Postdoc position in High Performance Computing per el CIMNE

The position is focused on the development of High Performance Computing (HPC) methods for the simulation of thermally coupled magnetohydrodynamic (MHD) flows in complex geometries. This problem is important for the development of fusion reactor technologies. Fusion is called to radically change the energy production worldwide, as it provides a large-scale, sustainable, carbon-free and, unlike fission, inherently safe source. A key component of fusion reactors is the breeding blanket which bound large part of the plasma chamber providing radiation shielding from free neutrons, extracting heat and regenerating the tritium burnt in the plasma. Its design is a complex task that requires advanced shape optimization based on MHD flows evaluation.

Existing tools are not mature enough to efficiently simulate MHD flows in the presence of strong magnetic fields, they are not able to properly capture complex blanket geometries, and therefore shape optimization methods are not available. The goal of the position is to contribute to the development of computational methods that include: fully monolithic solution schemes able to deal with the strongly-coupled problems associated with strong magnetic fields, embedded finite element methods to facilitate the discretization of complex blanket geometries, and new shape and optimization methods to optimize critical outputs of the MHD flow such as the pressure drop.

The developed methods will be implemented in the Julia programming language, using the Gridap ecosystem as a basis, will be released as free and open source and will exploit HPC resources. The resulting code, GridapMHD.jl, will be applied to simulate and analyse real-world breeding blanket concepts and to find optimal designs. This will be done in collaboration with the Fusion technology department at CIEMAT.

The deadline for registration to the offer ends on March 26th, 2024 at 12 noon.

Més informació

Més posts de Recerca