1507. Hu, A.G., 2007. China's medium-and long-term integrated demographic development strategy (2000–2050). J. Tsinghua Univ. 22 (5), 84–91. Huo, H., Wang, M., 2012. Modeling future vehicle sales and stock in China. Energ Policy 43, 17–29. http://dx.doi.org/10.1016/j.enpol. 2011.09.063. Huo, H., He, K.B., Wang, M., Yao, Z.L., 2012a. Vehicle technologies, fuel-economy policies, and fuel-consumption rates of Chinese vehicles. Energ Policy 43, 30–36. http://dx.doi.org/10.1016/j. enpol.2011.09.064. Huo, H., Yao, Z.L., Zhang, Y.Z., Shen, X.B., Zhang, Q., Ding, Y., et al., 2012b. On-board measurements of emissions from light-duty gasoline vehicles in three mega-cities of China. Atmos.
95
Environ. 49, 371–377. http://dx.doi.org/10.1016/j.atmosenv. 2011.11.005. Jong, G.D., Fox, J., Daly, A., Pieters, M., Smitb, R., 2004. Comparison of car ownership models. Transp. Rev. 24 (4), 379–408. http://dx.doi.org/10.1080/0144164032000138733 . Lee, S.C., Chiu, M.Y., Ho, K.F., Zou, S.C., Wang, X.M., 2002. Volatile organic compounds (VOCs) in urban atmosphere of Hong Kong. Chemosphere 48 (3), 375–382. Lei, Y., Zhang, Q., He, K.B., Streets, D.G., 2011. Primary anthropogenic aerosol emission trends for China, 1990–2005. Atmos. Chem. Phys. 11 (3), 931–954. http://dx.doi.org/10.5194/ acp-11-931-2011. Li, S.T., Hou, Y.Z., Zhai, F., 2003. Potential and projection of China's economic growth in the next fifty years. Rev. Econ. Res. (2), 51–60. Liu, Y., Shao, M., Fu, L.L., Lu, S.H., Zeng, L.M., Tang, D.G., 2008. Source profiles of volatile organic compounds (VOCs) measured in China: part I. Atmos. Environ. 42 (25), 6247–6260. http://dx.doi.org/10.1016/j.atmosenv.2008.01.070. Liu, X.H., Zhang, Y., Xing, J., Zhang, Q., Wang, K., Streets, D.G., et al., 2010. Understanding of regional air pollution over China using CMAQ, part II. Process analysis and sensitivity of ozone and particulate matter to precursor emissions. Atmos. Environ. 44 (30), 3719–3727. http://dx.doi.org/10.1016/j. atmosenv.2010.03.036. MECA, 2011. Written comments of the Manufacturers of Emission Controls Association on the U.S. Environmental Protection Agency's proposal concerning widespread use for onboard refueling vapor recovery and stage II waiver docket ID No. EPA-HQ-OAR-2010-1076. Available at:, http://www.meca.org/ galleries/files/EPA_ORVR_091311.pdf (Date accessed: 01 October 2014). National Bureau of Statistics, 2012a. China Statistical Yearbook. China Statistic Press, Beijing, China. National Bureau of Statistics, 2012b. China Energy Statistical Yearbook. China Statistic Press, Beijing, China. Reddy, S.R., 2010. Mathematical models for predicting vehicle refueling vapor generation. SAE Technical Paper http://dx.doi.org/10.4271/2010-01-1279. Song, Y., Shao, M., Liu, Y., Lu, S., Kuster, W., Goldan, P., Xie, S.D., 2007. Source apportionment of ambient volatile organic compounds in Beijing. Environ. Sci. Technol. 41 (12), 4348–4353. http://dx.doi.org/10.1021/es0625982. Srivastava, A., Sengupta, B., Dutta, S.A., 2005. Source apportionment of ambient VOCs in Delhi City. Sci. Total Environ. 343 (1), 207–220. http://dx.doi.org/10.1016/j.scitotenv. 2004.10.008. US EPA, 1991. Technical guidance — stage II vapor recovery system for control of vehicle refueling emissions at gasoline dispensing facilities. Available at:, http://www.epa.gov/ttn/ atw/gasdist/technica.pdf (Date accessed: 01 October 2014). US EPA, 2001. Modeling hourly diurnal emissions and interrupted diurnal emissions based on real-time diurnal data. Available at:, http://www.epa.gov/otaq/models/mobile6/r01019. pdf (Date accessed: 01 October 2014). US EPA, 2004. Stage II vapor recovery systems issues paper. In: Driscol (Ed.), U. S. EPA Office of Air Quality Planning and Standards, Emissions Monitoring and Analysis Division, Emissions Factors and Policy Applications Group (D243 02). Research Triangle Park (Available at: http://www.epa.gov/ttn/ naaqs/ozone/ozonetech/stage2/isspaper081204.pdf. Date accessed: 01 October 2014). US EPA, 2008. Transportation and marketing of petroleum fuels. Available at:, http://www.epa.gov/ttnchie1/ap42/ch05/final/ c05s02.pdf (Date accessed: 01 October 2014). US EPA, 2012a. Development of evaporative emissions calculations for the motor vehicle emissions simulator MOVES 2010. Available at:, http://www.epa.gov/otaq/models/moves/ documents/420r12027.pdf (Date accessed: 01 October 2014).
96
J O U RN A L OF E N V I RO N ME N TA L SC IE N CE S 3 3 (2 0 1 5) 88 – 9 6
US EPA, 2012b. Air quality: widespread use for onboard refueling vapor recovery and stage II waiver. Available at:, http://www. federalregister.gov/articles/2012/05/16/2012-11846/air-qualitywidespread-use-for-onboard-refueling-vapor-recovery-andstage-ii-waiver (Date accessed: 01 October 2014). US NRC, 2002. Effectiveness and impact of Corporate Average Fuel Economy (CAFE) standards. Available at:, http://www.nhtsa. gov/cars/rules/cafe/docs/162944_web.pdf (Date accessed: 01 October 2014).
Walsh, M.P., 2014. PM2.5: global progress in controlling the motor vehicle contribution. Front. Environ. Sci. Eng. 8 (1), 1–17. Wang, T., 2008. Study on Oil Consumption and Pollutants Emission Modeling of China On-road Vehicles. Master thesis. Tsinghua, Beijing, China. Yao, Y.F., He, J.H., 1994. Outlook of economic and social development in China (1990–2050). Rev. Econ. Res. (36), 24–35.
