Variable Turbulent Schmidt Number Research Articles

A fully dynamic mixing time‐scale model for the sparse Lagrangian multiple mapping conditioning approach

Multiple mapping conditioning / large eddy simulation (MMC-LES) is a promising model for turbulence-chemistry interactions, successfully applied to simulate different combustion regimes involving gaseous, liquid, and solid fuels. MMC-LES attempts to produce accurate molecular mixing by localising mixing in an independent, composition-like reference space. In the MMC mixing model, mixing between two stochastic particles is defined based on the Lagrangian mixing time-scale τL, which is defined as the ratio of the Lagrangian subgrid-scale mixture fraction variance (fL′2˜) and the scalar dissipation rate (χL), typically modelled algebraically, based on fixed values of global constants at all spatial locations. Accurate modelling of τL is essential towards predicting the conditional scalar fluctuations, which is otherwise highly challenging if the same set of static constants are used at all spatial locations for complex flames with varying extinction/re-ignition. Herein, a fully dynamic mixing time-scale model (dyn-aISO) is proposed within the MMC-LES framework, which dynamically evaluates fL′2˜ and χL, essentially based on a local variance coefficient and variable turbulent Schmidt number, respectively, eliminating the need for specifying tuneable model constants. This novel dyn-aISO model is used to simulate the pilot-stabilised Sandia flame series with increasing local extinction levels. The conditional and unconditional profiles of compositional scalars from the new dynamic model are compared with those from a fully static model (st-aISO) and a model which uses a static constant for fL′2˜ (stCf-aISO). Numerical results from different timescale models deviate from each other for Sandia flame F having a high jet Reynolds number. With the proposed improvement of the micro-mixing time-scale model, the dyn-aISO model accurately captures the transient behaviour of the complete flame series, which could otherwise be obtained by the tuning of static constants. The role of molecular and turbulent mass diffusivities, coefficient of residual variance of mixture fraction, mixing distance, and the mixing time-scale on the predictions of the reactive scalars is also explored.

CFD simulation of dense gas dispersion in neutral atmospheric boundary layer with OpenFOAM

In this study, Monin–Obukhov similarity theory is used to specify the profiles of velocity, turbulent kinetic energy (k), and eddy dissipation rate ($$\epsilon$$) in atmospheric boundary layer (ABL) flow. The OpenFOAM standard solver buoyantSimpleFoam is modified to simulate neutrally stratified ABL. The solver is able to obtain equilibrium ABL. For gas dispersion simulation, buoyantNonReactingFoam is developed to take into account fluid properties change due to temperature, buoyancy effect, and variable turbulent Schmidt number. The solver is validated for dense gas dispersion in wind tunnel test and field test of liquefied natural gas vapour dispersion in neutrally stratified ABL.

A variable turbulent Schmidt number formulation by numerical simulation of atmospheric plume dispersion

Turbulent Schmidt number as an important parameter in computational fluid dynamic (CFD) simulations is strongly dependent on height, whereas it is mostly considered to be constant in the literature. This paper presents a new variable turbulent Schmidt number formulation which can calculate the relative concentrations (RCs) in neutral atmospheric conditions more accurately. To achieve this aim, RCs from continuous releases are calculated in different distances by the analytical Gaussian plume mode. CFD simulations are carried out for single stack dispersion on a flat terrain surface and an inverse procedure is then applied so that different turbulent Schmidt numbers are used as inputs to determine the RCs to select the “best-fit” turbulent Schmidt number value. This process is continued for different heights to fit a curve to obtain the new formulation for turbulent Schmidt number varying with height. The values are compared with experimental results. The comparison indicates that the new formulation for turbulent Schmidt number is more accurate and reliable than previous research works.

Scalar Fluctuation Modeling for High-Speed Aeropropulsive Flows

A Reynolds-averaged Navier-Stokes-based scalar-variance model is described that extends a previous low-speed nonreacting jet model to more generalized high-speed compressible reacting flows. The model is cast in a k-e turbulence model framework. Transport equations for energy variance and its dissipation rate are solved to predict temperature fluctuations and provide a thermal time scale for use in calculating a variable turbulent Prandtl number. For multispecies problems, mixture-fraction variance and dissipation rate equations are solved that predict species concentration fluctuations and provide a species mixing time scale for use in calculating a variable turbulent Schmidt number. The formulation accounts for compressibility and near-wall damping effects. A series of high-speed flow simulations are presented for both nonreacting and reacting configurations and the predictions are compared to available measured data and companion LES calculations. Results demonstrate the models' capabilities over a range of conditions and suggest that the proposed formulation will provide improved predictions in practical high-speed aeropropulsive configurations of interest, such as scramjet combustors, where turbulent Prandtl and Schmidt numbers vary substantially.

Diffusion from a line source in a neutral or stably stratified atmospheric surface layer

Diffusion of pollutants downstream of a crosswind line source in a turbulent boundary layer is investigated. A modified Van Driest model of turbulence is used to represent the turbulent diffusion. Effect of thermal stratification on turbulent diffusion is represented by a flux Richardson number. Numerical solutions of the governing boundary-layer partial differential equations are obtained using an implicit finite difference method. Results are presented and compared with experimental data and show good agreement. Both constant and variable turbulent Schmidt numbers are considered.