Development of semi-parametric models of Raleigh-Taylor instability by means of neuronet processing of the results of numerical experiment. (Russian. English summary) Zbl 1131.76329

Nuzhny, A. S.; Rozanov, V. B.; Stepanov, R. V.; Shumsky, S. A.

Mat. Model. 16, No. 7, 21-30 (2004).

In recent times theoretical investigations of the turbulent flows are realizing in two directions: the direct numerical modeling on the base of the hydro-dynamic equations and the parametric model construction based on a statistical averaging of initial equations and on an application of semi-parametric relations closing the system. Here the fundamental hydrodynamic problem – the investigation of turbulent flows arising due to Raleigh-Taylor instability – is considered by use a variant of the second ’parametric’ approach. More exactly, the principally new approach for construction of a semi-parametric model by means of a neuronet analysis of the experimental (numerical) data is proposed. A principle possibility to predict a development of the turbulent flows is demonstrated.

Reviewer: Sergei Georgievich Zhuravlev (Moskva)

MSC:

76F25	Turbulent transport, mixing
76M99	Basic methods in fluid mechanics

Keywords:

turbulent transport; turbulent flows; mixing; neuronets