Advances in Scientific Machine Learning for Coupled Fluid Flow and Transport
This chapter reviews recent advancements in Scientific Machine Learning (SciML) for modeling complex fluid flow and transport phenomena. It surveys methods like Singular Value Decomposition, Dynamic Mode Decomposition, Physics-Informed Neural Networks (PINNs), and $\beta$-Variational Autoencoders ($\beta$-VAEs) for creating efficient surrogate models. The work combines these with High Performance Computing strategies and presents new contributions in modeling turbidity currents via PINNs and extracting disentangled nonlinear modes from thermal flows using $\beta$-VAEs. The chapter demonstrates how SciML can provide fast, accurate approximations of complex coupled systems at a substantially reduced computational cost. AI
IMPACT Enhances the efficiency and accuracy of complex fluid dynamics simulations through advanced machine learning techniques.