Abstract

This talk aims at presenting research topics in a broad sense. The story line always begins by some uncertainty analysis and Gael will present how, from the interpretation of some uncertainty propagations, he has worked - on accelerating hypersonic reentry simulations using machine learning based hybridization (first part of the talk), - on a suggestion to cancel what is commonly known ad the teleportation error for the Monte-Carlo resolution of the radiative transfer equations (second part of the talk). I chose to present the first topic because the difficulty I encountered for chemical reactions in reentry problem is certainly close to the one encountered when dealing with neutronic reactions in nuclear ones. This part mainly deals with model errors. He has chosen to present the second topic because it gives a good idea of his work on discretisation errors.

Join us for an interesting talk with tea and cakes to start!