Programs in Physics & Physical Chemistry
|[Licence| Download | New Version Template] acse_v1_0.gz(71 Kbytes)|
|Manuscript Title: A finite element program package for axisymmetric vector field problems.|
|Authors: A. Konrad, P. Silvester|
|Program title: AXISYMM-VECTOR-HELMHOLTZ-FINTEL6|
|Catalogue identifier: ACSE_v1_0|
Distribution format: gz
|Journal reference: Comput. Phys. Commun. 9(1975)193|
|Programming language: Fortran.|
|Computer: IBM 360/75, IBM 370/158.|
|Operating system: OS/360 WITH HASP-II.|
|Program overlaid: yes|
|RAM: 50K words|
|Word size: 32|
|Keywords: Plasma physics, Electromagnetics, Axisymmetric, Poisson, Heat transfer, Electrostatics, Electron optics, Vector field, Helmholtz, Linear accelerator Cavities, Bessel, Laplace, Finite element, Collisionless plasma.|
|Classification: 10, 19.3.|
|ACSF_v1_0||VECTR-FINTEL6 BLK-DATA-GENERATOR||CPC 9(1975)194|
Nature of problem:
The following types of problems can be solved with the program:- electromagnetic field problems requiring solution to the one-component vector Poisson or Helmholtz equation in cylindrical co-ordinates with no azimuthal variation (e.g. transformer problems, unsaturated magnetic lens problems , linear accelerator cavities), -electric field problems requiring azimuthally periodic solutions to the scalar Laplace , Poisson or Helmholtz equation, i.e. solutions to the generalized Bessel equation of order m), -a host of mathematically related problems which fall in the above two categories (e.g. heat and fluid problems).
The generalized Bessel equation of order m, derivable from the vector Helmholtz equation from m=1 under the assumption of one vector component in the azimuthal direction and no azimuthal variation, or derivable from the scalar Helmholtz equation under the assumption of azimuthal periodicity, is solved by high-order polynomial triangular finite- element method using an approximation technique which avoids singularities of the 1/r type at the axis of symmetry.
Since the AXISYMM-VECTOR-HELMHOLTZ-FINTEL6 program is closely modelled after the previously published scalar finite element programs, the same restrictions apply. The computations, as in the scalar case, are carried out in single precision arithmetic. For large problems round- off errors may accumulate during the solution of the matrix eigenvalue equation. This of course depends on the word length of the computer used.
Most program features closely resemble those of the scalar programs. Since computers may not accept the block data subprograms supplied with the program, a FORTRAN program (VECTR-FINTEL 6-BLK-DATA-GENERATOR) which punches out the complete block data subprograms is made available to the users.
The running times are similar to those cited in Comp. Phys. Commun. 5(1973)437 and can be estimated from the running times of the matrix equation solving subroutines.
|Disclaimer | ScienceDirect | CPC Journal | CPC | QUB|