This research enhances a novel finite element physics-informed neural network (FE-PINN) framework in order to optimize efficiency and results. The enhancements include tuning hyperparameters and consi..
This research enhances a novel finite element physics-informed neural network (FE-PINN) framework in order to optimize efficiency and results. The enhancements include tuning hyperparameters and considering new methodology in constructing the model architecture. This study achieved near convergence of model prediction to actual data and successfully incorporates finite element discretization into a neural network model.