A Deep Learning Capacity Estimation Method Based on Incremental Capacity Analysis and Differential Thermal Voltammetry

Abstract

Precise estimation of Lithium-ion battery capacity is critical for the battery management system (BMS). This paper proposed an innovative method that combinates convolutional neural network and feature curves which are incremental capacity analysis (ICA) and differential thermal voltammetry (DTV). Rather than extracting feature parameters of the IC curve as done in available research, the present method uses the whole IC and DTV curves in a certain range of voltage as the input to avoid complicated annual feature extraction and correlation analysis. The result shows that the max error of capacity estimation is less than 4.46 %, the mean absolute percentage error is less than 1.29 % for each battery, and overall MAPE is below 1.19 %.