Column by Professor Henrik Madsen published in Energy Supply, March 2024
We will only achieve the green transition if we rethink our entire energy system. Of course, it's about integrating energy systems—sector coupling—which is vital. This requires digitalization at all stages and using AI to manage processes and data. I won't comment any further on sector coupling, but instead focus on the rethinking of two key areas: flexibility and price structure—as well as the use of AI for this purpose. This applies to the entire energy area, but in the following I will use examples from the electricity sector which is at the most advanced stage furthest in the area.
About flexibility. With an ever-increasing share of renewable energy sources, we have long known that the energy system can no longer be balanced by the producer, but instead must be moved to the end user. It is you and me, businesses, and all other users who must provide the necessary flexibility. Basically, many consumers were aware of the electricity prices during the energy crisis in autumn 2022, when we learned to charge our electric car and run the washing machine at night. This flexibility must be utilized and systemized.
Solutions that automate consumption are already out there, but if AI is to help us make energy consumption sufficiently flexible, we need solutions that work in both spatial and temporal hierarchies, for example at all geographical levels and in time. This means that AI must be integrated into solutions both for the individual device in our household or business, for our street, for our entire area, our city, region, etc. In addition, it must be able to operate in time—both with historical data and with predictions of future consumption. And, most importantly, AI and hierarchies must be used to couple and exchange data between all the different levels at once.
This should take into account the special needs and priorities of each individual consumer. For example, as a private household, we may require our car to be fully charged at 8 a.m. the next morning before going to work. Businesses also have specific needs that need to be addressed in the distribution of energy. In a specific research project, we have worked with solutions adapted to wastewater treatment plants where the price of electricity is not the main priority, but rather preventing spills and overflows in connection with cloudbursts—no matter how expensive power may be at the time. The next priority is maintaining the wastewater treatment function (active sludge), and the desire to buy the energy at low cost is only the third priority. The priorities and wishes of consumers and businesses must, of course, be addressed in the solution that the local distributor, the Distribution System Operator (DSO), and the other stakeholders offer consumers and the treatment plant, respectively.
The other area that we need to have turned upside down and innovate is the overall energy pricing. Again, taking the electricity sector as an example, the current price structure is far too rigid. Today, our electricity prices are set by the daily prices fixed for the next 24 hours in the electricity market. This is done in collaboration between countries, for example Nord Pool between Denmark, Sweden, Finland, and Norway. Having fixed prices overall at country level is fine, but when electricity is closer to individual businesses and consumers, and when we approach the time of use, using market mechanisms does not make sense. Instead, prices must be much more dynamic and adapted to the needs, priorities, and flexibility of electricity system stakeholders, such as DSOs, TSOs (Transmission System Operators), and balance managers.
Therefore, we must rethink the entire price structure with an AI-based solution that will ensure transparency and fairness, and which often will make electricity significantly cheaper than we as end users experience today, where it follows the prices for 'day-ahead' markets and the established electricity tariffs.
It may sound like something of the future, but in reality these solutions can be implemented very quickly—both in terms of flexibility and a dynamic pricing structure. We have already developed the tool. A 'smart energy operating system'—just like the operating system you know from your computer. A system that works the same way by building different functions, aggregates, on top of the operating system, just as you know it with, for example, the Office suite, web browsers, etc. on your PC.
In order for AI to be used, a large amount of data is obviously needed from all parts of the energy system. Much of this data is also already available, for example in the Center Denmark data platform, which collects data from energy production and consumption from a large number of stakeholders. A platform that has shown that it is possible to assemble all links in the consumption chain and thus enable rethinking the energy area using AI. The alternative is very expensive infrastructure investments. We have shown that using AI and data-driven hierarchical solutions based on our Smart Energy OS to enable end-user flexibility, we can save up to 80 per cent of this investment requirement.
I therefore recommend that we in Denmark take the final necessary steps to get the legislation on setting energy data free in place, so we can design the final data-driven AI solutions for implementation based on the energy operating system. And it's somewhat urgent. The EU has come far advanced in terms of linking digitalization and the green transition with countries such as Denmark, but also the Netherlands, Austria, and Belgium at the forefront. Now that we have the solutions to be able to move quickly, we must ensure that Denmark can maintain a leading position. The solutions, which can therefore ideally be developed and tested in Denmark, will subsequently be the key to effectively accelerating the necessary transition to renewable energy on a global scale.
