Microscopic Driving Parameters-Based Energy-Saving Effect Analysis under Different Electric Vehicle Penetration

Abstract

Due to the rapid motorization over the recent years, China's transportation sector has been facing an increasing environmental pressure. Compared with gasoline vehicle (GV), electric vehicle (EV) is expected to play an important role in the mitigation of CO₂ and other pollution emissions, and urban air quality improvement, for its zero emission during use and higher energy efficiency. This paper aims to estimate the energy saving efficiency of EV, especially under different EV penetration and road traffic conditions. First, based on the emission and electricity consumption data collected by a light-duty EV and a light duty GV, a set of electricity consumption rate models and gasoline consumption rate models are established. Then, according to the conversion formula of coal equivalent, these models are transformed into coal equivalent consumption models, which make gasoline consumption and electricity consumption comparable. Finally, the relationship between the EV penetration and the reduction of energy consumption is explored based on the simulation undertaken on the North Second Ring Road in Beijing. The results show that the coal equivalent consumption will decrease by about 5% with the increases of EV penetration by 10% and the maximum energy-saving effect can be achieved when the traffic volume is about 4000 pcu/h.

1. Introduction

The transport sector, a major greenhouse gas (GHG) emitter and oil consumer, accounted for about 23% of world's energy-related GHG emissions and 26% of energy use in 2004. China's transportation sector has also been facing an increasing environmental pressure due to the rapid motorization over the recent years [1 –4]. However, compared with developed countries, the vehicle ownership in China (58 per 1000 people as of 2010) is still very low [5]. There is no doubt that the China's vehicle population will continue to increase and the vehicle ownership is forecasted to reach about 120 per 1000 people by 2020 and 300 per 1000 people by 2030 [6]. Considering that the road transportation almost entirely depends on oil-derived fuels, it is highly vulnerable to possible oil price shocks and supply disruptions [2]. Therefore, if effective measures could not be taken timely, such a rapid growth and huge vehicle population in the next several decades will certainly pose severe challenges to air pollution emissions mitigation, urban air quality improvement, and energy security in China.

EVs are often labeled “green” for zero emission during use and higher energy efficiency, which implies that substitution of GVs with EVs can reduce pollution emission and energy consumption significantly [7, 8]. However, it is still difficult to evaluate the accurate and detailed energy-saving efficiency of EV, especially under different traffic condition. Recent researches have shown that lots of factors such as driving environment and operation modes (i.e., accelerating, decelerating, cruising, and idling) have impacts on the energy consumption rate of EVs. In order to estimate the energy-saving efficiency of EV accurately, the impacts of these factors on the energy consumption rate need to be explored.

The electricity consumption of EV has been studied in some researches [9 –11]. He et al. adopted a method to analyze the EV's energy consumption, in which the driving range, energy consumption per kilometer, and the specific energy consumption are selected as the evaluation parameters. Based on experimental data collected on urban roads, the energy consumption of BJD6100-EV was analyzed. However, the method could only estimate the average energy consumption per kilometer and could not apply to other driving conditions [10]. With the automobile running equations, Liu et al. took the ECUV vehicle as an object and tested the energy consumption per 100 kilometer at different vehicle speeds and under different driving cycles. Energy consumption economic characteristic curve was depicted and the influence of vehicle weight on the energy consumption was analyzed by testing the energy consumption per 100 kilometers for EVs. Because the data are collected just under cruising status, the presented method cannot be applied in practice [11].

A lot of work has also been done to estimate the fuel consumption of internal combustion engine (ICE) vehicles. Wang et al. explored the influence of driving parameters on fuel consumption using a portable emissions measurement system on 10 passenger cars and found that fuel consumption per unit distance is optimum at speeds between 50 and 70 km/h. Fuel consumption factor increases significantly with acceleration. This paper also developed Vehicle Specific Power-(VSP-) based models to calculate fuel consumption rate [12, 13]. Frey et al. analyzed the influence of key factors (speed, acceleration, and road grade, etc) on fuel consumption rate for diesel and hydrogen fuel cell buses under real-world operating conditions [14]. This paper used a VSP-based approach to model fuel consumption for both types of buses.

Some scholars have made a quantitative comparison between electricity consumption and gasoline consumption by transforming them into coal equivalent consumption. Howey et al. presented the measured energy consumption of a range of “fuel efficient” vehicles over a 57 mile urban/extraurban route [8]. The results show that the electric vehicles had the highest energy efficiency and used the least amount of energy (0.62 MJ/km average), followed by the hybrid vehicles (1.14 MJ/km) and internal combustion engine vehicles (1.68 MJ/km). However, the quantitative methods were put forward under specific experimental conditions and the diversity of traffic conditions was not taken into consideration. Moreover, it is impossible to predict the energy consumption reduction under different EV penetration using these methods.

In order to overcome the limitations of current researches, new models using instantaneous speed and acceleration as input variables are developed in this paper to predict gasoline and electricity consumption rate, which makes it possible to accurately evaluate the energy efficiency under different traffic condition. To explore the energy-saving effect of EVs compared to GVs, both gasoline consumption and electricity consumption are transformed into coal equivalent consumption.

This paper is organized into four sections. The first section describes the background information and summarizes the previous work on energy consumption for EV and GV. In the second section, based on the data collected by chassis dynamometer, the fuel consumption and electricity consumption rate models are constructed and both of them are transformed into coal equivalent consumption rate models. Based on the second by second operation data of more than 4000 vehicles simulated by VISSIM, the energy-saving performance under different EV penetration is predicted in the third section. Finally, a conclusion of the findings and prospects for future work are provided.

2. Methodology

To estimate the energy-saving efficiency of EVs, especially under different driving condition, it is necessary to establish a series of microscopic energy consumption rate models for light duty EVs and light duty GVs, in which the relationship between energy consumption rate and driving condition parameters is described.

2.1. Data Source

The data used to establish electricity and gasoline consumption rate models were collected by chassis dynamometer test. An EV and a GV ran over the New European Driving Cycle (NEDC), respectively, and the electricity consumption data and emission data were collected. The NEDC is shown in Figure 1 [15].

Figure 1:

The driving cycle of data collection.

The EV data include time, vehicle speed, and battery working current The GV data contain time, vehicle speed, and emission data (hydrocarbon, carbon dioxide and carbon monoxide emission rate).

2.2. Energy Consumption Rate Models

Electricity consumption rate of EV is calculated by multiplying the working current and battery voltage using (1), and gasoline consumption rate of GV is estimated by using (2), which is based on the carbon balance method listed in the consider national standards of China (GB/T 19233-2003) [16]

E_{e} = I \times U,

(1)

E_{f} = 0.01154 [(0.866 \times {ER}_{HC}) + (0.429 \times {ER}_{CO}) + (0.273 \times {ER}_{{CO}_{2}})],

(2)

where E_e is the electricity consumption rate, J/s; E_f is the gasoline consumption rate, g/s; ER_HC is the HC emission rate, g/s; ER_CO is the CO emission rate, g/s; ER_CO₂ is the CO₂ emission rate, g/s; I is the battery working current, A; U is the battery voltage, V.

Microscopic gasoline consumption rate and electricity consumption rate models under different driving conditions (accelerating, decelerating, cruising, and idling) represented by a function of instantaneous speed and acceleration are established by the multiple linear regression approach as follows [17]:

E_{e} or E_{f} = {\begin{cases} \sum_{i = 0}^{3} \sum_{j = 0}^{3} (L_{i, j} \times v^{i} \times a^{j}), & a > 0, \\ \sum_{i = 0}^{3} \sum_{j = 0}^{3} (M_{i, j} \times v^{i} \times a^{j}), & a < 0, \\ \sum_{i = 0}^{3} (N_{i} \times v^{i}), & a = 0, v \neq 0, \\ \overset{—}{e}, & a = 0, v = 0, \end{cases}

(3)

where v is the instantaneous speed of vehicle, km/h; a is the instantaneous acceleration, km/h/s; $\overset{—}{e}$ is the average gasoline or electricity consumption rate for idling, g/s or J/s; L_{i, j}, M_{i, j}, and N_i are coefficients for accelerating, decelerating and cruising, respectively, i, j = 0, 1, 2, 3.

2.3. Parameters Calibration

The coefficients of the variables for the microscopic gasoline consumption rate and electricity consumption rate models are estimated with sequential regression approach, and the results are shown in Tables 1 and 2, respectively. The statistical results indicate a good fit for energy consumption rate models for the two types of vehicles (the adjusted R² values exceed 0.89 for EV model and exceed 0.80 for GV model). All the model coefficients are significantly not zeros (absolute t-values exceed 1.96), which indicates that the estimated results are validated.

Table 1:

Results of parameters calibration for light-duty EV.

Variable	Accelerating	Decelerating	Cruising	Idling
Variable	coefficients (t-value)	coefficients (t-value)	coefficients (t-value)	coefficients

Constant	2165.816 (13.344)	1449.163 (19.390)	1893.049 (21.769)	1439.743
v	−94.132 (−6.258)	30.917 (9.842)	—	—
v ²	4.985 (12.733)	—	1.810 (96.194)	—
v ³	−0.025 (−8.773)	0.021 (33.873)	—	—
a	−418.69 (−3.323)	—	—	—
a ²	126.145 (5.862)	—	—	—
a ³	—	—	—	—
va	223.883 (19.714)	—	—	—
va ²	−14.482 (2.133)	2.442 (3.835)	—	—
va ³	0.444 (11.321)	—	—	—
v ² a	−1.543 (−9.717)	0.368 (3.800)	—	—
v ² a ²	—	−0.082 (−2.308)	—	—
v ² a ³	—	0.010 (29.991)	—	—
v ³ a	—	0.018 (11.210)	—	—
v ³ a ²	−0.001 (−2.277)	0.005 (11.410)	—	—
v ³ a ³	—	—	—	—
Adjusted R²	0.973	0.898	0.965	—

Note: v is the instantaneous vehicle speed, km/h; a is the acceleration, km/h/s; R² is the coefficient of determination.

Table 2:

Results of parameters calibration for light-duty GV.

Variable	Accelerating	Decelerating	Cruising	Idling
Variable	coefficients (t-value)	coefficients (t-value)	coefficients (t-value)	coefficients

Constant	0. 5 (36.149)	0. 269 (7.192)	0. 4 (30.249)	0. 265
v	—	0.014 (6.68)	—	—
v ²	—	—	9.217 × 10⁻⁵(31.999)	—
v ³	9.219 × 10⁻⁷ (29.653)	2.554 × 10⁻⁶ (10.143)	—	—
a	—	0.025 (2.661)	—	—
a ²	—	0.004 (17.182)	—	—
a ³	1.712 × 10⁻³ (1.992)	—	—	—
va	7.910 × 10⁻³ (12.650)	—	—	—
va ²	− 8.110 × 10⁻⁴ (−2.692)	—	—	—
va ³	—	—	—	—
v ² a	—	5.934 × 10⁻⁵ (5.134)	—	—
v ² a ²	—	—	—	—
v ² a ³	—	—	—	—
v ³ a	—	− 2.246 × 10⁻⁷ (−2.092)	—	—
v ³ a ²	—	—	—	—
v ³ a ³	—	—	—	—
Adjusted R²	0.902	0.809	0.992	—

Note: v is the instantaneous vehicle speed, km/h; a is the acceleration, km/h/s; R² is the coefficient of determination.

2.4. Conversion Formula of Coal Equivalent

In order to ensure the comparability of energy consumption between EV and GV, the gasoline consumption rate and electricity consumption rate are transformed into coal equivalent consumption rate.

According to monthly analysis report of Chinas electric power industry in January 2013, the coal equivalent consumption of power plants is 0.326 kg/(kW·h) and the line loss is 6.62% in 2012. Because one kilogram of coal equivalent can give off 29271 kJ heat and the electrical equivalent is 3600 kJ/(kW·h), the average power generation efficiency of coal-fired power plants of China in 2012 can be estimated using the following:

P = \frac{(1 \times 3600)}{(0.326 \times 29271)} \times 100 % = 37.73 %,

(4)

where P is the average power generation efficiency of coal-fired power plants of China.

Due to the superior performance of lithium battery, its charge-discharge efficiency could reach to 97%. So the efficiency of electric cars can be calculated by (5), and the electricity consumption rate can be transformed into coal equivalent consumption rate by (6):

l = P \times (1 - 6.62 %) \times 97 %,

(5)

E_{e - s} = \frac{E_{e}}{(l \times 29271 kJ / kg)},

(6)

where l is the efficiency of electric cars and E_{e – s} is coal equivalent consumption rate of EV, g/s.

The value of the coal equivalent coefficient of gasoline is 1.4714, so the gasoline consumption rate can also be transformed into coal equivalent consumption rate by:

E_{f - s} = 1.4714 \times E_{f},

(7)

where E_{f – s} is coal equivalent consumption rate of GV, g/s.

3. Simulation and Analysis

Based on the models established in Section 2 and the dynamic traffic data of Beijing, a simulation is designed to evaluate the coal equivalent consumption reduction under different EV penetration in this section.

3.1. Data Collection and Simulation

The dynamic traffic data is collected by microwave traffic detectors in Beijing with 907 detectors included. This paper is based on the data collected by the microwave traffic detector number HI7033d on 23 August 2007 and data is composed of six parts and the format is shown in Table 3.

Table 3:

Dynamic traffic data collected by microwave traffic detectors.

Detector	Date	Time stamp	Flow (pcu)	Velocity (km/h)	Occupancy (%)

HI7033d	2007/08/23	1	73	58.04	7.00
HI7033d	2007/08/23	2	70	57.86	6.33
HI7033d	2007/08/23	3	86	55.40	8.67
⋮	⋮	⋮	⋮	⋮	⋮
HI7033d	20070823	718	59	58.42	5.67
HI7033d	20070823	719	84	55.43	8.67
HI7033d	20070823	720	73	56.10	7.00

The study road, a part of the North Second Ring Road of Beijing, is marked by the red line in Figure 2 and the length is about one kilometer. As for the traffic composition, since buses just account for less than 1% of traffic volume on the road section, then they buses are excluded from the vehicle set that can be powered with substituted electric power in this paper.

Figure 2:

The study road in Beijing.

Figure 3 shows the temporal variation of traffic volume and average speed in 24 hours, and obvious differences can be observed between the night periods and daytime periods. The traffic is heavy during most of the day and the traffic volume keeps above 3000 pcu/h from 6:00 to 23:00. The average speed is less than 55 km/h from 7:00 to 24:00 and more than 55 km/h from 0:00 to 7:00.

Figure 3:

The variation of traffic volume and average speed.

In order to estimate the energy-saving effect under different EV penetrations, the period from 6:00 to 7:00 is selected, and the number of cars detected by the microwave traffic detector in this hour is 4142. The cumulative probability and probability distributions of average speeds for 6:00–7:00 period are shown in Figures 4 (a) and 4 (b), respectively. All the speeds detected by the detector are ranging from 35.0 km/h to 70.1 km/h. The probabilities when speed is lower than 47.6 km/h and 54.7 km/h are 20% and 40%, respectively, and the probabilities when speed is faster than 56.3 km/h and 60.5 km/h are 80% and 60%. With the analyzed speed distribution characteristics, the parameters in VISSIM are calibrated.

Figure 4:

Distribution of average speed during the period from 6:00 to 7:00.

By using calibrated VISSIM simulation platform, more than 370,000 records of instantaneous operation parameter data for 4124 cars are generated (the data example is shown in Table 4).

Table 4:

Vehicle operation parameters generated by VISSIM.

Time (s)	Vehicle code	Vehicle type	Velocity (km/h)	Acceleration (km/h/s)

3	1	Car	52.03	1.12
3	2	Car	55.02	0.65
4	2	Car	55.67	0.65
4	1	Car	53.16	1.13
4	3	Car	65.47	0.79
5	4	Car	54.67	1.19
⋮	⋮	⋮	⋮	⋮
3600	4124	Car	23.93	−0.36
3600	4123	Car	28.83	−0.14
3600	4021	Car	57.90	0.43

3.2. Energy-Saving Effect Analysis under Different EV Penetration

To estimate the energy reductions in terms of coal equivalent consumption under different EV penetrations, gasoline powered cars are selected randomly and replaced by EVs in different proportions. The coal equivalent consumption for each EV and GV can be calculated using the following:

S_{e x} = \sum_{i = 1}^{n} E_{e - s}^{i},

(8)

S_{f x} = \sum_{i = 1}^{n} E_{f - s}^{i},

(9)

where S_ex is the coal equivalent consumption of EV with Number x, g; S_fx is the coal equivalent consumption of GV with Number x, g; n is the time span that vehicle runs on the road section with 1-second-time granularity, s; E_{e – s}ⁱ is the coal equivalent consumption rate of EV at ith second, g/s, E_{f – s}ⁱ is the coal equivalent consumption rate of GV at ith second, g/s.

The total coal equivalent consumption of all cars can be estimated using (10). In order to make the results more objective, random selections are repeated 10 times under each EV penetration and the average value will be used to evaluate the effect of coal equivalent consumption reduction as shown in (11) and (12):

W_{i}^{y} = \sum_{x = 1}^{4000 y} S_{e x} + \sum_{x = 1}^{4000 (1 - y)} S_{f x},

(10)

W^{y} = \frac{1}{10} \sum_{i = 1}^{10} W_{i}^{y},

(11)

p^{y} = \frac{W^{0} - W^{y}}{W^{0}},

(12)

where W_i^y is the total coal equivalent consumption of the random selection i with EV penetration y; W^y is the average value of total coal equivalent consumption when the EV penetration is y, g. p^y is the coal equivalent consumption reduction rate (CECRR), %; y is the EV penetration, equals to the percentage of EV to the total vehicle composition, %.

The relationship between CECRR and EV penetration (when traffic volume is 4142 pcu/h) is illustrated in Figure 5. It is clear that if the EV penetration increases by 10%, the CECRR will increase by about 5%, and the maximum CECRR will reach to 58% if all of the fuel powered cars are replaced by EVs.

Figure 5:

Relationship between CECRR and EV penetration.

3.3. Energy-Saving Effect Analysis under Different Congestion Level

In order to explore the relationship between energy-saving effect and congestion level, simulations under different traffic volumes (950 pcu/h, 1646 pcu/h, 2043 pcu/h, 2605 pcu/h, 3071 pcu/h, 3694 pcu/h, 4507 pcu/h, and 5150 pcu/h) have been executed. Using the method recommended in the previous section, the maximum CECRRs under different traffic volumes are calculated (shown in Figure 6). According to the figure, it is clear that as the traffic volume increases the maximum CECRR generally tends to increase when the traffic volume is less than 4000 pcu/h. The maximum CECRR decreases with the increase in traffic volume when the traffic volume exceeds 4000 pcu/h. For all the maximum CECRRs under different traffic volumes that are above 51%, it can be highly expected that developing EVs can contribute significantly to the energy reduction in the road traffic sector and energy conservation.

Figure 6:

The maximum CECRR under different traffic volumes.

4. Conclusions and Further Study

This study presents the microscopic coal equivalent consumption models for both EVs and GVs, which require instantaneous vehicle speed and acceleration as input variables. Based on a case study on an expressway road section in Beijing, the relationship between the EV penetration and the reduction of energy consumption is explored. Although the energy consumption reduction efficiency may be different for different types of vehicles, it is clear that EVs have a notable advantage over GVs in reducing energy consumption. The results show that if the EV penetration on the Second Ring Road of Beijing increases by 10%, the coal equivalent consumption will decrease by about 5% for the tested traffic condition. For all traffic conditions, the maximum CECRRs are above 51%, and the maximum energy-saving effect can be achieved when the traffic volume is about 4000 pcu/h. It can be highly expected that developing EVs is able to contribute significantly to the energy reduction and the development of low-carbon transportation system in Beijing.

Instead of the data collected by chassis dynamometer tests with NEDC driving circle, the data collected in real driving conditions will be used to establish EVs' energy consumption rate models. Moreover, an integrated traffic status indicator instead of current traffic volume will be used and the traffic status-based energy-saving effect analysis will be conducted in the next-step study.

Conflict of Interests

The authors declare that there is no conflict of interests regarding the publication of this paper.

Footnotes

Acknowledgment

This research is supported by the National 973 Program of China (no. 2012CB725403).

References

Intergovernmental Panel on Climate Change, “IPCC Fourth Assessment Report: Climate Change,” Tech. Rep., 2007.

Offer

G. J.

Howey

Contestabile

Clague

, and Brandon

N. P.

, “Comparative analysis of battery electric, hydrogen fuel cell and hybrid vehicles in a future sustainable road transport system,” Energy Policy, vol. 38, no. 1, pp. 24–29, 2010.

Zhang

, and Chang

, “Alternative fuel buses currently in use in China: life-cycle fossil energy use, GHG emissions and policy recommendations,” Energy Policy, vol. 38, no. 1, pp. 406–418, 2010.

Guo

Wang

, and Bubb

, “Modeling of traffic behavior in traffic safety using a reliability approach,” Computational Intelligence for Traffic and Mobility, vol. 8, pp. 243–265, 2013.

Hamadi

and Chittajallu

, World Road Statistics 2008, International Road Federation, Washington, DC, USA, 2008.

Yang

Lin

, “Energy consumption and CO₂ emission impacts of vehicle electrification in three developed regions of China,” Energy Policy, vol. 48, pp. 537–550, 2012.

Sandy Thomas

C. E.

, “How green are electric vehicles?” International Journal of Hydrogen Energy, vol. 37, no. 7, pp. 6053–6062, 2012.

Howey

D. A.

Martinez-Botas

R. F.

Cussons

, and Lytton

, “Comparative measurements of the energy consumption of 51 electric, hybrid and internal combustion engine vehicles,” Transportation Research D, vol. 16, no. 6, pp. 459–464, 2011.

Wang

Liu

Cui

, and Zhang

, “Research on quantitative models of electric vehicle charging stations based on principle of energy equivalence,” Mathematical Problems in Engineering, vol. 2013, Article ID 959065, 10 pages, 2013.

10.

H.-W.

Sun

F.-C.

, and Yu

X.-J.

, “Analysis of energy consumption of BJD6100-EV bus in urban driving,” Transaction of Beijing Institute of Technology, vol. 24, no. 3, pp. 222–225, 2004.

11.

Liu

, and Zong

, “Study on the energy consumption economy of electric vehicle based on test bench simulation,” Acta Scientiarum Natralium Universitatis Sunyatseni, vol. 50, no. 1, pp. 44–52, 2011.

12.

Wang

Zhou

, and Li

, “Modelling of the fuel consumption for passenger cars regarding driving characteristics,” Transportation Research D, vol. 13, no. 7, pp. 479–482, 2008.

13.

Song

Yao

Zuo

, and Lang

, “Emissions and fuel consumption modeling for evaluating environmental effectiveness of ITS strategies,” Discrete Dynamics in Nature and Society, vol. 2013, Article ID 581945, 9 pages, 2013.

14.

Frey

H. C.

Rouphail

N. M.

Zhai

Farias

T. L.

, and Gonçalves

G. A.

, “Comparing real-world fuel consumption for diesel- and hydrogen-fueled transit buses and implication for emissions,” Transportation Research D, vol. 12, no. 4, pp. 281–291, 2007.

15.

Karlsson

H. L.

, “Ammonia, nitrous oxide and hydrogen cyanide emissions from five passenger vehicles,” Science of the Total Environment, vol. 334–335, pp. 125–132, 2004.

16.

Yao

and Song

, “Study on eco-route planning algorithm and environmental impact assessment,” Journal of Intelligent Transportation Systems, vol. 17, no. 1, pp. 42–53, 2013.

17.

Ahn

Rakha

Trani

, and Van Aerde

, “Estimating vehicle fuel consumption and emissions based on instantaneous speed and acceleration levels,” Journal of Transportation Engineering, vol. 128, no. 2, pp. 182–190, 2002.