Granular Flow

Granular flow research focuses on understanding and simulating the complex behavior of granular materials, such as sand or grains, to improve predictions of phenomena like landslides and debris flows. Current research heavily utilizes graph neural networks (GNNs) to create efficient and generalizable simulators, often incorporating techniques like reduced-order modeling and dimensionality reduction to accelerate computations. These advancements offer significantly faster and more accurate simulations compared to traditional methods, impacting fields like robotics (e.g., legged locomotion on granular terrain) and geotechnical engineering through improved hazard assessment and mitigation strategies.