25T超导磁体优化中线圈数量影响分析

Effects of different coil combinations on the optimal design of a 25 T superconducting magnet

20 T以上强磁场在高场科学工程中有着不可替代的作用.电工研究所正在研制一个25 T全超导磁体系统,包括15 T背景磁场和10 T高温超导内插磁体.在磁体的设计和优化中,线圈的数量和种类对于最终优化结果十分关键.为了研究磁体数量和磁体相关参数的关系,计算了20组不同的线圈组合下磁体的优化结果.优化中除了考虑必要的限制条件以外,还采用了一种结合局部优化算法和全局优化算法的方法.通过对比分析发现,线圈数量和磁体造价之间,存在一个"V"形的关系.更进一步地,本文分析了不同超导体在磁体中应该贡献的最佳磁场,以及背景磁体统一供电给优化结果带来的影响.

High field above 20 T is required in diverse physical programs and nuclear magnetic resonance（NMR） systems.For intended science program requirements, as a demonstration of the development in high field superconducting magnet technology, a 25 T（4.2 K） 52 mm cold-bore all-superconducting magnet consisting of a 10 T high-temperature superconducting insert magnet and a 15 T low-temperature superconducting background magnet, is being developed at the Institute of Electrical Engineering, Chinese Academy of Sciences. The development of such a magnet requires its optimization, and the choosing the number and type of coils is crucial to the final optimal design. However there are few researches focusing on the effect of coil combinations. To study the relationship between the number of coils and the magnet parameters, we first discuss the magnet optimization. The objective function of the optimization is defined as the weighted function of coil volume according to the costs of different superconductors, and the following constraint conditions are taken into considerations： center field, YBCO conductor characterization, hoop stress in Nb3 Sn coils, and the critical performances of these wires. All those constraint conditions are taken in the analytical form, and the magnetic field, stress results are verified with the finite element method. To guarantee the reliability of the optimal results,in addition to consider the constraint conditions, a method of combining global optimization and local optimization is adopted. 20 different coil combinations are selected according to the investigation of superconducting wires, and their optimal results are calculated. The following conclusions are drawn from the analyses of these results. Firstly, in the design of high field magnet, the number of coils and magnet cost demonstrate a ＂V＂-shaped relationship, that is, there exist an optimal number of coils. Secondly, when the objective function demonstrates good values, Nb3 Sn coils generate fields in a range of 6–7 T, whereas Nb Ti coils generate fields in a range of 8–9 T. Finally, the objective functions under two different situations, i.e., Nb3 Sn coils and Nb Ti coils are powered together and separately, are calculated. From the comparisons we find that the effect of reducing one power supply is acceptable when the number of coils is not too big.