Dalton Schnack, Scott Kruger, and Alfonso Tarditi,
Science Applications International Corp., San Diego
Ahmet Aydemir, Institute for Fusion Studies, University of Texas, Austin
Ming Chu, General Atomics, San Diego
Thomas Gianakon, Richard Nebel, and Carl Sovinec,
Los Alamos National Laboratory
Charlson Kim and Scott Parker, University of Colorado, Boulder
Jean-Noel Leboeuf, University of California, Los Angeles
Steve Plimpton, Sandia National Laboratories, Albuquerque
Nina Popova, Moscow State University

Research Objectives

The goal is to develop a fusion plasma simulation code which provides both flexibility in the physics-by using two-fluid or magnetohydrodynamic (MHD) models with analytic or gyrokinetic closures-and flexibility in the geometry-allowing studies of any axisymmetric fusion concept, no matter how complicated the cross-section. NIMROD is designed to be user-friendly (but not simple), and is available to the entire fusion community. Because NIMROD was designed to include massively parallel constructs, the code can take advantage of the most powerful MPP machines to tackle the largest problems in fusion.

Computational Approach

	Closed flux surfaces generated in NIMROD simulations of spheromaks. NIMROD computations of z-pinches with embedded axial magnetic field produce sustained spheromaks as the nonlinear saturation of an MHD instability. Weakly driven cases yield flux surfaces threaded by a helical current column. Two views of a surface, with color indicating radial position, illustrate the helical distortion on the inboard side. (Carl Sovinec, John Finn, and Diego del Castillo Negrete, LANL)

NIMROD uses the extended MHD model to simulate electromagnetic plasma behavior. The code has a time-split, semi-implicit advance and a combined finite element/Fourier series spatial representation. This algorithm has been designed to run on massively parallel computers, while being able to handle the extreme stiffness of MHD problems in fusion plasmas. Normal modes of the system propagate across the domain in times that are orders of magnitude smaller than the behavior we wish to study. Therefore, we have paid particular attention to avoiding numerical dissipation in the part of the algorithm associated with wave propagation. We have also paid considerable attention to ensure that truncation errors do not lead to unphysical coupling of compressional and shear responses.

Accomplishments

The NIMROD code, currently at version 2.3.2, is maturing rapidly. The code performance has more than doubled in the past year, and we are actively working on further improvements. The physics models are being extended, with two analytic closure schemes implemented in the past year, as well as further use of the two-fluid equations. The code is broadening its user base, with simulations done for General Atomics, UCLA, the University of Washington, and the University of Wisconsin. In the past year the team has continued its extensive validation campaign, including the numerically challenging kink-ballooning case. In addition, challenging first-of-a-kind problems have exercised the code: neoclassical tearing modes in DIII-D geometry, spheromak sustainment and production, field-reverse configurations, and toroidal reversed field pinch (RFP) simulations. The breadth of these simulations testifies to the flexibility of the code.

Significance

Publications

A. H. Glasser, C. R. Sovinec, R. A. Nebel, T. A. Gianakon, S. J. Plimpton, M. S. Chu, D. D. Schnack, and the NIMROD Team, "The NIMROD code: A new approach to numerical plasma physics," Plasma Phys. and Control. Fus. 41, A747 (1999).

http://www.nimrodteam.org/