July 24, 2024: A recent study from Canterbury Christ Church University, in New Zealand published in the Journal of Energy Storage, shows research introducing an AI method that improves the accuracy of measuring lithium-ion battery charge levels.

Initial findings reveal that the AI approach outperformed current techniques. By integrating so-called ‘clustering algorithms’ with neural networks, the new method offers better precision in predicting battery charge levels — vital for the performance and safety of electric vehicles and renewable energy storage systems.

Mohammed Khalifa Al-Alawi, head of the research team, said the new AI method adapts to the changing conditions batteries face, enhancing reliability. “This advancement is crucial for the future of electric vehicles and renewable energy,” he said.

Using data from a Tesla Model 3 battery, the team demonstrated a major improvement in accuracy, reducing error rates by over 60%. This ensures that batteries operate more efficiently and have a longer lifespan.

The researchers also assessed the impact of this AI enhancement on second-life batteries, which are repurposed for home energy storage and other applications. The results indicate that the improved SOC estimation method reduces the degradation rate of these batteries. This leads to substantial cost savings, potentially amounting to thousands of dollars for operators of large-scale energy storage systems.

“By reducing the degradation rate of second-life batteries, our AI method can make energy storage systems much more cost-effective and sustainable,” said Al-Alawi. “This advancement offers economic and environmental benefits.”