 |
|
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
|
Physics Informed Neural Network 
Last Updated: 16 September 2024
People: Adhika Satyadharma, Chiang Wen-Yao, Yu Chun-Ying, M.J. Chern
Our study focused on utilizing both data and physics in physics informed neural network (PINN) to solve various fluid flow problems. By using data, the network would be easier to train compared to a fully physics based network. Also, as the network utilizes physics, the amount of required training data is significantly less than a data driven neural network. Currently, we are testing the PINN to predict fluid solid interaction, predict turbulence and for mesh refinement.
Tel: (886)-2-2737-6496
Fax: (886)-2-2737-6460
E-mail: mjchern@mail.ntust.edu.tw
| Copyright @ Ming-Jyh Chern , 2014-2020. All right reserved. |