Life Cycle Emissions and Lifetime Costs of Medium-duty Diesel and Alternative Fuel Trucks. A Case Study for Toronto

Transcription

1 Life Cycle Emissions and Lifetime Costs of Medium-duty Diesel and Alternative Fuel Trucks. A Case Study for Toronto Mo Zhou A thesis submitted in conformity with the requirements for the degree of Master of Applied Science Civil Engineering University of Toronto Copyright by Mo Zhou 2016

2 Life Cycle Emissions and Lifetime Costs of Medium-duty Diesel and Alternative Fuel Trucks. A Case Study for Toronto Abstract Mo Zhou Master of Applied Science Civil Engineering University of Toronto 2016 This thesis analyzes a Class 6 diesel truck, compressed natural gas truck, hybrid-electric truck and battery-electric truck, in terms of energy consumption, life cycle greenhouse gas (GHG) emissions and total cost of ownership. The energy consumption simulation and life cycle emissions for the trucks are based on payload, operating temperature and two drive cycles in Toronto, the University drive cycle, reflecting congested city driving, and the Freeway drive cycle, reflecting high speed driving conditions. The energy consumption simulations are performed using Autonomie, a simulation software for vehicles. The simulation data are used in GHGenius, a Canadian life cycle tool for fuels/vehicles, to calculate the life cycle GHG emissions for the trucks. The cost of GHG emissions abatement for each alternative fuel truck is determined based on the differences in life cycle cost and GHG emissions between the alternative trucks and the diesel (reference) truck. The results of the analyses indicate that there is no dominant technology that reduces both GHG emissions and ownership cost under all driving conditions, a conclusion that can inform transportation climate change policies. ii

3 Acknowledgments I would like to thank my supervisors, Professor Matthew Roorda and Professor Heather MacLean for giving me the opportunity and guidance to explore this topic. I am very grateful that they provided me with valuable inputs and insightful suggestions, and make the process of research more precious. I would also like to thank Jason Luk for assistance with Autonomie and vehicle modeling techniques. I am also grateful to my friends in the Department of Civil Engineering for making my Master program a very enjoyable experience. Finally, I would like to thank AUTO21 Network Centres of Excellence for financial support. iii

4 Table of Contents Table of Contents... iv List of Tables... vii List of Figures... viii 1 Chapter 1 Introduction and Research Motivation Introduction Medium-duty trucks Well-To-Wheel analysis Drive Cycle impacts energy consumptions and GHG emissions Temperature impacts on energy consumption Research motivation Research objective Thesis organization Chapter 2. Life Cycle GHG Emissions and Lifetime Costs of a Medium-duty Diesel and Battery-electric Truck. A Case Study for Toronto Abstract Introduction Methodology and data Truck specifications Autonomie Drive cycles Temperature Life cycle assessment Lifetime total cost of ownership and sensitivity analysis Five year cost of ownership Cost of GHG emissions abatement iv

5 2.4 Results and discussion Energy consumption by temperature, payload and drive cycles Life cycle GHG emissions Total cost of ownership Cost-effectiveness of GHG emissions abatement Conclusions Chapter 3 Life Cycle Emissions and Lifetime Costs of Medium-duty Diesel and Hybridelectric truck. A Case Study for Toronto Abstract Introduction Method and data Vehicle specifications Autonomie Drive cycles Temperature LCA scope Lifetime total cost of ownership and sensitivity analysis Cost of GHG abatement Results Energy consumptions Life cycle GHG emissions Total cost of ownership Conclusion Chapter 4. Life Cycle GHG Emissions and Lifetime Costs of Medium-duty Diesel and Compressed Natural Gas Truck. A Case Study for Toronto Introduction Methods and data v

6 4.2.1 Scope of the LCA Total cost of ownership Result and discussion Energy consumption Life cycle GHG emissions Total cost of ownership Conclusions Chapter 5 Conclusion Energy consumption Life cycle GHG emissions Lifetime total cost of ownership Cost of GHG emissions abatement Policy implication Future research References vi

7 List of Tables Table 2-1 Material breakdown by weight percentage (%) for fuel storage system (S&T Consultants Inc., 2013) Table 2-2. Price estimate of the BET Table 2-3 Ranges of parameters for sensitivity analysis for lifetime and five-year ownership cost analyses Table 3-1 Diesel truck and HET characteristics used in Autonomie to determine the vehicle energy consumption Table 3-3 Values of parameters for sensitivity analysis Table 4-1 Diesel Truck and CNG truck characteristics Table 4-2 Material breakdown by weight percentage for fuel storage system (S&T Consultants, 2013) Table 4-3 Best estimates and ranges for parameters for sensitivity analysis vii

8 List of Figures Figure 2-1. University and Freeway Drive Cycle for Medium-duty Trucks in Toronto (Amirjamshidi & Roorda, 2014) Figure average Ontario electricity generation sources (%) (IESO, 2014) Figure 2-3 Toronto diesel fuel price projection (2014$/Liter) (based on data in U.S. Energy Information Administration, 2015) Figure 2-4. Diesel truck and BET energy consumption by temperature and payload in the University drive cycle Figure 2-5. Diesel truck and BET energy consumption by temperature and payload in the Freeway drive cycle Figure 2-6. Life cycle GHG emissions of diesel truck and BET under different drive cycle and payload conditions Figure 2-7. Life cycle GHG emissions of diesel truck and BET under different temperatures and payloads on the University Drive Cycle Figure 2-8. Life cycle GHG emissions of diesel truck and BET under different temperatures and payloads on the Freeway Drive Cycle Figure 2-9. Lifetime total cost of ownership for the diesel truck and BET Figure year total cost of ownership for the diesel truck and BET Figure Sensitivity analysis of difference in lifetime total cost of ownership between BET and diesel truck Figure Sensitivity analysis of difference in 5-year total cost of ownership between BET and diesel truck Figure Cost-effectiveness of GHG emissions abatement for the BET based on lifetime total cost of ownership viii

9 Figure Cost-effectiveness of GHG emissions abatement for the BET based on 5-year total cost of ownership Figure 3-1 University and Freeway Drive Cycle for Medium-duty Trucks (based on data in Amirjamshidi et al. 2014) Figure 3-2 Material breakdown by weight percentage (%) for fuel storage systems (based on S&T Consultants Inc., 2013) Figure 3-3. Energy consumption comparison between the diesel truck and the HET on the University drive cycle Figure 3-4 Energy consumption comparison between the diesel truck and the HET on the Freeway drive cycle Figure 3-5. Life cycle GHG emissions at 20 C for diesel truck and the HET by payload and life cycle phases Figure 3-6. Life cycle GHG emissions comparison between the diesel truck and the HET in the University drive cycle Figure 3-7. Life cycle GHG emissions comparison between the diesel truck and the HET in the Freeway drive cycle Figure 3-8. Total Cost of Ownership for the diesel truck and the HET based on base value Figure 3-9. Sensitivity analysis results for the lifetime total cost of ownership comparison between the HET and the diesel truck Figure Sensitivity analysis results for the lifetime total cost of GHG emissions abatement of the HET Figure 4-1. University and Freeway Drive Cycles for Medium-duty Vehicles in Toronto. (Prepared based on data in Amirjamsidi & Roorda, 2014) Figure 4-2. Energy consumption of diesel and CNG trucks under different payloads from Autonomie simulation ix

10 Figure 4-3. Life cycle GHG emissions of the diesel and CNG trucks Figure 4-4. Lifetime total cost of ownership of diesel and CNG trucks Figure 4-5. Sensitivity analysis results for difference in lifetime total cost of ownership between CNG and diesel trucks Figure 5-1.Comparison of energy consumption of the CNG truck and the diesel truck Figure 5-2. Comparison of energy consumption of the HET and the BET on the University drive cycle Figure 5-3. Comparison of energy consumption of the HET and the BET on the Freeway drive cycle Figure 5-4. Temperature impact on energy consumption for diesel truck, HET and the BET on the University and Freeway drive cycles, at 10% payload Figure 5-5. Life cycle GHG emissions breakdown by lifecycle phase assuming 10% payload, 20 C and the University drive cycle Figure 5-6. Life cycle GHG emissions for the diesel truck and CNG truck Figure 5-7. Life cycle GHG emissions for the diesel truck, HET and the BET on the University drive cycle Figure 5-8. Life cycle GHG emissions for the diesel truck, HET and the BET on the Freeway drive cycle Figure 5-9. Total cost of ownership for the diesel, CNG, HET and BET for the base case Figure Cost of GHG abatement and mass of emissions reduced (base case) for CNG truck, HET and the BET compared to reference diesel truck Figure Cost of GHG abatement for the BET Figure Cost of GHG abatement for the HET compared to the diesel truck x

11 Figure Cost of GHG abatement for the CNG truck xi

12 1 1 Chapter 1 Introduction and Research Motivation 1.1 Introduction Freight movement is a significant component of economic growth but it also has negative impacts on the environment. The transportation sector accounts for 23% of the total greenhouse gas (GHG) emissions in Canada in 2013 (Government of Canada, 2015). Additionally, the rapid growth of retail e-commerce, which is about five times more than the overall pace of retail growth, increases the demand of last-mile delivery (Statistics Canada, 2014), Medium-duty trucks (Classes 3-6), trucks often used for urban delivery, are energy intensive (Bachmann et al., 2014). UPS has a fleet of over 100,000 vehicles and spends over $3 million on fuel costs annually (Burke & Zhu, 2014). Despite average diesel fuel consumption per unit distance traveled (L/100km) decreasing by 17% between 2011 and 1990, total energy use and GHG emissions of medium-duty vehicles in Canada have increased by 19% and 21%, respectively (Natural Resource Canada, 2014). The increasing tonne-kilometers traveled, and the rising sales and market share of medium-duty vehicles, may have contributed to the significant growth in energy consumption and GHG emissions, despite the improved efficiency ((Natural Resources Canada, 2011; Statistics Canada, 2014). The U.S. Energy Information Administration (EIA) estimates that medium-duty vehicles will have the highest average annual growth rate (2.2%) in energy use from among other commercial transportation modes (e.g., buses and heavy-duty trucks) (EIA, 2011). Criteria air containment emissions from diesel fuel use in the transportation sector are a significant health and environmental concern, and impact the approximately 80% of Canadians who work and/or live in urban areas (Transport Canada, 2006). Purolator, a Canadian courier company, added 115 new hybrid-electric trucks (HETs) for their curbside delivery fleet in 2005 and claimed to save more than 2,100 tonnes of GHG emissions and 645,000 liters of fuel (Purolator, 2010). Courier companies such as UPS, have adopted alternative fuel trucks, such as compressed natural gas (CNG) trucks, in their fleets to reduce fuel costs and GHG emissions (UPS, 2016).

13 2 Electrified powertrain trucks, such as battery-electric trucks (BETs), which use only electric power, and HETs, which use both diesel and electricity but do not need to be plugged in, are popular technologies to reduce energy consumption and life cycle GHG emissions in commercial vehicles (National Research Council, 2010). Both BETs and the HETs have some advantages compared to diesel trucks, including lower tailpipe emissions, higher fuel economy, less noise, and lower fuel cost. There are few companies in the world manufacturing mediumduty BETs. Smith Electric Vehicles produces Class 4-6 BET - the Smith Newton battery-electric truck - which has a payload up to 7,400 kg and is used by many companies (e.g., Staples and Fedex) and the U.S. government (U.S. Army) (Smith Electric Vehicles, 2016). According to Smith Electric Vehicles (2016), all Smith Newton BETs have a top speed of 88km/h and offer a range of km per charge. Staples has more than 100 Smith Newton electric trucks in their fleet in California, Missouri and Ohio (Smith Electric Vehicles, 2011). Medium-duty HETs are another alternative to diesel trucks that have the potential to reduce fuel costs and GHG emissions. The National Renewable Energy Laboratory conducted a 36-month evaluation of six model year (MY) 2007 Freightliner P70H HETs operating in Phoenix, Arizona (Ragatz et al., 2014). The Freightliner P70H HET is 544 kg heavier than the comparator diesel truck. Compared with the conventional diesel truck, electrified trucks have the advantage of regenerative braking to capture the energy during braking events (Karabasoglu & Michalek, 2013). Studies of medium-duty electric vehicles indicated that medium-duty BETs and HETs can reduce GHG emissions, especially in congested driving conditions with low average speed, and a high number of stops (Lee et al., 2013; Tong et al., 2015). Many studies have shown that electrification of vehicles results in higher reduction of GHG emissions in city, rather than highway, driving conditions compared with conventional vehicles (Raykin et al., 2012; Karabasoglu & Michalek, 2013; Lee et al., 2013). Another alternative fuel is compressed natural gas (CNG). Natural gas has lower emissions rates of particulate matter and GHGs than diesel fuel. In addition, the natural gas resource is abundant in North America, which encourages the lower fuel price of CNG and the growth of the market for CNG vehicles (U.S. DOE, 2015; Heffernan & Dawson, 2010). The Boston Consulting Group forecasted that sales of CNG vehicles would have an annual growth rate of 17 percent in the U.S., highly concentrated in the medium-and heavy-duty sectors (Nath et al., 2014).

14 3 1.2 Medium-duty trucks Medium-duty trucks, classes 3-6, have gross vehicle weight ratings between 6,351 kg and 11,793 kg. These trucks are primarily used for vocational purposes, such as emergency, delivery and dump trucks (Freightliner, 2016). The environmental performance and fuel consumption of medium-duty vehicles are heterogeneous because they are highly dependent on drive cycle, operating temperature, payload, and truck configuration (Taptich & Horvath, 2014; U.S. National Research Council (NRC), 2014; Lohse-Busch et al., 2013; Fleetcarma, 2013; Yuksel & Michalek, 2015). 1.3 Well-To-Wheel analysis Well-to-wheel (WTW) analysis is a method of analyzing emissions from vehicles. It allows one to examine the emissions associated with a vehicle from a life cycle perspective, from fuel and materials that are used to manufacture and power the vehicle, to tailpipe emissions. WTW analysis includes emissions from three phases of the vehicle life cycle: 1) fuel cycle (wellto-tank); 2) vehicle operation (tank-to-wheel); and 3) vehicle cycle (vehicle production and transportation). There are few WTW analyses of medium-duty battery-electric, hybrid-electric or CNG trucks. Tong et al. (2015) compared the life cycle GHG emissions for various medium-duty trucks using different fuels and found that BETs can achieve GHG emissions reductions of 31% to 41% compared with diesel trucks. Lee et al. (2013) compared the life cycle GHG emissions of a medium-duty BET and a diesel truck and concluded that the BET reduces GHG emissions by 19-61%, depending on drive cycle, compared with a diesel counterpart. Bachmann et al. (2013) studied life cycle emissions for a HET and a diesel truck. Existing WTW studies of commercial vehicles are generally focused on transit buses and heavy-duty trucks, and these studies have different conclusions. Rose et al. (2012) concluded that a CNG refuse truck reduced GHG emissions by 24% compared to a diesel truck. Tong et al. (2015) found that a CNG truck achieved small life cycle GHG emissions reductions (0-6%) compared to a diesel truck. Kliucininkas et al. (2012) found that a CNG vehicle had higher energy consumption and more GHG emissions than a diesel bus in Kaunas, Lithuania.

15 Drive Cycle impacts energy consumptions and GHG emissions Many studies have shown that electrification of vehicles results in higher reduction of energy consumption and GHG emissions in city, rather than highway, driving conditions compared with conventional vehicles (Raykin et al., 2012; Karabasoglu & Michalek, 2013; Lee et al., 2013). Karabasoglu and Michalek (2013) found that in the New York City Cycle, a battery-electric vehicle can reduce life cycle GHG emissions by up to 60% relative to conventional vehicles; while GHG emission reductions are marginal under highway driving conditions. This is because the GHG emissions of hybrid and plug-in vehicles are relatively insensitive to drive cycle compared to conventional vehicles. Studies have indicated that medium-duty BET can reduce GHG emissions by a considerable amount, especially in congested driving conditions with low average speed, and a high number of stops (Lee et al., 2013; Tong et al., 2015). Bachmann et al. (2013) conducted a LCA study comparing the GHG emissions between a diesel truck and the HET. They used energy consumption data from the HET manufacturer and concluded that the HET reduces 23% of life cycle GHG emissions compared to the diesel truck in city driving conditions and 8% in highway driving condition Temperature impacts on energy consumption Along with driving cycle, operating temperature is another crucial factor affecting energy consumption of both conventional, battery-electric, and hybrid-electric vehicles. Researchers from Argonne National Laboratory found that the fuel consumption for conventional vehicles operated in low (-7 C) and high (35 C) temperatures increased by 7% and 27%, respectively, compared with operating at 20 C in the Urban Dynamometer Driving Schedule drive cycle; and the energy consumption for battery-electric vehicles increased by 91% and 20% in cold and warm temperatures, respectively; and that for the hybrid-electric car increased by 65% and 56%, respectively (Lohse-Busch et al., 2013). The impact of temperature on energy consumption in highway driving condition is smaller than in city driving conditions. For example, the same study demonstrated that in the Highway Fuel Economy Test cycle, the energy consumption for the battery-electric vehicle increased by 42% and 5% in -7 C and 35 C, respectively, compared to in 20 C. Yuksel and Michalek (2015) studied energy consumption for light-duty vehicles at different temperatures in different regions in the United States. They found that life cycle emissions associated with battery-electric vehicles may increase by up to 22% in places with cold climates, even when operated off of the same electricity grid. This is because conventional

16 5 vehicles use the engine heat for cabin heating in cold weather while the battery-electric vehicle motor is relatively efficient and generates little waste heat to be used for heating purposes. To my knowledge, no study has investigated the temperature impact on CNG trucks. Due to data unavailability and vehicle simulation software limitations, it was not possible to model the temperature impact on energy consumption for the CNG truck. 1.4 Research motivation Currently in Ontario, there are no fuel efficiency regulations for medium-duty trucks, and trucks are excluded from financial incentives for alternative fuels (Ontario Ministry of Transportation, 2016). Potential exists for the government to create a more stringent fuel efficiency standard or to incentivize medium-duty alternative fuel trucks to promote these vehicles in commercial fleets in the future. But before policy action takes place, it is important to understand the costs and benefits of alternative fuel trucks, in a more comprehensive way. Current studies have not fully captured the interactions of weight, drive cycle, and temperature impact on fuel consumption. Studies, such as Yuksel & Michalek (2015) and Lee et al. (2013) used average fuel/energy consumption data obtained from the literature, or simulated fuel consumption on standard test drive cycles, which may not reflect real traffic conditions (Amirjamshidi & Roorda, 2014). The aim of this study is to address these gaps in the literature. 1.5 Research objective The objectives of this thesis are to: 1. simulate energy consumption for the diesel truck, BET, HET, and the CNG truck based on varies driving conditions 2. develop GHG emissions model for selected trucks and analyse the impact of driving conditions, such as payload, drive cycles and temperature on lifecycle GHG emissions 3. develop the total cost of ownership model for the selected trucks, and, 4. compare the cost of GHG abatement of the BET, HET and the CNG truck. 1.6 Thesis organization This thesis consists of three papers, which are presented in Chapters 2, 3 and 4. Chapter 2 is the analysis of the BET, Chapter 3 is the analysis of the HET and Chapter 4 is if the analysis of the

17 6 CNG truck. Finally, Chapter 5 compares and provides discussion of all alternative fuel truck results, including the conclusions and research limitations.

18 7 2 Chapter 2. Life Cycle GHG Emissions and Lifetime Costs of a Medium-duty Diesel and Battery-electric Truck. A Case Study for Toronto 2.1 Abstract Battery-electric trucks (BET) are an alternative to medium-duty diesel trucks and have the potential for lower life cycle greenhouse gas (GHG) emissions and lifetime costs. This study compares a Class 6 medium-duty battery-electric truck with a Class 6 diesel truck in terms of energy consumption (MJ/kg-km), life-cycle GHG emissions and total cost of ownership. The energy consumption simulation and life cycle assessment are performed based on payload weight, operating temperature and two drive cycles in Toronto (University: reflecting congested city driving and Freeway: reflecting high speed driving). The life cycle cost and 5-year cost of ownership are estimated. The BET is found to use 68% to 85% less energy and to have 41% to 86% less life cycle GHG emissions than the diesel truck in the University drive cycle; and to use 60% to 75% less energy and to have up to 22% more and up to 74% less GHG emissions, depending on the payload and operating temperature, in the Freeway drive cycle. For both lifetime and 5-year costs, depending on the conditions, the BET is found to have the possibility of higher or lower costs than the diesel truck (life cycle costs of BET range from 26% higher to 12% lower than the diesel truck; 5-year from 16% higher to 17% lower). The life cycle cost is most sensitive to lifetime vehicle kilometers traveled, energy consumption and discount rate; while the 5-year ownership cost is most sensitive to diesel price, battery cost and annual vehicle kilometers traveled. 2.2 Introduction Medium-duty trucks, Classes 4-6, have gross vehicle weight ratings between 6,351 kg and 11,793 kg. These trucks are mainly used for vocational purposes, such as emergency, delivery and dump trucks (Freightliner, 2016). The total energy used by medium-duty trucks in Canada increased by 19% from 1990 to 2011, despite average diesel consumption decreasing from 27.6 L/100 km in 1990 to 22.8 L/100 km in The U.S. Energy Information Administration (EIA) estimated that medium-duty trucks had the highest average annual growth rate (2.2%) in energy use from among commercial transportation modes (e.g., buses and heavy-duty trucks) (EIA, 2011). The increasing delivery demand and the rising sales and

19 8 market share of medium-duty trucks may have resulted in the growth in energy consumption despite the efficiency improvements (Natural Resources Canada, 2011; Statistics Canada, 2013). The battery electric powertrain is a promising technology to reduce energy consumption and life cycle greenhouse gas (GHG) emissions in both passenger vehicles and commercial trucks (National Research Council, 2010). Previous studies have demonstrated that the specifics of the drive cycle, among other issues, determine the fuel consumption of a vehicle (Sharer et al., 2007; Fontaras et al., 2008). Many studies have shown that electrification of vehicles results in higher reduction of GHG emissions in city, rather than highway, driving conditions compared with conventional vehicles (Raykin et al., 2012; Karabasoglu & Michalek, 2013; Lee et al., 2013). Karabasoglu and Michalek (2013) found that in the New York City Cycle, a batteryelectric vehicle can reduce life cycle GHG emissions by up to 60% relative to conventional vehicles; while GHG emission reductions are marginal under highway driving conditions. This is because the GHG emissions of hybrid and plug-in vehicles are relatively insensitive to drive cycles compared to conventional vehicles. Studies have indicated that medium-duty batteryelectric trucks (BET) can reduce GHG emissions by a considerable amount, especially in congested driving conditions with low average speed, and a high number of stops (Lee et al., 2013; Tong et al., 2015). Tong et al. (2015) compared the life cycle GHG emissions for various medium-duty trucks using different fuels and found that BETs can achieve GHG emissions reductions of 31% to 41% compared with diesel trucks. Lee et al. (2013) compared the life cycle GHG emissions of a medium-duty BET and a diesel truck and concluded that the BET reduces GHG emissions by 42-61% in the New York City Cycle, and by 19-43% in the City-Suburban Heavy Vehicle Cycle compared with a diesel counterpart. Along with driving cycle, operating temperature is another crucial factor affecting energy consumption of both conventional and battery-electric vehicles. Researchers from Argonne National Laboratory found that the fuel consumption for conventional gasoline vehicles operated in low (-7 C) and high (35 C) temperatures increased by 7% and 27%, respectively, compared with operating at 20 C in the Urban Dynamometer Driving Schedule drive cycle; and the energy consumption for battery-electric vehicles increased by 91% and 20% in cold and warm temperatures, respectively (Lohse-Busch et al., 2013). Yuksel and Michalek (2015) studied energy consumption for light-duty vehicles at different temperatures in different regions in the United States. They found that life cycle emissions associated with battery-electric vehicles may

20 9 increase by up to 22% in places with cold climates, even when operated off of the same electricity grid. This is because conventional vehicles use the engine heat for cabin heating in cold weather while the battery-electric vehicle motor is relatively efficient and generates little waste heat to be used for heating purpose. As shown in many studies, the electricity generation mix has a major effect on life cycle GHG emissions of battery-electric vehicles. Studies have assumed the average U.S. regional electricity generation mix and found that battery-electric vehicles emit less GHG emissions than the average gasoline vehicle, but the degree of reduction depends on where the vehicle is charged (Anair & Mahmassani, 2012; Yawitz, Kenward, & Larson, 2013). Yuksel and Michalek (2015) found battery-electric vehicle life cycle GHG emissions vary with U.S. regional grid mix, which has three times the GHG intensity in the Upper Midwest compared to the Pacific Coast. The GHG emissions and fuel consumption of medium-duty trucks are heterogeneous because they are highly dependent on drive cycles, operating temperature, payloads, and truck configurations (Taptich & Horvath, 2014; U.S. National Research Council (NRC)., 2014; Lohse- Busch et al., 2013; Fleetcarma, 2013; Yuksel & Michalek, 2015). However, no previous studies of medium-duty trucks that we are aware of comprehensively capture the impacts of drive cycle, operating temperature and payload on energy consumption of diesel and battery-electric medium-duty trucks. The studies of medium-duty trucks either use average fuel/energy consumption data obtained from the literature, or use standard test drive cycles, which may not reflect real traffic conditions (Amirjamshidi & Roorda, 2014; Tate, Harpster, & Savagian, 2008). These studies fail to capture the interactions of drive cycle, payload and temperature on energy consumption and life cycle GHG emissions. There is a financial incentive, up to $14,000, for light-duty battery-electric vehicles in Ontario but medium-duty vehicles are excluded (Ontario Ministry of Transportation, 2016). Potential exists for the government to incentivize medium-duty BET to promote these vehicles in commercial fleets in the future. But before such policy actions are taken it is important to investigate whether a BET is expected to be a good alternative to a diesel truck to reduce fuel consumption, GHG emissions and operating cost in the province. The aims of this study are; 1) to address these gaps by investigating impacts of fuel/powertrain type, payload, drive cycle, and

21 10 operating temperature on energy consumption and life cycle GHG emissions, and, 2) to estimate lifetime and five-year costs of ownership for medium-duty diesel and battery-electric trucks for a case study of Toronto. 2.3 Methodology and data The medium-duty diesel and battery-electric trucks are simulated using Autonomie (Argonne National Laboratory, 2016), a simulation tool, on different drive cycles, payloads, and temperatures. Then, energy consumption data are used to model life cycle GHG emissions using GHGenius (Natural Resources Canada, 2013), a Canadian life cycle assessment model for vehicles. Calculations of lifetime and 5-year cost of ownership, as well as sensitivity tests for two trucks are performed based on different scenarios of truck lifetime, truck purchase price and fuel prices, energy consumption, electric grid mix, etc. Finally, the cost of GHG abatement of the BET is calculated based on the incremental lifetime and 5-year ownership costs, as well as the GHG emissions reduced Truck specifications Truck specifications for the vehicles in this study are listed in Table 1. The diesel truck is similar to the Ford F-650 Super Truck (Ford, 2009) and the BET is similar to the SmithNewton BET (Smith Newton, 2013). Both vehicles are Class 6 trucks. The curb weight of the BET is 587 kg more, and the BET carries 317 kg less load, than the diesel truck. The BET is equipped with four 21 kwh batteries with a total weight of 720 kg. The drag coefficient of the diesel truck is not known and so is assumed the same as that of the BET. Table 1 Diesel truck and BET characteristics used in Autonomie to determine the vehicle energy consumption Fuel type Diesel Battery-electric Curb Weight without fuel storage (kg) Fuel storage system (kg)

22 11 Max payload (kg) Gross vehicle weight rating (kg) Drag coefficient Engine/motor Caterpillar 3126E UQM PowerPhase Engine/motor max power (kw) Autonomie Autonomie (Argonne National Laboratory, 2016), developed by Argonne National Laboratory, provides simulations that are suitable for automotive control system design (Vijayagopal et al., 2010). It performs customized vehicle simulations for particular vehicle specifications, drive cycles and temperatures. Autonomie is used in the U.S. Department of Energy FreedomCAR and Vehicle Technologies Program, and is used by more than 140 companies and research facilities (Karbowski et al., 2012). In Autonomie, a user can select a variety of commercially available vehicle components and adjust the peak power of components by linearly scaling the operating maps and component weights accordingly (Lewis et al., 2012). We rely on Autonomie in this study to simulate energy consumption for both the diesel truck and BET, reporting the estimates in MJ/kg-km. The diesel fuel conversion is based on its lower heating value, 35.8 MJ/Liter (U.S. Department of Energy, 2014). Lower heating values are used primarily for comparing alternative motor fuels because vehicles are typically not designed with the capability of condensing the moisture from combustion (MacLean & Lave, 2003) Drive cycles We adopt two real world drive cycles from Amirjamshidi et al. (2014) that reflect the traffic conditions experienced by medium-duty trucks in Toronto in the AM peak hour (8:00-9:00AM): University Ave drive cycle and Toronto Freeway drive cycle (Figure 2-1. University and Freeway Drive Cycle for Medium-duty Trucks in Toronto (Amirjamshidi & Roorda, 2014)). The University Avenue drive cycle ( University for short) has a high number of stops,

23 12 accelerations and decelerations, and low average speed; while the Freeway drive cycle ( Freeway for short) is the opposite. Figure 2-1. University and Freeway Drive Cycle for Medium-duty Trucks in Toronto (Amirjamshidi & Roorda, 2014) Temperature Temperature affects the energy consumption for both trucks in two main ways: it changes the efficiency of the powertrain and the energy consumption for the cabin climate control system. Thermal simulations are performed in Autonomie to examine these effects on energy consumption. There are several assumptions used in the medium-duty truck thermal models: 1) The change of efficiency of the vehicles' components due to variation in temperature is linearly scaled by the component s size and weight; 2) The BET and the diesel truck have the same heater/ac power capacity output; the maximum air conditioning and heater power output is 6 kw (Brotherton & Gallo, 2014). The climate control consumption varies with temperature; the more extreme the operating temperature, the more power the system consumes; 3) The truck operator uses the temperature control similarly to its use in a light-duty vehicle (i.e., turns heater/air conditioner on only when needed). Eight different temperatures (-20 C,-10 C, -5 C, 0 C, 10 C, 20 C, 30 C,40 C) are assessed to represent cold and warm weather. Vehicles of different weight, indicated by maximum payload percentage of 10%, 50% and 100%, are simulated in both the University and Freeway drive cycles for the eight temperatures Life cycle assessment After determining the energy consumption of both trucks using Autonomie (Argonne National Laboratory, 2016), the data are used in the well-to-wheel (WTW) analysis of GHG

24 13 emissions in GHGenius (Natural Resources Canada, 2013). The WTW analysis includes GHG emissions from three phases of the vehicle life cycle: 1) fuel cycle (well-to-tank); 2) vehicle operation (tank-to-wheel); and 3) vehicle cycle (vehicle production and transportation). GHGenius (version 4.03a) (Natural Resources Canada, 2013) is used for the analysis. Specifics of the analysis as well as modifications to the software to reflect the specific vehicles that are studied are described below. We focus on five GHGs: Carbon dioxide (CO2), Methane (CH4), Nitrous oxide (N2O), Chlorofluorocarbon-12 (CFC-12), and Tetrafluoroethane (HFC-134a),. These gases are converted to CO2 equivalent emissions using 2007 Intergovernmental Panel on Climate Change 100-year Global Warming Potential values (IPCC, 2007). The functional unit in this paper is gco2e per cargo kg per km travelled (gco2e/kg-km). For the diesel truck, the fuel cycle covers direct and indirect emissions associated with the diesel fuel from the following stages: fuel dispensing, fuel distribution and storage, fuel production, feedstock transmission, feedstock recovery, feedstock upgrading, land-use changes, fertilizer manufacture, gas leaks and emissions displaced by co-products of alternative fuels. For the BET, the fuel cycle emissions depend on both electricity generation sources and upstream emissions for each source for electricity generation. The geographic scope of this study is Toronto, Ontario. We use the 2014 Ontario average electricity generation mix (IESO, 2014) in this study (Figure average Ontario electricity generation sources (%) (IESO, 2014)). This information replaces the Ontario electricity mix 2014 forecast in GHGenius. The coverage of upstream emissions associated with each fuel source, including nuclear, natural gas, hydro, wind, and biomass, is similar to that of the diesel fuel. The vehicle operation phase considers the emissions from operating the vehicle. The BET has zero emissions in this life cycle phase. The vehicle cycle considers emissions from the production of the vehicle, e.g. material extraction, transportation and manufacturing. GHGenius does not include medium-duty BET in the model. To model GHG emissions from the vehicle cycle for the BET, we assume both trucks have similar material composition except for the fuel storage system, i.e., battery packs for the BET and fuel tank for the diesel vehicle. For the BET we replace the diesel fuel tank from the diesel truck with an 80 kwh lithium-ion battery pack. We assume that the pathway for the lithium-ion battery in the medium-duty truck is the same as that in the light-duty battery-electric vehicle. The material breakdown of fuel storage systems is shown in Table 2-1 Material breakdown by weight percentage (%) for fuel storage system (S&T Consultants Inc., 2013). The assumed efficiencies of battery and battery charging are 90% and 93% (S&T Consultants Inc., 2013a).

25 14 Figure average Ontario electricity generation sources (%) (IESO, 2014) Table 2-1 Material breakdown by weight percentage (%) for fuel storage system (S&T Consultants Inc., 2013) Material Liquid fuel tank (%) Lithium-ion battery (%) Virgin plain carbon steel Virgin stainless steel Recycled plain carbon steel 18 7 Advanced composites 0 10 Other plastics 0 20 Virgin aluminum 0 1 Virgin copper 0 1 Lithium 0 20 Total Lifetime total cost of ownership and sensitivity analysis The lifetime total cost of ownership (total cost of ownership), expressed in 2014 USD 1, is a function of annual vehicle kilometers traveled (VKT), fuel price, maintenance cost, energy average exchange rate: 1 CAD = 0.9 USD (

26 15 consumption (based on Autonomie simulation), and discount rate. For the BET, the total cost of ownership is also a function of the battery cost, the battery replacement cost, and the cost of electric vehicle supply equipment (EVSE) including installation. We also perform a sensitivity analysis on the total cost of ownership. Each parameter, mentioned above, has three values: lower estimate, base case, and higher estimate. In the sensitivity analysis, each variable, one at a time is varied from low to high, with all others held constant at the base case value. The purchase cost for the diesel truck is estimated to be $60,000 (Lee et al., 2013). The price for the BET is estimated to be $79,000+battery cost. The calculation for estimating the cost of the BET is shown in Table 3, in which the price is calculated based on vehicle weight (Lee et al., 2013); however, the price will vary depending on purchasing context because, as Lee et al pointed out, the prices for commercial vehicles can vary depending upon negotiation between the manufacturer and customers. The cost of the battery is estimated to be between $290 and $500 per kwh in 2014 (Nykvist & Nilsson, 2015). The specific range of battery prices is summarized in Table 2-2.

27 16 Table 2-2. Price estimate of the BET Non-battery components Parameter (unit) Value Source/note Wheel Base (mm) 3900 Payload (kg) 7558 (Smith Newton, 2011) Curb Weight (kg) 4432 Gross vehicle weight (kg) Non-battery (chassis & cab) vehicle price per unit GVW (2014$) GVW based non-battery parts Non-battery (chassis & cab) vehicle price per unit payload (2014$) 6.56 (Lee et al., 2013) $78, (Lee et al., 2013) Payload-based non-battery parts (2014$) $77,856 Battery The price range of nonbattery parts is small and it is similar to what was reported by (Davies, 2014). Thus, we use the average as an estimate for non-battery parts price Battery pack capacity (kwh) 80 (Smith Newton, 2011) Battery cost estimate in 2014 ($/kwh) Lower Base Case Higher $200 $350 $500 battery cost estimate in 2014 (Nykvist & Nilsson, 2015) 80 kwh battery pack cost estimate (2014$) Lower Base Case Higher $23,200 $28,000 $40,000 calculation

28 17 Continue Table 2-2 Overall (non-battery + battery) BET cost estimate (2014$) Lower Base Case Higher $101,426 $106,226 $118,226 calculation The annual VKT, ranging between 8,000 km and 32,000 km for medium-duty trucks (Natural Resources Canada, 2010), is assumed to be the same for both vehicles. This is equal to km per day assuming 260 working days per year (Feng & Figliozzi, 2011). The maximum advertised range for the Smith Newton BET is 160 km per day. We assume that driving behavior is the same for both vehicles. Fuel and electricity cost, in $/Liter or $/kwh, is the average diesel or electricity price in Toronto. The historical diesel price in Toronto in 2014 is collected from Natural Resources Canada (2014). The diesel price projection from 2015 to 2030 is adopted from the Annual Energy Outlook, 2011 (U.S. Energy Information Administration, 2015) U.S. Energy Information Administration, We assume that the Toronto diesel price has the same rate of change as the diesel price projected. The maintenance cost for the diesel truck is assumed to be $0.17/km in (Barnitt, 2011); and that of the BET is assumed to be between 30-50% of that of the diesel truck (Lee et al., 2013). The battery replacement cost for the diesel truck is relatively small and included in the maintenance cost. For the BET, we estimate that the lifespan for the battery is between 7-11 years based on the battery life span according to the daily traveled distance ( km) and energy consumption and battery manufacturer lifetime estimates -- between 1000 and 3000 cycles at 100% depth of discharge (A123 battery, 2015; Valence, 2015; Fleet News, 2010). The cost of battery replacement depends on when the battery is replaced. Smith Newton offers battery warranties for 5 years (Smith Newton, 2013), so there is no replacement cost for owners if replacement is within the first 5 years. In other cost scenarios it is assumed that the battery is $/km in 2011, converted to 2014 dollars, discount rate=6.5%

29 18 replaced after 7-13 years when the price of battery is between $ /kWh (Nykvist & Nilsson, 2015) Five year cost of ownership We calculate the five year cost of ownership by incorporating the depreciation rate and resale value estimates for both vehicles at the end of the fifth year of ownership. Since the battery and the EVSE will not have reached their end-of-life at this point, we exclude the battery and EVSE replacement costs in the analysis of the BET. We assume that the diesel truck depreciation rate is 24% for the first year and 15% afterwards based on (EEI, 2014). The depreciation rate for the BET is difficult to obtain. Light-duty battery electric vehicles are expected to have a 24-48% lower depreciation rate than conventional vehicles after 5 years (Propfe et al., 2012). Gipe from Smith Electric Vehicles assumed that the BET has the same depreciation rate as the diesel truck in their analysis but the battery depreciates faster (Gipe, 2013). In this study, we use Gipe s assumption: the BET truck body has the same depreciation rate as the diesel truck and the battery has an 8-year life with 10% residual, or 25% depreciation per year. Based on these assumptions, we assume that after five years, the diesel truck retains 40% of its original value and the BET retains 34-38% depending on the battery price. The ranges of annual and lifetime VKT, battery package cost, battery replacement cost, diesel fuel price, energy consumption, and BET maintenance cost are summarized in Table 2-2.

30 19 Table 2-3 Ranges of parameters for sensitivity analysis for lifetime and five-year ownership cost analyses Variable Lower Base case Higher Justification and Source Annual Vehicle Kilometers Travelled/VKT (km) 8,400 19,000 23,000 Lifetime VKT (km) 160, , ,000 Battery cost in 2014 (2014USD/kWh) Battery replacement cost: (2014USD/kWh) Toronto diesel price:(2014usd/l) Diesel truck energy consumption (L/km) Average annual VKT of Prince Edward Island (province with lowest annual VKT), Ontario, and Nova Scotia (province with highest VKT), based on Canadian Vehicle Survey 2009 (Natural Resources Canada, 2010); Base case from Lee et al. (2013); higher value is from (Antich, 2008); lower is base case minus 8000 km (the interval between the base case and higher) below base case Low/base case/high estimates in 2014 (Nykvist & Nilsson, 2015) Low/medium/high battery cost estimates in 2020, assuming 7 years of battery life (Nykvist & Nilsson, 2015) Assumption: 1CAD=0.9USD (Bank of Canada, 2016); the price range is obtained from (Natural Resources Canada, 2014); the diesel price projection (Figure 2-3) is obtained from Annual Energy Outlook, 2011, assuming the same rate of change (U.S. Energy Information Administration, 2015). Autonomie simulation results: lower 0% payload on Freeway driving cycle at 20 C; Base case: average of 50% payload on university and freeway drive cycles at 20 C; Higher: 100% payload on University drive cycle at - 20 C

31 20 Cont d BET energy consumption (kwh/km) BET maintenance cost ($/km) Autonomie simulation results: lower 0% payload on Freeway driving cycle at 40 C; Base case: average of 50% payload on University and Freeway drive cycles at 20 C; Higher: 100% payload on University drive cycle at -20 C BET maintenance cost is 30% to 50% of the diesel truck, $0.15/km in 2014, converted from $0.14/km in 2011 based on Lee et al., Inflation rates are from (World Bank, 2016) EVSE (installation + replacement after 6 years) Lower assumes Level 1 EVSE (zero cost); base case-and higher assume lower and higher range Level 2 EVSE from Lee et al., Discount Rate (%) Obtained from U.S. DOT guidance on cost-benefit analysis, Circulars A-4 and A- 94. (U.S. Department of Transportation, 2015) Electricity price (USD/kWh) Off-peak, mid-peak and onpeak prices (Toronto Hydro, 2015) Ontario electricity mix 100% natural gas 2014 Ontario average 100% Nuclear Selected two largest shares of the Ontario electricity grid: nuclear ~60%, and natural gas ~23%; the assumed natural gas plant efficiency is 40% (S&T Consultants Inc., 2013); Ontario average electricity generation mix is from IESO (2014) Diesel truck depreciation rate (only used for 5-year total cost of ownership) (%) 23 for 1st year and 15 for following years (EEI, 2014); Resale value is original price less depreciation

32 21 Cont d BET depreciation rate (only used for 5-year total cost of ownership) Straight line depreciation; assumes 10-year life on truck with 15% residual; 8-year life for battery with 10% residual BET depreciation rate assumption based on Gipe (2013) Figure 2-3 Toronto diesel fuel price projection (2014$/Liter) (based on data in U.S. Energy Information Administration, 2015) Cost of GHG emissions abatement The cost-effectiveness of reducing GHG emissions is the cost of reducing one tonne of life cycle GHG emissions (CO2e) using the BET. It is calculated as the difference in lifetime total cost of ownership between the BET and the diesel truck (2014 USD) divided by the difference in life cycle GHG emissions of the BET and diesel truck (tonnes CO2e). 2.4 Results and discussion Energy consumption by temperature, payload and drive cycles In line with common knowledge, we found that energy consumption of the diesel truck is higher than the BET under all drive cycles, temperature and payload conditions because the electric motor is more efficient than the diesel engine. Energy consumption for both types of trucks is considerably higher at 10% payload compared to 50% and 100% because the load weight (denominator) is much smaller at 10% payload despite the lower fuel being consumed.

33 22 Figure 2-4. Diesel truck and BET energy consumption by temperature and payload in the University drive cyclefigure 2-4 and Figure 2-5 report the energy consumption of the diesel truck and BET under various payloads and temperatures for the University and Freeway drive cycles. The figures also show the energy reduction (%) of the BET compared to the diesel truck. Comparing the figures, the diesel truck is more sensitive to the drive cycle than the BET: it consumes more energy in the University than in the Freeway drive cycle; while the energy consumption for the BET is similar in both drive cycles. As a result, the BET consumes less energy than the diesel truck in the University drive cycle than in the Freeway drive cycle. The colder temperatures increase the energy consumption of the BET more than warmer temperatures, and it is more pronounced in the University than in the Freeway drive cycle. The impact of temperature on the diesel truck is less prominent. Compared with the energy consumption at 20 C, the energy consumption of the BET in cold weather (-20 C) conditions increases by 23% to 91% in the University and 16% to 63% in Freeway drive cycle, while the diesel truck energy consumption only increases 3% to 17% and 1% to 13% in University and the Freeway drive cycle, respectively. Compared to operation at 20 C, in hot weather (40 C), the energy consumption of the diesel truck increases by 5% to 13% in the University and 2% to 5% on the Freeway drive cycle. The warm weather impact on the BET energy consumption is more complicated: the BET consumes 5% to 8% more energy in the University drive cycle when it is 10% loaded, or 5% to11% less energy for other payloads in the Freeway drive cycle. This is because the vehicle components are more efficient in warm weather and, thus lose less energy. In cold weather, the energy consumption of the vehicle increases not only due to use of heaters, but also due to additional energy loss from vehicle components. As a result, warm temperatures contribute to the higher energy reduction of the BET compared to the diesel truck.

34 23 Figure 2-4. Diesel truck and BET energy consumption by temperature and payload in the University drive cycle Figure 2-5. Diesel truck and BET energy consumption by temperature and payload in the Freeway drive cycle

35 Life cycle GHG emissions As stated before, the WTW analysis includes three phases: the fuel cycle (fuel extraction, production and distribution), the vehicle cycle (material extraction and production, as well as vehicle assembly), and finally the vehicle operation cycle. The breakdown of life cycle GHG emissions by life cycle phase, in terms of grams of CO2 equivalent (gco2e/kg-km), is reported in Figure 2-6. The BET has higher GHG emission than the diesel truck in the vehicle cycle but lower overall GHG emissions in most cases. Most of the life cycle GHG emissions for the diesel truck come from the vehicle operation cycle, while for the BET most come from the vehicle cycle. The BET is 104% to 508% more GHG intensive to produce than the diesel truck due to the battery pack depending on the payload. The BET reduces GHG emissions by 41% to 84% in the University drive cycle compared with the diesel truck with same cargo weight; and 27% to 73% in the Freeway drive cycle at 10% and 50% payload. However, the GHG emissions for the BET are 14% higher than those of diesel truck at 100% payload in the Freeway drive cycle. This is due to the notable decrease of GHG emissions in the vehicle operation cycle for the diesel truck, caused by the lower energy consumption at 100% payload. Sharer et al. (2007) pointed out that the engine efficiency increases at higher payload, which partially offsets the increased energy required at higher payload. In addition, the emissions from the vehicle cycle for the diesel truck is close to zero due to the high payload; while the GHG emissions from the vehicle cycle for the BET is still high, despite that the vehicle operation and fuel cycle emission are nearly zero. As a result, the GHG emission for the BET exceeds that for the diesel truck at 100% payload on the Freeway drive cycle.

36 25 Figure 2-6. Life cycle GHG emissions of diesel truck and BET under different drive cycle and payload conditions Figure 2-7 and Figure 2-8 report the life cycle GHG emissions of the diesel truck and the BET, as well as the amount of GHG emissions reduced by the BET, in different temperatures in the University and Freeway drive cycle. We use the energy consumption data by temperature from Autonomie simulations to calculate the life cycle GHG emissions, as if the vehicle is operating in one temperature, payload, and drive cycle for the entire life. In most Canadian/U.S. jurisdictions, temperatures would be varying throughout the year, so the life cycle GHG emissions could be weighted based on the percentage of time that a particular temperature would be experienced throughout the year. Because our aim is to examine the difference in life cycle emissions at different temperatures, instead we assume that the vehicle operates its entire lifetime under one temperature (and payload), and discuss key differences in the results, so as to essentially bound the life cycle estimates for these conditions.

37 26 Figure 2-7. Life cycle GHG emissions of diesel truck and BET under different temperatures and payloads on the University Drive Cycle Figure 2-8. Life cycle GHG emissions of diesel truck and BET under different temperatures and payloads on the Freeway Drive Cycle

38 27 Vehicles operating in colder and warmer climates may have increased life cycle GHG emissions due to increases in energy consumption, but the degree of change varies depending on fuel type, weight, and drive cycle. The life cycle GHG emissions for the BET are more sensitive to temperature than those of the diesel truck. In cold weather, the life cycle GHG emissions for the BET increase by 1% to 50% in the University drive cycle compared with operation at 20 C, or 1% to 40% in the Freeway drive cycle. In warm weather (30 C to 40 C) the GHG emissions for the BET show little increase (less than 2% in both drive cycles). Compared to at 20 C, for the diesel truck, the life cycle GHG emissions increase by 3% to 19% and 2% to 12% in cold temperatures in the University and Freeway drive cycles, respectively. Compared to at 20 C, in warm weather, the GHG emissions for the diesel truck increase by 6% to 14% and 1% to 4%, in the University and Freeway drive cycles. The life cycle GHG emissions reduction of the BET compared to the diesel truck varies under different temperatures and drive cycles. In the University drive cycle, the BET reduces emissions by 82% to 87% in all temperatures when it is 10% loaded; however, it only reduces emissions by 29% to 38% at 100% payload. In the Freeway drive cycle, the BET has less GHG benefits: it reduces emissions 72% to 79% when 10% loaded, but the emissions increase by 18% to 24% when fully loaded. As shown in Figure 6, this is due to the increased engine efficiency and reduced energy consumption at higher payload on the Freeway driving condition, which leads to a notable GHG emission reduction in the life cycle emissions for the diesel truck Total cost of ownership Figure 2-9 and Figure 2-10 show the lifetime and 5-year total costs of ownership of the diesel truck and BET based on the base case parameter values in Table 2-3 Ranges of parameters for sensitivity analysis for lifetime and five-year ownership cost analyses.

39 28 Figure 2-9. Lifetime total cost of ownership for the diesel truck and BET Figure year total cost of ownership for the diesel truck and BET The purchase cost accounts for 36% and 64% of the lifetime total cost for the diesel truck and BET, respectively. However, the BET has lower lifetime fuel cost and maintenance cost than the diesel truck. Based on the truck depreciation assumptions described above, we found that the resale value of the BET, $49,800 is higher than that for the diesel truck, $26,500, at the end of fifth year. Overall, the BET costs about the same as the diesel truck in both the lifetime total cost of ownership and 5-year total cost of ownership, respectively. We perform sensitivity analysis on the lifetime total cost of ownership (Figure 2-11) and 5-year total cost of ownership (Figure 2-12). In the base case, the lifetime total cost of ownership of the BET is 4% higher than that of the diesel truck. The total cost of the BET ranges from 26% more to 12% less than the diesel truck. The difference of lifetime total cost of ownership is most

40 29 sensitive to the energy consumption, annual VKT, the battery cost, and the EVSE installation costs. Figure Sensitivity analysis of difference in lifetime total cost of ownership between BET and diesel truck Figure Sensitivity analysis of difference in 5-year total cost of ownership between BET and diesel truck In terms of 5-year total cost of ownership (Figure 12), the cost difference is most sensitive to the annual VKT, energy consumption, the EVSE installation and battery cost parameters. It is least sensitive to fuel price and BET maintenance cost. In summary, the BET has a financial advantage over the diesel truck in city operating conditions in Toronto where the diesel truck consumes considerable fuel, and when annual VKT are low.

41 Cost-effectiveness of GHG emissions abatement The cost-effectiveness of GHG emissions reductions for the BET, assuming base case values, are $42/tonne-CO2e and -$153/tonne-CO2e based on lifetime ownership costs and 5-year ownership costs, respectively (Figure 2-13 and Figure 2-14). A negative cost of GHG abatement indicates that the BET not only reduces GHG emissions but also has a lower ownership cost. From a lifetime cost perspective, it costs money to reduce GHG emissions with the BET except when the energy consumption for both trucks is high (e.g., heavy payload in congested city driving conditions) (Figure 2-12). On the other hand, from a 5-year cost perspective, money can be saved with GHG abatement in most cases (Figure 2-13). Ontario will soon implement a capand-trade system (Canadian Press, 2015). If carbon is priced at $30/ton (Beaty et al., 2014) and the fleet industry is subject to this regulation, purchasing a BET would be expected to be financially attractive to fleet owners if the cost of GHG emissions abatement using the BET is lower than the market price. This result indicates that the BET has additional financial benefit in the short term if commercial fleets are subject to the cap-and-trade program. Figure Cost-effectiveness of GHG emissions abatement for the BET based on lifetime total cost of ownership

42 31 Figure Cost-effectiveness of GHG emissions abatement for the BET based on 5-year total cost of ownership 2.5 Conclusions The contribution of this paper is that it investigates the fuel consumption and life cycle GHG emissions for medium-duty diesel and battery-electric trucks using real world drive cycles and temperatures variations; It also assesses lifetime and 5-year total costs of ownership, and cost-effectiveness of GHG emissions reduction for the diesel truck and BET. Generally, for our Toronto case study, the BET has lower life cycle GHG emissions than the diesel truck, but this does not hold across all conditions. Emissions of the BET are higher than those of the diesel truck with 100% payload and in the Freeway drive cycle. We find that the BET reduces energy consumption by a notable amount compared with the diesel truck in most cases. For the BET, cold weather has a higher impact on energy consumption and life cycle GHG emissions than warm weather; and the temperature impact is more prominent in the University than in the Freeway drive cycle. Assuming there are no government incentives, the lifetime cost of the BET is higher than that of the diesel truck under most conditions, but the 5-year net ownership cost is lower due to the higher assumed resale value of the BET. These results are only for the trucks being modeled and should not be generalized to other diesel and BET trucks.

43 32 If a truck travels mostly in the Freeway drive cycle, especially at heavier payload in cold temperatures, the diesel truck is preferred over the BET because the diesel truck has both lower cost of ownership and lifetime GHG emissions. On the other hand, if a truck travels mostly in city driving condition, the BET is preferred over the diesel truck at all temperatures and payloads because it saves both GHG emissions and lifetime costs. However, this conclusion only applies to Toronto, where the electricity grid is mostly nuclear and hydro. The BET operating in other locations may not achieve the same life cycle GHG emissions reduction or lifetime costs of ownership. In 2011, the U.S. government released the first fuel economy standard for medium- and heavy-duty trucks. The standard (Phase 1) required new (Model Year ) vocational vehicles (delivery trucks, buses, garbage truck) to improve the corporate-average fuel economy by 10 percent by 2018 (NHTSA and EPA, 2010). This standard, however, has not yet been adopted in Ontario. As a result, there is an opportunity in Ontario to create more stringent standards for fuel economy in medium-duty trucks to promote the BET, especially for those making intra-urban movements, given the environmental and financial benefits under city driving conditions. The government could coordinate with local fleet operators and give incentives for the purchase of BETs. Last-mile delivery vehicles are good candidates for adopting BETs because; 1) they travel mostly in city driving conditions, where the greatest GHG reduction occurs; 2) they have fixed routes for delivery and travel relatively short distance and, 3) they return to the depot, where they can be charged, at the end of the day (Duran et al., 2014). Business owners of commercial vehicles are generally sensitive to fleet operating costs. Therefore, the research results suggest that BET and other alternatives to diesel trucks should be investigated further to determine the full range of conditions under which they are expected to lead to cost-effective reductions of GHG emissions and other negative impacts on the environment.

44 33 3 Chapter 3 Life Cycle Emissions and Lifetime Costs of Medium-duty Diesel and Hybrid-electric truck. A Case Study for Toronto 3.1 Abstract This research analyzes life cycle GHG emissions and lifetime costs of a class 6 medium-duty diesel truck and a class 6 hybrid-electric truck. The energy consumption (MJ/kg-km), lifecycle GHG emissions, and lifetime ownership costs for two trucks are simulated and compared based on various payload weights, operating temperatures and two real-world drive cycles in Toronto: the University drive cycle, reflecting congested, city driving condition, and the Freeway drive cycle, reflecting high speed freeway driving condition. We found that under our driving conditions specific to Toronto, the HET consumes 1% - 21% less energy than the diesel truck and, with the exception of the Freeway drive cycle with 100% payload, has 2% - 45% lower life cycle GHG emissions. Operating temperature has more impact on energy consumption and GHG emissions of the HET than the diesel truck, and the impact is more prominent for the University than for the Freeway drive cycle. We also found that the lifetime ownership costs for the HET varies from 3% less to 40% more than that of the diesel truck. The cost of GHG abatement using the HET is high, up to $1180/tonne CO2e, due to its higher lifetime cost of ownership, which is due to the higher purchase cost of the HET. 3.2 Introduction Freight transportation by truck was responsible for 32% of Canada s transportation sector greenhouse gas (GHG) emissions in 2014 (Environment and Climate Change Canada, 2016). As reported in Bachmann et al. (2014), medium-duty trucks for interurban movements are the least efficient amongst other freight vehicles. The energy use of medium-duty trucks in Canada increased by 19% from 1990 to 2011, due to increasing vehicle kilometers traveled (Statistics Canada, 2010; Natural Resource Canada, 2013a). To reduce energy consumption and GHG emissions from ground transportation, freight operators such as UPS practice fuel reduction strategies, such as reducing vehicle miles travelled by using advanced vehicle routing systems, and deploying alternative fuel trucks, such as hybrid-electric trucks (HETs) (UPS, 2016). Purolator, a Canadian courier company, added 115 new HETs for their curbside delivery fleet in

45 and claimed to save more than 2,100 tonnes of GHG emissions and 645,000 liters of fuel (Purolator, 2010). Many studies have shown that vehicles that employ hybrid-electric or battery-electric technology have lower energy consumption compared to diesel vehicles, especially for city driving conditions (Raykin et al., 2012; Karabasoglu & Michalek, 2013; Lee et al., 2013; Lave & MacLean, 2002). This is mainly due to engine start/stop patterns and regenerative braking (Azure Dynamics, 2010). Ragatz et al. (2014) evaluated UPS diesel hybrid-electric delivery vans and found that, compared with a diesel truck, the fuel economy of the HET was 13% to 36% higher and this resulted in annual fuel savings of 176 to 610 gallons (666 to 2309 liters) per vehicle. The study also showed that more fuel saving advantage could be realized by the HET in drive cycles with more frequent stop-and-go events and lower average speed. Studies on life cycle GHG emissions for light-duty diesel and hybrid-electric vehicles are abundant (e.g., Lave & MacLean, 2002; Spatari et al., 2005 Karabasoglu & Michalek 2013). Karabasoglu and Michalek (2013) studied the influence of driving patterns on life cycle GHG emissions for light-duty electrified vehicles. They found that under the New York City Cycle, hybrid-electric vehicles can reduce energy consumption by up to 54% compared with conventional vehicles, and can reduce life cycle GHG emissions by up to 58%. Under a highway drive cycle they found that the emission reduction is marginal, only 5%. Life cycle assessments (LCAs) of medium-duty trucks are less common than those of light-duty vehicles. Bachmann et al. (2013) conducted a LCA comparing the GHG emissions of HET and diesel medium-duty trucks. They used energy consumption data from the HET manufacturer and concluded that a HET has 23% lower life cycle GHG emissions than the diesel truck in city driving conditions and 8% lower in highway driving conditions. Tong et al. (2015) compared life cycle GHG emissions between a HET and an equivalent diesel truck, using energy consumption data from Lammert and Walkowicz (2012), and concluded that the HET reduced life cycle GHG emissions by 20% compared to the diesel truck. Recent studies show that operating temperature also influences the energy and environmental performance of both HETs and diesel trucks. Researchers from the Argonne National Laboratory tested the impact of ambient temperature on fuel consumption for both light-duty conventional (gasoline) vehicles and hybrid-electric vehicles (Lohse-Busch et al.,

46 ). They found that temperature impacts energy consumption in city driving conditions more than in highway driving conditions. In the Urban Dynamometer Driving Schedule drive cycle (city driving conditions), the energy consumption of the conventional vehicle (2012 Ford Focus) increased by 91% and 20% at -7 C and 35 C, respectively, compared with energy consumption at 22 C. For the hybrid-electric car (2010 Prius) energy consumption increased by 65% and 56% for the low and high temperatures, respectively. For the highway driving conditions, energy consumption for the conventional vehicle increased by 10% and 14%, and for the hybrid-electric vehicle increased by 18% and 20% in the low and high temperatures. No study we are aware of has compared the life cycle GHG emissions of diesel and hybrid-electric vehicles under different temperature conditions. However, Yuksel and Michalek (2015) studied the temperature impact on light-duty battery-electric vehicle life cycle emissions across regions in the United States. They found that the life cycle emissions associated with battery-electric vehicles may increase by up to 22% in places with cold climates, even if charging from the same electricity grid. Studies such as Yuksel & Michalek (2015) and Tong et al. (2015) that have used average fuel consumption data collected from surveys, are limited in their ability to determine the impacts of specific driving conditions, such as drive cycle, payload weight, and temperature, on energy consumption. On the other hand, studies such as and Lee et al. (2013), which used data from simulations on standard test drive cycles, may not reflect real-world traffic conditions and could result in under- or overestimating fuel consumption and GHG emissions (Amirjamshidi & Roorda, 2014). Prior research of the authors, Zhou et al. (2016), aimed to address these research gaps, and estimated life cycle GHG emissions and lifetime costs for a battery-electric truck and a diesel truck. This research is an extension of the previous study and focuses on the HET. The objectives of this study are; 1) to investigate impacts of fuel type, drive cycle, weight, and temperature on energy consumption and life cycle GHG emissions of diesel and HETs, and, 2) to estimate the lifetime total cost of ownership for a medium-duty diesel and hybrid-electric truck for a case study of Toronto, Canada. 3.3 Method and data Autonomie (Argonne National Laboratory, 2016), an industry standard vehicle simulation tool, is used to simulate the fuel consumption for the diesel truck and HET. Simulation data are then used in GHGenius (Natural Resources Canada, 2013), a Canadian life cycle model for fuels and

47 36 vehicles, to model life cycle GHG emissions. Cost of ownership estimates are developed based on scenarios of energy consumption, fuel price, lifetime, and purchase costs. Lastly, costs of GHG abatement are derived based on the differences in lifetime cost of ownership and life cycle GHG emissions Vehicle specifications Vehicle specifications in this study are listed in Table 3-1. The diesel truck is similar to a Ford F-650 (Ford, 2009) and the diesel HET is similar to the Utilimaster P100H truck, which has a series configuration powertrain (Ragatz et al., 2014). Both vehicles are Class 6 trucks. The curb weight of the HET is 266 kg more and thus, it is assumed the HET carries 266 kg less payload than the diesel truck. Table 3-1 Diesel truck and HET characteristics used in Autonomie to determine the vehicle energy consumption Fuel type Diesel Hybrid-electric Curb Weight without fuel storage (kg) Fuel storage system (kg) Max payload (kg) Gross vehicle weight rating (kg) Drag coefficient Engine/motor Max power (kw) 188 Engine: 188; Motor: Autonomie Autonomie (Argonne National Laboratory, 2016), developed by Argonne National Laboratory, is a simulation tool used in various vehicle technology programs in the U.S. Department of Energy, and by more than 140 companies and research facilities (Karbowski et al., 2012). It allows the user to modify the vehicle components (e.g., engine power, payload weight) by linearly scaling the operating maps and component weights accordingly (Lewis et al.,

48 ). Users can also change the operating conditions (e.g., drive cycle and temperature) to generate a more accurate fuel consumption estimate. In our study, the energy consumption is expressed in MJ/kg-km. The lower heating value of diesel fuel, 35.8MJ/Liter (U.S. Department of Energy, 2014), is used for conversion from diesel fuel consumption to energy consumption for the comparison of alternative motor fuels (MacLean & Lave, 2003) Drive cycles Drive cycles reflecting real world traffic and road conditions were used to estimate energy consumption reflecting conditions in Toronto, Canada. Amirjamshidi et al. (2014) developed Toronto specific drive cycles for light-, medium-, and heavy-duty vehicles based on GPS data and microsimulation. Two medium-duty vehicle drive cycles from Amirjamshidi et al. (2014) are selected in this study: the University drive cycle and Toronto Freeway drive cycle (Figure 3-1). The University drive cycle represents city driving conditions in Toronto (on University Avenue), with a low average speed and a large number of stops-and-starts. The Toronto Freeway drive cycle is the opposite, reflecting highway driving conditions. Figure 3-1 University and Freeway Drive Cycle for Medium-duty Trucks (based on data in Amirjamshidi et al. 2014) Temperature The powertrain efficiency as well as the energy used to power the air-conditioner or heater may change as the ambient temperature increases or decreases. The thermal model in Autonomie is used to simulate the temperature impact on energy consumption in various drive cycles and payload weights. We assume that; 1) the change in powertrain efficiency due to variation in temperature is proportional to powertrain size and weight; 2) the maximum air conditioner (AC) and heater power is 6 kw for both the HET and diesel truck (Brotherton & Gallo, 2014). The energy consumption varies with temperature: the greater the difference

49 38 between the operating temperature and the desired cabin temperature, the more energy is consumed; 3) the Heater/AC is only turned on when needed. This study analyzes eight temperatures from -20 C to 40 C, to represent very cold to very warm weather. We also test the temperature impact on energy consumption for different truck weights, 10%, 50% and 100% of maximum payload weight, and drive cycle, the University and Freeway drive cycles LCA scope Fuel consumption outputs from Autonomie are used in the well-to-wheel (WTW) analysis to determine life cycle GHG emissions. The WTW analysis is a method to analyze life cycle GHG emissions for vehicles from three phases of the life cycle: 1) fuel cycle (well-to-tank), which covers direct and indirect emissions from fuel dispensing, fuel distribution and storage, fuel production, feedstock transmission, feedstock recovery, feedstock upgrading, land-use changes, fertilizer manufacture, gas leaks and emissions displaced by co-products of alternative fuels; 2) vehicle operation (tank-to-wheel), which includes emissions associated with operating the vehicle; and 3) vehicle cycle, which considers emissions from the production of the vehicle. We assume both trucks have similar material composition except that the HET has additional GHG emissions resulting from battery pack production. The material breakdown of fuel storage systems (obtained from S&T Consultants Inc., 2013) is shown in Figure 3-2 To conduct WTW analysis, we use GHGenius (version 4.03a) (S&T Consultants Inc., 2013). GHGenius has the capability of estimating life cycle energy use, GHG and other emissions for vehicles of all classes and many transportation fuels. Five GHGs are assessed in this study: Carbon dioxide (CO2), Methane (CH4), Nitrous oxide (N2O), Chlorofluorocarbon-12 (CFC-12), and Tetrafluoroethane (HFC-134a),. These gases are converted to CO2 equivalent emissions using 2007 Intergovernmental Panel on Climate Change 100-year Global Warming Potential values (IPCC, 2007). The functional unit in this paper is gco2e per cargo kg per km traveled (gco2e/kg-km).

50 39 Figure 3-2 Material breakdown by weight percentage (%) for fuel storage systems (based on S&T Consultants Inc., 2013) Lifetime total cost of ownership and sensitivity analysis The lifetime total cost of ownership is calculated based on purchase costs, vehicle kilometers traveled (VKT), fuel price, maintenance cost, fuel costs, energy consumption (based on Autonomie simulation), and discount rate. Sensitivity analysis for the total cost of ownership is completed for each of these parameters. In each case, one parameter is varied (from lower to higher estimate), while other parameters are held at their base case values. Table 2 summarizes the estimates for each parameter in the sensitivity analysis. The parameter ranges for vehicle operating characteristics (annual and lifetime VKT), diesel fuel price, diesel truck energy consumption and discount rate and are the same as those selected in our prior work, Zhou et al.(2016) The purchase price for the HET is estimated to be 30%, 45%, and 60% more (representing lower, base case, and higher estimates, respectively) than that of the diesel truck (Mims, 2009), which is $63,174 in 2014 USD according to Lee et al. (2013). This price range is also shown in the alternative fuel vehicle dataset from GSA (2012). However, it should be noted that this is an estimated price, considering that purchase prices vary depending on truck model and negotiations between manufacturers and customers (Lee et al., 2013). The annual medium-duty truck VKT is obtained from the Canadian Vehicle Survey 2009 (Natural Resources Canada, 2010). We assume the lower, base case and higher estimates for

51 40 medium-duty truck VKT in Canada are 8,000, 19,000, and 32,000 km/year for both the HET and diesel trucks. The lowest annual VKT is the average medium-duty truck VKT for Prince Edward Island; while the base case and higher estimates are those of Ontario and Nova Scotia. The Toronto 2014 diesel fuel price is obtained from Natural Resources Canada (2014) and converted to 2014 USD. We assumed the average lower, base case and higher diesel fuel prices in Toronto in 2014 are 1.03, 1.2 and 1.6 USD per liter, based on an exchange rate of 1 CAD=0.9 USD. To model the lifetime fuel costs, we assume that the average diesel fuel price in Toronto from 2015 to 2030 has the same rate of change as that projected in the Annual Energy Outlook, 2011 (US Energy Information Administration, 2015). The maintenance cost for the diesel truck is assumed to be $0.17/km in (Barnitt, 2011); and that of the HET is assumed to be the same (Lammert & Walkowicz, 2012). Battery degradation for the HET is less sensitive to capacity fade than for plug-in hybrid or batteryelectric vehicles (Karabasoglu & Michalek, 2013), thus we assume no battery replacement cost for the HET. The lower (10% payload on the Freeway drive cycle), base (the average of the 50% payloads on the University and the Freeway drive cycles) and higher (100% payload on the University drive cycle) estimates of energy consumption for both trucks are obtained from the Autonomie simulation at 20 C. Because the high or low energy consumption for the HET and the diesel trucks occur at different temperatures, we do not include the temperature impact in the sensitivity analysis. The discount rates used in this study are 3%, 5%, and 10% based on the U.S. Department of Transportation guidance on cost-benefit analysis, Circulars A-4 and A-94. (U.S. Department of Transportation, 2015). The range of annual and lifetime VKT, battery package cost, battery replacement cost, diesel fuel price, energy consumption, HET maintenance cost are summarized in Table 3-2. Table 3-2 Values of parameters for sensitivity analysis $/km in 2011, converted to 2014 dollar value, discount rate=6.5%

52 41 Variable Low Base case High HET purchase costs (2014$USD) Annual Vehicle Kilometers Travelled (km) Lifetime Vehicle Kilometers Travelled (km) Toronto Diesel Price in 2014 (USD$/L) Diesel Truck Energy Consumption (L/km) HET Energy Consumption (L/km) 82,126 91, ,078 8,400 19,000 23, , , , Discount rate 3% 7% 10% Cost of GHG abatement The cost of GHG abatement is the additional cost of the HET compared to the diesel to abate one tonne of life cycle GHG emissions (CO2e) using the HET. It is useful for policy makers and other stakeholders to understand the cost of reducing GHG emissions. In this study, the cost of GHG abatement is the difference in ownership costs of the HET compared to the diesel truck divided by the difference in life cycle GHG emissions of the two trucks 3.4 Results Energy consumptions The energy consumption (MJ/kg-km) of the diesel truck and the HET under different payloads and temperatures on the University and Freeway drive cycle are shown in Figure 3-3 and Figure 3-4, respectively. Both types of truck consume more energy in the University drive cycle than in the Freeway drive cycle when the operating conditions (payload and temperature) are same.

53 42 The HET consumes less energy per kg-km in both drive cycles in all payloads and temperatures. The difference in energy consumption is larger in the University drive cycle than in the Freeway drive cycle, due to the benefit of regenerative braking. In the University drive cycle, the HET consumes 7% to 48% less energy than the diesel truck depending on payload and temperature; while in the Freeway drive cycle, the HET consumes 1% to 21% less energy. The reduction in energy consumption of the HET, compared to the diesel, decreases as payload increases, but rises as temperature increases. For instance, in the University drive cycle, the HET consumes 36% (cold) to 48% (warm) less energy than the diesel truck with a 10% payload. Temperature plays an important role in energy consumption for both trucks, and the HET is more sensitive to temperature change than the diesel truck. The impact of temperature is more prominent in the University drive cycle than in the Freeway drive cycle. In cold temperatures (- 20 C to -10 C), the energy consumption for the diesel truck increases by 1% to 17% compared to the energy consumption at 20 C for the University drive cycle and 3% to 13% in the Freeway drive cycle. For the HET the increases with temperature change are greater. In cold temperatures, the HET energy consumption increases by 3% to 38% (University drive cycle) and 4% to 20% (Freeway drive cycle). Warm temperatures (30 C and 40 C) affect the diesel truck more than the HET. In the University drive cycle, the energy consumption of the diesel truck increases by 5% to 12% compared to the energy consumption at 20 C, and the HET consumes 0% to 6% more energy. In the Freeway drive cycle, the diesel truck consumes 3% to 5% more energy, but the HET consumes 1% to 5% less energy in the warm temperature than at 20 C.

54 43 Figure 3-3. Energy consumption comparison between the diesel truck and the HET on the University drive cycle Figure 3-4 Energy consumption comparison between the diesel truck and the HET on the Freeway drive cycle Life cycle GHG emissions The breakdown of life cycle GHG emissions (in gco2e/kg-km), by life cycle phase, payload and drive cycle is shown in Figure 3-5. The HET has slightly, about 8%, higher GHG emissions in the vehicle cycle than the diesel truck, due to the energy intensive nature of battery production. However, the HET is more efficient during operation and thus results in less GHG emissions in both the vehicle operation cycle and the fuel cycle than the diesel truck. Overall the

55 44 HET reduces life cycle GHG emissions by up to 42% and 35% in the University and the Freeway drive cycles respectively. There is one exception, however, that the HET results in 7% more GHG emissions than the diesel truck in the Freeway drive cycle at 100% payload. This may be due to the fact that the diesel engine is more efficient at higher payload and on the Freeway drive cycle. Also, the regenerative braking has less benefit during freeway driving conditions and thus results in less GHG emissions advantage. Figure 3-5. Life cycle GHG emissions at 20 C for diesel truck and the HET by payload and life cycle phases. The operating temperature also has an impact on the life cycle GHG emissions for both types of truck. Cold temperatures (-20 C to 10 C) have more impact on the HET life cycle GHG emissions than warm temperatures. For instance, the GHG emissions for the HET in cold temperatures increase by up to 33% and 17% higher than at 20 C in the University and Freeway drive cycles, respectively. Warm temperatures (30 C and 40 C), however, have less impact on life cycle GHG emissions. Lifetime GHG emissions increase by up to 5% for the University drive cycle but decrease by up to 5% for the Freeway drive cycle, compared with at 20 C. For the diesel truck, GHG emissions increase by up to 11% and 19% in cold temperatures in the

56 45 University and Freeway drive cycles. Warm temperatures have less impact on GHG emissions, which increase by up to 11% and 5% in the University and Freeway drive cycles, respectively. Figure 3-6 and Figure 3-7 show that, compared to the diesel truck, the HET reduces life cycle GHG emissions more in warmer temperature than in colder temperatures. For instance, for the University drive cycle at 10% payload, the HET reduces emissions by 30% and 45% in -20 C and 40 C, respectively. There are similar upward trends of life cycle GHG emissions reduction in the Freeway drive cycle, but lesser emissions reduction. Figure 3-6. Life cycle GHG emissions comparison between the diesel truck and the HET in the University drive cycle Figure 3-7. Life cycle GHG emissions comparison between the diesel truck and the HET in the Freeway drive cycle

57 Total cost of ownership The purchase cost accounts for 45% and 58% of the lifetime cost of ownership for the diesel truck and the HET, respectively (Figure 3-8). The HET has lower fuel cost due to a slightly more efficient powertrain. The overall lifetime total cost of ownership for the HET is 12% higher than that of the diesel truck. The lower fuel cost for the HET does not offset the higher purchase cost. Figure 3-8. Total Cost of Ownership for the diesel truck and the HET based on base value Figure 3-9 shows the sensitivity analysis results for the difference in lifetime total cost of ownership between the HET and the diesel truck. In most cases, except for when the discount rate is low, the HET has a higher lifetime cost than the diesel truck. In the base scenario, the lifetime cost of the HET is 12% more than that of the diesel truck. The HET can range from 5% less to 35% more costly than the diesel truck. The cost difference between the two trucks is most sensitive to discount rate, purchase cost and energy consumption.

58 47 Figure 3-9. Sensitivity analysis results for the lifetime total cost of ownership comparison between the HET and the diesel truck Figure 3-10 shows that in the base case, the cost of GHG emissions abatement for the HET is $406/tonne-CO2e. Since in the sensitivity analysis the HET has lower lifecycle GHG emissions than the diesel truck (the 100% payload in the Freeway drive cycle is not included in the sensitivity analysis), a negative value of cost of GHG abatement indicates that the HET not only reduces GHG emissions but also has a lower ownership cost. The cost of GHG abatement is found to be positive except when the discount rate is low. Ontario will soon implement a capand-trade system (Canadian Press, 2015). If carbon is priced at $30/tonne (Beaty et al., 2014) and the freight and logistics industry is subject to this regulation, then the HET may not be financially attractive because of the high cost premium and high cost of GHG emissions abatement.

59 48 Figure Sensitivity analysis results for the lifetime total cost of GHG emissions abatement of the HET 3.5 Conclusion We estimate and compare energy consumption, life cycle GHG emissions and lifetime costs of a diesel truck and a HET. Our results indicate that the HET consumes less energy than the diesel truck and, with the exception of the Freeway drive cycle with 100% payload, has lower life cycle GHG emissions. Operating temperature has more impact on energy consumption and GHG emissions of the HET than the diesel truck, and the impact is more prominent for the University than for the Freeway drive cycle. The cost of GHG abatement using the HET is high due to its higher lifetime cost of ownership, which is due to the higher purchase cost of the HET. It is important to note that the analysis focused on two example trucks and drive cycles reflective of conditions in Toronto and specific results should not be generalized to other trucks/conditions. Future work will examine a broader set of truck models and operating conditions. In spite of this limitation, there are several insights for policy that can be taken from the study. First, in 2011, the U.S. government released the first fuel economy standard for mediumand heavy-duty trucks. The standard (Phase 1) required a 10% fuel economy improvement for all new (Model Year ) vocational vehicles (delivery trucks, buses, garbage truck) by 2018 (NHTSA and EPA, 2010). Canada has not yet implemented a fuel efficiency requirement. Since Canada has a long history of harmonizing policies and standards with U.S, there is potential that Canada may adopt the U.S. standard to improve the energy efficiency of the medium-duty truck sector. Given the results in this analysis, the HET may not have GHG emissions or ownership cost benefits under all driving conditions when compared with a diesel truck. Policy makers should be aware of these results when considering HETs as an alternative to reduce energy use and GHG emissions from the commercial vehicle sector. Second, policy makers can play a role in informing fleet owners of potential benefits and risks of adopting HETs into their fleets. Based on our analysis, the GHG emissions and costs vary with driving conditions, such as payload, drive cycle and temperature. Under certain circumstances, operation of a HET could result in higher ownership costs and/or lifecycle GHG

60 49 emissions. An HET may be a good alternative to a diesel truck for intra-urban travel conditions or last mile delivery; while a diesel truck may be preferred for inter-urban travel, freeway driving conditions. Third, while there is significant life cycle GHG reduction potential for the HET, the cost of GHG abatement is high. There are currently no government incentives for the purchase of hybrid-electric medium-duty trucks in Ontario. The government could consider incentives for fleet owners given the high ownership costs and costs of GHG abatement of HETs.

61 50 4 Chapter 4. Life Cycle GHG Emissions and Lifetime Costs of Medium-duty Diesel and Compressed Natural Gas Truck. A Case Study for Toronto 4.1 Introduction Life cycle assessment (LCA) for transportation fuels and vehicles, also known as well-towheel analysis, is a common technique to evaluate life cycle greenhouse gas (GHG) emissions for trucks. The energy consumption and GHG emissions of medium-duty trucks are highly dependent on vehicle characteristics, such as truck configuration, payload and driving conditions (Tong et al., 2015). Few of the previous LCAs of compressed natural gas (CNG) trucks comprehensively capture the impacts of both weight and drive cycle on fuel consumption and life cycle GHG emissions. For example, the fuel consumption estimates in TIAX LLC (2008) and Kliucininkas et al. (2012), two LCAs of CNG trucks, do not reflect the impacts of driving conditions. The objective of this chapter is to estimate the energy consumption, life cycle GHG emissions, and the lifetime costs for a diesel truck and a CNG truck based on real-world medium-duty vehicle drive cycles in Toronto. 4.2 Methods and data We rely on Autonomie, developed by Argonne National Laboratory, to simulate fuel consumption for both the diesel and CNG trucks (Argonne National Laboratory, 2016). Vehicle specifications for the two trucks are listed in Table 4-1. The diesel truck model is similar to a 2009 Ford F-650 (Ford, 2009). The CNG truck model has the same truck body but with a dedicated CNG engine, JohnDeere PowerTech. The CNG truck costs 19% to 30% more than the diesel truck (Natural Petroleum Council, 2012). Table 4-1 Diesel Truck and CNG truck characteristics Model Diesel Truck CNG truck Chassis body mass (kg) Max payload (kg)

62 51 Cont d Gross vehicle weight rating (kg) Engine Caterpillar 3126E JohnDeere Engine peak power (kw) Note: It should be noted that both diesel and CNG engines are replaced by newer models and no longer being used for new trucks. However, these are the only engines available in Autonomie that suits medium-duty trucks. We adopt two simulated real world drive cycles from Amirjamsidi & Roorda (2014) that reflect the traffic demand estimation of 1518 medium-duty trucks in Toronto in the AM peak hours (8:00-9:00): the University drive cycle and the Freeway drive cycle. Figure 4-1. University and Freeway Drive Cycles for Medium-duty Vehicles in Toronto. (Prepared based on data in Amirjamsidi & Roorda, 2014) Scope of the LCA We use GHGenius (version 4.03a), a publicly available tool often used in Canada for wellto-wheel analyses of fuels and vehicles (S&T Consultants, 2013). Three phases of the analysis are included: 1) fuel cycle; 2) vehicle operation; and 3) vehicle cycle (vehicle production and transportation). We focus on five GHGs: Carbon dioxide (CO2), Methane (CH4), Nitrous oxide (N2O), Chlorofluorocarbon-12 (CFC-12), and Tetrafluoroethane (HFC-134a), which are converted to CO2 equivalent emissions using the IPCC year Global Warming Potentials (IPCC, 2007). The functional unit for the LCA is gco2e per km travelled per cargo kilogram (gco2e/kg-km). Payloads of 10%, 50% and 100% are studied to examine the impact of weight on energy consumption. We assume the trucks have the same material composition for the powertrain, except for the fuel tank, as shown in Table 4-2.

63 52 Table 4-2 Material breakdown by weight percentage for fuel storage system (S&T Consultants, 2013) Diesel Fuel tank (%) CNG fuel tank (%) Virgin plain carbon steel 42 0 Virgin high strength steel 0 10 Virgin stainless steel 40 5 Recycled plain carbon steel 18 0 Advanced composites 0 40 Other plastics 0 20 Virgin aluminum 0 18 Virgin copper 0 8 Total Total cost of ownership The annualized total cost of ownership, expressed in 2014 U.S. dollars (USD), is a function of annual vehicle kilometers travelled (VKT), fuel price, maintenance cost, energy consumption, and discount rate. The ranges of these parameters are presented in Table 4-3. Sensitivity analysis are performed based on the lower bound, best estimate and higher bound of each of the parameters. The parameters are varied one at a time with all others remaining at their best estimate value, and the implications on total cost are examined.

64 53 Table 4-3 Best estimates and ranges for parameters for sensitivity analysis Parameter Lower Best Estimate Higher Source CNG truck purchase price 75,177 78,967 82,126 Annual vehicle kilometers travelled 8,000 16,000 32,000 Lifetime vehicle kilometers travelled 160, , ,000 Toronto diesel fuel price (2014 USD/Liter) The CNG truck price is assumed to be 19%, 25%, and 30% higher than that of the diesel truck ($63,174 in 2014 USD, obtained from Lee et al., (2013)) (American Petroleum Council, 2012) Lower: average annual VKT of Prince Edward Island (province with lowest annual VKT);Best: Ontario;,Higher: Nova Scotia (province with highest VKT), based on Canadian Vehicle Survey 2009 (Natural Resources Canada, 2010) Best estimate from Lee et al. (2013); upper value is from (Antich) 2008; lower is 80,000 km less than base case (the interval between the base case and higher) Assumption: 1 CAD=0.9 USD (Bank of Canada, 2016); the price range is obtained from Natural Resources Canada (2014); the diesel price projection is obtained from Annual Energy Outlook, 2011 (US Energy Information Administration, 2015). I assume Toronto diesel price has same rate of change as that in the Annual Energy Outlook, 2011 CNG fuel price (2014 USD/ Liter diesel equivalent) U.S. DOE (2015) Diesel Fuel consumption (Liter/100 km) CNG Fuel Consumption (Liter diesel equivalent/100km) Discount Rate 3% 7% 10% Autonomie simulation results: Lower bound is 0% payload on Freeway driving cycle at 20 C; Best estimate: average of 50% payload on university and freeway drive cycles at 20 C; Higher bound: 100% payload on University drive cycle at 20 C Autonomie simulation results: lower is 0% payload on Freeway driving cycle at 20 C; Base case: average of 50% payload on university and freeway drive cycles at 20 C; High: 100% payload on University drive cycle at 20 C Obtained from U.S. DOT guidance on costbenefit analysis, Circulars A-4 and A-94 (U.S. Department of Transportation, 2015)

65 Result and discussion Energy consumption The energy consumption, obtained from the Autonomie simulations, for the diesel truck and the CNG truck is expressed in MJ/kg-km (Figure 4-2). Figure 4-2 Energy consumption of diesel and CNG trucks under different payloads from Autonomie simulation As expected, both trucks consume less energy per tonne-km when payload weights are higher, thus they are more efficient if they carry more payload. This is because although the total energy consumption (MJ/km) increases as the truck carries a larger load, the denominator (tonnes) is larger, thus the load weighted consumption is lower. Both the diesel truck and CNG truck are more energy intensive in the University drive cycle than in the Freeway drive cycle due to the stops, starts and idling that occurs during the University (city) driving cycle. The CNG truck consumes more energy than the diesel truck (26 49% higher in the University and 18 37% higher in the Freeway drive cycle) Life cycle GHG emissions For both the diesel truck and the CNG truck, the GHG emissions per tonne-km are higher at low payload and lower at high payload (Figure 4-3). This is because the energy consumption per tonne-km for both trucks at high payload is lower than at low payload, as indicated in Figure 4-2. The GHG emissions of the CNG truck are higher or lower than those of the diesel truck, depending on the drive cycle and payload. For example, when 10% loaded, the CNG truck has 11% and 33% higher life cycle emissions than the diesel truck in the University and the Freeway drive cycle respectively; but the CNG truck emits 3 to 10% less life cycle GHG emissions than the diesel when 50% and 100% loaded.

66 55 Figure 4-3 Life cycle GHG emissions of the diesel and CNG trucks Total cost of ownership Figure 4-4 shows the total cost of ownership of the diesel truck and CNG trucks based on the best estimate scenario. The fuel cost accounts for nearly half (45%) of the life cycle total cost of ownership of the diesel truck. The fuel cost for the CNG truck is lower, specifically 31% of the lifetime total cost of ownership. Overall, the lifetime total cost of ownership of the CNG truck is 11% higher than that of the diesel truck. The sensitivity analysis (Figure 4-5) shows the lifetime total cost of ownership for the CNG truck varies from 2% to 27% more than that of the diesel truck. The difference in total cost of ownership is most sensitive to the CNG fuel price, and to a lesser extent, the diesel fuel price and energy consumption. The difference in total cost of ownership is least sensitive to the lifetime vehicle kilometers travelled.

67 56 Figure 4-4 Lifetime total cost of ownership of diesel and CNG trucks Figure 4-5 Sensitivity analysis results for difference in lifetime total cost of ownership between CNG and diesel trucks 4.4 Conclusions Energy consumption, life cycle GHG emissions and lifetime total cost of ownership for the diesel and CNG trucks are estimated. This study shows the CNG truck to always consume more energy on a MJ/kg-km basis than the diesel truck, irrespective of drive cycle or payload. The CNG truck is found to have higher or lower life cycle GHG emissions than the diesel truck depending on the payload and drive cycle. The CNG truck has a higher lifetime total cost of

68 57 ownership in all scenarios. The lifetime total cost of ownership is most sensitive to CNG price, diesel price and the energy consumption. The results of this work suggest that the CNG truck may not be the best alternative to reduce GHG emissions from fleets. The GHG emissions reduction only occurs under lower payload; the emissions for both trucks are similar at 50% and 100% payload. However, these results are only for the modeled trucks and should not be generalized to other diesel or CNG trucks. Additionally, higher quality data and additional investigation into the parameters studied, as well as other environmental impacts such as those associated with air pollutant emissions should be carried out in future work. With the absence of incentives, we found that the CNG truck has a higher cost of ownership compared to the diesel truck. To promote the deployment of alternative fuel vehicles, the U.S. government provides incentive programs, such as tax credits or rebates, both at the federal and at the state levels (DOE, 2015). This study suggests that incentive programs are likely necessary for the alternative fuel vehicle industry s development; and if properly designed, will help the industry to survive in today s highly-oil-dependent economy (Brown & Breckenridge, 2001). The natural gas vehicle market in Canada is relatively small compared to in other counties; this may be due to a lack of government support and weak refueling infrastructure (Heffernan & Dawson, 2010). A webinar hosted by IMW Inc. suggested that the market potential for CNG commercial vehicles is significant because of the high oil price (at the time of the webinar) and associated quick payback period (Milner, 2014). While natural gas prices continue to be low, oil prices have fallen drastically since 2014 and so the financial case for CNG may be more challenging. However, Canada has a history of harmonizing government policy with the U.S., and along with the potential growth in the Canadian CNG truck market as Canada has some leading companies producing natural gas and natural gas engines, a comprehensive analysis of CNG trucks is needed to assess the environmental and economic benefits for informing near and long-term decision making.

69 58 5 Chapter 5 Conclusion In this chapter, we first combine and compare all findings of the diesel truck, CNG truck, the HET and the BET presented in previous chapters, in terms of energy consumption, life cycle GHG emissions, lifetime cost of ownership, and the cost of GHG abatement. We then discuss the policy implications, and finally end with limitations and future research. 5.1 Energy consumption The CNG truck has higher energy consumption than the diesel truck in both drive cycles and payload (Figure 5-1). Compared to the diesel truck, the energy consumption for the CNG truck is 26% to 49% higher on the University drive cycle, and 18% to 37% higher on the Freeway drive cycle. Both the HET and the BET consume less energy than the diesel truck across the drive cycles, payloads and temperatures. The BET has the lowest energy consumption amongst all studied trucks (Figures 5-2 and 5-3). The temperature impact on the CNG truck energy consumption is not considered because of Autonomie s limitation it does not have a thermal model for a CNG truck. Thus, we report the results of the CNG truck separately. Compared to the HET, the BET reduces energy consumption more in both the University and Freeway drive cycles. On the University drive cycle, the BET consumes 68% to 85% less energy than the diesel truck, while the HET consumes 7% to 48% less. On the Freeway drive cycle, the energy reduction benefit is a little less than on the University drive cycle, the BET consumes 60% to 80%, and the HET consumes 1% to 21%. This is because the benefit of regenerative braking, which lowers fuel consumption, is reduced in both the BET and the HET on the Freeway drive cycle.

70 59 Figure 5-1 Comparison of energy consumption of the CNG truck and the diesel truck Note: Temperature effects are not included due to a limitation in Autonomie. Figure 5-2 Comparison of energy consumption of the HET and the BET on the University drive cycle Figure 5-3 Comparison of energy consumption of the HET and the BET on the Freeway drive cycle

71 60 Figure 5-2 and Figure 5-3 also show that the HET and the BET reduce energy consumption more in warm temperatures (40 C) than in cold temperatures (-20 C). This is because the impact of temperature on energy consumption is different for different powertrains. As shown in Figure 5-4, the impact of temperature is most prominent for the BET, less for the HET, and least for the diesel truck on both drive cycles. For example, the energy consumption for the BET increases by 92% at -20 C compared to at 20 C, while that for the HET and the diesel truck increases by only 38% and 17%. As discussed in previous chapters, the energy consumption increases in extreme temperatures due to two main reasons: 1. Change in efficiency of the powertrain, and, 2. Increase in power required for cabin thermal control. In cold temperatures, the heat generated from the diesel engine can be collected to warm the cabin; while an electric heater is required for the HET. In warm temperatures, air-conditioning is required to maintain the cabin temperature at 20 C. As Lohse-Busch et al. (2013) pointed out the impact of climate control (heater or air-conditioning unit) on energy consumption is greater when the powertrain is more efficient. Figure 5-4 Temperature impact on energy consumption for diesel truck, HET and the BET on the University and Freeway drive cycles, at 10% payload 5.2 Life cycle GHG emissions Figure 5-5 shows an example of life cycle GHG emission, broken down by life cycle phase, for the different trucks, at 20 C and 10% payload on the University drive cycle. For the diesel truck, HET and the CNG truck, the vehicle operation cycle accounts for the largest portion of GHG emissions among all life cycle phases. For the BET (assuming the 2014 average Ontario electricity grid), the vehicle cycle accounts for the most emissions in its life cycle. Overall,

72 61 compared to the diesel truck, the HET and the BET reduce GHG emissions by 42% and 84%, while the CNG truck increases emissions by 11% at this particular driving condition. Figure 5-5 Life cycle GHG emissions breakdown by lifecycle phase assuming 10% payload, 20 C and the University drive cycle. Figure 5-6 compares the life cycle GHG emissions of the diesel and CNG trucks. The CNG truck has higher GHG emissions than the diesel truck at 10% payload; at 50% and 100% payload, the CNG truck reduces GHG emissions by around 10% on the University drive cycle and around 3% on the Freeway drive cycle. Figure 5-6 Life cycle GHG emissions for the diesel truck and CNG truck

73 62 Figure 5-7 and 5-8 compare the life cycle GHG emissions of the diesel truck, HET and the BET. We found that the both the HET and BET reduce life cycle GHG emissions in most cases except for at 100% payload on the Freeway drive cycle, where the GHG emissions from the HET and the BET increase by 3% to 7% and 16% to 24%. In other cases, the BET reduces GHG emissions more than the HET under the same driving conditions. The GHG emissions benefit of the HET and the BET are reduced as the payload increases. For example, on the University drive cycle, the BET emits 76% to 80% less GHG emissions at 10% payload, 52% to 59% at 50% payload, and 31% to 38% at 100% payload. The same trend is true for the HET. Figure 5-7 Life cycle GHG emissions for the diesel truck, HET and the BET on the University drive cycle

74 63 Figure 5-8 Life cycle GHG emissions for the diesel truck, HET and the BET on the Freeway drive cycle 5.3 Lifetime total cost of ownership Figure 5-9 compares the lifetime total cost of ownership for all trucks in the base case scenario (best estimates). We found that the lifetime cost of ownership for a diesel truck is about $140,000, and that for the CNG and the HET is 11% and 14% higher than the diesel truck. The BET on the other hand, has similar lifetime cost of ownership as the diesel, despite that it has the highest purchase cost of all trucks. Figure 5-9 Total cost of ownership for the diesel, CNG, HET and BET for the base case Based on our sensitivity analysis for the lifetime cost, we found that the CNG truck costs from 2% to 27% more than the diesel truck, and the cost difference is most sensitive to CNG fuel price, diesel fuel price, and energy consumption (Figure 2-5); the HET costs from 3% less to 40% more than the diesel truck, and is most sensitive to discount rate, HET purchase cost, and energy consumption (Figure 3-9); and the BET costs from 12% less to 26% more than the diesel truck, and is most sensitive to energy consumption, annual vehicle kilometers travelled, battery costs, and electric vehicle supply equipment costs (Figure 4-11). 5.4 Cost of GHG emissions abatement We compare the base case cost of GHG emissions abatement for each alternative fuel truck. The cost of GHG abatement might be useful for fleet owners to guide their decisions about