Sparse Nonlinear Models for Fluid Dynamics with Machine Learning and Optimization

Reduced-order models of fluid flows are essential for real-time control, prediction, and optimization of engineering systems that involve a working fluid. The sparse identification of nonlinear dynamics (SINDy) algorithm is being used to develop nonlinear models for complex fluid flows that balance accuracy and efficiency. We explore recent innovations related to several complex flow fields: bluff body wakes, cavity flows, thermal and electro convection, and magnetohydrodynamics. Papers in order: 1. Original SINDy paper: 2. SINDy for PDEs: 3. SINDy Autoencoders: 4. JC Loiseau’s Galerkin Regression Paper: ( arxiv: ) 5. Loiseau, Noack, SLB, SINDy on Lift and Drag: 6. Loiseau thermosyphon paper: 7. Guan, SLB, Novosselov, electrocon