From: VAXK::LAO Date: 8-OCT-1991 23:27:35

This memo is intended to clarify some of the questions you have concerning the proper form of current profile to use when running the EFIT code to reconstruct the MHD equilibrium using only external magnetic data.

The representation of the current profile is specified through the EFIT_ SNAP.DAT file. As some of you may be aware, several SNAP files have been developed to accomodate the wide range of plasma configurations and transport regimes in DIII-D. There are presently five SNAP files available in the EFIT: area. They are described below. All of them yield plasma stored energy and volume to within about 5%, li to within about 10-15%, and very similar shapes. However, the vertical position of the separatrix boundary at the Thomson location may differ by about 1-3 cm. A detail physics report summarizing the results is forthcoming. We should emphasize here that if you are interested in the detail of the separatrix boundary near the Thomson location or near the X point, a full equilibrium reconstruction using addtional kinetic profile, MSE, and SXR information is advisable, although it is tedious and can be very time consuming. When you use the SNAP files to run EFITD, please exercise with care, follow the recommendations described below, and contact us if you have questions. Briefly, the available EFIT_SNAP.DAT_* files are

_DEF

no edge current density is allowed, good description for Ohmic and L-mode plasmas, highly robust. Tends to give Z Thomson larger by about 2-3 cm for H-mode plasmas with low edge collisionlity comparing with other forms described below, which allow for some form of edge current density.

_J

finite current density allowed. P' and FF' each has two fitting parameters. Reasonable description of H-mode plasmas. Sometimes yields a Thomson location larger by about 1 cm. Less robust.

_JT

edge current density is constrainted to vanish weakly. 3 parameters in FF' and 2 in P'. Reasonable description of H-mode plasmas. Usually yields a smaller Thomson location than the others. Less robust. Been used mostly for timeslices in boundary database.

_31

edge current density allowed. 3 paramters in in FF' and 1 in P'. Recently developed to model VH mode plasmas. Appears to yield sightly better results than others for VH mode plasmas. Still being tested. Detailed comparison for the time evolution of shots 72220 and 73204 using SNAP files 2,3, and 4 is available if interested.

_SCRAPE

force free current is allowed in the scrape-off layer. Been used mostly for vertical disrupted and biased plasmas. Some experience is required to use it.

A brief description of the relevant input variables can be found near the end of the SNAP files. A general HELP file is avaible on the VAXs by typing HELP EFIT. A bulletin board on the VAXs will also be available soon.

For the GSLs in the YOKA data base, the timeslices are typically routinely analyzed using the top 3 SNAP files and then screened for the best results within the constraint of the available manpower. If you have question or suggestion about a particular time slice, please send us a mail message on the VAXs. The timeslice will be re-examined and then sent to the YOKA data base again if needed.

Obviously, we would like to have a single current profile parameterization to accomodate as wide a range of DIII-D operating parameters as possible. However, the situation is not unlike trying to develop a good single set of knot locations for use in the ENERGY code. Thus, it may not come easy. We are still attempting to work out the best strategy. Some possible solutions may be to incorporate the edge Thomson measurements and/or the upcoming multi-channel MSE into the routine analysis for a subset of timeslices.

In the mean time, if you have a question or suggestion please either contact us or send us a message on the VAXs.