1. Code Name: DEGAS 2
2. Category: V. Source/Sink Codes
3. Responsible Physicist: Daren P. Stotler
4. Others involved in code development: C. Karney, A. Y. Pigarov, R. J. Kanzleiter, S. Jaishankar, M. E. Rensink, D. P. Coster, H. Takenaga
5. One line description: Uses Monte Carlo techniques to compute transport of neutral atoms and molecules in the vicinity of material surfaces; general three-dimensional devices can be modeled.
6. Computer systems which code runs on: Sun OS, Solaris, Linux (Intel), IRIX (SGI), HP-UX, OSF1 (Dec Alpha), AIX (IBM), Unicos (Cray)
7. Typical running time: ~1 minute on UNIX workstation
8. Approximate number of code lines: 60,000
9.Does this code read data files from another code? Yes, output from the edge plasma code UEDGE can be read in, specifying the background plasma and geometry upon which the neutral transport calculation is based.
10. Does this code produce data files that can be read by another code? An output file containing plasma source information can be read by UEDGE.
11. 1-2 paragraph description of code: Neutral transport in fusion devices is computed using a Monte Carlo technique in general three-dimensional geometries. DEGAS 2 has been designed as a state-of-the-art code featuring optimized geometry / tracking, dynamic memory allocation, and built-in parallel processing. General methods for handling atomic and surface physics data have been developed to make physics modifications easy. Neutral elastic scattering processes have been included. The code is designed to be run in a coupled fashion with fluid plasma codes, as well as in a stand-alone mode. It is intended to be portable with thorough documentation.
12. Similar codes to this code, and distinguishing differences: DEGAS [e.g., see D. Heifetz, D. Post, M. Petravic, J. Weisheit, and G. Bateman, J. Comp. Phys. 46 309 (1982)] is the predecessor to DEGAS 2. It uses the pseudo-collision algorithm for scoring, as opposed to the track-length estimator in DEGAS 2. The former works better in a dense plasma; the latter in a vacuum. Unlike DEGAS 2, the surface and atomic physics interactions are hard-wired into the code and are difficult to modify. EIRENE [e.g., see D. Reiter and A. Nicolai, J. Nucl. Mater. 128-129 458 (1984)] is also a Monte Carlo neutrals code. It also employs a track-length estimator for scoring reaction rates. Original user interface is arcane and difficult to use.
13. Journal References describing code (up to 3): D. Stotler and C. Karney, Contrib. Plasma Phys. 34 392 (1994). D. P. Stotler, C. F. F. Karney, M. E. Rensink, and T. D. Rognlien, "Coupling of Parallelized DEGAS 2 and UEDGE Codes" in Proceedings of the Seventh International Workshop on Plasma Edge Theory in Fusion Devices, (Tajimi, Japan, October 1999) Contrib. Plasma Phys. (to appear 2000). Randall J. Kanzleiter, "The Modeling and Impact of Neutral Elastic Collisions in Low-Temperature High-Density Plasmas", Ph. D. Thesis, Rensselaer Polytechnic Institute (1999).
14. New code capabilities planned for next 1-2 years: Revised user interface to geometry. New output graphics tool. Addition of sputtered impurity sources. Coupling to B2.5.
15. Code users: H. Takenaga (JAERI), G. Porter (LLNL), D. Alman (UIUC), B. Xiao (U. of Tokyo), M. Joung (KBSI)
16. Present and recent applications of code: Modeling of divertor baffling experiments on Alcator C-Mod
17. Status of code input/output documentation. Check one: ( ) does not exist (X) incomplete ( ) exists
18 Year Code was first used and present frequency of use: 1995
19. Estimate of Man-Years invested in developing code: 10
20. Catagories of usage of Code (Check all that apply): (X) application code to do analysis and prediction of experiments ( ) numerical testbed of theoretical ideas (X) physics module to be used in integrated moddelling (X) code for machine design
21. Language code is writen in: FORTRAN with FWEB preprocessor
22. Results of intercomparisons with other codes and results of validation against experiments. Thoroughly benchmarked against original DEGAS and EIRENE. Also benchmarked in various limits against analytic models. Currently being benchmarked against Alcator C-Mod experiments.