Actionable workflows for fusion neutronics simulation.

Neutronics simulations are complex and typically require significant effort in setting up multiple pre- and post-processing steps, such as preparing data and geometry before the neutronics simulation can be carried out. From the end user point of view, many complex tools must be mastered in order to run the simulations, creating a barrier of entry to a field that needs to transition from science to engineering. Different versions of the tools used and customisation of processing steps can also lead to reproducibility issues, and the data produced could often be better managed.
We propose setting up standard packages (i.e. OpenMC and Paramak) as interoperable tools that can be linked up to create an automated simulation process, to be executed using a workflow engine (i.e. Galaxy). The integrated tools can then be (re)configured into scalable, actionable workflows that are FAIR; findable, accessible, interoperable and reusable. The chosen workflow engine provides a simple and accessible interface with many added benefits, such as capturing metadata, documenting what simulation has been executed, when, by whom, how and why. The selected workflow engine also enables automatic scheduling on distributed and high performance computing systems.
The presentation will use a spherical tokamak case study to show how this approach can be used to orchestrate neutronics simulations. The authors will first show how individual tools can be put together as automated workflows. By presenting the results of a neutronics simulation carried out in this way, the authors will then highlight the simplicity and added benefits of workflows.
The work is aimed at the neutronics community but especially newcomers or those outside the community (such as SMEs or young researchers) who wish to run basic simulations but are unfamiliar with the tools used in the sector.