Séminaires à venir

In this presentation, we will focus on the numerical simulation of liquid-solid systems with phase change, using physics informed neural networks (PINNs). We will start with a brief overview of the PINNs method and show its accuracy on toy models. We will present some heuristics to better understand the possible accuracy of this approach. In a second part, we will present its application to the two phase Stefan problem, which is a mandatory step for more complex phase change problems. If time allows, we will present recent numerical simulation of the melting of a phase-change material (PCM), where natural convection is also taken into account.