Optimisation of STEP Magnetic Configuration and In-vessel Components using the Bluemira Fusion Power Plant Framework
Fusion power plant design and integration requires tools for rapid production and investigation of different concept designs. Bluemira is an open-source, Python 3 framework built to enable this process. It makes use of interdisciplinary models and incorporates a range of design activities, each of which is parameterised and automated. The bluemira framework can be used to create fusion device workflows - a structured set of design procedures, made up of selected system models, with inputs (e.g., configuration file of user choices, output from a fixed plasma equilibrium model, etc.) and outputs (e.g., field or heat-flux maps, CAD of the integrated components, etc.) chosen to suit a design program’s requirements. Workflows can be updated, either by modifying the inputs or by adding new features, allowing a rapid, iterative approach to producing a stable and converged design. Our STEP (Spherical Tokamak for Energy Production) workflow includes both a magnetic configuration design procedure, which considers the plasma equilibrium and toroidal and poloidal field coils, and an in-vessel component design procedure, which includes the first wall and divertor design. In this work, we present an addition to the STEP workflow of a divertor design and optimisation feature which links divertor magnetic flux shaping optimisation, using objective functions that act as proxies for divertor target heat-flux reduction, with target placement and angle optimisation. This modification allows us to explore how the physical space available to the divertor component can be exploited, while also optimising for coilset and first wall position, satisfying key engineering constraints, and maintaining the input plasma equilibrium. Our objective is to facilitate design flexibility in the divertor region for the STEP bluemira workflow and allow quick alterations to be made to and investigated for the STEP divertor silhouette design.