On the development of Langevin models for turbulent pipe flow by 3D PTV and DNS

A stochastic model for fluid particle dispersion in turbulent pipe flow is presented. A direction and position dependent Kolmogorov parameter, for high Reynolds numbers equal to the familiar Kolmogorov constant, is proposed and the dependencies are determined from results from Lagrangian experiments and direct numerical simulations for a bulk Reynolds number of 5,300. The drift tensor in the Langevin equation is determined from the well-mixed criterion, which leads to a unique formulation with the aid of essential information from Lagrangian experiments and DNS results. Eulerian and Lagrangian particle velocity statistics are in good agreement with experimental and DNS results.