"My dream is that in 10 years' time electric cars will drive around with a display where you can see the battery life depending on different factors like temperature, driving style and traffic."
This is one of the visions for the research project Smart Battery that Remus Teodorescu, Professor in the Department of Energy Technology at Aalborg University, has just been awarded DKK 28.5 million for from VILLUM FONDEN. He will develop a new type of battery that uses artificial intelligence to extend battery life and reliability.
This will be done by developing battery packs where battery cells can be monitored and controlled individually. Currently, a battery pack stops working if one of its cells runs out of capacity. And because the cells have different capacities, the weakest cells determine the life of a battery. Smart Battery will change that.
"We’ll get the cells to consume power individually so that they can adapt to each other, letting the battery live longer. We also want to make it possible to assess the capacity and life of the battery more accurately,” Remus Teodorescu says.
According to the professor, the improved life and reliability of the new batteries could benefit the green transition in several ways. In particular, it may have an impact on electric cars, making them more attractive to buyers because the batteries will be better.
At the same time, an electric car with a Smart Battery will emit less CO2 than a car with a conventional battery pack. And the new technology will also make it easier to recycle the batteries once they have served their purpose in cars.
"Once a battery pack in a car has been used up, it has approximately 80 percent capacity left. Then the big question is: When will it break down? With a Smart Battery you can evaluate and control the life of a battery so it can be used for many years afterwards for other things, for example in the home,” he says.
ARTIFICIAL INTELLIGENCE FOR CONSTANT IMPROVEMENT
In a new development, researchers will use artificial intelligence in developing new batteries. This is because, according to Professor Remus Teodorescu, you would need a huge amount of processing power to calculate the life of a lithium battery in real time using a regular computer.
"So we need artificial intelligence that can be trained to evaluate battery life. It also allows data from different batteries to be used to constantly retrain and improve the calculation model. This can be done through a cloud-based updating model so models are constantly becoming more accurate and tailored to their users,” Remus Teodorescu says.