2026-09-18T22:20:43Zhttps://dspace.mit.edu/server/oai/requestoai:dspace.mit.edu:1721.1/1516242023-08-01T03:20:02Zcom_1721.1_7582com_1721.1_7581col_1721.1_131023
Annaswamy, Anuradha
Sun, Andy
Barber, Adam
Massachusetts Institute of Technology. Department of Mechanical Engineering
Sloan School of Management
2023-07-31T19:53:56Z
2023-07-31T19:53:56Z
2023-06
2023-07-14T19:55:43.554Z
https://hdl.handle.net/1721.1/151624
Electric vehicles (EVs), with their potential to drastically reduce greenhouse gas emissions, pose a problem for energy distribution infrastructure which was not previously designed with hosting capacity capable of handling the additional demand generated by their mass adoption. Understanding when customers charge their EVs and how much energy they consume better enables electric utilities to provide more reliable and affordable energy to all customers while aiding the transition to clean transportation. The purpose of the research was to analyze passenger EV charging data from National Grid's Massachusetts EV Off-Peak Charging Program and determine whether generalizable and scalable machine learning models could be built to predict EV charging energy demand, and further determine the lowest possible geographic granularity of such models. This research was novel in its charge rate estimation methodology, normalization of charging energy on a per-vehicle basis, accounting for charging energy demand flowing into and out of the studied system, and the addition of ambient air temperature as a feature variable. Modeling employed supervised machine learning methods with random forests deemed optimal in terms of accuracy, complexity, and computational intensiveness. Ultimately, this research successfully created and operationalized an accurate service territory model and illuminated the challenges associated with utilizing telematics data for demand modeling.
M.B.A.
S.M.
Massachusetts Institute of Technology
In Copyright - Educational Use Permitted
Copyright retained by author(s)
https://rightsstatements.org/page/InC-EDU/1.0/
Modeling Passenger Electric Vehicle Charging Demand with Machine Learning Using Telematics Data and Temperature
Thesis
application/pdf
Master
Master of Business Administration
Master of Science in Mechanical Engineering
Publication
unknown
unknown
<Publication xmlns="https://www.openaire.eu/cerif-profile/1.1/" id="be1ce0ec-2c40-4a45-be5c-80912cb09608">
<Type xmlns="https://www.openaire.eu/cerif-profile/vocab/COAR_Publication_Types">http://purl.org/coar/resource_type/c_1843</Type>
<Title>Modeling Passenger Electric Vehicle Charging Demand with Machine Learning Using Telematics Data and Temperature</Title>
<PublishedIn>
<Publication>
</Publication>
</PublishedIn>
<PublicationDate>2023-06</PublicationDate>
<Authors>
<Author>
<DisplayName>Barber, Adam</DisplayName>
<Affiliation>
<OrgUnit>
</OrgUnit>
</Affiliation>
</Author>
</Authors>
<Editors>
</Editors>
<Publishers>
<Publisher>
<DisplayName>Massachusetts Institute of Technology</DisplayName>
<OrgUnit />
</Publisher>
</Publishers>
<License>https://rightsstatements.org/page/InC-EDU/1.0/</License>
<Abstract>Electric vehicles (EVs), with their potential to drastically reduce greenhouse gas emissions, pose a problem for energy distribution infrastructure which was not previously designed with hosting capacity capable of handling the additional demand generated by their mass adoption. Understanding when customers charge their EVs and how much energy they consume better enables electric utilities to provide more reliable and affordable energy to all customers while aiding the transition to clean transportation. The purpose of the research was to analyze passenger EV charging data from National Grid's Massachusetts EV Off-Peak Charging Program and determine whether generalizable and scalable machine learning models could be built to predict EV charging energy demand, and further determine the lowest possible geographic granularity of such models. This research was novel in its charge rate estimation methodology, normalization of charging energy on a per-vehicle basis, accounting for charging energy demand flowing into and out of the studied system, and the addition of ambient air temperature as a feature variable. Modeling employed supervised machine learning methods with random forests deemed optimal in terms of accuracy, complexity, and computational intensiveness. Ultimately, this research successfully created and operationalized an accurate service territory model and illuminated the challenges associated with utilizing telematics data for demand modeling.</Abstract>
<Access xmlns="http://purl.org/coar/access_right"
>
</Access>
</Publication>