Elsevier Science Home
Computer Physics Communications Program Library
Full text online from Science Direct
Programs in Physics & Physical Chemistry
CPC Home

[Licence| Download | New Version Template] abhj_v1_0.gz(177 Kbytes)
Manuscript Title: TUBE88, a code which computes magnetic field lines.
Authors: A.A. Mirin, D.R. Martin, N.J. O'Neill
Program title: TUBE88
Catalogue identifier: ABHJ_v1_0
Distribution format: gz
Journal reference: Comput. Phys. Commun. 54(1989)183
Programming language: Fortran.
Computer: CRAY-XMP.
RAM: 330K words
Word size: 64
Keywords: Magnetic field lines, Cylindrical coordinates, Plasma physics, Magnetic confinement.
Classification: 19.9.

Nature of problem:
TUBE88 computes magnetic field lines in cylindrical or toroidal geometry and calculates the intersections of those field lines with specified planes. A cylindrical coordinate system (r,phi,z) is used.

Solution method:
The magnetic field line equations are integrated using the fourth order Adams-Moulton method, initiated by a fourth order Runge Kutta algorithm. The stepsize is dynamically determined. The equations are transformed to Cartesian coordinates in the vicinity of r=O.

Restrictions:
The magnetic field may be computed in several ways: (a) through specification of currents flowing in specified helical and circular elements together with a "l/r" field (from a central conductor) and a vertical field; (b) as a Fourier series in Phi; the magnetic field B may or may not be specified in a form guaranteed to be divergence free; and (c) in a specific cooordinate system suited to a toroidally helical domain. In cases (a) and (c) an auxiliary magnetic field may be specified separately.

Unusual features:
A separate source, in which most of the output is plotted rather than printed, is provided for those users who have access to a CTSS system. Also, COMMON variables are assumed to be preset to zero.

Running time:
The computer time is highly dependent on the form of the magnetic field, the number of field lines traced, the length per field line, the specified accuracy, the change in step size, etc. Test Problem 1, which traces four field lines for ten thousand steps each in a torus whose magnetic field is computed from four helical windings, three circular coils and a central conductor, takes 1.38 minutes of CPU time on the Cray-XMP. Test Problem 2, which traces four field lines for four to ten periods (418 steps) along a periodic cylinder in which the magnetic field is simply specified, takes 0.04 minutes. Test problem 3, which traces six field lines for ten thousand steps in a toroidally helical domain in which the magnetic field is given by a double Fourier series, takes 0.57 minutes of CPU time.