Battery health evaluation using a short random segment of constant current charging

Area of Science:

• Battery Technology
• Data Science
• Electrochemical Engineering

Background:

• Accurate state of health (SOH) estimation is critical for lithium-ion batteries (LIBs) safety and efficiency.
• Complex degradation mechanisms hinder traditional SOH evaluation.
• Existing data-driven methods struggle with random and incomplete charging data from real-world applications.

Purpose of the Study:

• To develop novel data-driven methods for estimating LIB SOH.
• To utilize short random charging segments (RCS) for SOH estimation.
• To validate the proposed methods across diverse battery types and conditions.

Main Methods:

• Development of three distinct data-driven algorithms for SOH estimation.
• Utilizing short random charging segments (RCS) as input data.
• Validation using 75 commercial LIB cells under varied temperatures and discharge rates.

Main Results:

• The developed models achieve high-precision SOH estimation.
• Models trained under nominal conditions generalize well to different operating conditions.
• An RCS with a 10mV voltage window yields an average error below 5%, with further improvement as the window increases.

Conclusions:

• Data-driven SOH estimation is feasible using short random charging segments.
• The proposed methods offer a robust solution for real-world LIB monitoring.
• Optimizing the voltage window of RCS can significantly enhance SOH estimation accuracy.