Optimization of 1 T HTS Magnet for Extremity MRI using Real-Coded Genetic Algorithm

Abstract An optimal design of a 1 T magnet consisting of Second-Generation (2G) High-Temperature Superconductor (HTS) tape wound in the shape of double pancake (DP) coils operated at 65 K for Extremity Magnetic Resonance Imaging (E-MRI) scanners using Real-Coded Genetic Algorithm (RGA) is presented. The minimum length of HTS tape consumption is used as a fitness function in the RGA with field inhomogeneity, length of HTS tape per DP coil, and coil stress act as constraints. The above algorithm is implemented by using a finite element method (FEM) package (i.e., COMSOL Multiphysics) along with MATLAB LiveLink. The FEM package is used to calculate all the parameters related to the HTS magnet, i.e., maximum perpendicular field ( B ⊥ ), central magnetic field, field inhomogeneity, stresses, and 5 gauss-stray field, while the selection, crossover, and mutation operations are evaluated in MATLAB. An optimum tape length of 5.27 km is required for producing a 1 T magnetic field with an inhomogeneity of 16.5 ppm inside an 80 mm Diamter of Spherical Volume (DSV).