The figure shows three isosurfaces of the pressure from a 3D resistive MHD simulation of high

Parvez Guzdar, Bill Dorland, James Drake, Adil Hassam, Robert Kleva, and Barrett Rogers, University of Maryland

Research Objectives
The Maryland Theory and Computational Physics Magnetic Fusion Energy Program focuses on (1) 3D simulation of particle, ion, and electron energy transport in the core and edge region of tokamak plasmas using two-fluid Braginskii equations, gyrofluid equations, and Vlasov codes; (2) 3D simulation of high- disruptions, sawtooth crashes, and pellet injection for tokamak plasmas using a toroidal resistive magnetohydrodynamics (MHD) and two-fluid code; (3) 2D and 3D simulations of novel centrifugal confinement devices using MHD codes; (4) 2D and 3D hybrid simulations of collisionless reconnection; and (5) 3D gyrokinetic simulations of the levitated magnetic dipole experiment (LDX) at the Massachusetts Institute of Technology.

Computational Approach
For all the two-fluid Braginskii and MHD codes for the studies listed above, the basic codes solve a coupled system of convection-diffusion equations. For these codes we use a leapfrog trapezoidal algorithm for the time stepping and a fourth-order up-wind finite differencing scheme for the spatial convective derivatives.

The second-order accurate gyrokinetic algorithm we use is comprised of an implicit treatment of the linear dynamics; an explicit, pseudo-spectral treatment of the nonlinear terms; and an Adams-Bashforth integrator in time. Parallelization is accomplished with MPI and SHMEM libraries. The gyrokinetic problem involves the usual 3D spatial grid, as well as a 2D velocity space grid, for a total of five dimensions. We divide four of the dimensions at a time over processors to achieve a high degree of parallelization with good load balancing.

Our 3D gyrofluid code shares all communication and pseudo-spectral evaluation modules with the gyrokinetic code, as well as the design philosophy. Two of the spatial domains are spread among processors at any given time.

Accomplishments
During the last year we made progress on three areas which required large-scale computation: (1) nonlinear simulations of modes to understand electron energy transport, (2) hybrid simulation for the study of collisionless reconnection, and (3) high-resolution simulations of high disruptions of tokamaks.

Significance
These studies focus on the most important problems for understanding the issues of anomalous confinement and MHD, and collisionless disruption and reconnection processes that limit the parameter space of stable operation of tokamak devices. The codes that we have developed may also be used to study the confinement and stability properties of non-tokamak devices, such as the levitated dipole and centrifugal confinement configurations. The physics that we learn from these studies has motivated the study of a novel centrifugal confinement device which incorporates the positive features of shear flow suppression of microinstabilities and the associated reduction in anomalous transport.

Publications
M. Shay, J. Drake, B. Rogers, and R. Denton, "The scaling of collisionless, magnetic reconnection for large systems," Geophys. Res. Lett. 26, 2163 (1999).

A. Zeiler, J. Drake, and B. Rogers, "Magnetic reconnection in toroidal mode turbulence," Phys. Rev. Lett. 84, 99 (2000).

R. G. Kleva and P. N. Guzdar, "Nonlinear stability limit in high tokamaks," Phys. Plasmas 7, 1163 (2000).

http://www.ipr.umd.edu/Theory/research.htm