Thesis (Ph.D.)--George Mason University, 2008. / Vita: p. 178. Thesis director: Zafer Boybeyi. Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Computational Sciences and Informatics. Title from PDF t.p. (viewed June 30, 2008). Includes bibliographical references (p. 168-177). Also issued in print.

Thesis (Ph. D.)--University of Washington, 2005. / Vita. Includes bibliographical references (p. 155-170).

This thesis presents the implementation of several two-equation turbulence models into a finite difference, two- and three-dimensional Navier-Stokes Solver. Theories of turbulence modeling and the historical development of these theories are briefly investigated. Turbulence models that are defined by two partial differential equations, based on k-&amp / #969 / and k-&amp / #949 / models, having different correlations, constants and boundary conditions are selected to be adapted into the base solver. The basic equations regarding the base Navier-Stokes solver to which the turbulence models are implemented presented by briefly explaining the outputs obtained from the solver. Numerical work regarding the implementation of turbulence models into the base solver is given in steps of non-dimensionalization, transformation of equations into generalized coordinate system, numerical scheme, discretization, boundary and initial conditions and limitations. These sections of implementation are investigated and presented in detail with providing every steps of work accomplished. Certain trial problems are solved and outputs are compared with experimental data. Solutions for fluid flow over flat plate, in free shear, over cylinder and airfoil are demonstrated. Airfoil validation test cases are analyzed in detail. For three dimensional applications, computation of flow over a wing is accomplished and pressure distributions from certain sections are compared with experimental data.

This thesis aims to explore the potential for using large eddy simulation (LES) as a predictive tool for gas-turbine compressor flows. Compressors present a significant challenge for the Reynolds Averaged Navier-Stokes (RANS) based CFD methods commonly used in industry. RANS models require extensive calibration to experimental data, and thus cannot be used predictively. This thesis explores how LES can offer a more predictive alternative, by exploring the sensitivity of LES to sources of uncertainty. Specifically, the importance of the numerical scheme, the Sub-Grid Scale (SGS) model, and the correct specification of inflow turbulence is examined. The sensitivity of LES to the numerical scheme is explored using the Taylor-Green vortex test case. The numerical smoothing, controlled by a user defined smoothing constant, is found to be important. To avoid tuning the numerical scheme, a locally adaptive smoothing (LAS) scheme is implemented. But, this is found to perform poorly in a forced isotropic turbulence test case, due to the intermittency of the dispersive error. A novel scheme, the LAS with windowing (LASW) scheme, is thus introduced. The LASW scheme is shown to be more suitable for predictive LES, as it does not require tuning to a known solution. The LASW scheme is used to perform LES on a compressor cascade, and results are found to be in close agreement with direct numerical simulations. Complex transition mechanisms, combining characteristics of both natural and bypass modes, are observed on the pressure surface. These mechanisms are found to be sensitive to numerical smoothing, emphasising the importance of the LASW scheme, which returns only the minimum smoothing required to prevent dispersion. On the suction surface, separation induced transition occurs. The flow here is seen to be relatively insensitive to numerical smoothing and the choice of SGS model, as long as the Smagorinsky-Lilly SGS model is not used. These findings are encouraging, as they show that, with the LASW scheme and a suitable SGS model, LES can be used predictively in compressor flows. In order to be predictive, the accurate specification of inflow conditions was shown to be just as important as the numerics. RANS models are shown to over-predict the extent of the three dimensional separation in the endwall - suction surface corner. LES is used to examine the challenges for RANS in this region. The LES shows that it is important to accurately capture the suction surface transition location, with early transition leading to a larger endwall separation. Large scale aperiodic unsteadiness is also observed in the endwall region. Additionally, turbulent anisotropy in the endwall - suction surface corner is found to be important. Adding a non-linear term to the RANS model leads to turbulent stresses that are in better agreement with the LES. This results in a stronger corner vortex which is thought to delay the corner separation. The addition of a corner fillet reduces the importance of anisotropy, thereby reducing the uncertainty in the RANS prediction.

In this study, which is numerical in nature, direct numerical simulation (DNS) of the pipe flow is performed. For the DNS a solenoidal spectral method is employed, this involves the expansion of the velocity using divergence free functions which also satisfy the prescribed boundary conditions, and a subsequent projection of the N-S equations onto the corresponding dual space. The solenoidal functions are formulated in Legendre polynomial space, which results in more favorable forms for the inner product integrals arising from the Petrov-Galerkin scheme employed. The developed numerical scheme is also used to investigate the effects of spanwise oscillations and phase randomization on turbulence statistics, and drag, in turbulent incompressible pipe flow for low to moderate Reynolds numbers (i.e. $mathrm{Re} sim 5000$) ).