Scenarios for transitioning cars from ICEV to BEVs and PHEVs using household level GPS travel data

Highlights

• PEV adoption can reduce greenhouse gas emission by up to 65%

• Optimal PEV model assignment at household level is significantly different from vehicle level.

• BEVs with 120+ miles range can benefit the most travel demand.

• The achievable GHG emission reduction decreases by 2/3 if PHEV drivers never charge.

Abstract

There has been a continuous effort to quantify the expected greenhouse gas reductions from plug- in electric vehicles (PEVs). Most existing studies are based on standardized driving cycles to compare the performance of PEVs and conventional internal combustion engine (ICE) vehicles. This study is based on assigning the most efficient plug-in hybrid or full electric vehicle to a household based on the household travel behavior as collected by GPS tracking data of thousands of vehicles’ real trips in California, and it found that PEV adoption can reduce greenhouse gas emission by up to 65%. We further explore the market share difference in assigning optimal vehicle models by considering vehicle-based and household-based travel demand, and we found that, given currently available technology, in many cases adoption of Plug-in Hybrid Electric vehicles (PHEVs) will not necessarily result in significantly higher greenhouse gas (GHG) emissions (2.8% higher) compared to battery electric vehicles (BEVs), while PHEVs have advantages to satisfy household’s long-distance travel demand. However, if PHEV drivers never charge their vehicles, the achievable GHG emission reduction will be decreased by nearly 2/3.

Introduction

Plug-in electric vehicles (PEVs), including plug-in hybrid electric vehicles (PHEVs) and battery electric vehicles (BEVs), have an important role in reducing transportation-related greenhouse gas (GHG) emissions (Duvall, et al., 2007, Yang, 2009, Stephan and Sullivan, 2008, Kromer et al., 2010). It is a complex endeavor to quantify the benefit of PEV adoption in GHG emission reductions, especially for PHEVs, as they can be fueled by both gasoline and electricity. The amount of GHGs emitted per year of traveling is a function of the vehicle characteristics and of driving and charging patterns. For some driving patterns, a short-range (80 or 120 mile in our analysis) BEV will satisfy both travel needs and minimize GHGs, but for other travel needs either a long-range (200 mile) BEV or a PHEV will produce better results. Changes in daily driving routines from one day to another or between weekdays and weekends may change the household GHG emissions for any given vehicle. This paper compares the best BEV or PHEV range (based on battery size) vehicle that will produce the lowest amount of GHGs for California households based on their GPS travel patterns. We use household level data to optimize GHGs while controlling for the long and short-range vehicle mix.

A life cycle assessment (LCA) of vehicle GHG emissions usually consists of two parts: one looks at the life cycle of a vehicle (Moon et al., 2006, Lewis et al., 2014) including vehicle production (Samaras and Meisterling, 2008), vehicle operation, and after-life disposal/recycling (Ma, 2012); the other looks at the life cycle of fuel (Wang et al., 2005, Elgowainy, 2009) including well-to-tank fuel production and transmission and tank-to-wheel fuel consumption, which is accounted for in both the vehicle operation part of the vehicle life cycle and the fuel life cycle. The tank-to-wheel GHG emissions are also known as on-road emissions. There are many factors throughout the phases of a PEV’s life cycle that can influence the on-road emissions of the vehicle. For example, in the vehicle production phase, the powertrain design (Bradley and Frank, 2009, Letendre and Watts, 2009), battery capacity (Shiau, 2009), and energy-management strategies (Gonder and Markel, 2007, Wirasingha and Emadi, 2011) can all significantly influence the vehicle fuel consumption during operation, and thus influence the on-road GHG emissions. There are studies to model the performance of each power system component in order to understand the energy consumption of PHEVs (Lee et al., 2002). Although the on-road emissions of PEVs in all-electric mode is zero, the life-cycle emissions are determined by the upstream GHG emissions from electricity generation (Tamayao, 2015, Onat et al., 2015), which is subject to many uncertainties including the fluctuations in demand on the grid throughout the day and the power source used to satisfy the marginal electricity demand (Axsen, 2011, Zivin et al., 2014).

Beyond the vehicle design and the upstream GHG emissions from electricity generation, driving behavior is crucial for estimating the on-road GHG emissions of PHEVs as the engine may be on or off in charge-depleting mode (Bradley and Frank, 2009). There are generally two types of approaches to simulate driving behavior. One option is to use standard driving cycles to characterize the cycles (Samaras and Meisterling, 2008, Raykin et al., 2012, Silva et al., 2009) and to create various driving cycles to represent different types of driving behavior (Dynamometer, 2017, Yuan, 2015). Another common approach is to sample real-world driving behavior by either tracking vehicle performance through on-board diagnostics systems (Nicholas et al., 2016), quantifying variation of real-world driving behavior (Lee et al., 2011, Wang et al., 2015, He, 2016), or simulating vehicle performance based on speed trace data (Silva et al., 2009, Hamza and Laberteaux, 2016). Road grade (Wood, et al., 2014) is another factor that impacts a vehicle’s fuel economy. Charging behavior (Shiau, 2009, Neubauer et al., 2013, Tal, 2014) and weather conditions (Yuksel and Michalek, 2015) will also influence the amount of available electricity in the battery in real-world driving activities, therefore impacting the on-road GHG emissions.

There are many vehicle energy consumption models, and they all have their strengths and weaknesses in terms of the capacity to consider vehicle design, road grade, charging activities, etc. (Mahmud and Town, 2016). For example, the Advanced Vehicle Simulator (ADVISOR) is a popular vehicle energy analysis tool which takes a hybrid backward-forward approach to simulate the component performance (Markel, 2002). The Vehicle Technologies Office of the U.S. Department of Energy (U.S. DOE. VEHICLE TECHNOLOGIES OFFICE: MODELING AND SIMULATION., 2017) supported the development of Autonomie (Kim et al., 2010) and FASTSim (described in detail in Section 2) (Brooker, et al., 2015). There are also custom PEV energy consumption models, which are not publicly available, such as Muratori et al to analyze energy consumption of different vehicle models (Muratori, 2013), Doucette and McCulloch to compute CO₂ emissions from PEVs and ICE vehicles (Doucette and McCulloch, 2011), and Khayyer et al to analyze energy management strategies’ impact on energy consumption (Khayyer, 2012).

PEV market share is another essential factor in estimating the total GHG emissions reductions that can be achieved by implementing PEVs. Vehicle ownership is usually a household decision, and there are many studies that explored how household vehicle ownership is influenced by various factors including household size, age, education, income, number of workers, residential location and condition, vehicle utility, operating costs, etc. (Manski and Sherman, 1980, Mannering and Winston, 1985, Mannering et al., 2002). With the introduction of PEVs, electricity becomes a more significant part of the vehicle’s fuel so the electricity price plays an important role in the adoption of PEVs (Kumar and Alok, 2020 Apr, Bhat et al., 2009) and the improvement in fuel economy is an important motivation for PEV adoption (Kurani and Turrentine, 2004, Kurani et al., 2008). On the other hand, the limited battery range available in current technology is a primary barrier to the adoption of PEVs, especially BEVs (Lee et al., 2020 Feb, Tal and Nicholas, 2013).

This study explores the potential of reducing GHG emissions by wide-scale adoption of PEVs including a mix of BEVs and PHEVs in the household level. Recent studies explore the utility factor and GHG emissions of single PEV households (Nicholas et al., 2013, Tal and Nicholas, 2016) but not the electrification of all vehicles in the household based on travel needs and utilization restrictions. In this study, fuel consumption is simulated using FASTSim based on speed trace data from real trips rather than standard driving cycles or custom vehicle energy models which are more common approaches used in previous studies.

The rest of the paper is organized as follows: the second section introduces FASTSim and the configuration of vehicle models which are used for fuel consumption simulation; the third section introduces the speed trace data along with weights to produce representative statistics; and the fourth section shows the GHG emissions reduction in different PEV adoption scenarios with and without considering household vehicle ownership limitations.