Prof. Herman Clercx,
Department of Applied Physics,
Eindhoven University of Technology
When do rain drops form in clouds? Where do algae congeal in the sea? How fast do droplets form on the walls of the housing of a reactor? Prof. Herman Clercx of the Eindhoven University of Technology (TU/e) tries to answer these questions with simulations on the SURFsara supercomputer. The aim? To create computing models for forecasting purposes and to improve the modeling of small-scale processes. The result? More precise weather forecasts and safer and more efficient business processes.
Put briefly, Clercx studies currents, or turbulence: “A cloud is made up of turbulent air currents, called vortexes, which contain droplets. A typical cloud can have a diameter of one kilometer. But within the cloud there can be many smaller-scale turbulent vortexes, only meters or even centimeters in size, dragging along the droplets in the cloud. There is a correlation between large-scale dynamics and the smaller-scale dynamics. Therefore, to understand what is going on on the larger scales, we need to know exactly what is happening on the smaller scales. When we have that information, we can then create models for large-scale simulations.”
A challenge, according to Clercx: “For example, we want to be able to calculate how fast droplets grow until they are large enough to fall. If we know that, we’ll be able to predict when it’s going to start raining. Collisions between droplets are an important factor in this process because droplets have to come together first to form larger drops. But droplets are only likely to meet and collide if their number is extremely high. This presents us with a real challenge: only a combination of huge computational power and advanced algorithms will allow us to include all these particles in the simulation. And that is why SURFsara’s supercomputer is indispensible to our research.”
Clercx and his colleagues’ research is put to practice in a variety of fields: “We study turbulence in oceans and in the atmosphere, but also currents in industrial equipment. Take, for example, mixing processes. If we understand how currents of specific substances move, we can also better understand how to mix them. And that results in more efficient processes.”
Research into turbulence also plays a role in boosting safety: “Temperature differences will cause turbulence in the housing of a nuclear reactor. If there is leakage in the reactor, steam can appear in the housing of the reactor. That steam can escape, creating a dangerous situation. Therefore, it is important that we can predict how fast and where the droplets form on the walls. On the walls, the steam can no longer escape and no longer forms a direct threat to the environment.”
This type of research is only possible using a supercomputer: “Without SURFsara we would not be able to perform massive parallel computations. We have our own computer system for test computations, but for the really complicated computations we use SURFsara. We would also like to expand our research into particles in turbulence in the future as part of the Dutch Computing Challenge Project (DCCP). We have simulations planned with very high resolutions and massive numbers of particles. These simulations result in data files for statistical analysis. We need to use a significant portion of the supercomputer’s capacity for a limited period. Additional optimization of the software is required to get the computational process to run as smoothly as possible.” Clercx continues: “For research in the Netherlands, it is important that facilities like this one are easy to use. And that is certainly the case with SURFsara.”
A true-color image of algae in the English Channel to the south of Cornwall.