Mechanical and Aerospace Engineering Department
Cornell University
Abstract-
The modelling of turbulent flows remains a central challenge in fluid
mechanics, and one in which substantial progress has been made on several
fronts in recent years. An approach to the problem which has proved
particularly successful for turbulent reactive flows is based on solving
a modelled transport equation for the joint probability density function
(PDF) of relevant fluid properties, e.g., velocity and composition.
In the numerical implementation of such PDF methods, the fluid is
represented by a large number of computational particles which model fluid
particles (in a particular statistical sense). In this talk, examples are
given of PDF computations of turbulent non-premixed flames which, at a
detailed
level, are in excellent agreement with experimental data. The fluid
mechanics in the PDF approach appears principally in a model for the
velocity and/or acceleration of fluid particles. A new stochastic model
for acceleration is described which is based on an earlier proposal by
Sawford. It is shown that the model: accounts accurately for statistics
obtained from DNS; is consistent with the Kolmogorov hypotheses; and, at
high Reynolds number, tends to the generalized Langevin model for velocity.
In a natural way the model accounts for Reynolds-number effects, and the
tensor coefficients in the model can be obtained directly from DNS data.
GALCIT Home Page
|
|