Multiple layer superconducting magnet design for magnetic resonance imaging

A design method of multiple layers superconducting magnet for MRI is introduced. Considering the actual coils layout, the candidate domain is divided into several layers. The current density curves of each layer are solved by Quadratic Programming optimization with Minimum Stored Energy. The seed coils are arranged at the peak positions of current curves. Based on the regularization solution, the current of the seed coils can be solved. According to the coil's current, selecting the current density of the superconducting wires, the original sections for the coils can be obtained. Then a further optimization about the homogeneity based on constrained nonlinear multivariable optimization method is employed to determine the final coils' geometries. The method is especially suitable for short superconducting MRI magnet design.