Improving the Performance of Magnetic Resonance Systems using Feedback Control

Meng, Xinxing

Improving the Performance of Magnetic Resonance Systems using Feedback Control

Restricted

Author:: Meng, Xinxing
Graduate Program:: Electrical Engineering (PHD)
Degree:: Doctor of Philosophy
Document Type:: Dissertation
Date of Defense:: April 08, 2021
Committee Members:: Minghui Zhu, Major Field Member
Constantino Lagoa, Major Field Member
Patrick Lenahan, Outside Unit & Field Member
Thomas Neuberger, Outside Field Member
Jeffrey Schiano, Chair & Dissertation Advisor
Kultegin Aydin, Program Head/Chair
Keywords:: QR Detection
Reducing AM RFI
RF Pulse Sequence Manipulation
Feedback Control
Magnetic Field Fluctuations
Flux Regulation
Powered Magnet
35.2-T Series-Connected Hybrid Magnet
MRI
Abstract:: This dissertation is divided into two parts. The first part investigates techniques for reducing amplitude modulation (AM) radio frequency interference (RFI) in quadrupole resonance (QR) detection systems, while the second part focuses on magnetic field regulation in powered magnet towards enabling MR imaging. In the first part, this dissertation proposes a method of for determining the RF pulse spacing tau and pulse sequence separation Td for QR detection. It shows the AM RFI can be reduced by averaging if tau and Td are properly chosen prior to QR detection to enable destructive interference among the acquired AM signals. Experiments show tuning tau can reduce the AM noise by 20 dB. Simulations show that tuning both tau and Td can provide an additional 50 dB attenuation than using the nominal values. The second part of this dissertation improves flux regulation in powered magnets. While higher field strengths can enable cutting-edge research, the physical properties of low temperature superconductors (LTS) limit the field strength. Powered magnets can generate magnetic fields much higher than LTS persistent magnets, but the system suffers from magnetic field fluctuations. This dissertation quantitatively predicts the effect of field fluctuations in the powered magnets on MR images with the help of Structural Similarity Index (SSIM) metric. This dissertation identifies the model of an existing flux regulation system, and proposes an improved controller to further reduce the low-frequency field fluctuations in the 35.2-T Series-Connected Hybrid (SCH) magnet. Experiments show the improved controller can reduce field fluctuations by at least an additional 20 dB than the original controller to reject disturbances below 10 mHz in the SC magnet. The new control design is also augmented with an internal model principle (IMP) controller to attenuate a strong disturbance at 0.3 Hz caused by the magnet cooling system.

Tools