A Design-of-Experiments Approach to FEM Uncertainty Analysis for Optimizing Magnetic Resonance Imaging RF Coil Design

Published

October 10, 2014

Author(s)

Jeffrey T. Fong, Nathanael A. Heckert, James J. Filliben, Li Ma, Karl F. Stupic, Kathryn E. Keenan, Stephen E. Russek

Abstract

Using the RF module of COMSOL, we compute the magnetic flux density norm (BN) profiles for frequencies in the 76 to 100 MHz range, inside of a prototype birdcage coil, courtesy of Japan's National Institute of Radiological Sciences (NIRS), loaded with a cylindrical water phantom. At the first resonance of the lumped port impedance of the NIRS model, the BN profile was found to be highly non-uniform. A dimensionless metric for the non-uniformity of the profile is proposed as a parameter for assessing and improving the design of such a prototype bird cage coil. A statistical design of experiments (DOE) approach to the uncertainty analysis of the COMSOL solution of the NIRS model is used to minimize the BN non-uniformity metric. Significance and limitations of the DOE approach to uncertainty analysis as a design tool for MRI applications are presented and discussed.

Proceedings Title

Proc. 2014 International COMSOL Users' Conference, Boston, MA, U.S.A.

Conference Dates

October 8-10, 2014

Conference Location

Boston, MA

Pub Type

Conferences

Keywords

Birdcage coil design, design of experiments, design optimization, finite element method, magnetic resonance imaging.

Citation

Fong, J. , Heckert, N. , Filliben, J. , Ma, L. , Stupic, K. , Keenan, K. and Russek, S. (2014), A Design-of-Experiments Approach to FEM Uncertainty Analysis for Optimizing Magnetic Resonance Imaging RF Coil Design, Proc. 2014 International COMSOL Users' Conference, Boston, MA, U.S.A., Boston, MA (Accessed April 18, 2024)