2018 FALL LC MAGAZINE

Machine learning plays a different role in another project in Kulp’s lab, this one aimed at understanding the various states of a flame. With important industrial applications, the work attempts to discover the perfect balance that provides just enough fuel to keep a flame lit — something that could have implications on engines and fuel conservation. Although the project began six years ago, the most recent approach tasks a team of students with departing from traditional methods of analyzing the flame, and instead using machine learning to optically monitor it. The team snaps photos of a flame and uses sophisticated machine-learning algorithms to analyze the image, telling them if the flame is lean or rich. When it comes to identifying the air-fuel state of the flame, this method eliminates the need to solve time-consuming equations for applications where time is of the essence. Shawn McCollum ’18, an interdisciplinary major, spent the fall and spring semesters working with machine-learning algorithms to analyze the collected data. “I knew that Dr. Kulp was doing some research with data science and machine-learning algorithms, so I jumped at the chance to get some experience,” said McCollum, who has already secured a position as a technology analyst in the cybersecurity division of JPMorgan Chase. “This experience has given me an opportunity to apply my programming skills to a real-life problem.” For Narshini Gunputh ’20, physics-engineering and math major, and Lucas Wilkins ’19, physics major with a minor in astronomy, who are rebuilding the experiment that produces the data for McCollum, the project’s appeal is more physical. “I like the garage-feel of the work and learning to use different tools,” explained Wilkins, who plans to apply to graduate schools. “The research helps me familiarize myself with different lab techniques and equipment. In class we work a lot with numbers, but here I can actually get a taste for work I could be doing in grad school or the private sector.” Aspiring engineer and aircraft technician, Gunputh, also sees the value in gaining hands-on experience in Kulp’s lab. “This project is helping me gain the knowledge I will need to get into graduate school for physics or engineering,” said Gunputh, who would like to eventually return home to Mauritius to work on airplane engines, electronics, or mechanics. The work is being carried out in collaboration with Oak Ridge National Laboratory in Tennessee. Charles E. A. Finney, senior R&D scientist, and C. Stuart Daw, now retired from the energy and transportation science division, have served as combustion physics consultants, assisting with how to set up the experiment and interpret data, while Lycoming generates data and performs analyses. “The Center for Energy and the Future’s (CEF) mission includes helping students engage with the world of energy in all its forms. A CEF Capacity Grant enabled Dr. Kulp and his students to purchase flow meters to aid in their research of natural gas flame dynamics using machine learning,” said Jonathan Williamson, associate professor of political science and director of the CEF. “The research could someday result in significant advances in industrial energy efficiencies.” (L to R) McCollum, Gunputh, Kulp, and Wilkins 15 www.lycoming.edu