Sept. 27, 2021 — The Barcelona Supercomputing Center (BSC) is coordinating FusionCAT, an initiative that brings together seven Catalan institutions to collaborate in the field of research and development of fusion energy technology. The BSC is also participating in the three technical projects of the initiative: Project 1: “Towards the complete integrated modeling of a fusion reactor,” Project 2: “Neutronics, Tritium production and operational fuel cycle,” and Project 3: “Study of the fusion reactor.” In addition, the BSC helps FusionCAT partners disseminate and exploit the project’s results to the academic and industrial sectors to promote research and training in the field of fusion. The objective is to communicate, identify, and protect the projects’ outcomes authorship to create and stimulate a fusion community in Catalonia.

The BSC’s team involved in FusionCAT has grown from seven members at the project’s launch to the current 25 professionals. This growth has established a highly multidisciplinary working force with the objective of effectively facing the challenges posed by the project close to its half-point. The team comprises experts in physics, mathematics, computer science, engineering, project management and technology transfer, and it is led by the ICREA researcher Mervi Mantsinen with more than 23 years of experience in the field of fusion. “FusionCAT enables BSC to explore novel areas with great potential for advancement based on our unique combination of High-Performance Computing (HPC) and fusion expertise,” says Mervi Mantsinen. The methodology applied by the BSC team is widely recognized by the fusion community and has proven its validity in the past.

In the first 2 years of the project, the BSC has developed the first versions of the magnetism and neutron transport modules for its software Alya, has progressed in the coupled thermodynamics and fluid dynamics simulation of an ITER’s first wall segment, has validated several fusion community software against JET experimental data, is in the process of integrating them in ITER’s Analysis Suite IMAS, and has made progress in new materials atomic-level modeling. Other consortium members have progressed in tritium and lithium sensors, liquid metal simulations, catalytic membrane reactor for the recovery of hydrogen and fuel clean-up, particle acceleration components design, IFMIF facilities design, and Supercritical CO2 power cycles.

Plasma and fusion reactors are highly complex systems from the physical, numerical, and computational points of view. To model these, it is necessary to consider multiple physical phenomena at multiple spatiotemporal scales. Understanding the physics involved requires the integration of all these elements (for instance, to know how each factor of the reaction affects the others or how the cooling system captures the released energy), the validation of the results, and the subsequent integration into research tools for the research community. To achieve this, BSC has developed three tasks within Project 1: the implementation of multiphysics codes based on ALYA (parallel code of computational mechanics) to be able to model complex multiphysics systems, the experimental validation of codes for fusion, and the integration of codes in the production lines that will be used during ITER’s operation.

The future energy production reactors, such as DEMO, are committed to a fuel cycle based on an enormous production of neutrons from plasma and on the use of the breeding blanket to multiply further the number of neutrons produced to maintain the operational fuel cycle. However, to achieve efficient energy production, it is necessary to optimize the plant’s fuel cycle. Project 2 focuses on the fusion fuel cycle and its effects on the reactor and the components related to neutrons, lithium, and tritium. To this end, in close collaboration with CONICET (Argentina), BSC develops a new deterministic neutron transport code based on Boltzmann’s transport equation based on the finite element method. At present, Monte Carlo codes such as MCNP are commonly used in fusion neutronics, although they are more computationally expensive and focused on local analysis. Alternatively, the implementation of deterministic codes allows a global and less expensive analysis of the geometry, which is an advantage. Our development is based on the neutron transport code implemented in ALYA and aims to considerably increase the fidelity of this module to allow realistic predictions of fusion reactors, contributing to the development of a cutting-edge computational tool capable of tackling this complex multiphysics problem in existing and future fusion devices. Furthermore, this tool will potentially meet future challenges in fusion reactor design due to its HPC capabilities allowing the analysis of different variants of a design. Thus, the expected impact of this tool for the development of fusion as an energy source will be high and placed within the “Tritium self-sufficiency and the design of the Test Blanket Module (TBM)” mission of the European fusion energy roadmap.