Project_Machine_Learning_Algorithms

In this repository, there are three main working directories:

ReactionRateCoefficientsRegression
TransportCoefficientsRegression
PINN_Euler_1d_shock_STS

Pre-requisites

python 3 (with python 2 there could be some problems)
scikit-learn: https://scikit-learn.org/stable/install.html
Tensorflow 1.4 or 1.5 (superior may not work properly)
Keras

Reaction Rate Coefficients Regression

In this directory, we do the regression of reaction rate coefficients according to the state-to-state (STS) theory.

In Utilities there are few functions for plotting (not interesting for you). In docs there are some documents which may be useful for you to read.

How to run?

From Project_Machine_Learning_Algorithms/ReactionRateCoefficientsRegression/DR/src run:

./run_regression.sh

This will learn one vibrational level at the time.

Otherwise,

./run_regression_multioutput.sh

This will learn all the dataset at once.

In both cases, the model, scalers and figures will be saved in model, scaler and pdf folders, respectively.

Note

Only the scripts for dissociation are present. You can write the scripts for recombination (DR). In the same way, I only considered DR processes (datasets) here, neglecting VT, VV, VV2, and ZR. You can also try them.

Transport Coefficients Regression

In this directory, we do the regression of transport coefficients according to the state-to-state (STS) theory. In particular, we would like to predict:

shear viscosity
bulk viscosity
thermal conductivity
thermal diffusion
mass diffusion

For a rigorous definition of the transport coefficients, please refer to the book: Nagnibeda, E., & Kustova, E. (2009). Non-equilibrium reacting gas flows: kinetic theory of transport and relaxation processes. Springer Science & Business Media.

Physically Informed Neural Network (PINN)

In this directory, we do want to try to use PINNs to solve the Euler equations for a one-dimensional shock flow problem, according to the STS formuation.

There may be useful to take a look at some other repositories to better understand PINNs, for example:

lkampoli / project_machine_learning_algorithms Goto Github PK