In December 2025, the TRANS Research Group’s latest study, “A data-driven and physics-based model for assessing real-world usage behavior impacts on electric vehicle battery life,” was published in the Q1 (JCR) energy journal Journal of Energy Storage (Impact Factor: 9.8). This work develops an integrated, physics-informed and data-driven dynamic operation simulation platform for electric vehicles—Electric Vehicle Dynamic Operation Simulation (EV-DOS) model—enabling more accurate and personalized prediction of powertrain performance, as well as enhanced energy-efficiency management. The platform is built upon three key technical approaches:
- Bidirectionally coupled framework: A bidirectional coupling is established between a physics-based powertrain energy transfer model and an LSTM-based, data-driven battery lifetime model trained on real-world vehicle big data. This design combines the rigor of physical mechanisms with the adaptability of data-driven learning.
- High-accuracy simulation and validation: The platform is rigorously validated over long time horizons and under diverse driving conditions. The average monthly energy-consumption error is controlled within 0.53 kWh/100 km. Using battery state of health (SOH) as an example, the SOH estimation error is below 0.6%, demonstrating strong reliability in real-world scenarios.
- Multi-dimensional quantification of usage behaviors: Across driving conditions representing different regions and driving styles, the study systematically quantifies the long-term cumulative impacts of behaviors such as fast charging, deep charging, HVAC usage, and vehicle-to-grid (V2G) interaction on battery lifetime, providing actionable insights to optimize driving and charging practices.
Overall, this research offers important quantitative evidence and theoretical support for full-lifecycle energy management and battery performance management strategies in electric-vehicle powertrains.
The first author of the paper is Hao Jing, a Ph.D. student (Cohort 2024) in the TRANS Research Group. The corresponding authors are Prof. Shiqi Ou. Co-authors include Dr. Jianyao Hu (The Fifth Electronics Research Institute of MIIT), Prof. Xinwu Qian (Rice University, USA), Prof. Haobo Dong (South China University of Technology), as well as TRANS Research Group members Hao Qi (Ph.D. student) and Jiankuan Zhu (M.S. student). This study was supported by the National Key R&D Program of China and the Guangdong Provincial Talent Introduction Team Program.