Bernd Noack


Bernd Rainer Noack is a German physicist.
His research and teaching area is closed-loop flow control for transport systems. Focus is placed on machine learning control and model-based nonlinear control using reduced-order modelling and nonlinear closures. Currently investigated configurations include wakes, mixing layers, jets, combustor mixing and aerodynamic flows around cars and airplanes. Another main area is thermodynamic formalisms for turbulence modeling.

Life

He received his degree as diplom physicist from the Georg-August-University, Göttingen, in 1989. He stayed on to receive his physics doctorate in 1992 under Helmut Eckelmann. In the sequel, he had positions at the Max-Planck-Institut für Strömungsforschung, Göttingen, the German Aerospace Center, Göttingen, the Göttingen University, and the United Technologies Research Center before he joined the Berlin Institute of Technology. There, Professor Noack has headed the group "Reduced-Order Modelling for Flow Control" at the School V "Transport and Machine Systems". Currently, he is Director of Research CNRS at Laboratoire d'Informatique pour la Mécanique et les Sciences de l'Ingénieur, Paris-Saclay, France and Professor at the Technische Universität Braunschweig, Germany, Honorary Professor at the Berlin University of Technology and Visiting Professor at the Harbin Institute of Technology, China.

Scientific work

Focus of Noack's work are physical theories and mathematical methods for turbulence control. One direction is the development of control-oriented nonlinear models
and associated control design building on the Galerkin method, originally proposed by Boris Galerkin. He proposed the first mathematical Galerkin model for the two- and three-dimensional cylinder wake from Hilbert space considerations. Subsequent works employ the proper orthogonal decomposition and propose numerous enablers accounting for the pressure term, subscale turbulence and departures from the training set.
He has distilled three major facets of nonlinearity in dynamical least-order models:
The associated control laws can be derived from energy considerations and have been applied to streamlined and bluff bodies.
Recently, Noack works on implementing the powerful methods of machine learning in turbulence control. Major breakthroughs are learning the control law in real-world experiments with Machine learning control and an automated learning the control-oriented dynamical gray-box model from experimental data.
Supplementary projects include data visualization, phenomenological models, vortex models and entropy-based optimization in addition to a spectrum of model-free and model-based control approaches. The breadth of this research builds on a network of cross disciplinary collaborations at his affiliations in Paris-Saclay, Berlin, Braunschweig, Shenzhen and international teams.

Text books and review article

*