(263a) Computational Fluid Dynamic Simulations of Experimental and Natural Granular Flows: First Insights into the Flow-Wall Interaction Dynamics

Volcanic granular flows (some pyroclastic density currents, debris avalanches, debris flows) can be defined as gravity-driven currents of solid particles where the particle-particle interaction dominates the motion. The interaction with topography is another relevant factor controlling the dynamics and the evolution of these flows. In the last decade, granular flow modelling has become one of the main tools to replicate past events and to investigate potential hazards of volcanic granular flows. The research on the behaviour of volcanic granular flows is one of the main topics in present day geophysics and volcanology, due to the hazard that they pose in areas located in the vicinity of the volcanoes.

The aim of this research is to investigate the dynamics of channelised volcanic granular flows by combining large-scale experiments with multiphase computational fluid dynamic simulations using the MFIX software (http://mfix.netl.doe.gov/). In particular, we focus on the dynamics that regulate the flow-wall interactions. MFIX provides a suite of models that treat the carrier phase (fluid phase) and the dispersed phase (volcanic particles) as interpenetrating continua. Sensitivity analyses of the boundary conditions and of the frictional models implemented in the numerical code were carried out to search for the most appropriate numerical set-ups to simulate experimental granular flows. These analyses have shown how the coefficients, which rule granular energy and momentum diffusion equations, are crucial to model the physics of the granular flow close to the wall. The optimal MFIX configuration was applied to large-scale experiments of a monodisperse system and comparisons between simulated and observed flow thickness and instantaneous and average velocities were carried out. Preliminary results show that the model predictions are significantly affected by the diffusivity of granular energy and momentum at the wall and, therefore, a detailed study of the latter would improve knowledge of the boundary conditions and hazard control.