Highly Available Li-Ion Batteries Sensors Readings Prediction Framework

Lithium-ion batteries can experience internal failures due to issues like overcharging and excessive discharging. They can also encounter external failures, primarily involving sensors that monitor temperature, voltage, and current. External failures are often seen as the primary causes of internal...

Full description

Saved in:
Bibliographic Details
Main Authors: Anas Tiane, Mohamad Alzayed, Chafik Okar, Safi Bamati, Hicham Chaoui
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11027060/
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Lithium-ion batteries can experience internal failures due to issues like overcharging and excessive discharging. They can also encounter external failures, primarily involving sensors that monitor temperature, voltage, and current. External failures are often seen as the primary causes of internal failures <xref ref-type="bibr" rid="ref1">[1]</xref>. Numerous battery health monitoring systems (BHMS) have been developed to ensure the proper functioning of batteries. Current BHMS frameworks lack robust mechanisms to handle sensor data loss or corruption. Without reliable recovery of sensor readings, these systems remain vulnerable to undetected failures and incomplete diagnostics. This creates a serious bottleneck in developing fault-tolerant, real-time battery monitoring systems. In this article, the main contribution is presenting a highly available sensor readings prediction framework to address situations where sensor readings are unavailable. The framework aims to accurately reconstruct missing or faulty sensor data using correlations among healthy sensors, thereby ensuring continuous and dependable battery monitoring, even in the presence of sensor faults. The generated reading sequences achieve high <inline-formula> <tex-math notation="LaTeX">$R^{2}$ </tex-math></inline-formula> accuracy, ranging from 94% to 99%. The proposed highly available sensor readings prediction framework is implemented using a lightweight long short-term memory (LSTM) network. The prediction model is trained and tested on batteries B6 and B7 from the NASA dataset and validated using the Toyota dataset, one of the largest available datasets.
ISSN:2169-3536