Estimation and reduction of temporal magnetic field fluctuations in powered magnets using inductive and NMR feedback control.

摘要

Powered magnets provide high magnetic fields that promise to significantly improve nuclear magnetic resonance spectroscopy (NMR). Higher fields increase NMR chemical shift resolution and signal-to-noise ratio (SNR) while decreasing quadrupolar line broadening in solids. High resolution NMR is typically performed using superconducting magnets, which are currently limited to 24 Tesla. Powered magnets can provide continuous fields up to 45 Tesla, significantly larger than that achievable by superconducting magnets. This will dramatically expand opportunities in the areas of material science, chemistry, and biology. However, temporal magnetic field fluctuations due to both the power supply and cooling water system currently render these magnets unsuitable for high resolution NMR.;The focus of this dissertation is to design, synthesize, and verify a feedback control system that reduces temporal field fluctuations so that powered magnets can be used for high resolution NMR. Earlier studies have shown that feedback control using inductive measurements significantly reduces higher frequency field fluctuations associated with power supply ripple, but are limited in their ability to reduce lower frequency field fluctuations associated with variations in the cooling water system. Conversely, feedback control using NMR measurements are more conducive to reducing lower frequency field fluctuations and less successful at higher frequencies. Feedback control systems which use NMR measurements are often referred to as field-frequency locks (FFLs). Earlier studies have shown that FFLs can estimate and reduce lower frequency field fluctuations in superconducting magnets, but have limited ability to do the same in powered magnets. This dissertation investigates why such FFLs are limited in powered magnets, and demonstrates some alternative methods for estimating lower frequency field fluctuations using NMR measurements in powered magnets. A digital sampled-data feedback control design using field fluctuation estimates from NMR measurements is combined with feedback control using inductive measurements, forming a cascade feedback control design that reduces both lower and higher frequency field fluctuations. This approach is experimentally verified in a powered magnet operating at 25 Tesla.