This section describes the namelist settings which configure the NUBEAM fast ion Monte-Carlo code. NUBEAM also contains models for MHD instabilities, toroidal field ripple and finite Larmor radius effects amongst other things. The settings for these additional models are not described here, please refer instead to the Princeton TRANSP website. Some of the settings described here are written automatically to the namelist by the nblist program. These are indicated where relevant.
NLBEAM : When set to TRUE the neutral beam Monte Carlo model will be used. This will be automatically switched on by nblist
for shots with active beams.
NLBFPP : Set this to TRUE to use the Fokker-Planck beam model as opposed to the normal Monte Carlo model.
(Default FALSE)
NBEAM : The number of active neutral beams during the run. This is automatically set by nblist based on the information
in the PPF system.
DTBEAM : The timestep to be used by the beam code. This is automatically set by nblist but can be overwritten by
manually setting a value for DT_SOURCES.
If DT_SOURCES is NOT set then this specifies the step size for all heating codes.
NPTCLS : This is the fixed number of marker particles which will be followed by the Monte-Carlo code during the simulation.
It is referred to as the Constant Census. A typical value is ~50,000.
GOOCON : This is the Monte-Carlo goosing parameter which controls the acceleration of collisional effects with respect
to orbit effects in order to improve computational efficiency. This effectively sets the number of bounce orbits to be completed between collisional processes.
GOOCON=5 : Good but expeditious jobGOOCON=10 : DefaultGOOCON=20 : Excellent representation of banana transport
EBDMAX : The maximum energy considered when constructing distribution functions. This should be somewhat higher than the input beam
energy to accommodate scattering leading to particles with a larger energy than they had initially. It is set automatically by nblist
DXBSMOO : This sets the half-width in r/a to use when smoothing the profiles derived from the Monte Carlo model. (Default=0.05).
WGHTA : This variable controls the radial distribution of monte carlo particles. Increasing it from 1 leads to an increased number
of reduced weight monte carlo particles being generated in the core region while a reduced number of increased weight particles is generated at the edge such that the
constant census is preserved. This can lead to reduced statistical variance on quantities in the core and may be useful in shots with high density and poor beam penetration.
WGHTA may take a value between 1 and the number of radial zones used in the simulation.
NMCURB : This variable sets the method used for the calculation of neoclassically shielded beam driven currents
(see also magnetics).
It is common at JET to use setting 3 which utilises the updated neoclassical model.
DN0OUT : This variable sets the neutral density seen by particles which orbit outside the plasma boundary. Setting this to a large number
(e.g. 1e11) while cause rapid charge-exchange loss of these ions.
DTN_NBI_ACC : This variable is the orbit timestep control (replacing DTN). It sets the accuracy criterion "per bounce" for orbit integration.
The code interprets this somewhat conservatively, and the default value of 1.0e-3 is usually sufficient. Reducing this to 1.0e-4 will slow NUBEAM significantly and in standard
regression tests there aren't any other change in observable outputs.
The following controls can be used to manually set the random seed for the Monte-Carlo code, however this is automatically set differently for each run based on the system
clock so these settings are usually omitted. They can be used in order to exactly reproduce a previous run. The seed value for a given run can be found in the
file 'runID'TR.MSG.
NLSEED : Set this to TRUE in order to manually set the random number seed using the value of the namelist variable
NSEEDI.
LEV_NBIDEP : This variable allows the user to choose the deposition model used by NUBEAM. The following options are available:
NDEPMOD : Setting this to 1 will enhance the thermal plasma stopping cross-sections by an anomalous multiplier given by the namelist variable XDEPMOD. This is by default set to 0. When used, it only affects beam deposition.
TRANSP includes two models for calculating an enhancement factor to the fast neutral deposition and charge exchange recapture cross sections due to excited states. The first model calculates this enhancement factor by extending to lower energies a fit described in the paper "Penetration of Energetic Neutral Beams Into Fusion Plasma", Nucl. Fusion, 29, No. 12 (1989), 2125-2139 by Janev, Boley & Post. The fit is a function of the target plasma electron density, temperature and ZEff. The ion temperature and fast ion distribution are only roughly accounted for. Note that the fit in the paper is not the one actually used in TRANSP. The calculated enhancement factor is then applied equally to all deposition and charge exchange processes (CX, ionisation, beam-thermal and beam-beam). For more detailed information on this contact M. Gorelenkova at PPPL.
The second model is based on the ADAS310 fortran library. This calculates the excited state populations
of the atoms for the hydrogen isotopes as well as the resulting effective ionisation and charge exchange coefficients. Note that beam-beam interactions are still calculated in the ground
state in this model.
The choice of which excited states model to use is made using the switch NSIGEXC :
The current recommendation is to use the ADAS ground state deposition model and the ADAS excited state model (LEV_NBIDEP=2, NSIGEXC=3), however the ADAS excited
states model does not currently work at JET, it is work in progress to fix this.
NLBBCX : Set this flag to .TRUE. (default) to use beam-beam contributions to deposition. If set to .FALSE. these will
not be considered.
NBBCX_BB : If set non-zero then contributions from beam-beam interactions to the enhancement factor will not be considered. The default is for this
to be set to 0. This does not apply to the ADAS excited states model, only Janev, Boley, Post.