Mechanical engineering graduate student wins national engineering competition

Kristyn Johnson took first place at the Oak Ridge Institute for Science and Education Ignite Off Competition. (Submitted Photo)

Kristyn Johnson, a graduate student in the Department of Mechanical and Aerospace Engineering in the Statler College of Engineering and Mineral Resources, took first place in the Oak Ridge Institute for Science and Education (ORISE)Ignite Off Competition for her research on reducing fuel consumption and emissions in gas turbine power plants.

By Adrianne Uphold
Graduate Assistant

Benjamin M. Statler College of Engineering and Mineral Resources

Johnson, of Williamson, West Virginia, presented research on the analysis of Rotating Detonation Engines (RDE) using optical techniques.RDE is an engine that uses a form of pressure gain combustion, where one or more detonation waves continuously travel around a circular channel.

She focused on conveying the concept and benefits of machine learning integration by relating complex ideas to more accessible concepts like race cars and stacks of pancakes.

To visualize the speed of detonation, Johnson “held a race” during her presentation between Usain Bolt, a race car, a jetliner, a fighter jet and a detonation wave. Each competitor doubled the speed of the previous example with the detonation wave traveling almost a mile in just two seconds.

“Since most people have a grasp on the speed of race cars, only the second fastest of the five competitors, this served as the anchor example for the necessity of high-speed imaging,” Johnson said.

Even analyzing short sampling periods of the frames without the machine learning techniques is almost impossible because of the overwhelming number of frames you would have to sort through. Johnson used the example of a child eating a stack of pancakes twice the size of their bodies to compare how difficult it would be to analyze the frames.

“Thinking of the mismatched proportion of the child to the pancakes is a perfect visualization of big data,” Johnson said.“Analyzing big data, which in our case is a massive collection of detonation images, becomes much more plausible through the integration of machine learning techniques.”

An RDE is intended to replace conventional combustors within gas turbines, which could potentially lead to the reduction of emissions in gas turbine power plants.Johnson explained that these machine learning techniques can be developed to automatically perform classifications on the waves traveling within the RDE annulus that otherwise wouldn’t be identifiable.

“Through high-speed imaging, we can monitor and quantify detonation waves traveling circumferentially within the RDE annulus,” Johnson said. “Thanks to the incredibly large volume of images, machine learning techniques can be developed to automatically perform image classifications that are otherwise unreliable or computationally expensive.”

The ORISE Ignite Off competition included research fellows from three United States Department of Energy National Laboratories—Argonne National Laboratory, Oak Ridge National Laboratory and the National Energy Technology Laboratory.

Johnson felt that participating may help her ability to talk about her research in a more encompassing way, without distracting information that often stands as a barrier between engineers and the public.

“This contrasted conventional research presentations that research fellows are accustomed to delivering,” Johnson said.“Because of that, restructuring highly technical and detailed content into a topical presentation that could enlighten a public audience was a unique challenge to say the least.”

---