Theory Weekly Highlights for January 2008

January 25, 2008

A new δf Eulerian code NEO-GK has been developed for numerical simulations of neoclassical transport as part of the Edge Simulation Laboratory (ESL) project. NEO-GK is based on a hierarchy of equations derived by expanding the drift-kinetic equation in powers of ρ*, the ratio of the ion gyroradius to the system size. Thus, unlike NCLASS, NEO-GK computes the neoclassical transport coefficients (particle flux, heat flux, bootstrap current, poloidal rotation, etc.) directly from solution of the distribution function. NEO-GK extends previous numerical studies by including the self-consistent coupling of electrons and multiple ion species and the calculation of the first-order electrostatic potential via coupling with the Poisson equation. Fully general geometry is included, using either a full numerical equilibrium or the Miller local parameterized equilibrium model. Three model collision operators have been implemented (the full Hirshman-Sigmar operator, the reduced Hirshman-Sigmar operator, and the Connor model) and comparisons of the resulting second-order transport coefficients are in progress.

January 18, 2008

Using peeling-ballooning stability calculations from the ELITE code, together with a developing working model of the pedestal width, we are able to make quantitative predictions for the DIII-D pedestal height based on control parameters which are known before an experiment is conducted. This working model of the pedestal height has been successfully tested against the DIII-D database and will be used to predict the pedestal height in upcoming DIII-D experiments.

January 11, 2008

The saturation rule for TGLF was extended to include an impurity ion species. The weightings of the intensity for different ion species were determined by comparing to GYRO simulations with helium and carbon. The results suggests that the ion charge density fraction provide the best weight. The kinetic impurity results were also compared with those based on the simple dilution model. For carbon simple dilution was found to be a good approximation, as long as the impurity density gradient length is not too different from the main ion density gradient length. Simple dilution is a poor approximation when the charge density fraction exceeds 50%.

January 04, 2008

A new coupling interface for the parallel global Newton method solver GCNMP was developed using Python procedure calls that allow a general client to request a service from the GCNMP server. The server can be on a remote host or reside on the same host as the client. It is expected that the client will run on a suitable single processor machine while the server will take full advantage of the parallel hardware it is running on. Tests using Linux workstations show that such an arrangement is satisfactory. During actual use, the client will make repeated calls to the server to advance the transport equations. All intervening inputs and outputs are through a state file. A development path was also initiated that will allow single source files to serve as common components for different projects.

Wenfung Guo of ASIPP-Hefei is visiting GA for 6 months to collaborate on integrated modeling development based on the EFIT equilibrium reconstruction and the ONETWO transport code.

These highlights are reports of research work in progress and are accordingly subject to change or modification