09:00
Invited Lecture: Prof. Steve Tobias, "Direct Statistical Simulation of Turbulent Astrophysical Flows"
Chair: Szymon Malinowski
09:00
60 mins
|
DIRECT STATISTICAL SIMULATION OF TURBULENT ASTROPHYSICAL FLOWS
Steve Tobias
Abstract: Flows in astrophysics are often turbulent owing to the extreme values of the Reynolds numbers. A description of these flows via direct numerical simulation (DNS) would therefore have to be able to resolve a huge range of spatial and temporal scales, which is clearly beyond the capability of current algorithms and computational resources (and indeed those of the immediate future). However
astrophysical objects often display remarkable organisation so that non-trivial mean flows and mean magnetic fields are apparent (which is convenient since often these are all that can be observed). The jets on Jupiter, the differential rotation of the Sun and the solar activity cycle are all examples of this "order from chaos". Owing to the presence of rotation, stratification and (usually) magnetic fields, astrophysical flows (at both large and small scales) are often (if not always) inhomogeneous and anisotropic. In this talk I shall give some examples of turbulent astrophysical flows and describe a new programme, which we term Direct Statistical Simulation (DSS) that attempts to calculate directly the low-order statistics of such flows (such as mean flows and two-point correlation functions). DSS respects the inhomogeneity and anisotropy of the astrophysical systems, and is predicated on an expansion in cumulants. I shall further discuss generalisations of the quasi-linear approximation that lead to new closure schemes for astrophysical flows. Finally I shall take my life in my own hands by suggesting that some of these methods may be of interest to those studying wall-bounded shear flows.
|
|