In the realm of energy storage, a team of researchers from the University of Oxford, the Rutherford Appleton Laboratory, and other institutions has been working on a novel approach to battery research that could significantly improve our understanding of battery degradation. Led by Emily Lu and Gabriel Perez, the team has proposed a new framework called Heuristic Operando experiments, which aims to address the limitations of current battery characterization methods.
The researchers point out that current methods for studying battery degradation, known as operando characterization, often struggle with issues of reliability, representativeness, and reproducibility, collectively referred to as the 3Rs. These methods typically rely on custom-built hardware and passive, pre-programmed approaches that are not well-suited to capturing the random and complex failure events that can occur in batteries. To illustrate this, the team used the multi-modal toolkit at the Rutherford Appleton Laboratory as a case study, demonstrating how conventional experiments often miss transient phenomena like the initiation of dendrites, which are thin, metallic structures that can grow inside a battery and cause it to fail.
To overcome these challenges, the researchers propose a new approach that leverages artificial intelligence (AI) and physics-based digital twins to actively steer the experimental process. In this framework, an AI pilot uses the digital twins to predict and capture rare failure events, rather than simply reacting to them. This proactive search is guided by an entropy-based metric that prioritizes scientific insight, allowing the experiment to focus on the most mechanistically decisive moments. By doing so, the approach not only mitigates beam damage but also drastically reduces data redundancy.
The team also emphasizes the importance of integrating their approach with FAIR data principles, which stand for Findability, Accessibility, Interoperability, and Reusability. This integration, they argue, could serve as a blueprint for the trusted autonomous battery laboratories of the future. The research was published in the journal Nature Communications.
The practical applications of this research for the energy sector are significant. By improving our understanding of battery degradation, this approach could help to extend battery life, reduce costs, and enhance the safety and reliability of energy storage systems. This, in turn, could facilitate the wider adoption of renewable energy sources and support the ongoing energy transition.
This article is based on research available at arXiv.