AME - Various Hypersonics Projects

The specific project a student takes on will be geared toward their interests.  Both computational (CFD) and experimental (wind tunnel testing) projects will be available.  The experimental facilities available for use include the Notre Dame STEM Mach 6 Ludwieg Tube, and the Notre Dame Quiet Mach 10 tunnel.  Projects that will begin in the Fall include:

(1):  Optimization:  Some work has been done resulting in a conference publication on the optimization of a Mach 6 static pressure probe's geometry in order to reduce the length of the static pressure probe (which in turn reduces vibrations in the sensor and improves the measurement quality).  Further work on this project can be done by an undergraduate student in the areas of improving or choosing a different optimization algorithm, and potentially designing a probe to test in either the Mach 6 or Mach 10 tunnel.   The optimization algorithm is paired with a computational fluid dynamics (CFD) solver. Currently, the CFD is being done by a commercial solver (Ansys Fluent), but there are plans to write our own research CFD code in Matlab and C++.  The student can choose which aspect of the project they would like to focus on (optimization, or writing an in-house CFD code).  

(2):  Machine Learning:  Machine learning is often employed as a black box "problem solver", where a solution is obtained without ever solving a set of physics-based equations.  A more appropriate use of the technology is to use ML to discover new solution techniques and then employ those solution techniques, ensuring that the end result has physical meaning.  Various projects in machine learning will be started in the Fall.  ML projects that a student could focus on include:
    (a): The application of machine learning to improve turbulence modeling.  Solving the Navier-Stokes equations with a fine enough mesh to resolve all scales of turbulence is too computationally expensive.  Therefore, turbulence is often "modeled" instead of "resolved".  These models trade physics for assumptions, often resulting in large errors.  Improving the state of the art in turbulence modeling can result in better CFD predictions for designers in terms of lift, drag, and heat transfer to a vehicle. 
    (b):  Measuring forces in short-duration hypersonic wind tunnels (which only run for less than 0.1 seconds) is a challenge due to the time it takes for a  sensor to converge to the true steady state force being applied.  A dynamic force reconstruction technique has been developed to more quickly measure loads in the ND STEM Mach 6 tunnel; however, the technique is sensitive to noise.  Machine learning offers a pathway to improving the technique such that it is less susceptible to noise.

(3):   Wind tunnel testing: Each project above will involve an experiment in either the Mach 6, Mach 10, or low speed wind tunnels on campus to compare to the CFD results.  If a student is more interested in designing physical systems, they can choose to focus on the design of the experiment itself that would result in the best comparison to the CFD study. Design of the experiment includes choosing a data acquisition technique (ie. infrared thermography, Schlieren, pressure measurements, etc), designing the physical model that will be tested, designing and building any circuits needed, assembling the system after manufacturing, and running the wind tunnel itself with assistance from a graduate student or PI.