Neutron Diffusion

Neutron diffusion, crucial for analyzing nuclear reactor behavior, is being tackled with novel computational approaches. Current research focuses on improving the accuracy and efficiency of solving the associated equations using physics-informed neural networks (PINNs), employing architectures like convolutional neural networks with skip connections and adaptive resampling to mitigate overfitting and enhance convergence. These advancements, along with the exploration of AI-based solvers leveraging existing libraries and optimized for various hardware, aim to provide more robust and computationally efficient solutions for reactor physics simulations and other applications requiring the solution of neutron transport equations.