Physics-guided neural network framework for elastic plates

This repository contains codes for our published work on Comput. Methods Appl. Mech. Eng.. If this work helps you by any chance, you are encouraged to cite the following paper

@article{li2021physics,
  title={A physics-guided neural network framework for elastic plates: Comparison of governing equations-based and energy-based approaches},
  author={Li, Wei and Bazant, Martin Z and Zhu, Juner},
  journal={Computer Methods in Applied Mechanics and Engineering},
  volume={383},
  pages={113933},
  year={2021},
  publisher={Elsevier}
}

Brief introduction of the four examples

The codes are shared in four folders corresponding to the four examples we presented in the paper, where you can find more details. A brief summary is provided as follows:

1. Uniaxial tension of rectangular plates under non-unfiormly distrubuted loading

This example demonstrates how to solve a 2D plane stress tension problem with neural network. The loading condition and boundary conditions are illustrated as in the figure below.

Physic informed neural networks with two different loss functions: PDE-based vs. Energy-based.

2. Uniaxial tension of rectangular plates with centrol hole

A more complex 2D case with central hole:

3. Out-of-plane deflection of rectangular plates

4. Out-of-plane buckling of rectangular plates