Modal Reduction of an Advection-Diffusion Model Using a Branch Basis

Abstract

We propose an original method to reduce an advection-diffusion model in which parameters, as well as boundary conditions, are time-dependent. This modal method uses a branch basis, which differs from the Fourier one by a Steklov boundary condition. The treated application is a disk rotating at a variable velocity, with time-dependent volume and superficial thermal inputs. Comparison between the detailed model and the reduced one gives a gain in computational time of 24 times with a maximal error of less than 10%, opening the way to real-time simulation.