(522c) Analysis of the Micro-Mixing Time in FDF Methods

In reactive systems, chemical reactions can only take place after reactants are mixed at molecular level. In many industrial and biological applications taking place in liquid phase, the ratio of the kinematical viscosity to molecular diffusivity of the scalar is much greater than one, so the scalar field holds finer structures than the velocity field. Measurements of instantaneous concentration in turbulent liquid flows are hard to achieve with good accuracy. Numerical simulation is an important tool for predicting difficult-to-measure information of flow and scalar fields. Direct Numerical Simulation (DNS) is the most direct way to get information about the fluid flow but the computational cost is prohibitive in most engineering applications. A more realistic alternative to DNS is the utilization of spatial filtering or temporal ensemble averages such as Large Eddy Simulation (LES) and Unsteady Reynolds-Averaged Navier Stokes equations (U-RANS). In both approaches, additional unknown terms appear in the final form of the balance equations and they must be modeled in order to close the numerical set of equations. Different approximations for closing the source term are available during LES of reactive systems. One is Filtered Density Function (FDF) which holds the advantage that the chemical reaction term appears in a closed form. The conditional diffusion term in the FDF transport equation accounts for transport in physical space and mixing in the composition space and it must be modeled by mixing models such as Interaction by Exchange with the Mean (IEM), Modified Curl (MCURL) and Eucledean Minimum Spanning Tree (EMST). Experimental results show that the subgrid mixing time constant is not universal, with measured values varying between 0.6 and 3.1. In addition, molecular transport properties are not taken into account in these models. The non-universality of this constant can affect the predictive capability of LES. In this work, a LES- FDF approach is carried out in order to reproduce the experimental results of a turbulent mixing process.

The solution of the FDF transport equation is done via Monte Carlo method. A sensibility analysis of the number of stochastic particles is evaluated in this work, as well as different strategies for the estimation of the mixing time. Simulation results are compared to experimental data from a confined turbulent liquid jet, discharging a conserved scalar into a low-velocity water stream