2017 journal article

PARTICLE-FILTER BASED UPSCALING FOR TURBULENT REACTING FLOW SIMULATIONS

INTERNATIONAL JOURNAL FOR MULTISCALE COMPUTATIONAL ENGINEERING, 15(1), 1–17.

By: S. Srivastava n & T. Echekki n 

co-author countries: United States of America πŸ‡ΊπŸ‡Έ
author keywords: turbulent reacting flows; particle filter; large-eddy simulation; multiscale modeling
Source: Web Of Science
Added: August 6, 2018

The particle filter is used to couple a coarse-grained (CG) deterministic solution for a reacting flow with a fine-grained (FG) stochastic solution. The proposed method investigates the feasibility of implementing a multiscale approach for turbulent reacting flows based on large-eddy simulation (LES) coupled with a low-dimensional fine-grained stochastic solution for the subfilter scales reaction and transport. In this study, a model for the turbulent transport in the FG solution is implemented using the linear-eddy model (LEM), which combines a deterministic implementation for reaction, diffusion, and large-scale transport with a stochastic implementation for fine-scale transport. The solution for the continuity and momentum (the Burgers' equation) equations are implemented in 1D. The filtered densities obtained through the FG and the CG solutions are combined using the particle filter to obtain an updated density for the coarse solution that combines the effects of heat release (from the FG solution) and flow dynamics (from the CG solution). The results demonstrate that the particle filter may be a viable tool to couple deterministic CG solutions and stochastic FG solutions in reacting flow applications.