Machine Learning for Turbulence Control (Prof. Bernd R. Noack) – Part 4
This lecture was given by Prof. Bernd R. Noack, Harbin Institute of Technology, Shenzhen, China and TU Berlin, Germany in the framework of the von Karman Lecture Series on Machine Learning for Fluid Mechanics organized by the von Karman Institute and the Université libre de Bruxelles in February 2020. Closed-loop turbulence control has current and future engineering applications of truly epic proportions, including cars, trains, airplanes, jet noise, air conditioning, medical applications, wind turbines, combustors, and energy systems. A key feature, opportunity and technica l challenge is the inherent nonlinearity of the actuation response. This lecture outlines model-based and model-free control taming the nonlinear dynamics. Modelbased control employs the POD Galerkin method lectured earlier. Many flow control results will be explained with a set of few dierent sparse human-interpretable models. Artificial Intelligence (AI) / Machine Learning (ML) has opened another game-changing new avenue: the automated model-free discovery and exploitation of unknown nonlinear actuation mechanisms directly in the plant. Variants of this machine learning control (MLC) will be discussed. The lecture concludes with a tutorial of MLC on a simple dynamical system with a freely available matlab/octave code.