A GPU Computational Code for Eddy-Current Problems in Voxel-Based Anatomy

Author

Bottauscio, Oriano ; Chiampi, Mario ; Hand, Jeff ; Zilberti, Luca

Author_Institution

Ist. Naz. di Ricerca Metrol., Turin, Italy

Volume

51

Issue

3

fYear

2015

fDate

Mar-15

Firstpage

1

Lastpage

4

Abstract

A hybrid finite element-boundary element method, developed to solve eddy-current problems in the frequency domain, is applied to the electromagnetic analysis of voxel-based human models. A specific procedure employs a massively parallelized algorithm implemented on a multiple Graphic Processing Units (GPU) code to speed up the solution of large systems whose matrix exceeds the RAM capability. The database structure used for the electromagnetic problem is also suitable for a successive thermal analysis to evaluate the distribution of the temperature elevation due to the energy deposited by the waves in the tissues. Finally, some examples of application are presented.

Keywords

biological effects of fields; biological tissues; biothermics; boundary-elements methods; eddy currents; finite element analysis; graphics processing units; random-access storage; temperature distribution; thermal analysis; GPU computational code; RAM capability; biological tissues; database structure; eddy-current problems; electromagnetic analysis; electromagnetic problem; frequency domain; hybrid finite element-boundary element method; multiple graphic processing units code; parallelized algorithm; temperature elevation distribution; thermal analysis; voxel-based anatomy; voxel-based human models; Biological system modeling; Coils; Electromagnetics; Graphics processing units; Magnetic resonance imaging; Mathematical model; Radio frequency; Biological effects of electromagnetic radiation; boundary element (BE) method; eddy currents; finite element (FE) method; magnetic resonance imaging (MRI);

fLanguage

English

Journal_Title

Magnetics, IEEE Transactions on

Publisher

ieee

ISSN

0018-9464

Type

jour

DOI

10.1109/TMAG.2014.2363140

Filename

7093431