Vianney Koelman

On 1 February 2022 Vianney Koelman, founding scientific director of the Center for Computational Energy Research, left TU/e and DIFFER. He looks back on his five years at the CCER with pride: ‘We have established a thriving organization where computational scientists work on the novel materials and technologies we so desperately need to make the energy transition a success.’

It was the summer of 2016 when Vianney Koelman, then Vice President Computational R&D at Shell Bangalore, India, decided the time had come for him to retire from his career in industry. ‘After having spent a lifetime in different positions at Shell around the world, I took early retirement to gain some more flexibility.’ A mere few weeks following his return to the Netherlands, the phone rang, he tells. ‘TU/e and DIFFER had decided to join forces in the field of computational energy research, and they asked me if I wanted to become the scientific director of this new center. Since this challenge fitted perfectly with my experience and ambitions, I did not hesitate for a moment and said yes.’

Koelman was no stranger to the Eindhoven computational research community. ‘I had been involved in the set up and management of the large scale NWO/Shell research program ‘Computational Sciences for Energy Research’. TU/e and DIFFER had been awarded several projects within the scope of this ten-year program. As a representative of Shell, I had visited Eindhoven several times to get acquainted with the people and ongoing projects here.’

Need for new technology
The physicist and TU/e-alumnus is a strong advocate of the importance of computational research for the energy transition. ‘What many people do not realize, is that for the energy transition to become reality there is a desperate need for research breakthroughs. Many of the solutions we need simply aren’t available, especially when it comes to energy storage. Computational sciences are crucial to explore the novel materials and processes we need to fill this void.’

The center’s first management team was assigned with the task to give the computational research at TU/e and DIFFER a boost by bringing computational scientists from different departments in contact with energy researchers. Here, the ambition is to use computer power to develop for example better batteries, solar cells, or solar fuels. ‘In essence the NWO/Shell research program had the same focus. But within CCER, the scope is wider, and we collaborate with a larger variety of companies.’

Make the connection
The scientific director soon noted that there was great enthusiasm among computational scientists for energy-related topics. ‘But many of them didn’t have the right contacts to step into this field. So that was our first goal: to physically connect computational scientists to experimentalists who work on energy systems.’ Koelman is a strong believer in the added value of such collaborations. ‘Take solar cells. You can start synthesizing new combinations of elements and investigate the properties of the resulting materials, but that takes months at the least, without any guarantee that the endeavor will become a success. By simulating different combinations with a computer first, you can get better leads and save time. And when you have found some promising new candidate materials on your computer, you can go to the lab and do the experiments. Since we have a rather tight deadline to mitigate climate change and its devastating effects, we cannot do without this acceleration of the fundamental research that still needs to be done.’

The CCER focuses on computational challenges critical to the exploration, identification and early screening of scalable energy conversion and energy storage technologies. The challenge common to almost all such computational energy research is to describe and model a variety of physical phenomena over the full range of relevant length and time scales. ‘This raises the challenge of combining the large-scale laws of classical physics with the small-scale laws of quantum physics. That it easier said than done though. I always compare it with chess: knowing the rules alone doesn’t automatically turn you into a chess grandmaster. As computational scientists, we need to figure out the best way to model for example the interactions between atoms and how those play out in the material’s properties on a macroscale.’

Introducing machine learning
Since its inception CCER has been very strong in molecular modelling, Koelman says. ‘And over the past years, we have built a second capability: the integration of machine learning in computational research. During my time at Shell, I had experienced how machine learning techniques can be of great added value for energy research problems. It is very important for researchers and students to get acquainted with such technologies. When I came to Eindhoven, machine learning was a rather unexplored territory in the field of physical modeling. I am very proud of the bachelor course I developed on machine learning in science, which brings students from many different departments and vastly different backgrounds in contact with machine learning.’

In a mere five years, the CCER has become a thriving organization that is able to sustain itself. ‘Prior to COVID, every Thursday all CCER involved researchers would meet to discuss recent results and plans. These physical connections led to many new projects being granted. That started with small scale projects for one PhD student each, but soon grew into larger scale projects such as the Wax+ project. This is an excellent example of how computational researchers can be of great added value to accelerate the development of solutions for the widespread need to store heat in a cost-efficient way.’

Promising future
After his initial five-year contract ended, Koelman felt it was the right moment to hand over his responsibilities at the CCER to his successor Peter Bobbert. ‘The timing is on point. CCER is at the verge of entering into a new, very promising collaboration with EIRES. Over the past years we have brought together computational scientists and experimentalists and managed to build a strong community. By integrating CCER into EIRES, we can expand our scope even further, and introduce computational expertise to this TU/e institute. That is a win-win for both. There are ample opportunities to make this collaboration a fruitful and useful one, both for science and for society.’