Volume 29: Closing Carbon Cycles – A Transformation Process Involving Technology, Economy, and Society: Part IV

Voltage-temperature Feature-based Capacity Estimation Method for Li-ion Battery Cell Combining Probability Density Function and Random Forests Difan Jia, Xuqi Fu, Zhanyao Ma, Xiaowu Zuo

Abstract

Accurate capacity estimation is crucial to ensure operational safety of Li-ion battery. In this paper, a novel capacity estimation approach is proposed for Li-ion battery cell. Two voltage-related features on probability density function based incremental capacity curve and average temperature are extracted as healthy indicators. Regression between healthy indicators and capacity is constructed using random forests. Results show that the capacity estimation error could be controlled within 2.5% throughout the whole lifecycle of the battery.

Keywords Li-ion battery, capacity estimation, data-driven, random forests