An intelligent fault detection (IFD) system for lithium-ion battery using machine learning approach

Area of Science:

• Electrical Engineering
• Materials Science
• Computer Science

Background:

• Electric vehicles (EVs) are gaining popularity due to environmental concerns and efficiency.
• Battery failures in EVs pose significant safety risks.
• Existing fault detection methods require enhancement for proactive safety.

Purpose of the Study:

• To introduce an Intelligent Fault Detection (IFD) system for EV batteries.
• To develop and implement a machine learning (ML) based mechanism for monitoring and safeguarding EV batteries.
• To address critical faults including thermal protection, under-voltage, and over-voltage.

Main Methods:

• Real-time data collection from embedded battery sensors during vehicle operation.
• Application of K-means clustering algorithm for data analysis and defining valid operational ranges for temperature and voltage.
• Utilisation of an ensemble approach for classifying safe and unsafe battery operating conditions.

Main Results:

• The proposed ML-based IFD system achieved an accuracy of 0.94 in identifying battery faults.
• The system effectively monitors and safeguards batteries against critical faults.
• Defined valid operational ranges for temperature and voltage through data clustering.

Conclusions:

• The developed IFD system significantly enhances the safety and reliability of EV batteries.
• This proactive fault detection contributes to the long-term sustainability and economic viability of electric mobility.
• Intelligent techniques offer a robust solution for managing EV battery health and performance.