We conducted an elastic-plastic deformation analysis using a surrogate model based on Physics-Informed Neural Networks (PINNs).
PINNs are trained using a loss function that incorporates both physical laws and reference data (e.g., simulation results). One of their key advantages is the ability to achieve accurate approximations with fewer reference datasets.
In this study, we validated a cantilever beam deflection scenario using an elastic-plastic material model embedded in the physical constraints of the PINNs framework.
The full report is available at the link below—we invite you to take a look.
We have also published the AI model we used on our demo site, so please try it out by clicking the link below.
https://demo4.astraea-soft.com/non_linear.html
Please watch the video below to learn how to operate the demo application.
