An integrated design framework for future nuclear fusion power reactors

Change log
Coleman, Matti 

Nuclear fusion has long been touted as a future source of low-carbon energy. Whether or not fusion energy will be feasible, let alone commercially viable, remains a subject of research and development. Tokamaks are the most mature path to fusion energy, and two key steps are broadly seen as necessary before a first-of-a-kind fusion power plant (FPP) can be built. The first is the demonstration of a reactor-relevant plasma physics scenario which ITER, the world’s largest tokamak (presently under construction), is intended to deliver. The second is the demonstration of the reactor technologies necessary for a future FPP. In particular, this includes the demonstration of a viable breeding blanket and closed tritium fuel cycle, and maintenance scheme. For this, it is broadly acknowledged that a demonstrational fusion reactor is required, commonly referred to as “DEMO”.

At present, conceptual DEMO reactors are generally designed with systems codes, tools which rapidly perform constrained optimisations on zero- and one-dimensional representations of all major fusion reactor systems to generate an initial design point. From such codes, a series of higher fidelity design activities are performed by invariably distributed teams, eventually leading to a relatively well-substantiated reactor design point — typically many months later.

In this thesis, the classical approach to conceptual future fusion reactor design is revisited, with the primary objective being the acceleration of the design cycle, bridging the gap between systems codes and more detailed and slower studies, thus paving the way for a more rigorous exploration of the parameter space. A novel fusion reactor design framework, BLUEPRINT, is presented, in which an existing systems code, PROCESS , is combined with a broad range of reactor design and analysis techniques to create conceptual fusion reactor design points. Several design and analysis problems are solved in sequence, including the design of the toroidal field coils, the design of plasma equilibria and the poloidal field system, the 2-D geometry of the reactor, up to the 3-D CAD models.

Benchmarks with other codes and comparisons to results for a single-null European DEMO reactor design are carried out. The flexibility of the reactor design framework is demonstrated by means of global parameter optimisations, and application to a negative triangularity fusion reactor concept.

The key finding of this work is that conceptual designs for future fusion reactors can be generated relatively automatically, to a higher degree of fidelity than with present systems codes, within a matter of minutes — as opposed to the several months required at present.

Parks, Geoff
Shwageraus, Eugene
Surrey, Elizabeth
Nuclear fusion, Tokamaks, Systems codes, Design optimisation
Doctor of Philosophy (PhD)
Awarding Institution
University of Cambridge
United Kingdom Atomic Energy Authority