Nonlinear droop compensation for current waveforms in MRI gradient systems.

Area of Science:

• Magnetic Resonance Imaging (MRI)
• Electrical Engineering
• Control Systems

Background:

• Accurate gradient current generation is crucial for MRI quality.
• Gradient system nonlinearities from amplifiers and power supplies cause current droop, distorting waveforms.

Purpose of the Study:

• To develop a novel characterization approach for gradient systems, including amplifiers and power supplies.
• To create a nonlinear model capable of compensating for current droop in gradient coils.

Main Methods:

• A modified state-space averaging technique was used to characterize the gradient power amplifier and power supply stages.
• The nonlinear model was inverted for feedforward control of gradient coil current.
• High-resolution pulse-width-modulation signals drove custom gradient coils with linear and nonlinear controllers.

Main Results:

• Simulations and experiments validated the nonlinear model's effectiveness in characterizing the gradient system.
• The nonlinear controller reduced integral waveform error by 14-fold compared to the linear controller (0.13% vs. 1.9%).
• Phantom MRI images demonstrated correction of droop-induced distortions.

Conclusions:

• The proposed nonlinear characterization and feedforward controller enable droop-free gradient current waveforms without feedback.
• This approach significantly improves gradient current accuracy and MRI performance.