A Comprehensive Exploration on Different Machine Learning Techniques for State of Charge Estimation of EV Battery - Conference Paper

Aug 4, 2026

State of Charge (SoC) represents the available energy in a battery as a percentage over time and strongly influences Battery Management Systems. SoC estimation relies on proportional relationships among battery electromotive force, current, terminal voltage, and temperature, yet sharp terminal-voltage drops near discharge end can introduce substantial calculation errors. The study applies and compares machine learning methods—including Random Forest, Decision Tree, Gradient Boosting, Support Vector Regression, Polynomial Regression, and Multilayer Perceptron—using two real-world lithium-ion datasets (LG Battery and Unibo Powertools Battery). K-fold cross validation and grid search are used to evaluate model skill.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages7

Reading Time18 min

AccessFree to download

Introduction

State of Charge (SoC) and Error Sources

Machine Learning Techniques for SoC Estimation

Random Forest

Decision Tree

Gradient Boosting

Support Vector Regression

Polynomial Regression

Multilayer Perceptron

Datasets Used for Evaluation

LG Battery

Unibo Powertools Battery

Model Validation and Hyperparameter Search

K-fold Cross Validation

Grid Search Cross Validation

Results and Comparative Findings

A comprehensive exploration on different machine learning techniques for state of charge estimation of EV battery Mithul Raaj, A. T. ; Ratnam, Justin; Niranjan Kumar, S. ; Gupta, Tanya; Balaji, Keerthi; Rani, C. ; Rajesh Kumar, M. ; Farrag, Mohamed Published in: 2023 58th International Universities Power Engineering Conference (UPEC) DOI: 10.1109/UPEC57427.2023.10294352 Publication date: 2023 Document Version Author accepted manuscript Link to publication in ResearchOnline Citation for published version (Harvard): Mithul Raaj, AT, Ratnam, J, Niranjan Kumar, S, Gupta, T, Balaji, K, Rani, C, Rajesh Kumar, M & Farrag, M 2023, A comprehensive exploration on different machine learning techniques for state of charge estimation of [EV battery. in](EV battery. in) 2023 58th International Universities Power Engineering Conference (UPEC). 58th International Universities Power Engineering Conference, UPEC 2023, Institute of Electrical and Electronics Engineers Inc. , 58th International Universities Power Engineering Conference, Dublin, Ireland, 29/08/23 . [https://doi.org/10.1109/UPEC57427.2023.10294352](https://doi.org/10.1109/UPEC57427.2023.10294352) General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. Take down policy If you believe that this document breaches copyright please view our takedown policy at [https://edshare.gcu.ac.uk/id/eprint/5179 for details](https://edshare.gcu.ac.uk/id/eprint/5179 for details)[ ](https://edshare.gcu.ac.uk/id/eprint/5179 for details)[of how to contact us.](of how to contact us.) Download date: 12. Jan. 2024 A Comprehensive Exploration on Different Machine Learning Techniques for State of Charge Estimation of EV Battery Mithul Raaj A T Department of Electrical Engineering Vellore Institute of Technology, Vellore, India [mithulraaj24@gmail.com](mithulraaj24@gmail.com) Tanya Gupta Department of Electrical Engineering Vellore Institute of Technology, Vellore, India [tanyagupta822101@gmail.com](tanyagupta822101@gmail.com) Rajesh Kumar M Department of Electrical Engineering Vellore Institute of Technology, Vellore, India [mrajeshkumar@vit.ac.in](mrajeshkumar@vit.ac.in) Justin Ratnam Department of Electrical Engineering Vellore Institute of Technology, Vellore, India [justinratnam@gmail.com](justinratnam@gmail.com) Keerthi Balaji Department of Electrical Engineering Vellore Institute of Technology, Vellore, India keerthi0kiran@gmail Mohamed Farrag School of Computing, Engineering and Built Environment, Glasgow Caledonian University, Glasgow, [UK mohamed.farrag@gcu.ac.uk](UK mohamed.farrag@gcu.ac.uk) Niranjan Kumar S Department of Electrical Engineering Vellore Institute of Technology, Vellore, India[niranjanniru572@gmail.com](niranjanniru572@gmail.com) Rani C Department of Electrical Engineering Vellore Institute of Technology, Vellore, India [crani@vit.ac](crani@vit.ac) Abstract—The State of Charge (SoC) is a measurement of the amount of energy available in a battery at a specific interval of time, mostly expressed as percentage. Proportional relationships between the electromotive force of a battery, current, terminal voltage and temperature determine the SoC. There can be a considerable error in the calculations due to a sharp drop of the terminal voltage at the end of discharge. This research has explored how important SoC is, as a factor in Battery Management Systems. The work focuses on using machine learning techniques to obtain an accurate and reliable status of battery charge, this includes Random Forest, Decision Tree, Gradient Boosting, Support Vector Regression, Polynomial Regression and Multilayer Perceptron. In this paper, these techniques are tested and compared with two real world cap

FAQs

What does SoC mean in battery systems, and why is it important?

SoC measures the energy available in a battery at a given time, usually as a percentage. It is important as a key factor in Battery Management Systems.

Why can SoC calculations have significant errors?

Errors can occur due to a sharp drop in terminal voltage at the end of discharge, which disturbs the relationships used for estimation.

Which machine learning models were compared for SoC estimation?

The study compares Random Forest, Decision Tree, Gradient Boosting, Support Vector Regression, Polynomial Regression, and Multilayer Perceptron, evaluated on lithium-ion battery datasets.

Aug 4, 2026

Upload by Emma Wilson