An Integrated Approach to Shim Coil Design and Current Regulation for NMR Systems
ID：111
Submission ID：116 View Protection：ATTENDEE
Updated Time：2025-10-13 11:25:12 Hits：77
Poster Presentation
Start Time：2025-11-09 09:10 （Asia/Shanghai）
Duration：1min
Session：[P] Poster presentation » [P6] 6.AI-driven technology
No files
Abstract
Magnetic field homogeneity is a key determinant of signal sensitivity and spectral resolution in high-field nuclear magnetic resonance (NMR) systems. To address the limitations of conventional active shimming techniques—namely, the restricted precision of shim coil design, the time-consuming iterative adjustment of coil currents, and the lack of robustness validation against engineering errors—this paper proposes an integrated approach that combines shim coil design with current optimization. First, first- to third-order shim coils are designed using the target field method and the stream function method. A genetic algorithm is then employed to explore a wide range of current combinations and identify a “near-optimal solution region,” followed by gradient-based refinement to obtain the optimal shim current set. To verify the practical engineering applicability of the proposed method, a 5T NMR system was selected as the research object. Monte Carlo simulation was used to model the wire dimension tolerance and coil position tolerance, generating 1000 sets of magnetic field drift data. The shimming results showed that this method can reduce the peak-to-peak magnetic field inhomogeneity within a 10mm-diameter spherical target volume (DSV) from an initial maximum of 937.7ppm to 7.2ppm, significantly outperforming traditional methods in efficiency and stability.
Keywords
superconducting magnet,active shimming,Nuclear magnetic resonance,Genetic Algorithm (GA)
Speaker
Comment submit