TGLF code called for EP driven local linear AE simulations
TGLFEP code is a program
to run TGLF with
energetic paritcle (EP) conveniently,
where using the code TGLF (inside GAcode)
with the version number
(on May 7th, 2016)
''TGLF stable_r6.0.0-31-gba49''
on NERSC ''EDISON_CRAY Linux x86_64''
or '' CORI Linux x86_64''.
TGLFEP_interface.f90 contains
the module of internal parameters
for TGLFEP code.
TGLFEP_tglf_map.f90 contains
the module of input profiles
and the subroutine to
read plasma and geometry profiles
from input file input.profile
as well as the subroutine
to map the parameters to TGLF
in the required formats.
TGLFEP_driver.f90 will read
the control parameters
from input file input.TGLFEP.
The other subroutines contain different processes for different simulation scheme.
Four different example profiles have been tested:
GA-std case,
DIII-D NBI case (with GYRO inputs),
EPtran profiles,
Two EP species case.
The existed input.profile is
the example profiles for
the DIII-D NBI case
with 50 radii (r/a = 0.0 excluded),
which are calculated by EPtran code.
The control parameters of TGLFEP code
are read from file input.TGLFEP.
The existed input.TGLFEP has been set
to get the EP critical beta for
the inner 40 radii.
The outputs are in the file out.TGLFEP.
For the outside 10 radii,
save the above output file at first
and just change 40 SCAN_N to 10 SCAN_N,
1 IRS to 41 IRS,
and 4.0 FACTOR_IN to 8.0 FACTOR_IN
in input.TGLFEP file.
Then run TGLFEP again.
SCAN_N is the amout of radii to calculate
(don't have to be the same with the total number of radii,
but should be less or equal to that).
IRS is the starting radius.
FACTOR_IN indicates
the upper boundary
of the EP density scale.
If the input EP density cannot drive AE modes,
a larger FACTOR_IN > 1.0 (up to 2., 10., or even 100.)
should be used.
When you want to run a new case with TGLFEP, for example still the DIII-D NBI case but with different profiles.
- At first,
you should generate a new
input.profilebased on your plasma and geometry profiles in the TGLF formats. - Then try to run TGLFEP code and get the new output file.
- Check the outputs.
If at some radii,
EP critical beta are not found
(that is,
SFminshows NOT correctly inout.TGLFEP), you need to changeFACTOR_INininput.TGLFEPto a larger value and try for these radii again.
By the way,
however it is possible that
for some cases,
the EP critical beta doesn't exist and
surely cannot be found
even if with a very large FACTOR_IN.
Recommend to use 25xSCAN_N CPUs (or more)
to run these cases.
It usually costs about 20 mimutes
on Cori (using ceil(25xSCAN_N /32) nodes)
or on Edison (using ceil(25xSCAN_N /24) nodes).
The
p_primeinTGLFis usually negative and should use the total pressure including the EP species.