Over the Spring and Summer 2025, I studied at ETH Zurich. I conducted my Bachelor's thesis under Professor Filippo Coletti, studying two-phase flows and experimentally finding the collision rates between bi-dispersed particles in quiescent air. We utilized two particle species of diameter 250 μm and 800 μm and cameras at 4000 fps to reconstruct their trajectories with Particle Tracking Velocimetry (PTV) software. I then developed an algorithm that detects collisions based on the reconstructed trajectories.
Tasks as a researcher, my contributions were:
* Literature review (read Turbulence by Uriel Frisch).
* Found the theoretical terminal velocities of the particles, then confirmed them experimentally.
* Ran experiments in a quiescent air chamber and a shaker to drop both particle species together. Captured the dynamics using Phantom cameras at up to 6000 fps.
* Calibrated the PTV projection matrices, segmented images, optimized voxels, and reconstructed trajectories.
* Worked with the ETH Euler Cluster and developed Linux-compatible scripts to streamline workflow.
* Developed a collision detection algorithm based on Collision Theory, as denoted in Physics and Chemistry of
Clouds by Lamb et. al.
Clouds by Lamb et. al.
* Taught myself Machine Learning theory and implementation, then fine-tuned an existing Python ML model to detect collisions from videos.
This thesis is still in progress, so I will be able to post more of the results when the paper is released.
Real (Zoomed In) Image
Small Particles Segmented
