Affiliations: [a] Institute of Smart Sensors, University of Stuttgart, Germany | [b] Institut für Technische und Makromolekulare Chemie, RWTH Aachen University, Germany
Correspondence:
[*]
Corresponding author: Belal M.K. Alnajjar, Institute of Smart Sensor, University of Stuttgart, Pfaffenwaldring 47, 70569 Stuttgart, Germany. Tel.: +49711685 60868; Fax: +49711685 67222; E-mail: belal.alnajjar@iis.uni-stuttgart.de
Note: [†] These authors contributed equally.
Abstract: An H-shaped low-field permanent magnet is presented. It has been designed, simulated and optimized using the finite element method. The magnet mainly consists of two poles, each of which consists of 37 hexagonal magnetic blocks that are magnetically connected by an iron yoke. To achieve the magnetic field homogeneity in the ppm range required for NMR applications, mechanical shimming by a vertical adjustment of the hexagonal magnetic blocks has been applied. Here, the displacement of each block from the initial equidistant symmetric design has been optimized by applying a pattern search method. In addition, two pole pieces made of iron have been added to the magnet to further improve the magnetic field homogeneity. The dimensions of these pole pieces have been optimized using a genetic algorithm. The magnetic field homogeneity of 15 ppm (parts per million) measured inside the manufactured prototype verifies the accuracy of the proposed simulation based design approach.
Keywords: Finite element methods, genetic algorithms, mobile NMR, NMR magnet, permanent magnet design, pattern search algorithms, shimming
