This repository includes the implementation of the Physics Informed Neural Network and The Deep Energy Method on 1D, 2D boundary value and time dependent problems. Here We have used the Deep Neural Network as a function approximator and converted the problem of directly solving the governing equation into a loss function optimization problem. It works by integrating the mathematical model into the network and reinforcing the loss function with a residual term from the governing equation, which acts as a penalizing term to restrict the space of acceptable solutions.
haytey / meachine-learning-approach-to-solve-the-mechanics-problems Goto Github PK
View Code? Open in Web Editor NEWThis project forked from mohitksoni/meachine-learning-approach-to-solve-the-mechanics-problems
This repository includes the implementation of the Physics Informed Neural Network and The Deep Energy Method on 1D, 2D boundary value and time dependent problems. Here We have used the Deep Neural Network as a function approximator and converted the problem of directly solving the governing equation into a loss function optimization problem. It wo