Janosch completed his PhD in the CPDE group in 2022, working on numerical methods for wave propagation in stratified media, motivated by the challenging demands of computational helioseismology. He was supervised by Christoph Lehrenfeld, Thorsten Hohage and Damien Fournier, and was a member of the IMPRS for Solar System Science.

Research interests

The extreme conditions in the Sun amplify the classical problems of numerical wave propagation:

• Accurate high-order discretization methods
• Transparent boundary conditions
• Efficient preconditioners for iterative methods

His main contribution was the development of learned infinite elements: transparent boundary conditions that are learned from the stratified background medium to provide an optimal approximation of the Dirichlet-to-Neumann operator. Learned infinite elements can also serve as transmission conditions in a domain decomposition framework to construct efficient preconditioners.

Selected conferences

• WAVES 2022, Paris
• PRECOND 2022, Chemnitz
• Ants Workshop on Computational Helioseismology 2021 (online)
• GAMM 2021, Kassel (online)
• Ants Workshop on Computational Helioseismology 2019, Bordeaux — Learned infinite elements
• WAVES 2019, Vienna — Sweeping preconditioners for helioseismology
• ECCM-ECFD 2018, Glasgow — Higher order unfitted isoparametric space-time FEM on moving domains