Optimization of RF coil geometry for NMR/MRI applications using a genetic algorithm

A simulation method that employs a genetic algorithm (GA) for optimizing radio frequency (RF) coil geometry is developed to maximize signal intensity in nuclear magnetic resonance (NMR)/magnetic resonance imaging (MRI) applications. NMR/MRI has a wide range of applications, including medical imaging, and chemical and biological analysis to investigate the structure, dynamics, and interactions of molecules. However, NMR suffers from inherently low signal intensity, which depends on factors related to RF coil geometry. The investigation of coil geometry is crucial for improving signal intensity, leading to a reduction in the number of scans and a shorter total scan time. We have explored a better optimization method by modifying RF coil geometry to maximize signal intensity. The RF coil geometry comprises wire elements, each of which is a small vector representing the current flow, and GA chooses some of the prepared wire elements for optimization. The optimization of a substrate coil with a surface perpendicular to a static field was demonstrated for single-sided NMR system applications while considering various cylindrical sample diameters. A non-optimized and a GA-optimized substrate coil were compared through simulation and experiment to confirm the performance of the GA simulation. The maximum error between simulation and experiment was below 5%, with an average of less than 3%, confirming simulation reliability. The results indicated that the GA improved signal intensity by approximately 10% and reduced the necessary total scan time by around 20%. Finally, we explain the limitations and explore other potential applications of this GA-based simulation method.