Step Induced Separation of a Turbulent Boundary Layer: Part I. Experiments in Supersonic Flow. Part II. A Model for Incompressible Flow in a Channel

Author: Gran, Robert Lee

Year: 1970

Degree: Dissertation (Ph.D.)

Advisors: Zukoski, Edward E.; Kubota, Toshi

Committee Member: Unknown, Unknown

Option: Mechanical Engineering

DOI: 10.7907/XZZS-AR89

Abstract

This thesis is concerned with the separation of a two-dimensional turbulent boundary layer caused by a forward facing step. The two major flow regimes considered here are in the realms of supersonic flow and incompressible flow in a channel.

For the case of supersonic turbulent boundary layer separation, a series of experiments were conducted. The upstream pressure field for step sizes in the range from 5% to 150% of the local boundary layer thickness was determined and correlated with "large" step data. The separated shear layer was found to be approaching a constant pressure and self-similar flow in a distance of around 6-10 initial boundary layer thicknesses. Fluctuation measurements were conducted near the similar flow region. In addition, the low-frequency unsteady behavior associated with the separation phenomenon was examined and is presumed to be caused by a standing wave acoustic in the subsonic separated region.

For the incompressible flow over a step in a channel, an inviscid model utilizing free streamline theory and based on experimental observations was constructed and solved with the aim of predicting the upstream flow field. Although the solution is not in "closed form" (two experimental parameters are required) it does show that the far upstream pressure field is predominately fixed by the flow geometry as opposed to viscous effects such as the Reynolds number or step height-to-boundary layer thickness ratio. Close to the step these effects do, however, become important and are unaccounted for here. The effect of finite blockage ratio (step height-to-channel height ratio) is shown to be substantial for quite modest values (greater than 2%). The precise values of the two experimental parameters are not required for accurate prediction of the upstream influence.

Files