Building a better energy system with AI

AI4Energy, taking place on November 11-12 in Stockholm, will be a meeting place for experts, researchers, innovators, and decision-makers engaged in shaping the future energy landscape.

Jenny Gustavsson, CIO at Öresundskraft, is one of the speakers in the program. She will focus on the practical applications that Öresundskraft is already working with today.

"A relatively simple thing is to streamline customer service with a chatbot that can better answer customers' questions. But it's within our core business that we find the truly value-creating applications. For us, it's about predictive maintenance of our district heating pipes, saving millions of kronor per month through better forecasts for district heating production, and understanding the charging behavior of heavy transport to reduce imbalances in the grid," she says.

When Christian Lind, an AI expert at Modulai, takes the stage, new opportunities to create more accurate forecasts will also be in focus. Together with colleagues, he has investigated how the same type of AI models used to create images can also be used to make time series forecasts and test different scenarios in energy systems.

"Research conducted in recent years indicates that so-called diffusion models show great potential for creating time series-based forecasts, and in some cases outperform commonly used reference models and architectures," says Christian Lind.

Using technology primarily known for creating images to forecast an energy system may seem far-fetched. However, this is an excellent example of machine learning's flexibility and generalization ability across different areas and problems.

"It will be exciting to see how these models develop in the future; there is still much to explore around diffusion models – especially in the time series domain. For example, you can do scenario analyses, which can help the industry transition to emission-free power grids."

Speakers from other sectors are also invited to the conference because of the possibility of transferring knowledge and technology between industries. One of them is Ulrika Jägare, Head of AI, Data & Architecture at Scania:

"Using data and AI internally at Scania and in our commercial portfolio within digital services, for example, is vital to understanding and improving our customers' profitability while helping them achieve their sustainability goals. Succeeding with an AI initiative is much about understanding your company's role in the ecosystem, but also about being open and daring to learn from other actors in other industries," says Ulrika Jägare, Head of AI, Data & Architecture at Scania.

Filip Kjellgren at AI Sweden sees enormous potential in increased collaboration around data sharing and artificial intelligence in the Swedish energy industry:

"Our energy system is built up of 170 small and large electricity grid areas interconnected in real-time. Each actor can improve their systems using AI. Still, the huge gain for society occurs when actors implement AI solutions that reinforce each other locally, regionally, and nationally," he says and continues:

"There are many actors who are relatively similar in terms of mission and infrastructure and who all have limited expertise, money, and work hours. For these actors, it's advantageous to develop certain solutions together by 'pooling' the resources they have to get greater leverage from them."