Data-Driven Modeling of Haemodynamics in Vascular Flows

This lecture covers the Center for Intelligent Systems (CIS) at EPFL, focusing on research in intelligent systems. It delves into reduced order modeling for blood flow in large arteries, emphasizing the computational challenges and solutions for fluid-structure interaction (FSI) problems. The presentation showcases numerical assessments of reduced FSI models and the application of reduced basis methods. It also explores the use of neural networks for data-driven approximation of haemodynamics, with a particular focus on femoropopliteal bypass applications. The lecture concludes with discussions on enriching neural networks with PDE solvers and the development of physics-aware models for cardiac electrophysiology.