Turbulent boundary layer separation and its driving factors

Borra, Akhileshwar

Permalink

https://hdl.handle.net/2142/127219 Copy

Description

Title

Turbulent boundary layer separation and its driving factors

Author(s)

Borra, Akhileshwar

Issue Date

2024-12-02

Director of Research (if dissertation) or Advisor (if thesis)

Saxton-Fox, Theresa

Doctoral Committee Chair(s)

Saxton-Fox, Theresa

Committee Member(s)

• Goza, Andres
• Ansell, Phillip
• Chamorro, Leonardo

Department of Study

Aerospace Engineering

Discipline

Aerospace Engineering

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• Turbulent boundary layers
• pressure gradients
• separation
• history effects

Abstract

Turbulent boundary layers are common in nature and in many engineering applications where a solid and a fluid interact. In many such applications, pressure gradients influence the behavior of turbulent boundary layers leading to their acceleration or deceleration. This study experimentally examines the effects of spatially varying pressure gradients using two geometries: a symmetric Gaussian bump and a downward-facing ramp tested sequentially with a flat plate in a wind tunnel. These geometries create distinct pressure gradient patterns: the bump induces a sequence of mild adverse pressure gradient (APG), strong favorable pressure gradient (FPG), strong APG, and a recovery region, while the ramp produces a weak FPG, FPG, strong APG, and a recovery region. The bump’s maximum FPG and APG strength ranged from Clauser pressure gradient parameter, β0, values of -4.2 to -5.4 and 2.07 to 3.3, depending on the Reynolds number. For the ramp, the maximum APG values ranged between 1.9 and 3.4. The effects of varying pressure gradient history and freestream velocity on separation behavior were analyzed. Data were collected for both geometries at freestream velocities of 7.5, 10, and 15 m/s, resulting in six test cases. Planar particle image velocimetry was used to observe the evolution of the turbulent boundary layer, upstream and on the surface of the geometries, under the imposed pressure gradients. Pressure measurements were taken to quantify the strength of these pressure gradients. The momentum thickness Reynolds number, Reθ, ranged from 2347 to 3904 for the bump and 3378 to 6217 for the ramp, while the friction-based Reynolds number, Reτ , ranged from 889 to 1706 for the bump and 1526 to 2920 for the ramp. The incoming boundary layer thickness to bump and ramp height ratios in these experiments were near 1. The study examined the evolution of the turbulent boundary layer from statistical and structural perspectives. Mean streamwise velocity, Reynolds stresses, and backflow percentages were used to characterize the separated region and the detached shear layer. Spectral proper orthogonal decomposition (SPOD) and spectral analysis of the backflow region area were used to evaluate the dynamic behavior of the separated region. Similarities and differences in the region upstream of separation were analyzed using mean streamwise velocity, Reynolds stresses, and SPOD, providing insights into how the turbulent boundary layers were modulated. The study also explored the onset and progress of relaminarization and the development of an internal layer. The six cases were grouped into four categories based on the analysis. Group 1 included bump cases at 7.5 and 10 m/s, showing similar separation behavior with boundary layer relaminarization and an internal layer forming on the upstream half of the bump. Group 2 consisted of the bump case at 15 m/s where a smaller and weaker separation region and partial relaminarization were observed, along with an internal layer forming on the upstream half of the bump. Group 3 consisted of ramp cases at 7.5 and 10 m/s, which showed separation behavior similar to Group 1. However, relaminarization was not observed and the internal layer started at the ramp’s leading edge. Group 4 included the ramp case at 15 m/s featuring a smaller separation region than groups 1 and 3, but with similar backflow percentages, no relaminarization, and internal layer behavior comparable to group 3. Detachment was delayed as Reynolds numbers increased for the same geometry. At the same freestream velocity, detachment occurred further downstream on the ramp compared to the bump. Surface curvature was the primary driver of separation in groups 1 and 3, while differences between groups 1 and 2 and between groups 3 and 4 were attributed to variations in the conditions upstream of separation.

Graduation Semester

2024-12

Type of Resource

Thesis

Handle URL

https://hdl.handle.net/2142/127219

Copyright and License Information

Copyright 2024 Akhileshwar Borra

Owning Collections