Skip to main content

CGS Research Seminar: Nationwide Analysis of EV Charging behavior Using Machine Learning Techniques

Back to All Events
picture of School of Public Policy building

At this event, Safoura Safari will discuss her research on EV charging behavior.

Abstract: Electric Vehicles (EVs) offer a promising solution to reduce carbon emissions and promote climate-friendly transportation. However, the rapid adoption of EVs has raised concerns about the adequacy and equity of EV charging infrastructure. Challenges such as high charging costs, high electricity demand on grids, and the unequal distribution of EV Charging Stations (EVCS) limit accessibility, posing barriers to widespread EV usage. To address these challenges, a deeper understanding of EV charging behaviors, exploring the peaks and variations of daily EV charging patterns, is critical for stakeholders, including local governments, energy providers, and manufacturers, to balance supply and demand and improve grid resilience. The current study presents a comprehensive nationwide analysis of EV charging usage patterns, incorporating daily and seasonal variations across electricity market regions. By utilizing unsupervised machine learning techniques, such as the Clustering LARge Applications (CLARA) algorithm and agglomerative hierarchical clustering, we analyze real-time charging data to uncover distinct clusters of usage profiles based on a similarity metric that integrates both the magnitude and shape of usage profiles. Additionally, agglomerative hierarchical clustering is used to group charging stations based on the likelihood of typical daily usage profiles occurring at that station. The results are expected to reveal clusters of EV charging behaviors that differ in terms of peak usage times and magnitudes, providing insights into the capacity-in-use of charging stations across different energy market sectors. These findings will allow for a more detailed understanding of EV power demand, offering critical insights for optimizing the development of EV charging infrastructure to meet increasing demand. By leveraging real-time usage data across the U.S., this analysis provides a clearer picture of how to improve the resilience and equity of the EV charging network.


For Media Inquiries:
Gail Chalef
Director of Communications and Public Affairs
For More from the School of Public Policy:
Sign up for SPP News

You Also Might Be Interested In…