Anisotropy of acceleration in turbulent flows

The third-order Lagrangian stochastic models for fluid-particle hyperaccelerations, which account for anisotropy of acceleration variances in low-Reynolds-number turbulent flows, were investigated. A particle tracking algorithm, which employs the four-point Hermite interpolation in horizontal directions and Chebyshev interpolation in wall-normal direction was used to obtain Lagrangian statistics along fluid particle trajectories. It was observed that influence of inhomogeneities in statistical properties of flow is of secondary importance in determining Lagrangian properties. The results show that apparent universality of parameters in conventional Lagrangian stochastic models was a consequence of truncation at either first or second order.