The slowdown in the electric vehicles (EVs) market is leading EV battery manufacturers into the market for battery energy storage systems (BESS). Artificial intelligence (AI) company Monolith says it is seeing rapid growth in interest in BESS among its customers.
In an interview with BEST, CEO Richard Ahlfeld and VP for AI in batteries Marius Koestler said the energy storage business has started to take off only in the last year and comes on the back of worse-then-hoped-for EV sales by auto makers.
Ahlfeld said: “They haven’t had the sales of cars that they’ve wanted, but they have created a lot of expertise in understanding battery modelling. Because electric cars are essentially batteries on wheels and they are trying to utilise this knowledge and go into a domain where there is a huge amount of excess energy and demand.
“And since there is excess energy in the grid, there is demand. And so you can sell your battery solutions there and whether it is Volvo Energy, Porsche or Stellantis, they all have secondary battery projects. And so we’ve sort of slid into the ESS market through our automotive clients.”
He said Monolith is expecting the BESS market to grow exponentially and in line with forecasts of Robin Zeng, chairman of Chinese EV battery maker CATL: to approximately half the size of the automotive market within the near future.
Monolith already helps 10 of the world’s top 20 automakers test EV batteries using machine learning algorithms, including Mercedes-Benz, BMW and Honda. Koestler said the company’s understanding of battery degradation comes down to its experience in working with cell developers and cell integrators for the EV industry.
“And if you link that all the way to the financial markets that are underwriting the ESS that have a much, much higher lifetime, the warranty on the EV is seven years, 10 at most. That’s interesting. If you can start to include the degradation in your financial modelling and you could apply degradation penalties to different trading strategies, this can make a difference between a profitable and a non-profitable asset investment,” said Koestler.
Monolith applies physics-based AI to its tool, which it claims can halve the time taken to analyse data. It uses a combination of laboratory and field testing for greater insight. Traditional testing also contains many errors (such as varying temperatures).
It means its clients can sell a battery with a testing model that sets out expected degradation, he said.
“Typically ESS operators do not factor in degradation as a cost in the same way as they have their O&M cost as they have the capital cost, Capex, or the cost of energy,” he said. “They factor in the losses in energy, like to heat, when they do their modelling, but they don’t factor in degradation because they don’t have a good way of doing it.”
He thinks financial underwriters will start to think more about degradation and profitability.
Ahlfeld said most industrial style test laboratories for electric car companies will produce around 2TB of data per week in their R&D testing process, with tens of millions of points for review, “an insane amount of information.”
Monolith develops aluminium-air batteries as backup storage with a supplier in Canada and estimates it helps them get this done and validated two years earlier by deploying AI. That shortens time to market.
The future will increasingly see automation and autonomous laboratories, said Ahlfeld.
Photos: Monolith’s Marius Koestler (left) and Richard Ahlfeld. Monolith
