Associated team HAMSTER :

High order Adaptive moving MeSh finiTE elements in immeRsed computational mechanics

space

The project involves two main teams: the Inria team CARDAMOM, and the Civil and Environmental Engineering Department at Duke University . This team focuses on adaptive unstructured mesh finite element-type methods for fluid flows with moving fronts. These fronts may be interfaces between different fluids, or fluid/solid, and modelling or physical  fronts (e.g. shock waves) present in the flow. The two teams involved in the project have developed over the years complementary strategies,  one focusing more on an Eulerian description aiming at capturing fronts on adaptive unstructured grids,  the other  working more on Lagrangian approaches aiming at following exactly some of these features. Unfortunately, classical Lagrangian methods are at a disadvantage in the presence of complex deformation patterns, especially for fronts undergoing large deformations, since the onset of vorticity quickly leads to mesh rotation and eventually tangling. On the other end, capturing approaches, as well as Immersed Boundary/Embedded (IB/EB) methods, while providing  enormous flexibility when considering complex cases,  require a careful use of mesh adaptivity to guarantee an accurate capturing of interface physics. The objective of this team is to study advanced hybrid methods combining high order, adaptive, monotone capturing techniques developed in an Eulerian or ALE setting, with fitting techniques and fully Lagrangian approaches.

 

 

Research directions and keywords

• Embedded and shifted boundary methods
• ALE adaptive methods for flows embedding shocks and moving fronts
• Blending adaptive ALE and local remeshing
• Fitting and Lagrangian methods for shocks and interfaces
• Applications in geophysics, external compressible aerodynamics, material science