Performance of Multiple Alternatives to Reduce Carbon Emissions for Transit Fleets: A Real-world Perspective

(1)

Peer-review under responsibility of the organizing committee of CUE 2015 doi: 10.1016/j.egypro.2016.06.110

Energy Procedia 88 ( 2016 ) 908 – 914

ScienceDirect

CUE2015-Applied Energy Symposium and Summit 2015:

Low carbon cities and urban energy systems

Performance of multiple alternatives to reduce carbon

emissions for transit fleets: a real-world perspective

Hanyan Li, Haobing Liu, Yanzhi (Ann) Xu

*

_{, Michael O. Rodgers, Randall}

Guensler

Georgia Institute of Technology, School of Civil and Environmental Engineering, Atlanta, GA 30332 United States

Abstract

Transit agencies are always seeking opportunities to reduce fuel consumption, which typically provides simultaneous emissions reductions. However, the effectiveness and efficiency of fuel conservation and emission reduction alternatives differ as a function of local transit operating characteristics. The fact that emissions reductions from multiple alternatives are not necessarily additive further complicates these analyses. In selecting a set of alternatives, transit agencies need to evaluate multiple options simultaneously, and under agency-specific operating parameters. This paper presents the analysis of emissions reduction alternatives of eco-driving and purchasing new fleet with compressed natural gas, utilizing real-world operating data collected from the Metropolitan Atlanta Rapid Transit Authority (local transit) and the Georgia Regional Transportation Authority (express bus) fleets. Each individual emissions reduction alternative was first investigated for effectiveness individually and then in combination to assess cumulative impacts. The results show that the comparative performance of emissions reductions alternatives is highly dependent on vehicle operating characteristics, and thus the preferred alternatives, differ between different fleets. The analysis conducted for the Atlanta, Georgia region in the United States provides insight to how such modeling techniques can assist planning for future fleets given each agency’s unique local conditions.

Keywords: Eco-driving; Alternative Fuel; Bus; Fuel Consumption; Greenhouse Gases

Nomenclature

CNG Compressed natural gas

*_{Corresponding author. Tel: +1-404-723-0543}

E-mail address: [email protected]

(2)

CO2 Carbon dioxide

CO2e Carbon dioxide equivalent GGE Gasoline gallon equivalent GHG Greenhouse gas

GPS Global Positioning System

GREET Greenhouse gases, Regulated Emissions, Energy, and Transportation model GRTA Georgia Regional Transportation Authority

GT Georgia Institute of Technology

MARTA Metropolitan Atlanta Rapid Transit Authority MOVES MOtor Vehicle Emission Simulator

NTD National Transit Database QA/QC Quality assurance/quality check STP Scaled tractive power

1. Introduction

Eco-driving and switching from diesel to CNG are two popular alternatives among transit agencies as they seek to conserve fuel and lower GHG emissions [1]. Each emissions reduction alternative offers different return-on-investment (ROI), depending upon the local conditions and operational characteristics of each agency. While many studies have assessed eco-driving (e.g., [2-4]) and CNG (e.g., [5-7]) separately as viable carbon-reducing alternatives for transit, little research has been conducted to simultaneously evaluate these alternatives under agency-specific operating characteristics. This paper focuses on transit fuel and GHG emissions savings from eco-driving and CNG for two transit agencies. The analyses in this report are based upon real-world operations data collected from MARTA, a local transit agency, and GRTA, offering express services.

Adopting the modeling framework of the Fuel and Emissions Calculator for Transit Fleets [8], this paper quantifies potential emissions reductions and fuel savings using MOVES-Matrix [9], which integrates MOVES load-based emission rates [10] with real-world modal operating data, for pump-to-wheel emissions and GREET [11] for well-to-pump emissions . On-road operating data of local transit and express bus fleets were collected from MARTA and GRTA, respectively. Eco-driving impacts were modeled using a new algorithm based on the structure of the MOVES STP operating mode bins. Effectiveness was first investigated for each emissions reduction alternative individually and then in combinations to assess cumulative impacts.

2. Data

To evaluate the potential emissions and fuel consumptions associated with eco-driving for transit operations in the Atlanta metropolitan area, second-by-second transit operations data were collected from local transit operations and regional express buses. MARTA and GRTA operations data were collected on

(3)

buses using the GT Trip Data Collector [12] and Qstarz BT-Q1000eX GPS in this sampling effort, respectively. In all, real-world transit operations data were collected for over 68 thousand miles.

The GPS data underwent QA/QC and post processing before being used in the analyses. Data points with poor GPS quality, off-network activities, and extremely short trips were excluded from the analysis.

Table 1 Summary of Analytical Data Set

Type of Operation Agency Time Range Number of Buses Number of Trips Total Distance (miles) Total Duration (hours) Average Speed (mph) Local Transit MARTA 2004~2005 13 9,984 61,247 3,716 16.5 Express Service GRTA 2013~2014 13 852 3,637 84 43.3

3. Methodology

3.1. Eco-driving Cycle Development

To model the impacts of eco-driving, this study proposes a new method for optimizing each vehicle trajectories based upon the structure of the MOVES STP operating mode bins. The methodology conserves cycle distance, maintains overall average speed, but prevents instantaneous STP from increasing significantly by setting acceleration limits within each MOVES speed grouping.

The first step in the eco-driving process was to set a STP limit value (STPL). For each speed , an

acceleration limit was set to prevent STP from exceeding STPL, according to the STP equation employed

in MOVES [10]. The principle of the computational method is to read each second of vehicle activity and when the STP reaches or exceeds the STPL, adjust the acceleration rate downward enough to lower the

STP of the next data point to the median value of the STP range that meets STPL.

It is important to set appropriate STP limits by driving cycle. If the rules are too lenient, the reductions will not be significant. However, if the rules are too stringent, the average speed of the trace may be too low for drivers to accept. Furthermore, a reduction in average speed leads to increased driving time, offsetting some of the fuel and emissions savings. This study established different STP limits by road type and speed after iterative testing. The resulting STP limits are summarized in Table 2.

Table 2 STP Limits for Local Roads and Freeways at Different Speed Levels

Road Type Local Road Freeway

Speed Level 0~25 mph 25~50 mph >= 50 mph 0~25 mph 25~50 mph >= 50 mph STP Limit <=6 <=6 <=6 <=6 <=9 <=12

To implement the eco-driving strategy, three iterative steps applied to each vehicle trajectory: 1) Maintaining Status Quo: When the STP of original cycle doesn’t reach or exceed STPL, the next data

point in the eco-cycle is the same as the data point from the original cycle, without modifications. 2) Smoothing: When the STP of original cycle reaches or exceeds the STPL, the acceleration rate is

adjusted downward such that the resulting STP for the data point equals the median value for the STP bin that does not exceed the STPL. Because the acceleration rate decreases, the speed of the next

(4)

subsequent points in the cycle are also set to achieve the median STP value for that STP bin. Smoothing of acceleration continues until the speed of eco cycle matches that of the original cycle. 3) Conservation of Distance: Once the speed of eco and original cycle align, the distance covered by the

eco cycle is less than that of the original cycle. To conserve distance travelled, the eco cycle cruise speed is extended until the distance traversed by the eco cycle matches that of the original cycle. This step assumes that the vehicle is not limited by the presence of a slower-moving vehicle in its path. An example of an observed cycle and its corresponding eco cycle is presented in Fig. 1. The modified eco cycle smoothed the sharp acceleration, especially during high speed operations.

Fig. 1. Example of Current and Corresponding Eco Cycles 3.2. Fuel and Emissions Analysis

The goal of the fuel and emissions analysis is to assess the impact of the eco-driving and switching to CNG at the transit agency level. To do so, real-world operations data collected from MARTA and GRTA buses are used and fuel consumption and emission rates are applied as if the entire fleet experiences the observed operating conditions. To provide a fair comparison between diesel and CNG, we evaluated the full fuel cycle, i.e. well-to-wheel emissions. Pump-to-wheel emissions were estimated using MOVES-Matrix [9], a multi-dimensional emission rate look up table derived directly from millions of MOVES emission rate runs for Atlanta. Well-to-pump emissions were estimated using the GREET model [11].

To estimate the emissions and fuel consumptions for the entire fleet, fleet size and annual mileage information from the NTD [13], as summarized in Table 3. The NTD does not provide information of operating mileage on different road types, but does differentiate between revenue and non-revenue (also known as deadhead) mileage. Therefore, proportions of freeway and non-freeway mileage were estimated separately for revenue and deadhead operations, using spatial analysis in ArcGIS.

Table 3. Mileage and Fleet Information [13]

Transit Agency Annual Mileage (1,000 Miles) Deadheading Percent (%) Number of Buses CNG Percent (%)

MARTA 25,850 12 508 69

(5)

4. Scenario Settings and Results

Annual life-cycle emission results under different scenarios in 2015 were calculated for the MARTA and GRTA fleets. On-road fuel consumption and life-cycle GHG emissions were estimated for three scenarios - eco-driving, replacing existing diesel buses with new CNG buses (denoted as CNG), and the combination of the two strategies (denoted as eco-driving+CNG). These three scenarios were compared with the baseline, characterized by existing fleet composition and driving behavior. Table 4 shows the annual fuel usage and life-cycle emissions for the base scenario against the analyzed scenarios. Figure 2 further breaks down CO2e emissions by fuel cycle components.

For MARTA (local service), eco-driving would reduce on-road fuel consumption and life-cycle GHG emissions by 5%. By contrast, replacing existing diesel buses with new CNG buses would increase fuel consumption by 4%, and provide no significant difference in CO2e emissions, due to lower well-to-pump emissions of CNG compared to diesel. Combining CNG with eco-driving would reduce fuel consumption by 1% and life-cycle CO2e emissions by 5%, compared to the baseline. For GRTA (express service), eco-driving and CNG both show reductions in life-cycle GHG emissions. However, CNG would increase on-road fuel consumption by 6%, whereas eco-driving would decrease fuel consumption by 7%. The combination of eco-driving with a CNG fleet would reduce fuel consumption by 4% and life-cycle CO2e emissions by 17%, compared to the baseline.

Table 4. Performance of Alternatives

MARTA GRTA

Base Eco-driving CNG Eco-driving+CNG Base Eco-driving CNG Eco-driving+CNG

CO2 (Metric tons) 65.6 -5% -3% -8% 9.2 -7% -16% -24%

CO2e (Metric tons) 81,233 -5% 0% -5% 9,511 -7% -8% -17%

Fuel (1,000 GGE) 7,226 -5% 4% -1% 859 -7% 6% -4%

(6)

Figure 2 Well-to-wheel CO2e Emissions Comparison

5. Conclusions

This paper presented the analysis of two emissions reduction alternatives for transit fleets, namely, eco-driving and purchasing new CNG fleet. Performance regarding energy consumption and GHG emissions was evaluated both individually and in combination. For local transit service, using MARTA as an example, eco-driving would reduce GHG emissions and fuel consumption by 5%, whereas replacing existing diesel buses with new CNG buses would slightly increase fuel consumption, and have negligible impacts on GHG emissions. If eco-driving was implemented in conjunction with the purchase of new CNG buses, MARTA would reduce fuel consumption by 1% and GHG emissions by 5%, compared to the baseline fleet mix and operational characteristics. For express bus service, as showcased through GRTA operations, both eco-driving and purchasing new CNG buses showed reductions in GHG emissions, and the combinations of the two showed compounded benefits in GHG reductions. However, a new CNG fleet would increase GRTA’s energy consumption by 6%, unless the CNG fleet adopts eco-driving.

The comparative performance of emissions reductions alternatives is highly dependent on vehicle operating characteristics, and that the preferred alternatives, or set of alternatives, differ across local transit and express services. From a low-carbon perspective, a new CNG fleet was shown to be advantageous for express bus service but not for local transit service. Regardless of service type, eco-driving was shown to reduce GHG emissions in the existing fleets and in the assumed all-CNG fleets.

Acknowledgements

This work was sponsored by the National Center for Sustainable Transportation, a U.S. University Transportation Center.

(7)

References

[1] TCRP. Current Practices in Greenhouse Gas Emissions Savings from Transit: A Synthesis of Transit Practice.

Transportation Research Board, 2010.

[2] Strömberg HK, Karlsson IM. Comparative effects of eco-driving initiatives aimed at urban bus drivers–Results from a field trial. Transportation Research Part D: Transport and Environment. 2013; 22, 28-33.

[3] Wåhlberg. AE. Long-term effects of training in economical driving: Fuel consumption, accidents, driver acceleration behavior and technical feedback. International Journal of Industrial Ergonomics 2007; 37(4), 333-343.

[4] Zarkadoula M, Zoidis G, Tritopoulou E. Training Urban Bus Drivers to Promote Smart Driving: A note on a Greek Eco-driving Pilot Program. Transportation Research Part D: Transport and Environment. 2007; 12(6), 449-451.

[5] Barnitt, R., In-Use Performance Comparison of Hybrid Electric, CNG, and Diesel Buses at New York City Transit. SAE Technical Paper 2008; 2008-01-1556.

[6] Turrio-Baldassarri L, Battistelli C, Conti L, Crebelli R, De Berardis B, Iamiceli AL, Gambino M, Iannaccone S. Evaluation of emission toxicity of urban bus engines: Compressed natural gas and comparison with liquid fuels. Science of the Total

Environment. 2006; 355(1), 64-77

[7] Tzeng GH, Lin CW, Opricovic S. Multi-criteria analysis of alternative-fuel buses for public transportation. Energy Policy. 2005;33(11):1373-83.

[8] Xu Y, Gbologah FE, Liu H, Rodgers MO, Guensler RL. Assessment of alternative fuel and powertrain transit bus options using real-world operations data: Life-cycle fuel and emissions modeling. Applied Energy. 2015;154:143-59.

[9] Guensler, R., H. Liu, X. Xu, Y. Xu, and M.O. Rodgers (accepted). MOVES-Matrix: Setup, Implementation, and Application. 95th Annual Meeting of the Transportation Research Board. Washington, DC. January 2016.

[10] EPA's Office of Transportation and Air Quality. MOVES (Motor Vehicle Emission Simulator).

[11] Argonne National Laboratory, 2013. The Greenhouse Gases, Regulatory Emissions, and Energy Use in Transportation (GREET) Model.

[12] Ogle, J., Guensler, R., and Elango, V. (2005). Georgia's Commute Atlanta value pricing program: Recruitment methods and travel diary response rates. Transportation Research Record: Journal of the Transportation Research Board, (1931), 28-37.

[13] National Transit Database. (2014). Retrieved from: http://www.ntdprogram.gov/ntdprogram/data.htm on November 11, 2015.

Biography

Dr. Yanzhi (Ann) Xu is Research Engineer II at the School of Civil and Environmental Engineering of the Georgia Institute of Technology in Atlanta, GA, USA. She received her doctorate degree in transportation systems engineering from Georgia Institute of Technology, and her bachelor’s degree in environmental science from Peking University in Beijing, China.