Eric Fowler
Senior
Major: Aerospace Engineering Minor: Biomechanics
I currently work on implementing Physics Informed Neural Networks (PINNs) to model physical systems. These PINNs using governing equations from physics to find general solutions to problems with incomplete boundary conditions. My most recent work has been exploring the use of transfer learning with these models. Transfer learning involves retraining an existing model for a new problem. For example, this could be training a model on a flow with a Reynolds number of 100 and then retraining it on a flow with a Reynolds number of 200. I conduct my research on HiPerGator, the UF supercomputer, where I use A100 GPUs to accelerate my script runtimes.
My sophomore year, I was the team lead for our NASA 2021 Big Idea Lunar Dust Challenge team. Our goal was to propose a solution to remove lunar regolith from astronaut spacesuits when they returned from a spacewalk. Our design combined the concept of cyclonic separation with the low-pressure atmosphere of the Moon to create a lightweight and inflatable dust filtration chamber.
During my freshman year, I used SolidWorks to create plasma actuator designs, which could be printed and then fabricated into plasma actuators using photolithography. I also hand built and soldered simpler actuator designs to collect data on their ozone distribution in a small-volume region.
In addition to primarily working in machine learning, I also machine parts, design and CAD airfoils, and organize lab presentations and social events.