Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)


Aeronautics and Astronautics

Committee Chair

Steven Schneider

Committee Member 1

Jonathan Poggie

Committee Member 2

Sally Bane

Committee Member 3

Hermann Fasel


Hypersonic boundary-layer transition can be caused by many different mechanisms. This research focuses on studying the growth and breakdown of the second-mode instability. Experiments were performed in the Boeing/AFOSR Mach-6 Quiet Tunnel located at Purdue University. This facility has freestream noise levels similar to conditions in flight making it an excellent facility to study laminar-turbulent boundary-layer transition.

Using a flared cone model, measurements were first made with a smooth wall to characterize natural second-mode transition. As the second-mode amplitude increased up to a maximum of nearly 30% of the mean static pressure, streamwise streaks of heating were observed forming around the circumference of the model. Once the maximum magnitude was reached and the second-mode began to break down, the amplitude of the streaks of heating decreased to near laminar levels. As the intermittency increased, streaks of heating once again formed around the model circumference resulting in a second increase in heating. This formed a characteristic hot-cold-hot pattern of heating.

Under noisy flow conditions, Marineau has shown that a linear correlation exists between the edge Mach number and the maximum second-mode amplitude prior to breakdown. He has used this relationship to develop a more physics-based transition prediction method. Results of 24 different experiments under quiet flow on the present flared cone with a smooth wall show the maximum amplitude is about 30%. This is the first breakdown magnitude obtained under fully quiet flow conditions, and it is more than two times larger than what would be predicted by the noisy flow correlation. Further testing is required with geometries resulting in different edge Mach numbers to determine if a similar correlation exists under quiet flow.

The effect of well controlled roughness elements interacting with the second-mode instability was investigated using the Rod Insertion Method (RIM) roughness inserts. Three different aspects of the roughness elements were studied. For RIM inserts with 30 evenly spaced elements with a diameter of 838 μm, it was found that elements 305 μm in height or less interacted with the instability without decreasing the maximum second-mode magnitude prior to breakdown. The pattern of heating was altered, but maximum second-mode magnitudes prior to breakdown were still approximately 30% of the mean surface pressure. Changes in the roughness element diameters and azimuthal spacing showed that when the element interacted with the second-mode wave and did not trip the flow the maximum second-mode magnitude remained relatively unchanged.

Experimental data for the flared cone at a stagnation pressure of 140 psia was compared to direct numerical simulations performed by Christoph Hader at the University of Arizona. The data compared well qualitatively showing many of the same trends. Quantitatively the results were not comparable due to the arbitrary forcing amplitudes used in the numerical simulation. Future simulations with a different forcing input may model the experiments better and lead to better transition prediction methods.

Experiments were performed under quiet flow conditions on a 2.5◦ and 3◦ half-angle straight cones to measure second-mode transition without the effect of surface curvature. The goal of these slender cones was to measure the maximum second-mode magnitude prior to breakdown and to see if the streamwise streaks observed on the flared cone also occur on the slender straight cones. Large second-mode waves were measured, but due to a reduction in the maximum quiet unit Reynolds number of the facility, the maximum pressure fluctuation magnitude prior to breakdown was not successfully measured. No streamwise streaks were observed, but since it is unclear if transition had occurred the results are inconclusive.