Researchers have developed an economical vehicle-side strategy for electric bus charging stations participating in vehicle-to-grid services, using reinforcement learning to optimize when and how long buses provide grid support. The study focuses on a real electric bus charging station in China?s Pearl River Delta region and evaluates how health-aware vehicle-to-grid operation can reduce lifecycle cost while extending battery life.
Vehicle-to-grid, or V2G, services allow idle electric vehicle batteries to act as distributed energy storage resources. In principle, electricity can flow from the grid to the vehicle during charging and from the vehicle back to the grid when support is needed. This could help improve grid flexibility, but the economics are complicated because discharging a vehicle battery may accelerate degradation, reduce battery life, and affect vehicle availability.
Many V2G studies focus on passenger vehicles, whose idle times can be highly random and difficult to coordinate at scale. The new study takes a different approach by examining electric buses. Electric buses often follow fixed routes and schedules, and bus charging stations may have more regular operating patterns than dispersed private vehicles. According to the article, this gives electric bus charging stations stronger practical application value for scalable V2G services.
The researchers collected a complete real-world dataset from a Pearl River Delta electric bus charging station, covering 25 routes and 377 electric buses. The dataset includes basic station information, bus batteries, charging records, driving behavior, bus departure times, and electricity prices. This real operational foundation is important because V2G profitability depends strongly on scheduling, pricing, battery condition, and whether vehicles can still meet departure requirements.
To evaluate the strategy, the study developed a power battery electrical-health-economic model. This model simulates electrical behavior, health degradation, and economic performance under different charging and discharging conditions. By combining electrical operation, battery degradation, and cost, the framework can assess not only whether a V2G strategy earns revenue or reduces charging expense, but also whether it damages the battery enough to offset those gains.
The charging station then uses reinforcement learning to implement a health-aware V2G strategy, referred to as V2G-H. The study?s highlights specify that an Actor-Critic reinforcement learning approach is used to optimize the dynamic V2G start time and duration. In practical terms, the algorithm learns how to schedule grid-service participation while considering station constraints, bus operation needs, electricity prices, and battery health.
The researchers compared several operating strategies, including no V2G, uncoordinated V2G, deep-discharge V2G, and V2G-H. This comparison is useful because V2G can appear attractive if only short-term energy revenue is considered, but may become less economical if it causes excessive battery aging. A health-aware strategy aims to avoid that problem by balancing grid-service benefits against long-term battery degradation.
The results reported in the paper suggest that the reinforcement learning-based V2G-H strategy provides substantial lifecycle benefits. Compared with the traditional approach, V2G-H saved $1,539 in total cost per bus over the entire lifecycle and extended battery life by more than 21 months. The study also analyzed the effects of electricity price fluctuations, departure frequency, and battery cost on the V2G strategy, providing additional insight into when such services may be most economically feasible.
Further validation will still be needed across other cities, tariff structures, bus fleets, battery chemistries, and grid-service markets. Even so, the study offers a strong indication that electric bus charging stations can be promising platforms for scalable V2G services when scheduling is battery-health-aware. As electric bus fleets grow, reinforcement learning strategies that coordinate station economics, grid needs, and battery degradation may help make V2G deployment more practical and financially sustainable.
Reference
Author:
Taotao Li a , Xiaoqi Zeng a , Xiao Qi a , Tianyang Zhao b , Zhengmao Li c , You Lv d , Yajun Qiao e , Zijian Tan e , Jizhen Liu a d , Jinghan He a , Weixiong Wu a
Title of original paper:
Economical vehicle-side strategies for an electric bus charging station in vehicle-to-grid services based on reinforcement learning
Article link:
https://www.sciencedirect.com/science/article/pii/S2773153725001379
Journal:
Green Energy and Intelligent Transportation
DOI:
10.1016/j.geits.2025.100387
Affiliations:
a Energy and Electricity Research Center, Jinan University, Zhuhai 519070, China
b School of Electrical Engineering, Xi'an Jiaotong University, Xi'an 710000, China
c School of Electrical Engineering, Aalto University, Finland
d State Key Laboratory of New Energy Power System, North China Electric Power University, Beijing 100096, China
e Guangzhou Power Supply Bureau, Guangdong Power Grid Co., Ltd., CSG, Guangzhou 510620, China
Green Energy and Intelligent Transportation
Experimental study
Not applicable
Economical vehicle-side strategies for an electric bus charging station in vehicle-to-grid services based on reinforcement learning
27-Jan-2026