Faculty of Medicine, Department of Radiology, Medical Physics, University of Freiburg, Freiburg, Germany.
Magn Reson Med. 2022 Sep;88(3):1465-1479. doi: 10.1002/mrm.29294. Epub 2022 May 8.
An automated algorithm for generating realizable MR gradient and shim coil layouts based on the boundary element method is presented here. The overall goal is to reduce postprocessing effort and thus enable for rapid prototyping of new coil designs. For a given surface mesh and target field, the algorithm generates a connected, non-overlapping wire path.
The proposed algorithm consists of several steps: Stream function optimization, two-dimensional surface projection, potential discretization, topological contour sorting, opening and interconnecting contours, and finally adding non-overlapping return paths. Several technical parameters such as current strength, inductance and field accuracy are assessed for quality control.
The proposed method is successfully demonstrated in four different examples. All exemplary results demonstrate high accuracy with regard to reaching the respective target field. The optimal discretization for a given stream function is found by generating multiple layouts while varying the input parameter values.
The presented algorithm allows for a rapid generation of interconnected coil layouts with high flexibility and low discretization error. This enables to reduce the overall post-processing effort. The source code of this work is publicly available ( https://github.com/Philipp-MR/CoilGen).
本文提出了一种基于边界元法的可实现磁共振梯度和匀场线圈布局的自动算法。总体目标是减少后处理工作，从而能够快速原型设计新的线圈设计。对于给定的表面网格和目标场，该算法生成一个连接的、不重叠的线路径。
所提出的算法由几个步骤组成：流函数优化、二维表面投影、势离散化、拓扑轮廓排序、开轮廓和连接轮廓，以及最后添加不重叠的返回路径。为了质量控制，评估了几个技术参数，如电流强度、电感和场精度。
该方法在四个不同的示例中成功得到验证。所有的示例结果都表明，在达到各自的目标场方面具有很高的准确性。通过生成多个布局并改变输入参数值，可以找到给定流函数的最佳离散化。
所提出的算法允许快速生成具有高灵活性和低离散化误差的互连线圈布局。这可以减少整体后处理工作。本工作的源代码可在（https://github.com/Philipp-MR/CoilGen）上公开获得。
