IJSEA Volume 15 Issue 12

Research on State of Health Prediction of Power Batteries for New Energy Vehicles

Hailan Zhao
10.7753/IJSEA1512.1004
keywords : new-energy vehicles; power batteries; state of health; data-driven; cloud-edge collaboration

PDF
The state of health (SOH) of power batteries is a core parameter that characterizes driving-range attenuation, power decline and safety risks of new-energy vehicles. Accurate SOH prediction is of great significance for vehicle energy management, warranty judgment and echelon utilization of retired batteries. This paper defines an SOH characterization system dominated by capacity retention rate and internal resistance growth rate, and analyzes aging mechanisms including solid-electrolyte interphase (SEI) film thickening, active-lithium loss and lithium plating, as well as the coupled influences of temperature, state-of-charge window and charge-discharge rate on degradation rates. Existing prediction methods are categorized into model-based, data-driven and mechanism-data fusion approaches, and their accuracy, computing-power requirements and engineering implementability are compared. Furthermore, a vehicle-level prediction framework of "on-board feature extraction - cloud-side model training - cloud-edge collaborative updating" is proposed. Key challenges such as working-condition generalization, label acquisition, battery inconsistency and computing-power constraints are discussed. The analysis reveals that the engineering bottleneck for SOH prediction has shifted from algorithm accuracy to data closed-loop performance and generalization capability. Constructing mechanism-constrained data-driven models calibrated with a small volume of real-vehicle calibration data is a feasible approach to improving practical performance.
@artical{h15122026ijsea15121004,
Title = "Research on State of Health Prediction of Power Batteries for New Energy Vehicles ",
Journal ="International Journal of Science and Engineering Applications (IJSEA)",
Volume = "15",
Issue ="12",
Pages ="13 - 16",
Year = "2026",
Authors ="Hailan Zhao"}