Speaker
Description
The numerical representation of fully developed high-Reynolds-
number turbulence, particularly in the magnetohydrodynamic approximation,
represents a formidable challenge. We report on an interesting modelling
approach employing a network-based representation that shares fundamental
characteristics with port-Hamiltonian structures known from model reduction.
We show that this technique is able to reproduce the basic but non-
trivial statistical properties of turbulence. This framework exhibits the
logarithmic scaling of classical shell-models without their strong constraints
regarding e.g. dimensionality or isotropy of the underlying physical system.
The model furthermore allows to study the nonlinear dynamics of turbulence on a fundamental level.
