Reconstruction of refractive index gradients in a reactive turbulent flow based on U-net supervised reconstruction dataset and python code

This research is the result of Austrian-German joint research project, funded by the Austrian Science Fund FWF within grant FWF-I5392-N and the Deutsche Forschungsgemeinschaft DFG within grant CZ 55/50-1. This project is a Lead-Agency D-A-CH project in cooperation between Graz University of Technology, Austria and Technische Universität Dresden, Germany, named "FOUR-DIMENSIONAL MEASUREMENT OF THERMOACOUSTIC OSCILLATIONS".

The underlying hypothesis of this project claims that four-dimensional detection of local thermoacoustic oscillations, based on the combination of a (camera-based) laser interferometric vibrometer with multidirectional background-oriented schlieren method, density tagging velocimetry and deep neural networks as binding element, will reveal local and coupled information on combustion, acoustics and fluid dynamics. 

As of Jan 2025, this repository contains the Python routines and sample data from this project.

This publication refers to the previous publication by

Tasmany, S., Kaiser, D., Woisetschläger, J. et al. Heterodyne background-oriented schlieren for the measurement of thermoacoustic oscillations in flames. Exp Fluids 65, 151 (2024). https://doi.org/10.1007/s00348-024-03890-1

Tasmany S, Woisetschläger J (2024) Heterodyne background-oriented schlieren for the measurement of thermoacoustic oscillations in flames. Graz Univ Technol. https://doi.org/10.3217/7k9f0-7wb03

repository.zip is composed of 2 subfolders:

gradient test

Neural Network code

gradient test folder contains 100 couples of numpy array files, which represents a sample of the image dataset used.

Neural network code contains the files main.py, Evaluate_model.py and net_model.pt.

main.py is the python code used to preprocess images and build and train the neural network

Evaluate_model.py is the python code used to test the network trained with main.py

net_model.pt is a pretrained pytorch network file (trained on the full dataset of 1000 images)