Transcript GTC development - University of California, Irvine

GTC Status: Physics Capabilities
& Recent Applications
Y. Xiao for GTC team
UC Irvine
Global Gyrokinetic Toroidal Code (GTC)
• Non-perturbative (full-f) & perturbative (df) simulation
• General geometry using EFIT & TRANSP data
• Kinetic electrons & electromagnetic simulation
• Neoclassical effects using Fokker-Planck collision operators
conserving energy & momentum
• Equilibrium radial electric field, toroidal & poloidal
rotations; Multiple ion species
• Parallelization >100,000 cores
• Global field-aligned mesh
• Parallel solver PETSc
• Advanced I/O ADIOS
[Lin et al, Science, 1998]
Lin, Holod, Zhang, Xiao, UCI
Klasky, ORNL; Ethier, PPPL;
Decyk, UCLA; et al
• Applications: microturbulence & MHD modes
General geometry and profiles
• General global toroidal magnetic geometry from GradShafranov equilibrium
• Realistic density and temperature profiles using spline fits
of EFIT and TRANSP data
• No additional equilibrium model
is needed
• Experimental validation
Realistic temperature and density profiles from DIII-D
shot #101391 [Candy and Waltz, PRL 2003]
GTC poloidal mesh
Full-f capability
• Non-perturbative full-f and perturbative d-f
models are implemented in the same version
full-f ITG
intensity
df ITG intensity
full-f zonal flows
df zonal flows
time
Kinetic electrons
• Hybrid fluid-kinetic electron model is used
• In the lowest order of electron-to-ion mass ratio
expansion electrons are adiabatic: fluid equations
• Higher-order kinetic correction is calculated by solving
drift-kinetic equation
Electromagnetic capabilities
• Only perpendicular perturbation of magnetic field
considered
• Parallel electric field expressed in terms of effective
potential, obtained from electron density
• Continuity equation for adiabatic electron density,
corrected by drift kinetic equation.
• Inverse Ampere’s law for electron current
• Time evolution for parallel vector potential
• Gyrokinetic Poisson equation for electrostatic potential
Structure of GTC algorithm
Dynamics
dne
dfi&dge
dA||
Fields
due
dA|| ZF dfes
dfind
Sources
dA|| dui
dni dne1 due1 dne
Equilibrium flows and neoclassical effects
• Equilibrium toroidal rotation is implemented
• Radial electric field satisfies radial force balance
• Neoclassical poloidal rotation satisfies parallel force
balance
• Fokker-Planck collision operator conserving energy and
momentum
Multiple ion species
• Fast ions treated the same way as thermal ion specie
• Energetic ion density and current non-perturbatively
enter Poisson equation an Ampere’s law
Numerical efficiency
•
•
•
•
Effective parallelization >105 cores
Global field-aligned mesh
Parallel PETSc solver
Advanced I/O system ADIOS
Recent GTC applications
• Electrostatic, kinetic electron applications
– CTEM turbulent transport [Xiao et al, PRL2009; PoP2010]
– Momentum transport [Holod & Lin, PoP2008; PPCF2010]
– Energetic particle transport by microturbulence [W. Zhang et al,
PRL2008; PoP2010]
– Turbulent transport in reversed magnetic shear plasmas [Deng &
Lin, PoP2009]
– GAM physics [[H. Zhang et al,NF2009; PoP2010]
• Electromagnetic applications
– Electromagnetic turbulence with kinetic electrons [Nishimura et al,
CiCP2009]
– TAE
[Nishimura, PoP2009; W. Zhang et al, in preparation]
– RSAE
[Deng et al, PoP2010, submitted]
– BAE
[H. Zhang et al, in preparation]
CTEM turbulent transport
The CTEM turbulent transport studies reveal
• Transport scaling---Bohm to gyroBohm with system size
increasing
• Turbulence properties---microscopic eddies mixed with
mesoscale eddies
• Zonal flow---Zonal flow is important for the parameter applied
• Transport mechanism
Xiao and Lin PRL 2009
Xiao et al, POP 2010
 electrons: track global profile of turbulent intensity; but contain a
nondiffusive, ballistic component on mesoscale. The electron transport
in CTEM is a 1D fluid process (radial) due to lack of parallel
decorrelation and toroidal precession decorrelation and weak toroidal
precession detuning
 ions: diffusive, proportional to local EXB intensity. The ions decorrelate
with turbulence in the parallel direction within one flux surface
Experimental validation
• Real radial temperature and density
profiles are loaded
• Zonal flow solver is redesigned for
the general geometry
• Heat conductivity uses the ITER
convention
• The measured heat conductivity
(preliminary) is close to Candy
Waltz 2003 value
q  


2
T

Toroidal momentum transport
• Simulations of toroidal angular momentum transport in ITG
and CTEM turbulence
• Separation of momentum flux components.
Non-diffusive momentum flux
• Intrinsic Prandtl number
Pr  0.2  0.7 (ITG)
Pr  0.5(CTEM)
Holod & Lin, PoP 2008
Holod & Lin, PPCF 2010