Fast Computation Method of Magnetic Field Homogeneity for NMR/MRI REBCO Pancake Coils
Fast Computation Method of Magnetic Field Homogeneity for NMR/MRI REBCO Pancake Coils
复制标题
NMR/MRI REBCO 薄饼线圈磁场均匀性快速计算方法
DOI：
10.1109/tasc.2016.2637937
复制
发表时间：
2017
影响因子：
1.8
通讯作者：
Noguchi So
中科院分区：
文献类型：
--
作者：
Miyao Ryosuke;Igarashi Hajime;Kim SeokBeom;Noguchi So
近年来，稀土钡铜氧化物（REBCO）带材缠绕的多层扁平线圈有望应用于核磁共振（NMR）/MRI磁体。由于REBCO带材非常昂贵，因此应优化设计REBCO磁体，以最大限度地减少绕组体积。当然，NMR/MRI磁体也需要高度均匀的磁场。为了实现REBCO磁体的精确均匀性，有必要逐个计算所有REBCO层的贡献，因为与整个磁体横截面相比，电流在非常薄的区域上传播。然而，产生高磁场的NMR/MRI磁体通常由多个堆叠的REBCO饼状线圈组成，并且每个饼状线圈具有>300匝。因此，为了评估由每个REBCO层贡献的场均匀性，有必要重复计算一万次以上作为一个磁体形状。在传统的优化算法中，需要对NMR或MRI磁体进行超过一百万次的优化设计。这样的迭代计算是不现实的。基于这一背景，我们提出了一种利用球谐函数位移算子的快速计算方法。此外，为了验证该方法的有效性，将其应用于1.5 T MRI磁体的形状优化。文中还给出了优化结果和计算时间。
Recently, multiple-stacked pancake coils wound with rare-earth barium copper oxide (REBCO) tapes are expected to apply for nuclear magnetic resonance (NMR)/MRI magnets. Since REBCO tapes are very expensive, REBCO magnets should be optimally designed so as to minimize the winding volume. Surely, a highly homogeneous magnetic field is also required for NMR/MRI magnets. To achieve the accurate homogeneity of REBCO magnets, it is necessary to compute the contribution of all REBCO layers one by one, because currents carry over very thin area compared with a whole magnet cross section. However, NMR/MRI magnets generating high-magnetic field usually consist of multiple-stacked REBCO pancake coils, and each pancake coil has >300 turns. To evaluate the field homogeneity contributed by every REBCO layer, therefore, it is necessary to repeatedly compute more than ten thousand times as one magnet shape. In a conventional optimization algorithm, it is necessary to iterate more than one million times for the optimal design of an NMR or MRI magnet. Such an iterative computation is not realistic. Based on the backgrounds, we present a fast computation method using the shift operator of spherical harmonics. In addition, to confirm the validity of the proposed method, it was applied to the shape optimization of 1.5-T MRI magnet. In this paper, the optimization result and computation time are also shown.