Dielectric Breakdown in a Simplified Parallel Model

MOVIE , code (CADMUS).

Publications:

• Howland A. Fowler, Judith Devaney, John G. Hagedorn, Francis E. Sullivan, "Dielectric Breakdown in a Simplified Parallel Model", NISTIR 6174, June, 1998, Computers in Physics, September-October, 1998.
• Howland A. Fowler, Judith Devaney, John G. Hagedorn, "User Guide to CADMUS, a Simplified Parallel Code for Laplacian Fractal Growth", NISTIR 6180, June, 1998.

Abstract
Howland A. Fowler*, Judith E. Devaney, and John G. Hagedorn
Mathematics and Computational Science Division, ITL
National Institute of Standards and Technology
Francis E. Sullivan, Director,
Center for Computing Sciences, Institute for Defense Analysis
*Guest researcher

The growth of streamer trees in insulating fluids (a submicrosecond process which triggers high-voltage breakdown) has been simulated with a combination of parallel-coding tools. Large grids and arrays display well the multifractal, self-avoiding character of the streamer trees. Two physical cases have been approximated by power-law weightings of the statistical growth filter: dense anode trees, in uniform field; and sparse cathode trees (a rarer experimental case).
The coding combines Fortran 90, NIST's DPARLIB extensions, and an underlying (invisible) MPI environment. Block-partitioning creates processes of reasonable size, which operate in parallel like small copies of the original code. The user needs only to express his model in transparent array-directed commands; parallel interfacing between blocks is handled invisibly. The Fortran 90 commands, extended by DPARLIB across partition boundaries, can address very large arrays without explicit looping. After the serial version was composed and tested on Sun and SGI workstations, the code was run on 9 nodes of the IBM SP-2, a multiple-processor machine.
Display is carried out in 3-D. Timing of the growth can be shown by color banding or by frame-animation of the results. The adequacy of the growth rules and size scaling are tested by comparing the simulations against experimental data.