New Method Helps to More Accurately Predict Battery Failure

Electricity is unavailable to millions of people across the globe. A key to overcoming this issue while avoiding carbon emissions and air pollution is to set up decentralized solar-battery systems.

New Method Helps to More Accurately Predict Battery Failure.

Image Credit: University of Oxford.

However, such solar-battery systems are hampered by high costs and remote sites that prohibit timely servicing. When the batteries in these systems die, it might be hard to replace them, leaving people without power.

Knowing when the batteries are likely to die is critical for planning maintenance procedures and reducing power outages. A new technique to determine battery failure, developed by the Faraday Institution’s Multiscale Modelling project, has been demonstrated to be 15%–20% more accurate than existing methods when applied to the same dataset. The collaborative study by the University of Oxford and the Faraday Institution was published recently in the journal Joule.

The researchers tested their method by teaming with Bboxx — a next-generation utility that provides clean energy by manufacturing, distributing and financially supporting decentralized solar-powered systems in developing countries — to evaluate their approach, which produced real-world operating data. This circumvented the drawbacks of previous battery health modeling research, which primarily relied on tiny datasets obtained in lab settings.

Bboxx acquired raw measured voltage, current and temperature data from over 1000 functioning batteries in Africa during a two-year period. This technology requires no additional sensors or other needs, thereby allowing energy systems to remain up at all times.

Our approach is unique in showing how physics-based machine learning can work in real-world battery applications at scale. We use advanced probabilistic machine learning techniques to infer battery internal resistance as a function of current, temperature, state of charge and time, enabling calibration to standard conditions.

Professor David Howey, Department of Engineering Science, University of Oxford

The success of the approach is due to the combination of a population-wide health model and a battery-specific health indicator that becomes increasingly informative towards end of life,” added Professor Howey.

The methodologies reveal the factors that cause battery aging, such as voltage and temperature extremes, and the method may be used for any battery that can be represented by a basic electrical circuit model.

These results are of interest to a wide audience of battery operators and customers and can be used to accelerate innovation in understanding battery performance, especially if organizations make operational data more widely available in the way Bboxx have pioneered here. We are delighted that this research paper is a first of its kind demonstration of a scalable approach for getting insights from field data.

Professor David Howey, Department of Engineering Science, University of Oxford

Bboxx has agreed to make the data publicly available, where the data includes over 600 million rows of operational metrics from genuine battery systems.

We hope this will prove to be a key resource for the community and kick start efforts to analyze field data for new insights into battery performance.

Professor David Howey, Department of Engineering Science, University of Oxford

Journal Reference:

Aitio, A & Howey, D A (2021) Predicting battery end of life from solar off-grid system field data using machine learning. Joule. doi.org/10.1016/j.joule.2021.11.006.

Tell Us What You Think

Do you have a review, update or anything you would like to add to this news story?

Leave your feedback
Your comment type
Submit

While we only use edited and approved content for Azthena answers, it may on occasions provide incorrect responses. Please confirm any data provided with the related suppliers or authors. We do not provide medical advice, if you search for medical information you must always consult a medical professional before acting on any information provided.

Your questions, but not your email details will be shared with OpenAI and retained for 30 days in accordance with their privacy principles.

Please do not ask questions that use sensitive or confidential information.

Read the full Terms & Conditions.