Remaining useful life prediction of lithium-ion batteries

SOH prediction has been a popular topic of discussion and research in recent years, with many new developments around Machine Learning and Artificial Intelligence. Leveraging Machine Learning techniques requires large datasets. As such, this project does not only aim to highlight the developments in...

Full description

Saved in:
Bibliographic Details
Main Author: Er, Harrick Yue Hui
Other Authors: Xu Yan
Format: Final Year Project
Language:English
Published: Nanyang Technological University 2022
Subjects:
Online Access:https://hdl.handle.net/10356/163604
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:SOH prediction has been a popular topic of discussion and research in recent years, with many new developments around Machine Learning and Artificial Intelligence. Leveraging Machine Learning techniques requires large datasets. As such, this project does not only aim to highlight the developments in Machine Learning that has not yet been implemented to current conventional Recurrent Neural Networks, but also participate in a battery aging experiment to generate an aging dataset based off both dynamic and static load profiles. This poses a huge hurdle in Machine Learning techniques, as datasets are scarce and severely time-consuming to procure. In this project, many current methods of SOH prediction and RUL prediction are discussed in this project, and SOH prediction models are trained to predict the RUL of a battery cell, using different Machine Learning techniques. Such techniques include the Regression model and the Long Short-Term Memory model, an improvement over the convention Recurrent Neural Network model.