Premium
Automatic Evaluation of Edge Transport Coefficients with B2‐SOLPS5.0
Author(s) -
Coster D.P.,
Kim J.W.,
Haas G.,
Kurzan B.,
Murmann H.,
Neuhauser J.,
Salzmann H.,
Schneider R.,
Schneider W.,
Schweinzer J.
Publication year - 2000
Publication title -
contributions to plasma physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.531
H-Index - 47
eISSN - 1521-3986
pISSN - 0863-1042
DOI - 10.1002/1521-3986(200006)40:3/4<334::aid-ctpp334>3.0.co;2-1
Subject(s) - divertor , enhanced data rates for gsm evolution , computational physics , physics , minification , position (finance) , plane (geometry) , grid , resolution (logic) , algorithm , computer science , statistical physics , mechanics , mathematics , tokamak , mathematical optimization , plasma , geometry , nuclear physics , telecommunications , finance , artificial intelligence , economics
One of the most critical parameters for the predictive capability of edge transport calculations is the description of the radial turbulent (anomalous) transport. Due to the complexity of the edge, highly sophisticated 2d numerical transport codes have been developed with run times per case of order of weeks, which limits their application to a small number of typical scenarios. Downgraded versions with a simpler neutral model, coarser grid (one half normal resolution) and eventually simplified radiation losses, can be fast enough (hours per run) to allow a routine mid‐plane profile analysis, at least for cases where divertor details are less important. Implementation of appropriate iteration and control loops allow an automatic fit of the anomalous transport coefficients with the new B2‐SOLPS5.0. The minimization algorithm is generalized such that additional experimental information such as neutral fluxes, divertor profiles or spectroscopic measurements, can easily be included for a better discrimination of different anomalous transport laws (constant transport coefficients, 1/n like scaled or Bohm‐like scaled). A special feature of the fit routine is the determination of the position of the separatrix from the power balance equations. This is necessary because the experimental uncertainty of about +/—0.5 cm is comparable to the experimentally observed gradient lengths.