Optimal Design of Hybrid Magnets

Continuous very high magnetic fields (over 30 T) can be generated using suitable combinations of superconducting and resistive magnets. Devices of such class are mainly used for research purposes, and are available at a few laboratories all over the world. The combined use of superconducting and resistive technologies calls for very demanding design effort, since the thermal and mechanical behavior of the two parts of the magnet present different characteristics, but both are pushed to their limits due to the very high field strength. As a matter of fact, a model of a hybrid magnet behavior to be used for computer aided design would require to keep into account the thermal aspects of the resistive part to dimension the cooling system, the possible start of quench in the superconducting part, the computation of mechanical forces for the stress analysis, and finally the connection to the power supply system, to properly include a protection system for the magnet. On the other hand, under suitable hypotheses, some simplifications are possible and analytical models are available, making the optimized design of such magnets a feasible task in reasonable computation times. The possibility of achieving the required performance eventually using different layouts calls for the adoption of broad search optimization algorithms able to search the whole design space without being trapped in local minima. As an example, stochastic, population-based algorithms such as Genetic Algorithms or Particle Swarm Optimization algorithms present the required characteristics. In the paper, an optimized design procedure, based broad search algorithms, for high field hybrid magnets, able to take into account the complex behavior of hybrid magnets, will be proposed.