A homogeneous superconducting magnet design using a hybrid optimization algorithm

Abstract

This paper employs a hybrid optimization algorithm with a combination of linear programming (LP) and nonlinear programming (NLP) to design the highly homogeneous superconducting magnets for magnetic resonance imaging (MRI). The whole work is divided into two stages. The first LP stage provides a global optimal current map with several non-zero current clusters, and the mathematical model for the LP was updated by taking into account the maximum axial and radial magnetic field strength limitations. In the second NLP stage, the non-zero current clusters were discretized into practical solenoids. The superconducting conductor consumption was set as the objective function both in the LP and NLP stages to minimize the construction cost. In addition, the peak–peak homogeneity over the volume of imaging (VOI), the scope of 5 Gauss fringe field, and maximum magnetic field strength within superconducting coils were set as constraints. The detailed design process for a dedicated 3.0 T animal MRI scanner was presented. The homogeneous magnet produces a magnetic field quality of 6.0 ppm peak–peak homogeneity over a 16 cm by 18 cm elliptical VOI, and the 5 Gauss fringe field was limited within a 1.5 m by 2.0 m elliptical region.

High Field Superconducting Magnet Technologies for Magnetic Resonance Imaging

“…The electromagnetic design includes the homogeneity in the over diameter spherical volume(DSV), the scope of fringe field, the safety of the superconducting magnet and the construction costs. The goal of design is to meet the requirement of homogeneity over the DSV region by arranging a plurality of superconducting solenoids within a predetermined coil region with rectangular cross-section, and 5 G fringe field should be constrained within a reasonable range for an active shielding MRI system [11][12][13][14][15][16][17][18][19][20][21]. A hybrid optimal model with a discrete current-carrying region to solve the electromagnetic inverse problems was proposed.…”

“…The local optimization method is based on previous experience of electromagnetic designers, and it searches on the pre-specified small area or route. The advantage is the high computational efficiency due to the small search range, and the drawback is difficult to ensure global optima [11].…”

“…In 1996, subsequently, in order to effectively reduce the effects of vibration coming from MRI system to image quality, pulse tube refrigerator successfully replaced GM refrigerator and is commonly used in magnetic resonance imaging systems. The IEE, CAS has been studying on all kinds of magnet and developed electromagnetic optimal design methods and software for MRI superconducting magnets [11][12][13][14][15][16][17][18][19][20][21]. The optimal objective function is the amount of superconducting wire which determines the cost of superconducting magnets.…”

“…This magnet structure overcomes technology problems of separation of coils in the high magnetic field. A 3 T superconducting magnet for animal MRI with horizontal warm bore [11] has been developed in IEE, CAS as shown in Fig.12. The magnet is about 1.2 m high and 0.8 m long.…”

Pareto-Optimization of HTS CICC for High-Current Applications in Self-Field

“…In the frame of superconductors, Parametric optimization was recently used to design the highly homogeneous superconducting magnets for magnetic resonance imaging [5], or in beam-optics design of a gantry for carbon ions in cancer therapy accelerators [6], or in the optimal design of saturated iron-core superconducting fault current limiter [7], or to support designers with a guideline for selecting a suitable machine topology for a 10 MW wind turbine application [8].…”