Thesis

2016

#### Degree Name

Master of Science

#### Department

Aerospace Engineering

#### Major

Aerospace Engineering

Paul Durbin

#### Abstract

In this thesis, a formal approach for inferring the functional forms of information that might be missing in Wilcox's $k-\omega$ model for turbulent channel flow has been presented. Different terms of $k-\omega$ transport equations have been \textit{modified} by including a \textit{spatial} parameter. The resulting inverse problem has been efficiently solved using \textit{continuous} adjoint method to accurately predict mean velocity. In many cases, a simplifying assumption of a frozen" eddy viscosity in the turbulence model has been made to arrive at the final set of adjoint equations. Good agreement between inferred solution and corresponding DNS data has been demonstrated, first, for plane channel flow and later, for channel flow with spanwise rotation where \textit{re-laminarization} of fluid takes place. Finally, some parameters have been explored that seem to scale the corrections.

#### DOI

https://doi.org/10.31274/etd-180810-5660

Kumar Vishal

en

application/pdf

56 pages

COinS