Degree Type

Dissertation

Date of Award

2017

Degree Name

Doctor of Philosophy

Department

Aerospace Engineering

Major

Aerospace Engineering

First Advisor

Paul A. Durbin

Abstract

The work described in this dissertation, follows the attempt made in Reddy et al. (2014a), to make Detached Eddy Simulation model more like traditional LES in eddy simulation region. Work done by Reddy et al. (2014a) proposed the l^2w DDES model that shares a similar formulation with Smagorinsky model in eddy simulation region.

In the present research, an adaptive procedure was devised (Yin et al., 2015), to allow automatic adjustment of a model coefficient CDES to flow condition and grid resolution. The adaptive method is based on the Germano identity, and on a lower limiting value that is a function of the grid resolution and the Kolmogoroff length scale. The function, being a gauge for grid resolution, allows the model coefficient to be computed dynamically, wherever suitable.

To extend adaptive DES to compressible flow and heat transfer, a passive scalar transport model is proposed for Hybrid RANS/LES (Yin and Durbin, 2016b). This too is an adaptive model. Adaptivity is based on computing test-filter fluxes. The formulation proves to be especially effective on coarse grids, as occur in DES.

Under the principle that DES should converge to wall-resolved LES as the mesh becomes fine near the wall, a modification is made on the adaptive DES model to make this limit feasible (Yin and Durbin, 2016a). The modification is to the limiting function. It is found that the RANS region shrinks to y+ ∼ 5 on fine meshes, thus allowing the model to be almost equivalent to wall-resolved LES. One place where the wall-resolved asymptote can play a role is in laminar to turbulent transition. Both the original and modified formulation are tested in orderly, bypass, and separation induced transition.

Three separated test cases are also included here: a series of 3-D diffusers, jet in cross flow, and rotating channel flows | for more elaborate testing of proposed model. The 3-D diffuser series, reveals that the adaptive method in Yin et al. (2015) has discrepancies with LES even on fine meshes, which partially motivates the revised model (Yin and Durbin, 2016a). The JICF test case validated both the passive scalar transport model in Yin and Durbin (2016b) and the revised model in Yin and Durbin (2016a). And rotating channel flow gives a more detailed assessment of the revised model.

In summary, adaptive DES and passive scalar transport models are proposed. They adapt to flow and geometry. The passive scalar transport model is compatible with both wall-resolved LES and hybrid RANS/LES models.

Copyright Owner

Zifei Yin

Language

en

File Format

application/pdf

File Size

164 pages

Share

COinS