PINN(s): Physics-Informed Neural Network(s) for Burgers equation

This is an implementation of PINN(s) on TensorFlow 2 to solve Burgers equation (1D Navier-Stokes eq. with no pressure gradient / external force) under Dirichlet boundary condition w/o training data (data to fit initial & boundary conditions need to be provided). This is keras-utilized version unlike other two of my repos (PINN_wave / PINN_von_Karman).


Simply type python to run the entire code. Basic parameters (e.g., network architecture, batch size, initializer, etc.) are found in and could be modified depending on the problem setup.


Tested on python 3.8.10 with the following:

Package Version
numpy 1.22.1
scipy 1.7.3
tensorflow 2.8.0


[1] Raissi, M., Perdikaris, P., Karniadakis, G.E.: Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations, Journal of Computational Physics, Vol. 378, pp. 686-707, 2019. (paper)


View Github