SPOKE VS3 - Earthquake and volcanoes
Description and objectives
Pyroclastic density currents (PDCs) associated with explosive volcanic eruptions is one of the most noticeable examples of sediment density currents. Such currents are geophysical surface flows that move over the Earth surface due to the density difference between a multiphase fluid–particle mixture and the surrounding lean fluid.
In this project we will mainly focus on PDCs generated by the collapse of an eruption column: upon impact with the ground, the column forms a shear current whose thickness increases laterally and develops as a turbulent boundary layer shear flow. These flows are hundreds of meters thick and move along the volcano slope and over the surrounding territory at speeds of tens of m/s and temperatures of hundreds of °C, with a relevant destructive power. For the aim of volcanic risk, flow dynamic pressure is useful to territory planners to assess building resistance to the lateral stress exerted by the current. Therefore, understanding the basic mechanism of column collapse is crucial to define models able to predict the hazard associated with PDCs.
The objective of PYROMESH is to generate efficient and effective computational grids for multiphase numerical simulations of pyroclastic density currents, able to resolve the flow–topography interaction, but also the detailed interaction of the current with man-made structures (buildings, infrastructures), which is crucial for an accurate quantitative evaluation of hazard associated with PDCs.
Two different mesh generation approaches are used, namely, body-fitted and non-body conformal grids, in order to obtain detailed information on the density and velocity gradients at the base of the current. The computational mesh should be able to capture all the relevant scales of PDCs, while keeping the computational cost acceptable.
The use of body-fitted meshes has the advantage of an easy treatment of boundary conditions and of a strict control of the mesh size and quality at walls, whilst requiring a well-established expertise along with a time-consuming procedure, particularly in the application under analysis, characterized by a large
computational domain with wide a range of scales. On the other hand, non-conformal Cartesian meshes are relatively easy to generate by using an automatic approach along with the advantage of allowing a local refinement procedure to improve the space accuracy.
Lead Partner
- Polytechnic University of Bari – Prof. Giuseppe Pascazio