Neural networks in feedback for flow analysis and control

In this work we propose a machine learning methodology for flow modeling and control design based on an iterative approach for training neural networks. We demonstrate that the methodology is able to achieve stabilization of complex nonlinear plants, such as an unstable confined flow past a cylinder. We also show that, through linearization of neural network models, we can use the methodology to conduct optimal sensor selection, as well as to perform unstable equilibrium estimation and stability analysis.