Future society: The utility company turning up the heat with AI

Leentje Chavatte
Read Time, 3 min.  

One Swedish utility company is undergoing a journey in energy innovation―empowered by dynamic collaborations and AI technology.

Frosty commutes from work in wintery Nordic cities are endured knowing that a toasty home awaits. Keeping city energy consumption steadily regulated is key, not just for an economical and warm home for residents, but to minimise environmental impact as global cities continue to grow.

Utility company Mälarenergi is taking an innovative approach to keeping homes reliably heated in a rapidly changing society. Providing vital infrastructure and core energy services to Västerås, Sweden’s fifth largest city, Mälarenergi aims to provide a more intelligent service to consumers. Using the AI-enabled Azure platform, Mälarenergi is digitising and automating district heating production to best manage the increasing demands on the grid and optimise their energy production.

“I used to say there’s only one constant in the universe―and that is change,” says Henrik Wickström, Project Manager IT at Mälarenergi.
“We need to rethink our industry and our purpose in society. We need to handle that change.”

Mälarenergi is leading the way in transformation of the energy market in the Nordics. Using Microsoft Azure as a platform, the Smart Flow project employs a holistic approach to the digitalisation of their services. Collecting three years of energy data across their entire network, they conducted user studies and collated many other data sources―like weather patterns, distribution times and social behaviors.

“Optimising our production processes, distribution and better understanding our customers’ needs requires powerful digital solutions,” says Henrik on the Mälarenergi company blog. “Microsoft Azure has the solutions we need to gather everything together in one place.”

Digitisation and collaboration, the key to change

Henrik Wickström places a great emphasis on the collaborative process that has been central to their digital transformation. Leveraging intelligent cloud technology is enabling external collaboration by unifying, collecting and managing this huge volume of data from multiple service providers, regardless of what third-party technology that service is using. Being able to collaborate with statistical and system experts, RICE SICS, has propelled the analytical depth of the project to the next level.

By harvesting over a billion data points, comprised of diverse and encompassing statistics, Mälarenergi has constructed predictive AI algorithms within Azure’s suite. The overall objective: intelligently optimise the production of district heating, so as to avoid overproduction while being prepared for high-usage times.

“AI technology is a game changer when it comes to making predictions.”
Henrik Wickström, Project Manager IT, Mälarenergi

The work is extremely complex, and is only achieved with a framework that can handle this enormous amount of data. The algorithms can generate predictions which will then directly inform the power plants, significantly lowering the chance of an oversupply or undersupply.

“The goal is for the customer to be able to purchase a service in the form of comfortable indoor heating that can be individualised.” says Henrik.

The statistical insights generated from the Smart Flow project are being used to empower better decision-making, including industrial customers using these services. With new ways to learn, adjust, improve and continuously work on their roadmap, technology is an important catalyst for Mälarenergi to enact wider industry improvements.

“Either we sit down, lean back in our chair, and let the industry change for us – or we be the one who changes the industry.”

Watch the Microsoft Pivot-Cast

Microsoft Pivot-Cast presents six experts in three conversations, each with its own perspective on digital transformation and artificial intelligence

Articles you may be interested in:

Innovate your business

The Artificial Intelligence podcast: Setting the scene