Artificial Intelligence is taking a deeper look into the problem of battery aging

Batteries degrade with age. Monitoring their condition is very important, but it is also kind of tricky. Now scientists from Cambridge and Newcastle Universities have developed a machine learning method to monitor batteries by sending electrical pulses into them and measuring the response. This technology could improve battery health and safety in electric cars.

With the growing demand for electric cars it is important to ensure the longevity of their batteries. Image credit: Mariordo via Wikimedia (CC BY-SA 4.0)

Most batteries are complex chemical devices. Over time their chemical composition changes through rogue reactions, which inevitably degrade battery performance. However, current methods of assessing the condition of the battery rely on measurement of current and voltage during battery charging and discharging cycles. This tells some information about the state of degradation, but not the processes involved. That is why scientists created this new battery monitoring system, which is non-invasive and easily added to existing battery systems.

This technology is based on Artificial Intelligence. A computer sends a short electric signal to the battery. The battery, of course, responds to that impuls and the computer measures that response. AI algorithm assesses different features of that reply and is able  to discover specific features that are the tell-tale sign of battery aging. This data-driven technology can accurately monitor and predict battery aging. Scientists performed over 20,000 experimental measurements to train the model, which provided AI with a huge data set to compare new information with. Interestingly, data gathered by this model can be used to research batteries too. Some information may encourage engineers to probe the battery to see what is happening and how it could be fixed.

Dr Yunwei Zhang, co-author of the study, said: “Machine learning complements and augments physical understanding. The interpretable signals identified by our machine learning model are a starting point for future theoretical and experimental studies”.

The advantage of AI is that it can analyse vast amounts of data very quickly and very efficiently. It can recognize processes happening in the battery and compare them to those it saw in the data set before. Hopefully, this will produce reliable results and provide a lot of new information, which could lead to improvement in battery technology.

Scientists are now using their AI technology to monitor processes in different battery systems. They want to see how the degradation happens and how it can be solved. They are also working on AI-based charging protocols, which could enhance battery life in some cases.


Source: Cambridge University