AUTOMATED TRANSPORT SEPARATION USING THE NEURAL SHIFTED PROPER ORTHOGONAL DECOMPOSITION

This repo contains the source code and jupyter notebooks for an automated transport separation and model reduction framework with the help of neural networks. The script (.ipynb) files for all the test cases are present here.

• Crossing_waves.ipynb describes the application and results for the application of our method to a synthetically generated crossing wave data set.
• Wildfire.ipynb performs the transport separation and model reduction for a 1D wildland fire model. The snapshot data are also provided for this example in the folder Wildfire_input.

The reader is encouraged to try out the examples on their own. We have however, provided the already trained weights for both the examples in the form Crossing_waves.pth and Wildfire_alreadyTrained.pth.