Fighting blackouts with mathematics
Florian Dörfler develops algorithms that keep our power grids stable using mathematics and, as he says, a strong willingness to take risks. He has now been awarded the Rössler Prize, the most prestigious honour for young professors at ETH Zurich.
“We’ve already pulled off some really big things. One example is the data-driven control of a 100-kilowatt inverter using AI,” says ETH professor Florian Dörfler. “If this control system fails, the inverter will overheat and the entire building where it is located could catch fire.” He finds such safety-critical applications that allow no margin for error particularly fascinating. “If a robot falls over, it doesn’t matter. But if a building burns down, it does.”
Dörfler’s field of expertise is control engineering. Yesterday, he was awarded the Rössler Prize, the most valuable funding award for outstanding young professors at ETH Zurich (see text box).
Before his controls are applied in inverters or other energy systems, Dörfler starts with the mathematical foundations. He is interested in the fundamental approaches: how systems can be optimised using models or data and AI; how disruptive oscillations in a system can be prevented through real-time feedback control; and how the subsystems of a highly complex system can be synchronised.
What begins as a theory with abstract mathematical proofs or an algorithm often ends up solving a real practical problem after several years. Dörfler feels proud, not only because he contributed to solving the problem, but also because he and his team successfully took it from theory all the way to practical application.
Algorithm keeps power grid stable
An example of such an application is the algorithm used by Aargauer Elektrizitätswerke, a regional energy provider in northern Switzerland, to feed solar power into the grid. It involves the distribution of active power and reactive power. If solar power were fed into the grid only as active power, voltage spikes would occur on very sunny days, causing the grid to overload.
In principle, building additional power lines or battery storage could provide a solution. But that costs money. Dörfler’s solution worked entirely without additional infrastructure, making it much more cost-effective. He integrated an additional algorithm into the control software at the electricity operator's control centre. His approach measures solar power output, current and voltage in real time, and then calculates the amount of active and reactive power to supply to the grid. His solution received the Watt d’Or last year, a Swiss federal award for excellence in the energy sector.
“The algorithm keeps the power grid in Aargau stable. It’s based on theoretical and mathematical research we began eleven years ago,” explains Dörfler. He is now in discussions with energy companies in Germany and France to implement his simple solution there too.
Taking risks to tackle big questions
Dörfler aims to address the big questions with his research. One of these questions is: How can the European power grid be made ready for a future with more and more electricity from wind and solar power using smart control systems?
He recently patented an algorithm for a fault-resisilient control system for inverters in wind and solar power plants. Dörfler says that this smart technology could greatly reduce the risk of large-scale blackouts like the one in Spain two months ago.
He encourages his team, including master’s students, PhD candidates and postdoctoral researchers, to tackle big research questions that others avoid because success is uncertain. “The best people in the world come to ETH, and I want to give them as much freedom as possible.” He tells them that for two years he expects nothing from them except that they think about the important big problem they want to solve and how to express it mathematically.
He is often asked whether this approach is too risky. Dörfler says he is someone who regularly takes risks both at work and when doing his mountain sports, which he pursues whenever he has free time.
Remaining close to the mountains
Dörfler grew up in Upper Bavaria. As a boy, he regularly went to the mountains to ski, ride his mountain bike and go rock and ice climbing. He describes himself on his social media profile as a “Professor at ETH Zurich and wannabe dirtbag climber,” someone who would love to travel the country in a camper van and dedicate his life to climbing. “If I lost my job overnight, I'd do exactly that.”
His social media profile shows him climbing North America’s most famous icefall, Widow’s Tears in Yosemite National Park. That was in 2013 when he was working towards his doctorate in Santa Barbara, California. He also climbed the Aiguille du Dru in the Mont Blanc massif, following a very obscure and difficult route. The route, called “American Direct”, climbs a 1,000-metre wall that rises almost to 4,000 metres above sea level.
There are parallels between science and mountain sports. Both attract similar characters: disciplined people who are willing to take risks, endure hardships and keep going even when it gets tough. “Risk-taking and persistence are needed in both science and mountain sports.”
Interestingly, many, though not all, of his team members have previously competed at a high level in sports such as skiing, climbing or other disciplines. The team even goes skiing or climbing together. “Many of the people I recruit have climbed before, but definitely all of them climb afterwards,” he says with a wink.
Dörfler is also aware that his risky research approach would not be possible at many other institutions besides ETH Zurich. This is because it requires solid basic funding for research, such as that provided by ETH. When other researchers apply for third-party funding, they already need to know in advance how they plan to use the money. They also spend a lot of time writing project reports and meeting project milestones. In his opinion, too much structure hinders creative, risky, and blue-sky research.
Florian Dörfler’s award of the Rössler Prize shows that his path has been successful and that taking risks has clearly paid off. He plans to use the prize money to do something special with his team. Will it be a trip to the mountains? He is, in any case, very pleased about the increased visibility that his research field and team have gained thanks to this prize.
The Rössler Prize
Max Rössler gifted 10 million Swiss francs to the ETH Zurich Foundation in 2008. He donates the interest from this sum in the form of an annual sponsorship award for ETH professors in the expansion phase of their research careers. The Rössler Prize is worth 200,000 Swiss francs, making it the most highly endowed research award at ETH Zurich. It is conferred annually at the ETH Foundation’s “Thanks Giving” event. The prize founder studied mathematics at ETH Zurich and wrote his doctorate in orbit calculations in space travel. Following a period as a visiting researcher at Harvard University, he returned to ETH Zurich, where he was Senior Scientist and Lecturer at the Institute for Operations Research from 1967 to 1978. He later worked in wealth management before retiring from business life. He was named an Honorary Councillor by ETH Zurich in 2013.
More information about the external page Rössler-Preis