State-of-Charge Estimation of a Li-Ion Battery using Deep Forward Neural Networks

TL;DR

This paper develops deep neural network models to accurately estimate the State of Charge of lithium-ion batteries, using a robust machine learning workflow to prevent overfitting and validate performance.

Contribution

It introduces a methodology for building and assessing deep forward neural networks for lithium-ion battery SOC estimation following best machine learning practices.

Findings

Deep networks accurately model battery drive cycles

Proposed workflow effectively prevents overfitting

Models demonstrate reliable SOC estimation performance

Abstract

This article presents two Deep Forward Networks with two and four hidden layers, respectively, that model the drive cycle of a Panasonic 18650PF lithium-ion (Li-ion) battery at a given temperature using the K-fold cross-validation method, in order to estimate the State of Charge (SOC) of the cell. The drive cycle power profile is calculated for an electric truck with a 35kWh battery pack scaled for a single 18650PF cell. We propose a machine learning workflow which is able to fight overfitting when developing deep learning models for SOC estimation. The contribution of this work is to present a methodology of building a Deep Forward Network for a lithium-ion battery and its performance assessment, which follows the best practices in machine learning.