Text Box: NCSX Memorandum

To:        T. Brown, A. Brooks, H-M. Fan, B. Nelson, D. Williamson, M. Cole, P. Goranson

CC:       J. Schmidt, H. Neilson, P. Heitzenroeder, J. Lyon, S. Hirshman, M. Zarnstorff, A. Reiman. L. Berry, D. Strickler, R. Simmons, J. Graham, C-H. Jun

From:    Wayne Reiersen

Date:     12/1/2000

Re:        Minutes of 9/6 Engineering telecon

Presentations: [1] CHJ_000906.pdf [2] MJC_000906_NBAccess.htm [3] EvaluationCriteria.htm

Our weekly telecon was conducted on Wednesday, 9/6.  The main topics were plasma surface quality, NB access, and evaluation criteria for selecting between the saddle and modular coil options.

Chang-Hoon Jun presented the current PIES results to calculate the plasma surface quality with saddle and modular coil designs [1].  The purpose of this work is to see if there is a clear distinction in the surface quality of plasmas produced by the two different coil types.  Previous fixed boundary calculations of the LI383 plasma configuration were presented by Reiman at the PAC meeting, as shown below.

 

The PIES reconstructions feature a large 3/5 island chain around s=0.8.  In addition, the zero beta reconstruction features a smaller 3/6 island chain around s=0.55.  The total island width is 15% at 4.2% beta and increases to 20% at zero beta.

Jun presented free boundary results for the saddle coils (page 9 in [1]).  The results are quite similar to the fixed boundary reconstruction above, except that:

*   An additional 10% of the toroidal flux is lost outside the last closed flux surface (s=0.64), which might be due to a 9/14 island chain

*   The 3/6 island chain, which was previously visible in the fixed boundary calculation only at zero beta, is also apparent at full beta

The resulting equilibrium, shown below for comparison with the fixed boundary calculations above, clearly does not meet the “less than 10% toroidal flux lost” criterion.  The volume inside the last closed flux surface is about 90% and about 20% of the toroidal flux inside of that is lost to islands.  The net result would be a lost of about 25-30% of the toroidal flux.

 

The results presented for the modular coils were suspicious.  It was agreed that the input data would be reviewed with Brooks and if changes were required, to re-run the modular case.

Action: Jun to review input data with Brooks and re-run modular case (if changes were in order).

In discussing the modular case, Lyon noted that the island structure for the 3/5 island chain looked unlike anything he had ever seen before.  There appears to be a doublet-shaped island (with an x-point) inside large islands also separated by x-points.  Jun agreed to discuss this observation with Monticello to determine if this was real or an artifice of the PIES code.

Another issue that came up in discussing the modular case was the need to document the basis of the calculation for all design basis calculations.  This discipline is essential.  The rule of thumb should be that the basis of the calculation should be documented and archived in such as way that we could come back a year from now and repeat the calculation, if necessary.

Action: ALL PROJECT PARTICIPANTS need to document results and archive data files for design basis calculations in such a way that the basis of the calculation (e.g. input coil data sets, plasma profiles, etc.) can readily be reviewed and the calculation repeated, if necessary.  This requirement should be applied to all design basis calculations, not necessarily to scoping or exploratory studies.  NCSX Management should establish appropriate standards.

Another issue raised in the context of the PIES reconstructions was the definition of the reference plasma for evaluating plasma performance.  The plasma reconstructed with PIES will likely have a volume inside the last closed flux surface that is different (probably smaller) than the volume calculated by VMEC.  In the case of the reconstruction with saddle coils, the volume is smaller by 10%.  This plasma could have significantly different properties than the full size plasma.  The toroidal flux assumed in VMEC should be adjusted to get consistency in volume between the VMEC and PIES calculations.  For this case in which the toroidal flux lost is unacceptably high, there is no real need to force consistency at this point.

Mike Cole reported on progress in evaluating NB access for the saddle coil option [2].  He pulled the solitary beams further back to avoid interference with the TF coils.  He indicated that the next step would be to explore options to avoid co- and counter-directed beams to point at each other.

We briefly discussed the evaluation criteria proposed by Reiersen [3].  The attached checklist is what we are proposing to fill out for the purpose of selecting between the saddle and modular coils.  It was noted that no compulsory requirements for flexibility have yet been defined.

The next engineering telecon will take place next Wednesday, September 13, at 1:30pm.  Please be prepared to provide updates on current tasks, especially those related to the upcoming downselection.  Thank you for your cooperation.