Data-Driven Electric Vehicle Performance Analysis: Integrating Large-Scale Telematics for Charging Duration, Battery Degradation, and Range Estimation

Abstract: This presentation explores data-driven models to advance electric vehicle (EV) performance analysis through large-scale telematics integration. Using Tesla Scope data from over 600 Tesla vehicles, we collect minute-level measurements of location, state of charge (SOC), odometer readings, temperature, and charging events. These variables enable the development of charging duration estimation models that capture the influence of ambient conditions, vehicle usage, and charging infrastructure characteristics. Battery degradation is analyzed using multivariate regression and time-series modeling to quantify the impacts of age, temperature cycles, mileage accumulation, and user charging patterns on long-term energy capacity. In parallel, high-resolution telematics data retrieved from OBD-II ports are used to estimate remaining driving range. This modeling framework incorporates driver acceleration dynamics, traffic congestion levels, ambient temperature, and terrain slope into predictive remaining driving range estimation models. The integration of these datasets and analytical approaches provides a robust methodology for understanding real-world EV operation, optimizing battery management strategies, and improving range estimation accuracy. By advancing predictive methodologies rather than static models, this work creates scalable analytical tools to accelerate EV electrification and support data-informed climate mitigation strategies.
Speaker(s): Brian,
Agenda:
Must RSVP. Dinner will be served.
Bldg: UVA RICE HALL, 85 Engineer's Way, Charlottesville, Virginia, United States, 22903