• Speaker : Prof. Zhibin Chen
• Date : July 6, 2026
• Affiliation : Assistant Professor, New York University Shanghai / Global Network Assistant Professor, New York University
• Category : Special Lecture (2026 Jeju Summer Camp)

On the Value of Orderly Electric Vehicle Charging in Carbon Emission Reduction

Abstract

This special lecture at the 2026 Jeju Summer Camp asks how much carbon reduction electric vehicles actually deliver, and argues that the answer turns on orderly EV charging rather than on adoption numbers alone. Grid carbon intensity swings by more than 100 times across the hours of a day, from roughly 5 to 15 gCO2/kWh at best to 800 to 1,000 gCO2/kWh at worst, so when a vehicle charges matters as much as whether it is electric.

Earlier work on optimized charging mostly built charging profiles from national travel surveys instead of observed EV behaviour, and treated power-plant emissions as an outside parameter rather than modelling how charging and generation dispatch affect each other. The lecture closes that gap with a bi-level model: a lower level that reschedules charging within each driver’s travel needs, and an upper level that minimizes power dispatch cost against electricity demand, solved by a sensitivity-analysis-based algorithm.

The model runs on real data from the Shanghai EV Data Center, which covers over 2 million registered EVs. The working sample is 3,777 battery electric vehicles with more than 5 million driving records and 1.5 million charging records across an 11-month period, combined with actual Shanghai power plant and grid data. Coordinating the whole fleet cuts emissions by up to 39 percent, and the talk traces how that figure responds to fleet penetration, battery capacity, plant ramp rates, and added wind capacity.

Presentation Overview

This presentation covers the following key topics:

• Motivation: transport’s share of global greenhouse gas emissions, the EV mandate landscape, and market growth in China, Shanghai, New York City, and Korea
• The core problem: why the carbon value of an EV depends on its charging schedule and on grid carbon intensity, and what earlier studies missed by using simulated charging data and approximate emission factors
• Data sources: Shanghai EV Data Center vehicle and charging records, the local generation mix and thermal generator characteristics, external supply through the Anhui to East China transmission project and Three Gorges and Gezhouba hydropower, and grid load simulation
• Modelling framework: a bi-level formulation minimizing total system emissions, rescheduling charging at the lower level and dispatching power at the upper level, solved with a sensitivity-analysis-based algorithm
• Basic results: original against optimized charging schedules, state-of-charge transitions, cumulative charging electricity, and how the emission saving shifts by season
• Sensitivity analysis: coordinated-EV fleet penetration reaching 39 percent mitigation at full coordination, battery capacity gains flattening beyond roughly a 40 kWh increase, low sensitivity to plant ramp rate, and the effect of adding wind capacity
• Restricted orderly charging: a more realistic scenario setting original, restricted, and fully optimized schedules side by side
• Key findings: baseline battery EV emissions near 9.44 kg CO2 per day and 73 gCO2/km against 170 gCO2/km for a gasoline vehicle, and savings of 0.47 percent at 1 percent coordination against 39 percent at full coordination, or roughly 5,554 against 458,889 tons of CO2 per day