Pore Network

Pore network modeling focuses on representing the complex interconnected spaces within porous materials like rocks and soils to understand fluid flow and transport processes. Current research emphasizes developing advanced algorithms, including novel neural network architectures like TransU-Net and Bayesian Physics-Informed Neural Networks, and efficient extraction methods such as the flashlight search medial axis, to improve the accuracy and speed of pore network generation from imaging data. These advancements are crucial for applications ranging from predicting subsurface fluid flow in reservoir engineering to simulating microbial activity in soil science, enabling more accurate predictions and improved understanding of diverse natural and engineered systems. The integration of uncertainty quantification into these models is also a growing area of focus, leading to more robust and reliable predictions.