UBC Math Department Colloquium: Anthony Wachs

Particle-laden flows are ubiquitous in many environmental flows and man-made processes. Their intricate dynamics is governed by momentum transfer between fluid and particles and is strongly multi-scale. I will discuss mathematical, computational and modelling issues related to the accurate, fast and realistic computing of particle-laden flows. Specifically, I will elaborate on the three following problems: (i) a massively scalable and fast Navier-Stokes solver on Cartesian grid, (ii) an extension of a distributed Lagrange multiplier/fictitious domain method to octree adaptive grid and (iii) an attempt to design realistic drag law models using a stochastic approach and machine-learning in the context of a multi-scale analysis of particle-laden flows.