PINN(s): Physics-Informed Neural Network(s) for von Karman vortex street

This is an implementation of PINN(s) on TensorFlow 2 to learn the flow field of von Karman vortex street, and estimate the fluid density and kinemetic viscosity.


Simply type python to run the entire code in src directory. This will load data in input and starts training to estimate the fluid density (rho), and kinematic viscosity (nu). 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