- Award ID(s):
- Publication Date:
- NSF-PAR ID:
- Journal Name:
- Proceedings of the Transportation Research Board (TRB) annual meeting
- Sponsoring Org:
- National Science Foundation
More Like this
Extracting User Behavior at Electric Vehicle Charging Stations with Transformer Deep Learning ModelsMobile applications have become widely popular for their ability to access real-time information. In electric vehicle (EV) mobility, these applications are used by drivers to locate charging stations in public spaces, pay for charging transactions, and engage with other users. This activity generates a rich source of data about charging infrastructure and behavior. However, an increasing share of this data is stored as unstructured text—inhibiting our ability to interpret behavior in real-time. In this article, we implement recent transformer-based deep learning algorithms, BERT and XLnet, that have been tailored to automatically classify short user reviews about EV charging experiences. We achieve classification results with a mean accuracy of over 91% and a mean F1 score of over 0.81 allowing for more precise detection of topic categories, even in the presence of highly imbalanced data. Using these classification algorithms as a pre-processing step, we analyze a U.S. national dataset with econometric methods to discover the dominant topics of discourse in charging infrastructure. After adjusting for station characteristics and other factors, we find that the functionality of a charging station is the dominant topic among EV drivers and is more likely to be discussed at points-of-interest with negative user experiences.
Detecting Behavioral Failures in Emerging Electric Vehicle Infrastructure Using Supervised Text Classification AlgorithmsThere is a growing interest in applying computational tools to the automatic discovery of social and economic behavior. For example, with decisions involving resource allocation related to public infrastructure, the ability to predict failures can allow for more efficient policy responses. In this paper, we use social data from a popular electric vehicle (EV) driver app to characterize the emerging EV charging station infrastructure. We introduce a typology of EV charging experiences collected from user reviews and deploy text classification algorithms, including convolutional neural networks (CNN), to automatically learn about potential failures. We use machine learning techniques as a pre-processing tool for econometric analyses on the quality of service delivery. After classifying the reviews into 9 main user topics and 34 subtopics, we find that the dominant issues in EV charging relate to station functionality and availability, which drive negative consumer experience. Contrary to the public discourse about EVs, range anxiety was not of large concern to existing EV drivers. Based on our findings, we move towards automated identification of failures in public charging infrastructure that can significantly reduce research evaluation costs through relatively simple computational solutions.
Greenhouse gas emissions embodied in electric vehicle charging infrastructure: a method and case study of Georgia, US 2021–2050
Electric vehicle (EV) charging infrastructure buildout is a major greenhouse gas (GHG) mitigation strategy among governments and municipalities. In the United States, where petroleum-based transportation is the largest single source of GHG emissions, the Infrastructure Investment and Jobs Act of 2021 will support building a national network of 500 000 EV charging units. While the climate benefits of driving electric are well established, the potential embodied climate impacts of building out the charging infrastructure are relatively unexplored. Furthermore, ‘charging infrastructure’ tends to be conceptualized in terms of plugs and stations, leaving out the electrical and communications systems that will be required to support decarbonized and efficient charging. In this study, we present an EV charging system (EVCS) model that describes the material and operational components required for charging and forecasts the scale-up of these components based on EV market share scenarios out to 2050. We develop a methodology for measuring GHG emissions embodied in the buildout of EVCS and incurred during operation of the EVCS, including vehicle recharging, and we demonstrate this model using a case study of Georgia (USA). We find that cumulative GHG emissions from EVCS buildout and use are negligible, at less than 1% of cumulative emissionsmore »
AbstractThis dataset contains information from 3,395 high resolution electric vehicle charging sessions as presented in "Electric vehicle charging stations in the workplace: high-resolution data from casual and habitual users ", including indicator variables for user types based on time of adoption, total sessions logged, and position held within the firm. The data contains sessions from 85 EV drivers with repeat usage at 105 stations across 25 sites at a workplace charging program. The workplace locations include facilities such as research and innovation centers, manufacturing, testing facilities and office headquarters for a firm participating in the U.S. Department of Energy (DOE) workplace charging challenge. The data is in a human and machine readable *.CSV format. The resolution of the data is to the nearest second, which is the same resolution as used in the analysis of the paper. It is directly importable into free software.
Agrivoltaic systems have the potential to meet energy demands of electric vehicles in rural Oregon, US
Electrification of the transportation industry is necessary; however, range anxiety has proven to be a major hindrance to individuals adopting electric vehicles (EVs). Agrivoltaic systems (AVS) can facilitate the transition to EVs by powering EV charging stations along major rural roadways, increasing their density and mitigating range anxiety. Here we conduct case study analyses of future EV power needs for Oregon, USA, and identify 174 kha of AVS viable agricultural land outside urban boundaries that is south facing and does not have prohibitive attributes (designated wetland, forested land, or otherwise protected lands). 86% highway access points have sufficient available land to supply EV charging stations with AVS. These AVS installations would occupy less than 3% (5 kha) of the identified available land area. Installing EV charging stations at these 86% highway access points would yield 231 EV charging stations with a median range of 5.9 km (3.6 mi), a distance comparable to driver expectations, suggesting that this approach would serve to mitigate range anxiety. AVS powered rural charging stations in Oregon could support the equivalent of 673,915 electric vehicles yr−1, reducing carbon emissions due to vehicle use in OR by 3.1 mil MTCO2yr−1, or 21%.