Electric Vehicles and Automotive Brand: How Has Tesla Shaped Consumer Perceptions?
Zoe Long, Simon Fraser UniversityShow Abstract
Jonn Axsen, Simon Fraser University
Inger Miller, Simon Fraser University
Christine Kormos, Simon Fraser University
Among the broad literature on consumer research for battery electric vehicles (BEVs), little emphasis is placed on the role of brand perceptions. Specifically, no studies have explored if and how Tesla may have shaped consumer perceptions of BEVs. Drawing from marketing literature, consumers might prefer the automotive brands that they are used to, due to brand loyalty. Alternatively, they might prefer Tesla for being more innovative, affording Tesla a pioneer advantage. We study the role of brand using survey data collected from a representative sample of 2,123 Canadian new vehicle-buyers in 2017. Respondents most frequently associate BEVs with Tesla (27%), Toyota (27%), Chevrolet (26%), and Nissan (13%). More than two-thirds of respondents are familiar with Tesla, and 40% select Tesla as a brand representing the “future of BEVs”. In contrast, when asked which brand they’d prefer to buy a BEV from, responses are more evenly spread across a variety of brands. Of the Tesla-familiar respondents, most indicate that Tesla has influenced them to perceive BEVs as more innovative, stylish, and environmentally beneficial. Further, we observe that respondents more strongly associate Tesla with several positive images (e.g., being stylish, high-performance, and attractive), relative to BEVs more generally, and that such perceptions are statistically associated with stated intention to purchase a BEV. These exploratory findings indicate that Tesla has contributed to shaping consumer perceptions of BEVs, where some consumers might be attracted to the company’s pioneering innovation, while others might be loyal to a conventional automaker.
Identifying Locations for Residential On-Street Electric Vehicle Charging Infrastructure: A Practical Methodology for Local Government Authorities
Matt Grote, University of SouthamptonShow Abstract
John Preston, University of Southampton
Tom Cherrett, University of Southampton
Neil Tuck, Southampton City Council, Civic Centre, Southampton, SO14 7LY, United Kingdom, Tel: +44-23-8083-3409; Fax: not available; Email: Neil.Tuck@southampton.gov.uk
Depending on electricity generation method, mass-market penetration of electric vehicles has the potential to dramatically reduce emissions of greenhouse gases and air pollutants, and to reduce dependency on fossil fuels. This paper presents a novel methodology for Local Government Authorities (LGAs) to identify suitable locations for the initial provision of residential on-street Plug-in Electric Vehicle (PEV) charging infrastructure in urban areas. Provision of such infrastructure removes a barrier to PEV uptake. The methodology is practical for LGAs to use within limited resources because it is based on simple Geographic Information System (GIS) analysis of routinely available census and parking data to identify the spatial overlaps between areas where residents are most likely to be PEV users and areas with a high reliance on residential on-street parking. The methodology has been implemented in practice by a LGA to determine a charging infrastructure installation strategy for the city of Southampton, UK, where 128 streets (out of 1,924 in total) were recommended as suitable locations. The list of recommended streets was reviewed by participants possessing a detailed familiarity with Southampton’s streets during a workshop and generally assessed as a sensible set of locations for initial installation of residential on-street charge points.
Are Consumers Learning About Plug-In Vehicles?: Comparing Awareness Among Canadian New Car Buyers in 2013 and 2017
Zoe Long, Simon Fraser UniversityShow Abstract
Jonn Axsen, Simon Fraser University
Christine Kormos, Simon Fraser University
Suzanne Goldberg, Simon Fraser University
Plug-in electric vehicles (PEVs) are increasingly receiving attention for their potential to reduce greenhouse gas emissions. However, consumer adoption of plug-in hybrid (PHEV) and battery electric vehicles (BEV) remains limited globally. While several factors inhibit their adoption, here we focus on consumer awareness. Specifically, we analyze how Canadian new vehicle-buyers’ awareness, familiarity, and experience with PEVs have changed over a four-year period. We compare survey responses from two cross-sectional, representative samples of new vehicle-buyers collected in Canada in 2013 (n = 2,922) and in 2017 (n = 1,808). In the 2017 sample, the vast majority of respondents have “heard of” the Nissan Leaf, Chevrolet Volt, or Tesla Model S. However, less than one-quarter of 2017 respondents report being familiar with PHEVs or BEVs in general, or with the Chevrolet Volt or Nissan Leaf in particular – similar to or below awareness levels in the 2013 sample. In terms of understanding, only about one-quarter of 2017 respondents know how the Nissan Leaf or the Chevrolet Volt is refueled, which is slightly lower than the 2013 proportion. Direct experience with PEVs is also low, with only a few percent having driven in or spoken with an owner of the Volt, Leaf or Model S – which also has changed little since 2013. In contrast, awareness of public chargers has more than doubled, with over 50% of 2017 respondents reporting have seen at least one charger. Overall, in contrast to expectations, PEV familiarity remains low and has generally not increased in recent years.
The Long Way to an Electric Car: Observations and Interviews
Anne Faxér, RISE Research Institutes of Sweden ABShow Abstract
Ellen Olausson, RISE Research Institutes of Sweden AB
Tommy Fredrik Magnus Fransson, RISE Research Institutes of Sweden AB
Ana Magazinius, RISE Research Institutes of Sweden AB
Oscar Enerbäck, RISE Research Institutes of Sweden AB
Jens Hagman, KTH Royal Institute of Technology
Jenny Janhager Stier, KTH Royal Institute of Technology
The aim of this study is to investigate how salespersons approach Electric Vehicle (EV) sales. Mystery shoppers visited twelve car dealerships and expressed user needs that suited an EV. Mystery shoppers were instructed to act as customers that had not decided on what vehicle to buy and ask to see vehicles of a certain size. The results show that the initial discussion often anchored around the brand’s bestselling internal combustion engine vehicle (ICEVs), only two out of twelve salespersons initially offered an EV. The turning point that lead the conversations towards an EV in the rest of the cases was either knowing that a car could be purchased as a company car or the customer specifically asking for an EV. Even though the mystery shoppers talked about the monthly costs the discussion focused mainly on (high) purchasing price. Changing the sales process to offer EVs as a viable option and focusing on monthly costs could make EVs more attractive to more customers than the ones that had already decided to buy one and thus possibly increase EV sales. The incentives for car dealerships, as well as individual salespersons to promote EVs however need to be further explored.
An Efficiency-Based Approach to Biofuel Facility Location-Routing Network Design Under the Risk of Facility Disruptions
Jae-Dong Hong, South Carolina State UniversityShow Abstract
Judith L Mwakalonge, South Carolina State University
This paper proposes an innovative procedure of combining a goal programming (GP) model into Data Envelopment Analysis (DEA) approach for a biofuel facility location-routing network (BFLRN) design problem under the risk of facility disruptions. The BFLRN considered in this paper consists of four echelons; harvest sites, storage facilities, biorefineries, and blending facilities. The location-routing of these four echelons is the main decision variable. Under the risk of facility disruptions, we consider four goals as the major performance measures. To accommodate these four goals in one objective function, we first use goal programming (GP) approach as a tool. Solving the GP model for various values of weights, which simultaneously considers multiple performance measures, generates various alternative options. Each option represents a BFLRN scheme. Considering each scheme as a Decision-Making Unit (DMU), we apply DEA to find efficient schemes and also apply two DEA-related methods, the stratification DEA method and the cross-efficiency method (CEM), to find the most efficient one. A case study using the biomass feedstocks data for South Carolina is conducted to evaluate the proposed procedure. We observe that the proposed procedure performs well for identifying the efficient and robust BFLRN schemes, which would attract potential investors who are interested in investing in the biofuel/bioenergy industry.
Electric Vehicle Charging Station Locations: The Case of User Equilibrium and Elastic Demand
Yantao Huang, University of Texas, AustinShow Abstract
Kara Kockelman, University of Texas, Austin
Electric vehicles (EV) generally offer better air quality through lowered emissions, along with energy savings. The issue of long-duration battery charging makes charging-station design and placement key for EV. This work provides a genetic algorithm framework for profit-maximizing station placement and design details, with applications that reflect costs of installing, operating, and maintaining EV service equipment, including land acquisition costs. EV Charging stations (EVCS) are placed subject to a stochastic demand for charging stations under a user-equilibrium traffic assignment. Random utility theory is used to provide the station choice of EV users, considering the road congestion and on-site charging queues, which provide for travel time feedbacks to route and station choices. The travel assignment with elastic demand problem is formulated as a convex program and is solved using a modified Frank-Wolfe algorithm. Various realistic costs for power delivery and elastic demand (for driver sensitivities to travel times, wait times and charging costs) are assumed. Results show that EVCSs locate mostly in the city center and along the highway, but specific pattern varies according to different background settings. For the investigated Sioux Falls network, assuming 10% of EV users need charging among 1.1% EV adoption rate of the automobile, a fare of $6 for a 30-minutes charging session is not enough to make a profit. Three cords limit for stations could accommodate the EV demand with fewer stations. Based on sensitivity analysis, profit increases with longer time horizon, less time of a charging session, higher fare and greater cord limit.
Building a Theory of Heavy-Duty Alternative Fuel Vehicle Fleet Adoption Behaviors in California: An Initial Theoretical Framework
Youngeun Bae, University of California, IrvineShow Abstract
Suman Mitra, University of California, Irvine
Stephen Ritchie, University of California, Irvine
Revealing heavy-duty fleet operators’ attitudes, preferences, and behaviors towards alternative fuel vehicles (AFVs) is essential to understand the demand-side aspects of the heavy-duty AFV fleet sector. Such understanding can contribute to more facilitated diffusion of heavy-duty AFVs which can lead to the mitigation of climate change and improved local air quality. This study aims to build a theory of heavy-duty AFV fleet adoption behaviors with the focus on the California heavy-duty sector. To this end, a comprehensive literature review was conducted to obtain an overview on the factors found to influence AFV fleet adoptions at the organizational level, especially from fleet purchase decision-makers’ point of view. Based upon existing frameworks centering on organizational innovation adoption behaviors, an initial theoretical framework was developed and elaborated to understand AFV fleet adoption behaviors. The proposed framework consists of two levels of sub-frameworks: at the decision-making unit level and the individual (e.g., fleet drivers) acceptance level. As a next step, in-depth qualitative interviews are being employed with the goal of building a theory which has a capability to explain revealed behaviors of heavy-duty AFV fleet adoptions in California. The data collection and analysis of these interviews are in progress. As the ultimate outcome of this study, the finalized theoretical framework will contribute to a deeper understanding of heavy-duty AFV adoption behaviors and help elicit demand-side policy measures to foster the transition towards zero-emission transportation in California.
A Preliminary Analysis on the Commute and Non-Commute Travel Patterns and Fuel Economy for Electric Vehicles Based on Longitudial Travel Data
Yu Chen, Tongji UniversityShow Abstract
Hao Li, Tongji University
Xuqi Ning, Tongji University
The proportion of private electric vehicles (EV) is increasing substantially in recent years in China, mainly attributed to the incentive policies from government. However, the charging and travel behavior of EV users, the commute features, and preferences of electricity usage are still unknown due to the lack of the real world longitudinal data. This research aims to investigate the differences in travel pattern between EV commuters and non-commuters and to analyze the fuel economy of Plug-in electric vehicles (PHEV). On the basis of revealed preference (RP) longitudinal travel data for 400 PHEVs and 300 Battery Electric Vehicles (BEV), each being tracked for roughly one year, EV commuters and noncommuters are identified and home/work places of travelers are extracted, from which the commute features are discussed. The fuel economy of the researched PHEVs is also analyzed. The electricity and fuel consumption rates of each PHEV are estimated respectively, by taking advantages of OLS methods. Results show that the proportion of longer distance travel of noncommuters is larger than that of commuters. The Petroleum Displacement Factor (PDF) varies among different PHEVs for several reasons. This study also provides new insights into the charging availability. All the results are helpful for optimizing the planning of the charging infrastructures and gives advices on measures to improve the electrified travel distances.
Exploring Heterogeneous Electric Vehicle Charging Behavior: Mixed Usage of Charging Infrastructure
Jae Hyun Lee, University of California, DavisShow Abstract
Debapriya Chakraborty, University of California, Davis
Scott Hardman, University of California, Davis
Gil Tal, University of California, Davis
This paper examined charging behavior of 7,979 PEV owners in California, focusing on their mixed usage of level 1 (L1), level 2 (L2), and DC fast chargers at different locations. Unlike most of the current models and simulations used to characterize electric vehicle charging, we identified different day of week charging behavior among different types of PEV owners. We were also able to classify seven different groups of charging behaviors based on mixed usage of charging locations, and identify factors associated to the heterogeneous charging behavior. We found socio-demographics, vehicle characteristics, commute behavior, and workplace charging availability as significant factors driving charging behavior. Among these factors access to workplace charging was the most important one.
A Network-Load Interaction Instruction Abnormal Detection Method of Electric Vehicles Networks
Qianmu Li, Nanjing UniversityShow Abstract
Yinhai Wang, University of Washington
Ying Jiang, University of Washington
Jun Hou, Jiangsu Police Institute
Yong Qi, Jiangsu Intelligent Transportation systems Co.,Ltd
With the rapid construction of electric vehicle charging facilities and the development trend of interconnection, the information security of charging facilities is related to the operation security of network operators, the privacy of car owners, the safety of car charging, and the safety of public transportation. In the case of frequent interaction between the network and the load, multiple layers of operational information and control commands in the electric vehicle network are subject to eavesdropping, tampering and interruption during collection, transmission and triggering. This paper presents an instruction anomaly detection method for network load interaction of electric vehicles networks command anomalies. The main steps of the method include packet capture, packet depth analysis, protocol feature modeling based on Markov time-varying method, feature tracking based on LSTM deep learning, and protocol instruction level field extraction. This method provides a solution for the rapid detection of command level anomalies in electric vehicle networks. The main feature of the method is firstly to apply the Markov state transition diagram to achieve state transition and abnormal initial screening of protocol messages. Secondly, LSTM (Long Short Term Memory Networks) depth learning algorithm is improved to mine command-level anomaly features. The method is extensible. If other new protocols are adopted for electric vehicle networks in different countries or regions, the analysis of the new protocol can be completed in the same way, which has strong application value and prospect.
A Time Series Association State Analysis Method in Smart Internet of Electric Vehicle Charging Network Attack
Qianmu Li, Nanjing UniversityShow Abstract
Yinhai Wang, University of Washington
Ziyuan Pu, University of Washington
Zichen Zhang, Jiangsu Intelligent Transportation systems Co.,Ltd
Weibin Zhang, University of Washington
A robust, integrated and flexible charging network is essential for the growth and deployment of electric vehicles (EVs). The State Grid of China has developed a Smart Internet of Electric Vehicle Charging Network (SIEN). At present, there are three main ways to maliciously attack SIEN: distributed data tampering, distributed denial of service, and forged command attacks. This paper analyzes the characteristics of these three attacks and proposes a time series association state analysis method. The method firstly analyzes alarm logs of different locations, different levels and different types, and establishes the temporal association of scattered and isolated alarm information. Secondly, it tracks the transition trend of SIEN main station layer, channel layer and sub-station layer abnormal event status, and identifies the real attack behavior. This method not only provides a prediction of security risks, but more importantly, it can accurately analyze the trend of SIEN security risks. Compared with the ordinary Markov chain model, this method can better smooth the fluctuation of processing values, with higher real-time performance, stronger robustness and higher precision. This method has been applied to the State Grid of China.
Effects of Charging Infrastructure and Non-Electric Taxi Competition on Electric Taxi Adoption Incentives in New York City
Jaeyoung Jung, Ford Motor CompanyShow Abstract
Joseph Chow, New York University
With major investments in electric taxis emerging around the world, there is a need to better understand resource allocation trade-offs in subsidizing electric vehicle taxis (e-taxis) and investing in electric charging infrastructure. This is addressed using simulation experiments conducted in New York City: 2016 taxi pickups/drop-offs, a Manhattan road network (16,782 nodes, 23,337 links), and 212 charging stations specified by a 2013 study from the Taxi and Limousine Commission. The simulation is based on a platform used to evaluate taxi operations in California and Seoul. Eleven scenarios are analyzed: a baseline of 7000 non-electric taxis, five scenarios ranging from 1000 e-taxis to 5000 e-taxis, and another five scenarios where the e-taxis have infinite chargers as an upper bound. The study finds that the number of charging locations recommended in the earlier study may be insufficient at some locations even under the 3000+ e-taxi scenarios. More importantly, despite an average revenue of $260/taxi for the 7000 non-electric taxis and about $247/taxi for electric taxis over the finite charger scenarios, the revenue gap between e-taxis and non-electric taxis in a mixed fleet increases significantly as the e-taxi share increases. This is because the increasing queue delay imposed on e-taxis gives the non-electric taxis an increasing competitive advantage, raising their average revenue from $260/taxi (1000 e-taxis) up to $286/taxi (5000 e-taxis, 150% increase in the revenue gap), all other operating costs being equal. This has implications on policies applied to individuals versus a whole fleet, as the individual-oriented policies may be less effective.
Implementation of a Self-Learning, Onboard, Geo-Clustering Platform for Reducing Emissions in Drayage Operations
Parthav Desai, Volvo Group North AmericaShow Abstract
Eddie Garmon, Volvo Group North America
Hoda Yarmohamadi, Zenuity
Jason Strait, Volvo Group North America
Pascal Amar, Volvo Group North America
Aravind Kailas, Volvo Group North America
The paper presents an approach for dynamically managing mode of operation on Class-8 plug-in hybrid electric trucks (PHETs) to increase zero emission (ZE) operation per duty cycle without increasing battery size requirements. The prototype platform records many vehicle-specific (e.g., speed) and location-centric (e.g., GPS coordinates) data during each operational run to manage the on-board energy usage and driving maneuvers. As an example, it is desirable to go into the “ZE mode” when operating at lower speeds and propulsion energy demands. Using lower first- and second-order statistics of the truck speed and power requirement, it is possible to identify and group such geographical locations (to create “geo-clusters”) along a route. The next time the truck is in a “geo-cluster,” it can automatically transition to electric operation. While it has been established that incorporating operational characteristics will enhance the performance of hybrid-electric vehicles, implementation is most feasible when operations are repetitive or on fixed routes, like transit buses (herein referred to as “static geofences”). On the other hand, the proposed platform enables predictive controls on a vehicle-to-vehicle basis without imposing any operational constraints or knowledge. Deployment of the concept in customer operations at the ports of Los Angeles and Long Beach has demonstrated the potential to almost double the ZE mode mileage compared to the “static geofence” implementations on PHETs.
Micro-Analysis of the Fueling Costs of Electric Vehicles in Consideration of the Range of Options Available to Electric Vehicle Users
Peter Weldon, Trinity College, DublinShow Abstract
Patrick Morrissey, Trinity College, Dublin
Margaret O'Mahony, Trinity College, Dublin
It is common knowledge that the costs of powering electric vehicles (EV) are less than the fuelling costs of internal combustion engine vehicles (ICEV) due to the relative per-unit prices of electricity and gasoline, respectively, which enhances the consumer acceptance of electromobility as a viable alternative to ICEV usage. However, whilst ICEV users do not have a choice regarding how to fuel their vehicles since they must make use of fuelling stations, EV users have a number of options with respect to where, when, and how to recharge their vehicle batteries. This paper presents a micro-analysis of the costs associated with the various choices available to EV users and the effects of these differing methods on the overall costs of running EVs. Comparisons are made to fuelling costs of ICEVs, wherein the potential effects of rising gasoline prices are estimated. Whilst the cost analysis is primarily conducted within an Irish context, fuelling costs in additional countries are also considered. The results show that the choices EV users make with respect to how to charge their vehicles have a large impact on the fuelling costs of EVs. Substantial savings can be accrued by EV users through their selection of charging method. It is recommended that the fuelling costs of EVs and comparisons to ICEV costs should be publicised to expand consumer awareness.
Content Analysis of Interviews with Hydrogen Fuel Cell Drivers in Los Angeles
Oscar Lopez, Arizona State UniversityShow Abstract
Rhian Stotts, Arizona State University
Scott Kelley, University of Nevada, Reno
Michael Kuby, Arizona State University
Hydrogen fuel cell vehicles (HFCVs) are zero-emission vehicles (ZEVs) and their widespread adoption may help to mitigate some of the issues arising from fossil-fuel usage in the transportation sector. Only in recent years have these vehicles become available for purchase or lease in the United States, and only within the state of California. To date, nearly 5,500 HFCVs have been sold or leased there, supported by a developing refueling infrastructure there. This population represents a unique opportunity, as previous studies on HFCV adoption have largely employed hypothetical stated preference surveys distributed to likely adopters. Seeking to investigate the real experiences of actual adopters from their own perspective, we conducted semi-structured interviews with twelve early adopters of HFCVs in the Los Angeles metropolitan area. We conducted thematic content analysis using these interviews to identify the prevalence of factors deductively derived from published literature. All respondents consider lifetime cost of vehicle ownership, engage in comparison shopping, and assess the adequacy of the refueling infrastructure by various geographic criteria. Environmental concerns motivated many respondents to pursue HFCV adoption, though only if it made financial sense. Respondents chose HFCVs over battery electric vehicles (BEVs)s after consideration of range, refueling time, and cost. Early HFCV adopters consistently cast their adoption of the technology as a contribution to a diverse ZEV marketplace. Strategies for the promotion of HFCV technology must account for this range of variation in adopter motivations, concerns, and behaviors early-adopter considerations that might complicate targeted HFCV promotion strategies.
Investigating Contributory Factors to Adoption of Electric Vehicles: Comparison of Logistic Regression with Genetic Programming
Mohammadreza Kavianipour, Michigan State UniversityShow Abstract
Sam Shojaei, Michigan State University
Iliya Miralavi, Michigan State University
Mehrnaz Ghamami, Michigan State University
Sharlissa Moore, Michigan State University
Wolfgang Banzhaf, Michigan State University
Annick Anctil, Michigan State University
Transportation in the United States heavily relies on fossil fuels which poses energy security and environmental challenges. Transition to alternative fuel vehicles (AFVs) can effectively decrease oil use, and be the solution to energy security, climate change and sustainable development. Electric vehicles (EVs) have a few notable superiorities compared to other AFVs, which makes EVs a thriving technology. Thus, identification of influential factors in public decision would contribute to prosperity of EV adoption. This study captures users’ perspective by conducting online surveys of both conventional vehicle and EV users from across the United States, which covers demographics, infrastructure, policy, and travel pattern factors. Two different modeling techniques are used in this study to model the vehicle choice of the users. The first model is a traditional Logistic Regression (LR) technique and the second modeling framework adopts Genetic Programming (GP) as an evolutionary computation (EC) method. The two methods are compared and results show that while evolutionary computation provides with the ability to test more combinations of variables and better predictions as a result, logistic regression models are easier to interpret.
Cost-Effectiveness Analysis of Plug-In, Hybrid Electric Vehicle Using Vehicle Usage Data Collected in Shanghai, China
Yan Xia, Southeast UniversityShow Abstract
Jie Yang, Southeast University
Zhiyuan Liu, Southeast University
Jing Dong, Iowa State University
This paper investigates the economic viability of plug-in hybrid electric vehicles (PHEV) in Shanghai, China based on a real-world in-use PHEV dataset. To quantify PHEV drivers’ gross profit compared to internal combustion engine vehicle (ICEV) owners, a total cost of ownership (TCO) model is adopted taking account of vehicle retail price, tax credits, subsidies, insurance, maintenance, energy prices, and resale value. The impacts of the determinants for gross profit are examined in terms of vehicle distance traveled (VDT) electrically, gasoline price, electricity price, and car-buying cost. It is found that (1) only 10% of the deployment of PHEVs (i.e. BYD Qin) is economically viable if the benefit from a free license plate is exempt; (2) the 100-kilometer gross profit of PHEVs increases linearly with the electric driving distance, while the saving of energy cost per kilometer decreases with the total VDT; (3) PHEVs’ profit could be significantly improved by reducing the car-buying cost – a decrease of 10% in car-buying cost makes 80% of the PHEV deployment feasible; (4) if switching the daytime charges to off-peak hours, 50% of the PHEV deployment will become feasible.
Comprehensive Transportation and Energy Analysis: A Price Sensitive, Time-Specific Microsimulation of Electric Vehicles
Felix Steck, Institute of Transport Research, German Aerospace CenterShow Abstract
John Anderson, DLR - German Aerospace Center
Tobias Kuhnimhof, Institute of Transport Research, German Aerospace Center
Carsten Hoyer-Klick, Institute of Transport Research, German Aerospace Center
Despite ambitious climate goals, the German transportation sector has failed to reduce emissions. As these emissions are dominated by personal vehicles, electric vehicles are central for achieving environmental objectives. To determine potential emission reductions from electric vehicles, a detailed analysis of the transportation and energy sectors is necessary. Thus we present a methodology to calculate charging demand of electric vehicles using a time and location specific microsimulation and probability estimation based on a utility function for charging behavior. The transportation model is coupled with a detailed energy model for Germany, which provides electricity generation per energy source on an hourly basis over a year. We apply the methodology and models to the case study of Germany in 2030 for five scenarios. The scenarios represent difference pricing schemes reflecting policy options for electric vehicles. The results show that charging demand can be shifted using market incentives. We find that charging subsidies can shift charging demand to or away from peaks. We then combine charging demand with the energy model to quantify the CO2 emissions. The results show that shifting charging demand can reduce emissions, albeit at a minimal level. For the entire year, shifting charging to the daytime can reduce emissions by 2%. New areas of research including bidirectional charging and hourly pricing are needed to ensure maximum emission reductions from electric vehicles.
Simulation-Based Scenario Analysis on the Growth of Medium-Duty and Heavy-Duty Hydrogen Fuel Cell Vehicles and Their Refueling Infrastructure Needs in California
Guozhen Li, University of California, DavisShow Abstract
Joan Ogden, University of California, Davis
This paper describes a predictive model that simulates the population growth of medium-duty (MD) and heavy-duty (HD) hydrogen fuel cell vehicles (FCVs) and estimate their hydrogen fuel demand with temporal and spatial granularity. It then finds optimal refueling facility layouts to meet the fuel demand. Additionally, several scenarios of MD/HD FCV growth pathways for California are simulated and compared. These scenarios suggest that California my need about 130 refueling facilities to meet the hydrogen fuel demand from the local MD/HD FCV market by 2030, and each facility needs to handle demands ranging from hundreds of kilograms of hydrogen per day to tens of thousands of kilograms of hydrogen per day.
Effect of Road Level of Service on Compact Car Fuel Economy
Rami Chkaiban, University of Nevada, RenoShow Abstract
Elie Hajj, University of Nevada, Reno
Gary Baily, Nevada Automotive Test Center
Muluneh Sime, Nevada Automotive Test Center
Hao Xu, University of Nevada, Reno
Seyed-Farzan Kazemi, Rowan University
Peter Sebaaly, University of Nevada, Reno
Several factors can influence transportation user costs including road level of service (LOS), road properties (grade, curvature, etc.), and vehicle engine fuel type. In this study, the impact of LOS on fuel economy of compact cars powered with gasoline, diesel, and ethanol (E85) was evaluated. Physics-based full vehicle simulation models were used to estimate fuel economy over a range of synthetically optimized (SO) driving cycles representing roadway with LOS A through LOS E. The SO driving cycles were determined based on a collected sample of SHRP 2 NDS time series data. The speed coefficient of variation for the SO driving cycles increased with the decrease in traffic LOS. Based on the data and assumptions used in this study, fuel economy was lowest for vehicle operating at LOS C, irrespective of engine fuel type. The speed variation due to traffic congestion and stop and go operating conditions at LOS D and LOS E resulted in a decrease in fuel economy in comparison to the steady state conditions (i.e., constant speed driving cycles). User costs, comprising of fuel and travel time costs, were the lowest for LOS A, irrespective of vehicle engine fuel type. The diesel powered compact car had the lowest user costs followed closely by gasoline powered compact car. The E85 powered compact car showed the highest user costs among the evaluated vehicles. The lower price for the E85 biofuel did not seem to compensate for the reduced fuel economy of E85 powered compact car operating at different LOS.
Deeply Integrated Vehicle Dynamic and Powertrain Operation for Efficient Plug-In Hybrid Electric Bus
Peng Hao, University of California, RiversideShow Abstract
Kanok Boriboonsomsin, University of California, Riverside
Guoyuan Wu, University of California, Riverside
Zhiming Gao, Oak Ridge National Laboratory
Tim LaClair, Oak Ridge National Laboratory
Matthew Barth, University of California, Riverside
The emerging connected and automated vehicle (CAV) technology has opened the door for developing innovative applications and systems to improve vehicle energy efficiency. While most of the recent research has been focused on optimizing vehicle dynamic (VD) and powertrain (PT) operation in isolation, there exists untapped potential to further improve vehicle fuel efficiency through a co-optimization of VD&PT control. In this paper, we develop an eco-operation solution for a plug-in hybrid electric bus (PHEB) which seamlessly integrates state-of-the-art CAV applications with advanced powertrain optimization strategies, aiming at improving vehicle energy efficiency and reducing tailpipe emissions. The proposed eco-operation system have 6 components, including traffic/signal timing information acquirement, information integration, scenario identification, powertrain, trajectory planning and a MATLAB/Simulink model for validation and fine-tuning. A deeply integrated vehicle dynamic and powertrain control algorithm is proposed in the paper to optimize the energy-efficiency. Based on the key logic of powertrain control strategy of PHEB, we develop a simplified PHEB powertrain model, and put it into our graph based optimization model as the edge cost to derive the optimal speed profile, which is further fine-tuned in the Simulink model. The proposed mode is validated in multiple numerical tests under Eco-Approach and Departure, Eco-Stop and Launch and Eco-Cruise scenarios, and shows significant performance (above 20%) in energy-saving.
Calculated Choices or Quick Decisions?: Modeling the Effect of Public Charging Opportunities on Plug-In Electric Vehicle Use and Charging Choices
Yanbo Ge, University of WashingtonShow Abstract
Don MacKenzie, University of Washington
The impact of plug-in electric vehicles (PEVs) on the electricity grid and gasoline displacement depends on the distance of the trips that is covered by electricity. It is therefore important to understand how PEV owners make decisions on which vehicles to use and when to charge, which can be influenced by multiple factors including the characteristics of the trips and charging opportunities. Complicating such analyses is the intertemporal dependence of choices: when using a PEV, decisions about charging depend on both prior choices and expectations about the future charging opportunities; and decisions on which vehicle to use depend on opportunities for charging. Based on the data from a stated preference survey among PEV owners, this paper compares two approaches to modeling PEV owners’ choices of vehicle choice for a home-based tour and charging choices at the subsequent stops: static discrete choice modeling which treats all choices as independent; and dynamic discrete choice modeling (DDCM) which explicitly accounts for the intertemporal payoffs associated with vehicle use and charging choices under uncertainty. The results indicate both models can help to understand how PEV users make decisions about which vehicle to use for a travel day and can inform charging demand forecasting of PEV users. The DDCM based on the intertemporal payoffs offers slightly better within-sample predictions. However, this improved predictive power comes at the cost of considerably higher computational time and a much more involved process for model development and estimation that may be substantially less accessible to many potential users.
Characterizing Plug-In Electric Vehicle Driving and Charging Behavior: Observations from a Year-Long Data Collection Study
Seshadri Srinivasa Raghavan, University of California, DavisShow Abstract
Gil Tal, University of California, Davis
The growth in plug-in electric vehicle (PEV) adoption poses numerous uncertainties in terms of their real-world usage and the gap between the potential and estimated environmental benefits. In order to facilitate the policymakers in shaping the trajectory of policies to encourage PEV uptake, The Advanced PEV Travel and Charging Behavior project was started in 2015 to understand PEV usage at a household level. As part of this 5-year project, in-vehicle GPS tracked data from 140 battery electric vehicles (BEVs), 235 plug-in hybrid electric vehicles (PHEVs), and 300 internal combustion engine vehicles (ICEs) from approximately 320 households will be collected. The focus of this paper is on analyzing a yearlong driving and charging data from 121 BEVs covering 5 BEV types with different driving ranges (24kWh Leaf, 30 kWh Nissan Leaf, 60 kWh Chevrolet Bolt, Tesla ModelS 60-80kWh and Tesla Model S 80-100 kWh). Analysis of Variance (ANOVA) and group means comparison tests on the daily vehicle miles traveled (DVMT) revealed no statistically significant differences associated with trips that are recurring and routine between all the BEV types. Clustering analysis using k-means algorithm indicated that the Bolts charging behavior aligned more with Leaf-24 compared to the ModelS vehicles, but on days when the BEVs were not recharged, the Bolts had a higher percentage of days when they drove for 50 miles or more compared to the Leafs.
Exploring the Value of Clean Air Vehicles High Occupancy Lane Access in California
Wei Ji, University of California, DavisShow Abstract
Gil Tal, University of California, Davis
This paper estimates the value of the California program that allows single-occupant use of high occupancy vehicle lanes by plug-in vehicles and to benefit from reduced tolls as high same as high occupancy vehicles. A survey was conducted that targeted PEV owners in California to understand their attitudes towards EV-related incentive policies, as well as their commute routes and the frequency of their access to high occupancy vehicle/toll (HOV/T) lanes. In San Francisco bay area for example, 32% reported that they are paying reduced tolls of about $540 dollars per year and 28% save on commute time while driving alone. The value of a program was estimated based on the toll savings reported and the value of travel time savings for each survey respondent while commuting, and the estimated value was compared with the corresponding Clean Vehicle Rebate for which the respondent was eligible. We also examined the spatial heterogeneity of CAV decal value across different regions and tested the impact of local HOV/T lane accessibility to the value of CAV decals.
Design Power Transportation Expansion Framework with Electric Vehicle Market Penetration
Mohannad Kabli, Mississippi State UniversityShow Abstract
Farjana Nur, Mississippi State University
John Usher, Mississippi State University
Mohammad Marufuzzaman, Mississippi State University
This study presents a novel two-stage stochastic programming model to facilitate power expansion and charging station establishment decisions. The optimal grid power and renewable power expansion plans are determined in the first stage. Additionally, the model suggests a list of discrete locations where power expansion needs to be done considering demand stochasticity. The second stage determines the number of charging stands of varying plug types to be installed in a charging station. We demonstrate a case study based on the current EV adoption rates and EV usage in Washington, D.C. to visualize and validate the modeling results. The results provide the appropriate distribution of charging station locations with the number of chargers by plug types to be installed for the selected region. The experimental results suggest that the increment of grid power expansion budget significantly impacts the grid and renewable power cell selection therefore the optimal configuration of the charging stations location is highly impacted by this decision. We also conduct multiple sensitivity analysis that reveals some notable managerial insights. These insights can significantly help in determining an appropriate charging infrastructure for a network so that it will favorably support increased EV adoption.
Network Equilibrium Model of Electric Vehicles with Stationary and Dynamic Charging Infrastructure on the Road Network
Xiasen Wang, University of WashingtonShow Abstract
Don MacKenzie, University of Washington
Battery Electric Vehicles (BEVs) have the potential to reduce the emission of air pollution and greenhouse gas. One difficulty in promoting the usage of BEVs is the availability of charging infrastructures. A new recharging method called dynamic wireless charging lanes is promising to solve this problem because it can charge the electric vehicles while driving on the road. This paper first analyzes the features attracting drivers to use charging lanes. Since the charging lane technology is still under experiment, we design a stated preference (SP) survey and distribute to 161 BEV drivers in China. The results show that income and travel distance significantly affect the charging method choice of drivers: with the distance and income increase, people are more likely to use charging lanes. We also compared three classification models to explore the method with the highest prediction accuracy. Second, given the locations of public charging stations and charging lanes, this paper developed a network equilibrium model to describe the effect of charging infrastructures on the route choice of drivers. At last, a numerical analysis is conducted on a simple road network. The results show that the location of charging lanes and charging stations has a significant influence on the route choice of drivers. When the income increases, or the investment of charging lanes decreases, charging lanes can be an economical effective charging method for drivers.
Investigating the Buyers of Electric Vehicles in California: Are We Moving Beyond Early Adopters?
Jae Hyun Lee, University of California, DavisShow Abstract
Scott Hardman, University of California, Davis
Gil Tal, University of California, Davis
This paper investigated multi-year changes of plug-in electric vehicle (PEV) buyers in California. The socio-economic profile of 11,559 PEV owners were used in a latent class cluster analysis. This revealed four heterogeneous groups of PEV buyers: 49% are High income families, 26% Middle income old families, 20% Middle income young families, and about 5 % are Lower income renters. With multi-year analysis (2012-2017) we found a gradually decreasing proportion of High income families and increasing proportions of Middle income young families and, more recently, Lower income renters. Bass diffusion models were used to investigate how the proportion of first time PEV buyers by socio-ecnomic cluster may change up to 2030. The models are traditioanl diffusion models in that they estimate the first time adoption of PEVs, not the size of the market. Based on this analysis High income families may not be the majority of first time PEV adopters due to there being limited number of consumers with their profile in California. However, the three remaining clusters have the potential to become larger groups of PEV adopters and may be where the majority of new PEV buyers emerge. The results of this paper will helpful for policymakers working on the market introduction of PEVs and may be able to inform incentive programs and infrastructure development.
Dynamic Wireless Charging Infrastructure Planning for Electric Vehicles
Sabya Mishra, University of MemphisShow Abstract
Huan Ngo, University of Memphis
Amit Kumar, University of Texas, San Antonio
Battery electric vehicles (BEVs) has been showing the promise of being energy efficient and the frequency of BEV adoption is on the rise. While BEVs have the potential to compete with the internal combustion engine vehicles in the future, the constraints include extended recharging time, shortened range, and insufficient number of charging locations. To avoid these constraints dynamic wireless charging (DWC) appears to be a plausible solution. However, the question remains how to develop a DWC infrastructure considering transportation planners objective of system level travel time and energy; and users intent route choice given the DWC infrastructure by the planner. In this paper, we propose a sequential two-level planning approach considering objectives of the planner and users. Two different planner objectives, total system travel time and total system energy consumption are considered; and for each objective road users select their choice of routes subjected to the type of DWC infrastructure provided. The proposed framework is first demonstrated with small test network and then it is applied to real scale network as a case study. The research will be helpful to the practitioners and planners on optimally locating DWC infrastructure to satisfy the need of different BEV’s range constraints within a given budget to minimize the public social cost and energy. The numerical results show that DWC implementation in a region with one million population is expected to reduce the overall energy consumption by more than 10% with an estimated cost savings of $30 billion, assuming an average gasoline price of $2 and electricity cost of 25 cents per unit.
Estimating Potential Demand for Long-Distance Electric Vehicle Travel in Washington State
Parastoo Jabbari, University of WashingtonShow Abstract
moein khaloei, University of Washington
Don MacKenzie, University of Washington
Our objective in this study is to use mobile phone apps data to identify high demand locations for charging electric vehicles (EVs) during long-distance trips. It is important to understand travel behavior of current and future EV owners and their destinations, to provide them with reliable and equitable access to charging network. High costs and low utilization of the charging infrastructure emphasizes the need for thorough study of travel patterns. Origin and destination (OD) of trips can be extracted from apps data and ultimately, we can use it to estimate OD matrix of EV trips and predicting their travel patterns. However, using this type of data has its own challenges. In this paper, we discuss the issues we faced using raw mobile phone apps data and how we addressed them. Using current travel patterns and EV registration data, we estimated potential statewide EV trips if they were not constrained by range and charging availability. We then assigned the trips to the network using the shortest path given by Google Maps. To make our results more accessible we developed a visualization tool to illustrate potential EV traffic on major road segments and ODs of the traffic flow. The framework can be readily extended to incorporate different levels of EV adoption throughout the state, and to account for EV owners’ choices of whether to use an EV or a conventional vehicle for a long trip. Keywords: GPS, data, long-distance, electric vehicle
A Spatial Analysis of Commuting Trips of Electric Vehicle Drivers: The Case of Maryland
Hyeon-Shic Shin, Morgan State UniversityShow Abstract
Z. Farkas, Morgan State University
Amirreza Nickkar, Morgan State University
This paper focuses on exploring the possible socio-demographic characteristics and factors that contribute to electric vehicle (EV) owners’ commuting trip patterns and travel behavior. The objective of the study is to recommend public policies to decision makers to prompt EV purchase and use, by identifying socio-demographic attributes that influence EV travel patterns and behavior. An online survey of EV owners was conducted from May 28, 2015, to February 19, 2016. In total, 1,257 EV owners in Maryland completed usable surveys. A set of statistical analysis methods was employed to analyze the data. A multinomial logistic regression model (MNL) was constructed to examine the associations between EV owner characteristics and their spatial commuting trip behavior, and the results have been compared to the spatial travel patterns of drivers of internal combustion engine vehicles (ICEV) in Maryland. The results of this study showed that socio-demographic factors including age, education, income, household size, the number of vehicles in the house, and political affiliation played a significant role in the commuting travel behavior and pattern of EV drivers. Furthermore, the results show there is a direct statistical correlation between the driving distance of EV drivers and their sociodemographic characteristics, such as family size, number of vehicles in the house, education and income.
Survey of Oregon Electric Vehicle Owners: Understanding Perceptions, Motivations, and Concerns
John Macarthur, Portland State UniversityShow Abstract
Michael Harpool, Portland State University
Even though electric vehicles (EVs) have been on the market for in the U.S. for almost ten years, they are still an emerging alternative fuel vehicle. In the 2017, EVs only represented approximately 1% of new vehicle sales in the U.S. This study looked to learn from people living in Oregon who either own or lease an EV about their views on EVs and their experiences purchasing, owning and driving their vehicle. This paper presents the findings of the first statewide survey of EV owners in the state of Oregon. While findings on motivation are fairly similar for EV types and non-EV owners, findings related to pre and post purchase concerns varied by vehicle type and rural vs. urban household locations. These initial findings from the analysis of this study can be a useful tool in how to inform hybrid owners and other populations who may be considering an EV that many of their concerns, such as battery range and longevity, charging time and costs, are unlikely to be realized. The study could also inform advocacy groups and educational programs.
The Effect of Substitutional Policies Against Purchase Subsidy Decrease on AFV Purchase Intention: A Case Study of Beijing, China
Tianwei Lu, Beijing Jiaotong UniversityShow Abstract
Enjian Yao, Beijing Jiaotong University
Fanglei Jin, Beijing Jiaotong University
Subsidy policy gives powerful support in alternative fuel vehicle (AFV) market penetration. However, the subsidy is also a huge financial burden so that it is only a transitional measure and will gradually be phased out. This paper aims to investigate the impact of future purchase subsidy phase-out on AFV purchase intention and explore substitutional policies to continue stimulating AFV purchase. To achieve such goals, a stated preference (SP) survey is conducted in Beijing, China and a binary logit (BL) model is established and calibrated to describe how various factors influence vehicle purchase preferences between conventional vehicle (CV) and battery electric vehicle (BEV) as well as analyze the BEV adoption rate variation when purchase subsidy decreases and purchase tax increases. In addition to factors related to vehicle technical features, we focus more on policies due to the unique external policy environment regarding the purchase and drive restriction in Beijing. Furthermore, we test several policies as substitutions against purchase subsidy regarding daily use subsidy and bus line permission, and also attempt to simulate their performances in BEV adoption rate. The results identify that the incentive policies can signiﬁcantly impact BEV purchase intention. If the purchase subsidy or purchase tax exemption policy is phase-out in Beijing, the BEV selection probability is reduced from 45.93% to 16.62% and 16.15%, and the daily use subsidy needs to be set at the level of $734 and $759 respectively per year to maintain the original selection probability. These results will give useful reference to future policy making.
Evaluating Electric Vehicle User Mobility Data Using Neural Network-Based Language Models
Kevin Alvarez, North Carolina State UniversityShow Abstract
Emerson Wenzel, Tufts University
Arielle Dror, Smith College
Omar Asensio, Georgia Institute of Technology (Georgia Tech)
The development of an efficient and reliable charging infrastructure is critical to reducing barriers to the adoption of electric vehicles (EVs). However, the availability of charging services can often be unreliable, even in dense urban centers. In this paper, we provide national evidence on how well publicly-owned and privately-owned charging stations are serving the needs of the rapidly expanding population of EV drivers in 651 core-based statistical areas in the United States. We apply natural language processing to automatically classify 127,257 user reviews from the world's most popular EV charging station locator app. We classify user sentiments into positive and negative charging station experiences. To do this, we apply a convolutional neural network (CNN) and compare the model's classifications to those obtained from human raters. We find that the CNN-based classifer approaches human accuracy in this classification task (accuracy = 84.3%) and learns domain specific terms at the level of human experts. Contrary to expectations about the private provisioning of public charging services, we also find that privately-owned stations (e.g. hotels, commercial shopping areas or car dealerships), do not outperform publicly-owned stations (e.g. parks, municipal and government buildings, or transit centers) from the perspective of the consumer. Additionally, we find higher negative sentiment in dense urban centers, where issues of charge rage and congestion may be the most prominent.
Filling the Gender Gap in Electric Vehicle Markets
Kenneth Kurani, University of California, DavisShow Abstract
Nicolette Caperello, University of California, Davis
The early market for “zero emission vehicles” (ZEVs)—including plug-in hybrid, battery, and fuel cell electric vehicles—included fewer women than expected based on their participation in the overall market. Using survey and interview data from new car buying households primarily from California, this research explores whether and how this gender gap can be closed. Qualitative analysis provides several hypotheses about potential differences between female and male respondents’ awareness, knowledge, experience, and assessments of ZEVs. These hypotheses are tested using a nominal logistic regression model on the drivetrain types of vehicles designed by respondents in an on-line survey. Simple frequencies by respondent sex may seem to confirm a smaller number of women are prospectively interested in a ZEV for their household, and thus to suggest an ongoing gender gap. However, diagnosis of the model results concludes two things. First, a variable for respondent sex identifier does not statistically significantly improve the multivariate model; neither alone nor crossed with several other hypothesized variables. Second, the difference between female and male respondents’ propensity to design ZEVs is explained by differences between them in measures specific to ZEVs. The most important difference seems to be the consistently higher likeliness of female respondents to state they don’t know about ZEV performance, charging and fueling, pricing and incentives. These results suggest closing this gender gap may be possible through tailored outreach messages, education information, and experience opportunities to increase engagement with, and purchase of, ZEVs by women—and men.