Premium
Remarks on computational magnetostatics
Author(s) -
Sussman M. M.
Publication year - 1988
Publication title -
international journal for numerical methods in engineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.421
H-Index - 168
eISSN - 1097-0207
pISSN - 0029-5981
DOI - 10.1002/nme.1620260415
Subject(s) - computation , stiffness matrix , finite element method , magnetostatics , scalar (mathematics) , stiffness , computer science , matrix (chemical analysis) , direct stiffness method , mathematics , conductor , mathematical optimization , static analysis , code (set theory) , computer program , algorithm , geometry , engineering , programming language , structural engineering , magnetic field , physics , materials science , quantum mechanics , set (abstract data type) , composite material
Abstract A two‐scalar‐potential formulation for magnetostatic computations has been recommended 4 based on a comparative study of several different formulations in terms of their accuracy and efficiency. A different study 3 challenged this recommendation. It is the purpose of this report to present results which confirm the practical necessity for the two‐potential formulation. The method used for the present study was to construct a simplified computer program designed to easily compare the alternative methods. This program uses the finite element method on a geometry suited for analysis of the ‘bifiliar conductor’ problem treated in both studies mentioned above. In this program, the alternative formulations are implemented with a great deal of common code. In particular, the input, stiffness matrix, matrix solution and editing routines are common to both formulations, with differences appearing only in the source terms for the stiffness matrix equation. The known analytic solution to this particular problem provides a reference with which the computed solutions is compared. The main conclusion of this study of the ‘bifiliar conductor’ problem is that for a given choice of mesh, the two‐potential formulation yields significantly more accuracy than the reduced‐potential formulation.