Anagrafe della ricerca

The simulation of real-life problems characterised by multiphysics phenomena presents numerous challenges. Some examples can be found in fluid-structure interaction, multiphase and coupled fluid heat transfer problems. In these cases, classical numerical approaches require -- at the very least -- large computational power, excellent meshing tools, and the satisfaction of hypotheses which are often too stringent for real life application. Consider, for example, fluid structure interaction and phase transition phenomena. Their simulation requires the solution of coupled problems on moving domains, where the solid, the fluid, and the interface itself obey different physical laws, each presenting its unique set of complex endeavours. The most used simulation strategy in this case is the Arbitrary Lagrangian Eulerian technique, where domains are continuously deformed and interfaces are matched exactly: the biggest limit of this approach is that it cannot handle topological changes and large deformations without remeshing. Embedded/Immersed/Non-matching methods have been proposed to overcome some of the limitations of body fitted techniques. This research project focuses on the development and analysis of several types of non-matching and immersed computational methods. Further, it aims at the integration of these methodologies with cutting edge reduced order and complexity reduction methods. Together, these approaches will allow for the treatment of extremely challenging multiscale and multiphysics problems out of reach with classical approaches, unlocking computational complexity reduction and reducing design-cycle times. The research units in Trieste, Torino and Brescia share a consolidated expertise on advanced discretisation schemes based on variational approaches, such as finite elements, including spectral and hp type finite elements, immersed methods, finite volumes and reduced order methods. The project is organised in three work packages (WP): WP1: Numerical Analysis of immersed, embedded, and non-matching methods for coupled problems WP2: Model order reduction and machine learning enhancement of scientific computing problems WP3: Open source software development