DYNAMIC SCHEDULING OF ELECTRIC VEHICLE CHARGING NETWORK: A GAME-THEORETIC APPROACH
Amir Mirheli, North Carolina State UniversityShow Abstract
Leila Hajibabai (firstname.lastname@example.org), North Carolina State University
This study proposes a dynamic scheduling scheme for electric vehicle (EV) charging facilities under uncertain charging demand, charger availability, and charging rate. The problem is formulated as a dynamic programming model that minimizes the (i) total travel cost and charging expenses and (ii) extra charging duration. A stochastic look-ahead technique is proposed that first models a generalized Nash equilibrium (GNE) to distribute a non-cooperative game-theoretic model to EV users. A consensus-based coordination scheme is incorporated into the GNE procedure to push the user-level solutions toward system-level optimality. Then, a Monte Carlo tree search algorithm, with an embedded tree policy and shooting heuristic, is implemented to efficiently capture the uncertainties and approximate the value function of the dynamic program. The tree policy evaluates the available actions and estimates the value functions over time, while the shooting heuristic predicts the value of recently added tree nodes to determine optimal charging schedules (i.e., number of time periods to charge and charging spot assignments at each time period). Numerical experiments on hypothetical and real-world networks confirm the quality and efficiency of solutions obtained from the proposed methodology. The GNE algorithm ensures the feasibility of solutions in a finite number of iterations.
Prioritizing Cities for Hydrogen Station Deployment in China
Pengfei Yang, FTXT Energy TechnologiesShow Abstract
Marc Melaina, FTXT Energy Technologies
When introducing passenger fuel cell electric vehicles into household markets, a key challenge is the lack of hydrogen refueling infrastructure. A prioritization method is applied to all major cities in China to identify which cities can be provided with convenience hydrogen refueling networks in the most cost-effective manner. The methodology relies upon two estimates: (1) early market potential for fuel cell electric vehicles, and (2) the number of stations required to provide convenient refueling for early adopters. The goal is to compare all cities using cost-effectiveness priority indicators to inform future infrastructure planning and investment decisions. Early market potential is estimated based upon both the percent of wealthy households and the total disposable income in each city. The number of hydrogen stations required in each city is estimated by applying a genetic algorithm to a unique and geographically detailed city population database. Station network trends are found to be consistent with previous studies characterizing an average 6-minute driving time as a sufficient level of refueling convenience for early adopters. Two cost-effectiveness metrics indicating market potential per station serve as the basis of the city prioritization ranking. Results suggest more developed coastal cities are included in the top priority tiers while less developed inland cities and those with lower population density are ranked lower. The results are an important contribution to the development of a more complete assessment of hydrogen station networks serving household vehicles in China.
How Can Vehicle-to-Grid Enhance the Business Model of Car Sharing Compared to Electric and Conventional Schemes? A Customer Perspective
Konstantin Krauß (email@example.com), Fraunhofer-Institut fur System und InnovationsforschungShow Abstract
Christine Gschwendtner, Swiss Federal Institute of Technology (ETH Zurich)
To support the decarbonization of the transport and the electricity sector, two innovations are emerging: shared mobility, i.e. car sharing in this context, and vehicle-to-grid (V2G). However, extant literature has not yet investigated the compatibility of these two innovations despite expected synergies. To better understand whether car sharing services with V2G-enabled electric vehicles would be attractive for travelers, we conducted a survey (n = 340) in Germany and Switzerland with a stated preference design. By estimating three models (MNL, SMNL, and GMNL type 2), we find that V2G car sharing is more attractive to people than E- (using electric vehicles with unidirectional uncontrolled charging) or conventional (using vehicles with diesel-/petrol engines) car sharing. In user decision-making, cost, egress and access time are the most important factors. Range anxiety is not specific to the electrified car sharing services but needs to be substantially longer than the trip length for any of the alternatives. Remuneration is not expected by the respondents and is of less importance to the respondents’ utilities. Thus, if regulatory barriers are overcome and bidirectional charging infrastructure is expanded, V2G stands a good chance of reducing financial pressure from car sharing business models whilst providing a car sharing service that is in demand of potential customers.
Powering the transition: An assessment of the influence of curbside charging stations on electric vehicle adoption in Montreal, Canada
Pénélope Renaud-Blondeau, Ecole Polytechnique de MontrealShow Abstract
Genevieve Boisjoly (firstname.lastname@example.org), Ecole Polytechnique de Montreal
Hanane Dagdougui, Ecole Polytechnique de Montreal
Sylvia Y. He, Chinese University of Hong Kong
Electric vehicles are increasingly included in government policies to reduce greenhouse gas emissions and improve air quality in urban areas. In recent years, the questions of how to depict the profile of electric vehicles (EV) consumers and to accelerate their mass market adoption have received a lot of attention in the literature. Yet, few of these studies have analyzed the influence of public charging infrastructure accessibility on EV adoption in an urban context. Using data from a survey (n = 647) across Montreal Island, this work investigates the influence of three measures of accessibility to charging stations: objective (number of charging stations near home), perceived (perceptions about the current public charging network), and prospective (expectations on the public charging network in 5 years). Two binary logistic models including accessibility measures are generated to assess the determinants of (i) EV ownership and (ii) EV purchase intention, while controlling for individual characteristics, household attributes, and attitudinal factors. The findings reveal that perceived accessibility is associated with EV ownership, while prospective accessibility is associated with EV purchase intention. It is also shown that non-EV owners underestimate their accessibility to curbside chargers. The results lastly confirm the positive and significant impact of environmental awareness, social influences, and interest in new technologies on EV purchase intention, as well as the importance of EV range and performance. By focusing on curbside EV chargers, this study provides a nuanced understanding of how accessibility-based policies might contribute to enhancing EV adoption in urban areas.
Close Enough to Buy in? How Early Hydrogen Fuel Cell Vehicle Adopters Evaluate a Network of Refueling Stations
Scott Kelley (email@example.com), University of Nevada, RenoShow Abstract
Aimee Krafft, University of Nevada, Reno
Michael Kuby, Arizona State University
Oscar Lopez Jaramillo, Arizona State University
Rhian Stotts, Arizona State University
Jingteng Liu, Arizona State University
Strategies differ on how best to arrange initial hydrogen stations in a region to maximize convenience and encourage hydrogen fuel cell vehicle (FCV) adoption. After a few years of initial sales, there is an opportunity to analyze how early FCV adopters evaluated the spatial arrangement of a network of stations before getting one. We distributed a web-based survey to 129 FCV adopters throughout California in 2019, asking them where they lived and traveled at the time of adoption, up to five stations they planned to use, and subjective reasons for considering those stations. Using network GIS analysis, we estimated travel times to respondents' home locations and other frequent locations, and deviations from frequently traveled routes. We compared differences in subjective and objective convenience for primary, secondary, and lower-ranked stations using multinomial logistic regression, and tabulated the different combinations of stations that satisfied various geographic criteria for FCV adopters. Over 80% of respondents planned to rely on a portfolio of multiple stations subjectively convenient to key activity locations. Estimated travel times to stations subjectively considered to be "near" home, work, and other location types vary greatly but consistently decay beyond 90 minutes. Primary stations are subjectively and objectively more convenient to home and work than 3rd-5th stations, and more associated with subjective convenience to home and objective convenience to work than secondary stations. Other destination types align with lower-ranked stations. These findings encourage the development of future station planning methods that balance many geographic criteria.
An Integrated Transportation Network and Power Grid Simulation Approach for Assessing Environmental Impact of Electric Vehicles
Xiaodan Xu (X-Xu@tti.tamu.edu), Texas A&M Transportation InstituteShow Abstract
Hanyue Li, Texas A&M University, College Station
Jessica Wert, Texas A&M University, College Station
Ju Hee Yeo, Texas A&M University, College Station
Komal Shetye, Texas A&M University, College Station
Alexander Meitiv, Texas A&M University
Thomas Overbye, Texas A&M University, College Station
Josias Zietsman, Texas A&M University
Yanzhi Xu, Texas A&M Transportation Institute
The wide adoption of electric vehicles (EVs) will fundamentally change the structure of the energy supply in the transportation sector, with a substantial amount of energy coming from the power generation and contributing to higher emissions at power plants. Numerous studies quantify the electricity generation units (EGUs) emissions due to EVs using aggregated emission rates multiplied by the EV charging load. However, actual emissions depend on dispatched EGUs, which further depend on the EV charging demand, and the characteristics of the power grid. Therefore, an integrated approach which includes EV charging and power generation is needed to assess the complex cross-sector interactions of vehicle electrification and its environmental impact. This study develops such an integrated approach. The charging load from on-road EV operation is developed based on a regional-level transportation simulation and charging behavior simulation, considering different EV penetration levels, congestion levels, and charging strategies. The emissions from EGUs are estimated from a dispatch study in a power grid simulation using the charging load as a major input. A case study of Austin, Texas is performed to quantify the environmental impact of EV adoption on both on-road and upstream emission sources. The results demonstrate the range of emission impact under a combination of factors. The interaction of factors accentuates the need for the integrated simulation approach in achieving air quality and climate change goals. Using the co-simulation approach, system planners, and operators from the transportation and electricity sectors can gain a comprehensive view of the environmental impact of transportation electrification.
Developing a Hybrid Electric Vehicle Eco-Cooperative Adaptive Cruise Control System at Signalized Intersections
Hao Chen, Virginia Polytechnic Institute and State University (Virginia Tech)Show Abstract
Hesham Rakha, Virginia Polytechnic Institute and State University (Virginia Tech)
This study develops an eco-driving strategy for hybrid electric vehicles (HEVs) in the vicinity of signalized intersections, entitled HEV Eco-Cooperative Adaptive Cruise Control at Intersections (Eco-CACC-I). The proposed system computes real-time, energy-optimized vehicle trajectories using HEV vehicle dynamics and energy consumption models. In the proposed system, a simple HEV energy model is used to compute the instantaneous fuel consumption. This HEV energy model is selected since it is general, transferable, and can be easily used to compute instantaneous energy consumption levels for HEVs without the additional input of vehicle engine data or complicated power control strategies. In addition, a vehicle dynamics model is used to capture the relationship between speed, acceleration level, and tractive/resistance forces on vehicles. The energy-optimum problem is formulated as an optimization problem with constraints, which is solved using a moving-horizon dynamic programming approach. The proposed HEV Eco-CACC-I system was tested to evaluate its performance for various speed limits, roadway grades, and signal timings. Lastly, the proposed HEV controller was implemented in a microscopic traffic simulation software to test its network-wide performance. The test results from an arterial corridor with three signalized intersections demonstrate that the proposed system can effectively reduce stop-and-go traffic in the vicinity of signalized intersections producing savings of 7.4% in energy consumption, 5.8% in traffic delay and 23% vehicle stops, respectively.
A Metanetwork-Based Optimization Approach to Locating Charging Stations for Electric Vehicles in Intercity Networks
Jiapei Li, Tongji UniversityShow Abstract
Chi Xie, Tongji University
Zhaoyao Bao, Shanghai Jiao Tong University
Adding en-route charging stations into a road network increases the charging opportunities and extends the driving ranges of electric vehicles. This paper develops a metanetwork-based approach for modeling and solving an optimal en-route charging station location problem for electric vehicles commuting along intercity highways. This problem takes into account the strategic decisions of the station planners (charging station locations) and operational decisions of individual vehicle drivers (route and charge choices), the purpose of which is to reduce on the network level the frequency and extra mileage of making travel detours due to charging needs and ultimately encourage the adoption and utilization of electric vehicles. By explicitly incorporating the charging requirement caused by limited driving ranges, a mixed linear integer programming model based on the station-subpath metanetwork topology, is formulated and solved, as the result of a two-phase process by decomposing a routing decision into two parts. The two phases of the problem formulation and optimization are made on the original network and metanetwork levels, respectively. The first phase is done by repeatedly solving a network-based distance-constrained shortest path problem, while the second phase is tackled by implementing the classic branch-and-bound algorithm encapsulating a metanetwork-based shortest path algorithm. Numerical results demonstrate the superiority of the metanetwork-based approach in computational efficiency for solving the en-route charging station location problem, especially for large-size networks.
ACRP Graduate Student Paper - Economic Feasibility Analysis of Charging Infrastructure for Electric Ground Fleet in Airports
Laura Soares, Rutgers UniversityShow Abstract
Hao Wang (firstname.lastname@example.org), Rutgers University
Many airports are converting ground fleet to electric vehicles to reduce greenhouse gas (GHG) emissions and increase sustainability of airport operations. Although this paradigm shift is relevant to the environment, it is necessary to understand economic feasibility of this change in order to justify tje decision. This study used lifecycle cost analysis (LCCA) to compare the economic performance of electrified gournd fleet in the airport as compared to conventional fossil fuel option. Three different charging systems (plug-in charging, stationary wireless charging, and dynamic wireless charging) for pushback tractors and intraterminal buses at a major hub airport were considered in the analysis. Although the conventional fossil fuel options present the lowest initial cost for both feets, it costs most in 30-year analysis period. Among three electric charging infrastructres, the plug-in charging station shows the least accumulative cost for pushback tractors; while their cost differences are negligible for intraterminal buses. Although electric ground fleet is proved to show economic benefits, the most cost-effective charging infrastructure may vary depending on driving mileage and system design. The use of LCCA to analyze new systems and infrastructures for decision-making is highly recommended.
Estimating the Total Number of Workplace and Public EV Chargers in California
Bingzheng Xu, University of California, DavisShow Abstract
Adam Davis, University of California, Davis
Gil Tal, University of California, Davis
Allocating charging infrastructure to support the Plug-in Electric Vehicle (PEV) market transition is a major challenge for policy makers and planners seeking to promote PEV adoption. Most electric vehicle charging away from home takes place at or near the workplace, but existing research is focused on public charging. This study develops a method to combine charging location data from a variety of sources including users’ self-reports in order to produce a combined database of workplace and public charging locations. By matching survey responses to charging locations reported in other datasets using spatial clustering methods, we are able to estimate the number of charging locations and chargers in California. This analysis finds that charger location datasets from the Alternative Fuels Data Center and Plugshare both significantly underrepresent the total number of chargers in California. We find 34,577 chargers by combining AFDC and Plugshare and between 17,000 to 25,000 chargers near commute destinations by adding locations reported in a survey of EV drivers.
Robust Design of Electric Charging Locations under Travel Demand Uncertainty and Driving Range Heterogeneity
Mohammad Miralinaghi (email@example.com), Purdue UniversityShow Abstract
Mahmood T. Tabesh, Purdue University
Gonçalo Homem de Almeida Correia, Technische Universiteit Delft
Sania E. Seilabi, Purdue University
Amir Davatgari, Purdue University
Samuel Labi, Purdue University
In the past decade, electric vehicles (EVs) have proved to be a viable option to replace internal combustion engine vehicles (ICEVs) and thereby to mitigate the environmental adversities associated with ICEVs. The rising demand for EVs is expected to lead to an increased number of EV charging stations, and vice versa. Therefore, to promote EVs, governments seek guidance on continued investments in EV charging infrastructure development. Such investment decisions, which include EV charging station locations and capacities, and the timing of such investments, require robust estimations of the future travel demand and EV battery range constraints. This study aims to determine the optimal schedule and location of constructing new charging stations and decommissioning refueling stations over a long-term planning horizon, considering the uncertainty in future travel demand forecast and the driving range heterogeneity of EVs. A robust mathematical model is proposed to solve the problem by minimizing not only the worst-case total system travel cost but also the total penalty for unused capacities of charging stations. This study uses an adaptation of the cutting-plane method to solve the proposed model. The results indicate, based on two key decision criteria (travelers’ cost and sufficient charging supply) that the robust scheme outperforms its deterministic counterpart.
Maximum Utilization and Deployment Prioritization of public charging infrastructures – Insights from Revealed Travel Patterns
Wan Li, Oak Ridge National LaboratoryShow Abstract
Zhenhong Lin (firstname.lastname@example.org), Oak Ridge National Laboratory
For Battery Electric Vehicle (BEV) to be as usable as gasoline vehicles, a certain level of public charging availability is needed. Direct current fast charging (DCFC) is yet to achieve profitability, while more expensive extreme fast charging (xFC) is being promoted. To inform deployment decisions and avoid wasteful investments, it is important to better understand the potential utilization and deployment priority of different charging technologies. Based on 2017 National Household Travel Survey data that includes daily trip sequence, trip distance, dwell times, the study develops Cumulative Public Recharging model to estimate the daily maximum charging potentials and the resulting maximum daily electric range under different types of charging speeds, battery capacity and charging behavior constraints. The results suggested that more advanced public chargers and high charging opportunities increase the daily maximum driving range. Residential charging is sufficient for most daily short-distance trips while public chargers are still needed for middle and long-distance trips. xFC may not be necessary for people with home charging but could be more useful for people without and for situation in needs of urgent charging. xFC fast charging becomes even less important with longer BEV ranges. In other words, it is most useful for short range BEVs, assumed to be capable of handling the xFC high power of charging. As a conclusion, high market penetration of Level 2 chargers and medium market penetration of DCFC should be considered primarily for deployment to serve all short-, mid-, and long-distance trips.
Statistical Modeling for Charging Behavior of Plug-in Electric Vehicles using Large-scale Charging Data
Choudhury Siddique, Argonne National LaboratoryShow Abstract
Zhaomiao Guo (email@example.com), University of Central Florida
Fatima Afifah, University of Central Florida
Yan Zhou, Argonne National Laboratory
Chaitanya Kaligotla, Argonne National Laboratory
The market share of plug-in electric vehicles (PEVs) are increasing over the past decade. Understanding how the PEV drivers charge their vehicles is imperative in charging infrastructure planning, power system operation, and energy impacts analyses. This research aims to quantify the impacts of different factors that play significant roles in PEV charging behavior. In particular, we develop statistical models to understand when, where, and how long do PEVs plugin, and identify predictive factors, such as vehicle types, driver characteristics, charger features, and charging prices. Data for more than 189,000 charging sessions in the State of Illinois for a 13-month period were collected from ChargePoint, one of the leading charger network providers in Illinois. This paper provides an in-depth statistics analysis of this data by developing six linear and logistic regression models to estimate the impacts of different influencing factors on charging behavior, including the location of a charging session, state of charge (SOC) at the beginning of the session, charger type used, etc. Our results based on a unique dataset from supply-side and statistical models highlight the essential roles that different charger and vehicle attributes play in PEV’s charging behavior. For example, the battery capacity of vehicles affects choices of charging locations, charger types, length of charging sessions, as well as whether PEVs will be fully charged at the end or not. PEV types, charging date/time, charger types have an influence on start SOC and charging duration.
Investigating the sensitivity of electric vehicle out-of-home charging demand to changes in fleet makeup and vehicle usage, a case study for California 2030
Adam Davis, University of California, DavisShow Abstract
Gil Tal, University of California, Davis
Accurately predicting the spatial distribution and charging demand of future electric vehicles is vital to directing investment in charging infrastructure and planning policy interventions. To date, this expansion has been heavily focused in wealthy cities and suburbs, among commuters, and among households able to charge their vehicles at home. The expansion of EV ownership will include both changes in where the vehicles are owned as well as how they are used and charged. This paper demonstrates methods to to generate projections of predict where the expansion of electric vehicle ownership is likeliest to occur under current market characteristics and allow for testing of scenarios of future characteristics. These methods are demonstrated with an analysis of California, using a scenario of 3 million battery electric vehicles and 2 million plugin-hybrid electric vehicles, to match the state’s goal of 5 million zero-emission vehicles by 2030. These projections are combined with a model for charging behavior to generate scenarios of demand for charging away from home under various fleet characteristics.
Integrating Plug-in Electric Vehicles (PEVs) into Household Travel- Factors Influencing PEV Use in California
Debapriya Chakraborty (firstname.lastname@example.org), University of California, DavisShow Abstract
Scott Hardman, University of California, Davis
Gil Tal, University of California, Davis
The topic of household vehicle miles traveled (VMT) has been traditionally studied in the context of gasoline vehicles. In this study, we analyze the VMT of PEVs as part of the household travel demand to understand how much PEVs are being used relative to gasoline cars and the factors that influence their VMT in a household. We use data from a unique repeat survey of PEV owners in California who were originally surveyed when they first bought their vehicle. The key results suggest that in addition to the traditional factors like population density, build environment, or cost per mile, in the case of PEVs, electric range, commute patterns, access to infrastructure, and vehicle sharing within a household has a major influence on PEV VMT. As PEV penetration goes up, understanding the factors that influence the driving behavior of PEV drivers will help to refine the emissions impact assessment of these alternative fuel technology vehicles that depends on assumptions of average annual VMT.
DISCLAIMER: All information shared in the TRB Annual Meeting Online Program is subject to change without notice. Changes, if necessary, will be updated in the Online Program and this page is the final authority on schedule information.