ISPIV 2023: Convolutional Neural Networks for Particle Diffusometry Predictions

Authors: Pranshul Sardana and Steven Wereley
Affiliation: School of Mechanical Engineering, Purdue University

Connect to Pranshul Sardana via LinkedIn: https://www.linkedin.com/in/pranshulsardana/
Pranshul Sardana's Google Scholar: https://tinyurl.com/psardanaPublications
Steve Wereley's Google Scholar: https://tinyurl.com/swereleyPublications
Link to GitHub: https://github.itap.purdue.edu/psardana

Simulation type vs. Diffusion coefficient

No Flow	D = 0.30	D = 1.20	D = 2.10	D = 3.00
2D Simulation
Defocused Simulation

Flow type vs. Diffusion coefficient

8 Frame Average	D = 0.30	D = 1.20	D = 2.10	D = 3.00
No Flow
Uniform Flow
Couette Flow
Poiseuille Flow

Flow type vs. Image average

D = 0.30	1 Frame	2 Frame Average	4 Frame Average	8 Frame Average
No Flow
Uniform Flow
Couette Flow
Poiseuille Flow

Image average vs. Diffusion coefficient

No Flow	D = 0.30	D = 1.20	D = 2.10	D = 3.00
1 Frame
2 Frame Average
4 Frame Average
8 Frame Average