This document contains *very brief* descriptions of new TRANSP
features. Generally, detailed descriptions of namelist controls
and/or new code outputs associated with the new feature, are
described in the main TRANSP help file.
The date in the subtopic refers to the approximate date of availability
of the new feature.
Old news will be purged out of this document from time to time.
The namelist variable NCMODEL selects the NCLASS code base used
for the neoclassical analysis. Previously NCMODEL=1 was the default
which used Wayne Houlberg's original fortran 77 code. The default
is now NCMODEL=2 which uses Wayne's f90 NCLASS module. There are
a few additional features available with NCMODEL=2 described in
tranp.hlp along with the capability for executing NCLASS through
a stand alone program. The driver for NCMODEL=2 is based on
Wayne's forcebal program. Also the default for NLSQUEEZ has been
changed from .TRUE. to .FALSE. so squeezing is turned off. This
change is based on the recommendation to Bob Budny by Wayne.
To improve consistency of internal treatment of sources and current
drive terms, all internal arrays representing profiles of particle
sources, momentum sources, and power sources, have their area or
volume normalizations taken out. (Previously this was done for beam
heating but not other sources; now it is done for *all* sources, torques,
and powers). (Exceptions: neutral gas model profiles with neutral source
index; neutral gas model particle, momentum and power balance terms,
poloidal field diffisuion equation energy balance check, and certain
otherm sums for "output only" such as particle, momentum, and power
balance checks).
The motivation here is to treat all auxilliary heating sources and
related terms the same way as the beam sources are treated; also to
facilitate generation of an MKS-based communications mechanism btw
TRANSP and external predictive solvers.
The output seen by the user is unchanged.
For quantities appearing in "outcor/plotgen.i", volume normalizations
and area normalizations are re-inserted for output, with the expressions
(item)/DVOL(#1,2) or (item)/DAREA(#1).
The TRANSP internal COMMON has been replaced with a fortran-90 module
with dynamically allocated components. This means the following
former restrictions on code use have been removed:
(a) no limit on size/amount of input data
(b) "no limit" other than size of output file on number of radial
zones. A soft limit of 800 is still enforced; if the code is
run at this resolution some output utilities may need conversion.
NZONES up to 800 (or slightly more) may be set, but is not
necessarily recommended.
(c) "no limit" other than size of memory and output files on the
number of fast ion distribution function spatial zones-- i.e
NZONES_FB -- i.e. NZONES_FB > 20 may be set. Since the size
of this grid scales as NZONES_FB**2, NZONES_FB =~50 is probably
still a limit in practice. NZONES has to be an integer multiple
of NZONES_FB, as before.
For traditional runs (NZONES=20 or 50; NZONES_FB=10), the code will run
~10% faster, because it is not carrying statically allocated arrays far
bigger than needed to meet these modest resolution requirements.
A new feature has been added to the code, for simulation of MSE signals.
Please look for "N_MSESIM" in the "Transp Help" webpage:
http://w3.pppl.gov/~pshare/help/transp.htm
Miscellaneous changes to TRANSP due to requests received by D. McCune
on visit to JET & MAST (Culham) and IPP/Garching:
=> default value of DANOM_QLIM0 reset to 0.0
(no anomolous diffusion in "current hole" regions by default)
=> NLBEAMCX switch added; NLBEAMCX=.FALSE. to suppress Monte Carlo
fast ion charge exchange losses; NLBEAMCX=.TRUE. by default.
=> TABORT default changed, to avoid spurious crashes when TINIT=0.0.
The default value of the namelist control TAUPMN has been changed
from 0.0001 to 0.002 seconds -- this variable puts a lower limit
on ion species confinement times and hence an upper limit on recycling
flux at any given time in the simulation.
The file i/o of the xtranspin GUI has been redone, so that namelist
changes invoked through the GUI are edited into the namelist in a
"minimalist" fasion, and all user comments are preserved in place.
The Electron Cyclotron Resonance Heating (ECRH) package TORAY-GA
is now available in TRANSP. It can be invoked by the switch
NLTORAY=.TRUE. in the TRANSP namelist. In addition to this switch,
other parameters such as number of gyrotrons, on-off times,
and their positions and launch angles need to be specified.
Complete documentation on TORAY-TRANSP is available in Transp Help
( http://w3.pppl.gov/transp/).
The TRANSP output profile quantity PMHD_IN now contains the plasma
pressure as used by the MHD equilibrium solver (for TRANSP runs
which do equilibrium reconstruction, i.e. LEVGEO.NE.8). This is
a change from the old TRANSP where the profile PTOWB was always
the pressure profile input for equilibrium solution.
For most TRANSP configurations, PMHD_IN = PTOWB = the sum of pressure
components as defined in source/doc/beta.doc. However, some namelist
configurations allow these two to be different. For example, in
"current hole" experiments with zero or minimal current near the
plasma core, setting grad(PMHD_IN)=0 is often required in order to
obtain an equilibrium solution. For modeling purposes in such runs
users can set a lower limit on the toroidal current using
NLQLIM0 = .TRUE. ! use QLIM0 to force some toroidal current
QLIM0 = <upper limit on q near axis>
then in regions near the axis where q >= QLIM0 + DQ_PLIM0, the
pressure profile (PMHD_IN) is flattened. DQ_PLIM0 is a namelist
control with a default value of (-0.5d0).
TRANSP post-processing codes which used to read PTOWB may want to
consider reading PMHD_IN instead.
JET scientists have contributed a new option for TRANSP neoclassical
and bootstrap current: an all-collisionalities general geometry
model, using a Zeff parametrization for speed and simplicity. The
model was developed by O. Sauter, C. Angioni and Y. R. Lin-Liu.
Controls for the model are described in TRANSP HELP -- see the
"magnetics" "shared controls" subtopic under the namelist description.
reference: O.Sauter and C.Angioni, Phys of Plas 6(7), 2834 (1999)
Increasingly, C++ is used in certain TRANSP components. The gcc
C++ compiler (g++) has advanced significantly in recent years.
Older versions of gcc do not support C++ features that are now
widely expected in the software development community.
Therefore, TRANSP has now been upgraded and built and tested to run
with gcc3.2 (previously it ran with gcc2.96). (It has been tested
under the following configurations: SUN, Intel-Linux Redhat7.1 with
Lahey-Fujitsu fortran, ditto with Portland-Group fortran, and,
Alpha-Linux running Compaq fortran).
Sites building TRANSP on their local machines are urged to make gcc3.2
available on these systems. The TRANSP doc file "setcpp.login" shows
the environment variable definitions maded to support mixed fortran-C++
linkage on PPPL machines. These are examples only; other sites will
need different definitions depending on local system configuration.
There will be compile-time and load-time errors at sites still running
gcc2.X. Although no "essential" software is affected by such errors
at the moment, this could change in the near future.
The entire TRANSP kernel (solvers, internal geometry representation)
has been upgraded to real*8 (64 bit) floating point precision. Much
of the kernel was formerly coded for default real (32 bit) precision.
The maximum radial resolution of the code has been increased to 400
zones. Future upgrades will improve the efficiency of high resolution
operation.
TRANSP can now be run as a "computational service" over the network.
This capability, now being used by GA and MIT as well as PPPL to share
a common TRANSP production facility, has been created as part of the
SciDAC Fusion Collaboratory project ( http://www.fusiongrid.org ).
Potential Collaboratory users need to apply for a DOE Science Grid
identity, to meet the network security/authentication requirements
of this service. For more information see:
http://w3.pppl.gov/transp/ComputeServerAccess.html
The number of radial zones in the neutral gas model (FRANTIC)
is now independent of the number of zones NZONES in the main
TRANSP calculation. Set
NZONES_FRANTIC = <no. of evenly spaced radial zones in FRANTIC>
Note also that the name NZONES_FRANTIC is longer than 8 characters!
Names of length up to 16 characters are now permitted in the TRANSP
COMMON and namelists.
When measured Ti profile data is being used in TRANSP (Ti TRANSP
"analysis mode", NLTIPRO=NLTI2=.TRUE.), there are new options which
allow the Ti data to be suppressed for regions near the magnetic
axis, instead creating a region where Ti-Te is held constant or
forced to have a simple behavior. This might be useful in cases
where the Ti measurement is only valid in regions bound away from
the very core of the plasma.
Namelist switches: TIXLIM0, TIDXSW0, and TIFACX0 ... see
"Transp Help" namelist subtopic on ion temperature.
Wed Jan 14 10:38:43 EST 2004 -- pshare update
Thu Jan 22 17:47:53 EST 2004 -- pshare update
Mon Feb 2 02:17:15 EST 2004 -- pshare update
Fri Feb 6 11:44:57 EST 2004 -- pshare update
Thu Feb 26 19:04:25 EST 2004 -- pshare update
Mon Mar 1 15:10:24 EST 2004 -- pshare update
Mon Mar 22 19:00:15 EST 2004 -- pshare update
Tue Mar 23 11:37:17 EST 2004 -- pshare update
Wed Mar 31 16:42:53 EST 2004 -- pshare update
Tue Apr 6 17:25:25 EDT 2004 -- pshare update
Wed Apr 14 13:20:54 EDT 2004 -- pshare update
Tue Apr 20 15:27:30 EDT 2004 -- pshare update
Fri Apr 23 00:45:35 EDT 2004 -- pshare update
Fri Apr 30 17:49:14 EDT 2004 -- pshare update
Mon May 10 19:41:40 EDT 2004 -- pshare update
Sun May 23 21:19:32 EDT 2004 -- pshare update
Mon May 24 12:21:27 EDT 2004 -- pshare update
Wed Jun 2 17:31:09 EDT 2004 -- pshare update
Tue Jun 8 23:31:19 EDT 2004 -- pshare update
Thu Jun 10 10:48:21 EDT 2004 -- pshare update
Thu Jun 10 17:27:43 EDT 2004 -- pshare update
Fri Jun 25 14:09:40 EDT 2004 -- pshare update
Tue Jul 6 17:20:57 EDT 2004 -- pshare update
Thu Sep 9 19:01:49 EDT 2004 -- pshare update
Sun Sep 26 21:30:18 EDT 2004 -- pshare update
Tue Nov 9 11:59:13 EST 2004 -- pshare update
Mon Nov 15 16:40:53 EST 2004 -- pshare update
Wed Dec 8 13:16:34 EST 2004 -- pshare update
Wed Dec 15 11:20:44 EST 2004 -- pshare update
Fri Dec 17 19:28:32 EST 2004 -- pshare update
Wed Feb 9 16:28:28 EST 2005 -- pshare update
Wed Mar 16 12:31:57 EST 2005 -- pshare update
Tue May 3 22:00:22 EDT 2005 -- pshare update
Thu May 26 19:10:36 EDT 2005 -- pshare update
Thu Jun 2 00:17:05 EDT 2005 -- pshare update
Fri Jul 1 17:07:20 EDT 2005 -- pshare update
Thu Aug 11 17:18:13 EDT 2005 -- pshare update
Fri Aug 19 16:09:28 EDT 2005 -- pshare update
Tue Sep 6 18:35:23 EDT 2005 -- pshare update
Tue Sep 27 12:52:09 EDT 2005 -- pshare update
Fri Oct 28 13:11:26 EDT 2005 -- pshare update
Wed Jan 11 20:56:52 EST 2006 -- pshare update
Thu Feb 16 17:31:32 EST 2006 -- pshare update
Fri Mar 3 12:58:38 EST 2006 -- pshare update
Mon Mar 20 11:05:41 EST 2006 -- pshare update
Wed Apr 5 12:50:38 EDT 2006 -- pshare update
Wed Apr 5 19:52:49 EDT 2006 -- pshare update
Tue May 2 18:33:15 EDT 2006 -- pshare update
Tue May 9 11:00:00 EDT 2006 -- pshare update
Fri May 26 14:24:11 EDT 2006 -- pshare update
Mon Jun 12 21:02:41 EDT 2006 -- pshare update
Fri Jun 23 17:02:49 EDT 2006 -- pshare update
Tue Jul 18 21:40:39 EDT 2006 -- pshare update
Tue Jul 25 13:14:42 EDT 2006 -- pshare update
Wed Aug 23 20:09:56 EDT 2006 -- pshare update
Thu Sep 28 16:29:31 EDT 2006 -- pshare update
Mon Nov 27 14:18:32 EST 2006 -- pshare update
Mon Jan 8 19:39:04 EST 2007 -- pshare update
Fri Jan 19 19:10:03 EST 2007 -- pshare update
Tue Jan 30 21:05:30 EST 2007 -- pshare update
Tue Feb 6 13:30:02 EST 2007 -- pshare update
Sat Feb 10 13:10:41 EST 2007 -- pshare update
Tue Feb 27 18:56:21 EST 2007 -- pshare update
Tue Mar 6 13:51:08 EST 2007 -- pshare update
Tue Mar 6 17:52:42 EST 2007 -- pshare update
Tue Apr 3 22:44:33 EDT 2007 -- pshare update
Fri Apr 13 18:53:47 EDT 2007 -- pshare update
Tue May 22 19:33:48 EDT 2007 -- pshare update
Tue May 29 19:45:40 EDT 2007 -- pshare update
Thu Jun 21 21:24:34 EDT 2007 -- pshare update
Mon Jul 2 19:21:56 EDT 2007 -- pshare update
Fri Jul 6 15:41:19 EDT 2007 -- pshare update
Mon Jul 9 19:23:55 EDT 2007 -- pshare update
Wed Jul 25 20:20:44 EDT 2007 -- pshare update
Tue Jul 31 14:00:51 EDT 2007 -- pshare update
Thu Aug 9 18:17:18 EDT 2007 -- pshare update
Tue Aug 14 14:59:54 EDT 2007 -- pshare update
Fri Sep 7 03:21:33 EDT 2007 -- pshare update
Fri Sep 28 13:31:14 EDT 2007 -- pshare update
Tue Oct 30 15:45:24 EDT 2007 -- pshare update
Fri Nov 30 20:16:49 EST 2007 -- pshare update
Mon Dec 17 10:53:39 EST 2007 -- pshare update
Tue Jan 8 10:53:23 EST 2008 -- pshare update
Wed Jan 16 16:26:43 EST 2008 -- pshare update
Wed Feb 6 15:16:59 EST 2008 -- pshare update
Mon Mar 17 15:50:12 EDT 2008 -- pshare update
Wed Mar 19 12:39:53 EDT 2008 -- pshare update
Fri Mar 21 17:03:37 EDT 2008 -- pshare update
Fri Apr 25 16:09:39 EDT 2008 -- pshare update
Mon Apr 28 23:28:37 EDT 2008 -- pshare update
Tue Jun 3 14:31:46 EDT 2008 -- pshare update
Fri Jun 27 18:05:14 EDT 2008 -- pshare update
This Document was created by hlptohtml
Written By:Manish Vachharajani(mvachhar@pppl.gov)