Title :
Performance of Low-Rank QR Approximation of the Finite Element Biot-Savart Law
Author :
White, Daniel A. ; Fasenfest, Benjamin J.
Author_Institution :
Lawrence Livermore Nat. Lab., CA
Abstract :
We are concerned with the computation of magnetic fields from known electric currents in the finite element setting. In finite element eddy current simulations it is necessary to prescribe the magnetic field (or potential, depending upon the formulation) on the conductor boundary. In situations where the magnetic field is due to a distributed current density, the Biot-Savart law can be used, eliminating the need to mesh the non-conducting regions. Computation of the Biot-Savart law can be significantly accelerated using a low-rank QR approximation. We review the low-rank QR method and report performance on selected problems
Keywords :
eddy currents; electromagnetism; finite element analysis; magnetic fields; conductor boundary; distributed current density; finite element Biot-Savart law; finite element eddy current simulations; magnetic field computations; nonconducting regions; Boundary conditions; Computational modeling; Conductors; Eddy currents; Electrostatics; Finite element methods; Magnetic fields; Matrix decomposition; Partitioning algorithms; Sparse matrices;
Conference_Titel :
Electromagnetic Field Computation, 2006 12th Biennial IEEE Conference on
Conference_Location :
Miami, FL
Print_ISBN :
1-4244-0320-0
DOI :
10.1109/CEFC-06.2006.1633142