Prestigious EU funding for TU Ilmenau: Artificial intelligence to study solar turbulence
The fluid dynamics researcher at Technischen Universität Ilmenau Professor Jörg Schumacher has been awarded an ERC Advanced Grant, the European Union's most prestigious research funding, for his excellent cutting-edge research on fluid turbulence. With the funding of 2.5 million euros for five years, the scientist can now use super and quantum computers to calculate and analyze flow patterns, such as those that occur near the surface of the sun. The increased use of artificial intelligence in this area of research could lead to improved models of solar activity and more precise weather and climate forecasts in the future.The MesoComp project ("Order at the Mesoscale: Connecting supercomputing of compressible convection to classical and quantum machine learning") at TU Ilmenau is expected to start in January 2023.
Mesoscale convection is the buzzword of the cutting-edge research of Prof. Jörg Schumacher,Head of the Fluid Mechanics Group at the TU Ilmenau and Fellow of the American Physical Society. In meteorology, we speak of a mesoscale convective system when, for example, thunderstorms combine to form a larger thunderstorm complex that exists for several hours. Because the turbulence in the atmosphere is seemingly chaotic and changes rapidly, it is still difficult today to predict such gigantic weather phenomena quickly and reliably. In global circulation and climate models that describe the long-term evolution of the entire Earth's atmosphere, these mesoscale convection processes still fall through the coarse computational grid with which scientists have spanned our globe. In the global circulation and climate models, they are incorporated, greatly simplified, as additional heat fluxes. It is precisely this modeling of the processes that the MesoComp project aims to improve. To do this, however, the scientists first have to find out the cause of the seemingly orderly flow patterns.
What thunderstorm clusters or cloud streets are in the Earth's atmosphere are near the surface of the sun so-calledgranules. Together they form huge, 30,000 kilometer large flow cells with a lifetime of one day. At the edge of such supergranules, thick bundles of magnetic field lines erupt from the interior of the sun into the solar corona. During solar storms, large amounts of energetic particles then enter the solar wind in the direction of the earth and can cause great damage there: Satellites can be destroyed and even entire electrical grids can be brought down. The occurrence of these solar activities also cannot be reliably predicted at present, since the formation of the supergranules by mesoscale convection is not yet understood.
To enable faster and more accurate predictions of the behavior of such flow turbulence, Prof. Jörg Schumacher is bringing order to the chaos. Using numerical research models, he uses high-performance computers to analyze what experts call "ordered convection patterns" clouds and granules. But suchmesoscale convection modeling is still the largest source of uncertainty for more accurate predictions of global warming and, consequently, the frequency of extreme events. For computational tasks of the dimension that the Ilmenau scientist wants to solve, even today's high-performance computers reach their physical limits.
The use of artificial intelligence and machine learning now offers fluid dynamics research completely new ways of not only thinking about effective modeling, but also carrying it out - and thus gaining a whole new understanding of fundamental turbulent flows. Theoretically, quantum computers, which are many times more powerful, can handle more complex tasks in less time than conventional computing clusters. For his research work, Prof. Schumacher will also use the quantum computer installed by the IT corporation IBM in Ehningen, Baden-Württemberg, and operated by the Fraunhofer Competence Network Quantum Computing - the "most powerful quantum computer in Europe" according to IBM.
The European Research Council (ERC) Advanced Grant supports scientists in all disciplines who have already established themselves as leading researchers and have a recognized track record of research achievements. Unlike most other research funding, it is awarded to individuals rather than to research groups. ERC grants are 100 percent grants that cover everything that goes into a project: personnel costs, equipment, consumables, travel, publications, subcontracting, etc. Flow researcher at TU Ilmenau Prof. Jörg Schumacher prevailed with his MesoComp project in a two-stage selection process against more than 1,700 submissions from all over the EU from all scientific fields. Only one third of the submissions were admitted to the second stage, and only 253 projects will now be funded - including Prof. Schumacher's project.
Whether the high expectations in the new technology quantum computing can be fulfilled is another question that Professor Jörg Schumacher wants to clarify in the MesoComp project. For although the computer chips are already cooled below minus 273 degrees Celsius to prevent all possible noise effects, quantum computers are still very susceptible to interference. But technical progress in this field is rapid. The German government alone is investing more than two billion euros in this technology of the future over the next few years. Prof. Schumacher is confident that he will be able to use artificial intelligence methods to develop data-driven models of mesoscale convection that will enable more precise and faster predictions: "Will climate models be running on quantum computers in ten years' time? The technology is still in its infancy and the research involved is risky. But I wouldn't try it if I didn't believe in it."
Prof. Jörg Schumacher
Head of Fluid Mechanics
+49 3677 69-2428