Prime pagine RA2010FUS:Copia di Layout 1 - ENEA - Fusione

magnetic confinement (cont’d.)
progress report
2010
013
% CPU usage
100
EFDA-ITM batch cluster yearly usage
0
Jan 01 Mar 01 May 01 Jul 01 Sept 01 Nov 01
Total CPU load % Current: 2% Average: 9% Maximal: 100%
Figure 1.3 – CPUs yearly usage of the
Gateway batch cluster
yearly usage of the Gateway batch cluster in 2010 and shown in figure 1.3. The upgrade activity has been
carried out involving the base software (operating system, parallel filesystem, distributed filesystem) and
scientific tools as well.
The Storage Data Area used in 2010 is as follow: 4.3 TB for users and software repositories stored under the
distributed filesystem, 14 TB for simulation data stored under the parallel filesystem.
The RAM resources of Front–End nodes have been updated to 32 GB for each one, to take into account the
heavy CPU usage due to the interactive sessions during Code Camp meeting.
Several EU Code Camp 2010 meetings were supported by using hardware/software resources of the Gateway.
The Inter–operability between ITM Gateway and High Performance Computing For Fusion (HPC–FF) has
been implemented fully. It allows both HPC facilities to be used in a common operating environment for
sharing codes, data exchange and user access, as well as joint tools and libraries to facilitate the creation of
community fusion codes.
1.2 FTU Experimental Activity
Lower hybrid
Lower hybrid current drive experiments in plasma regimes of interest for ITER. The need for non–ohmic
steady–state current drive is a fundamental weakness of the tokamak as a magnetic fusion reactor. Low
frequency microwaves, in the range of a few GHz, have been successfully used to drive current in tokamak
core plasmas at relatively low densities [1.1–1.3], but extrapolation of these results to the higher densities of
ITER and subsequent reactor grade plasmas has proved to be problematic, because the coupled LH power
stubbornly localizes in the plasma edge region [1.4,1.5].
A thermonuclear plant needs to maximise the bootstrap current for saving current driven by external power
sources. For this reason ITER will need to operate with almost flat radial profiles such that high pressure
gradient occur in the external region of the plasma column with high densities even at the edge (at normalized
minor radius r/r LCMS
∼0.9: n e_0.9
≥0.7×10 20 m –3 , and in the centre: n e0
≥0.8×10 20 m –3 , where r LCMS
is the
plasma minor radius at the last closed magnetic field surface (LCMS)). The operation with relatively high
plasma density in the region of the periphery is thus an important constraint to be considered for designing a
tool for driving actively non–inductive current in the plasma, which is essential for obtaining the necessary
requirements of stability and confinement.
Theoretical and experimental works have been extensively carried out on FTU, which have evidenced the role
of parametric instabilities (PI) in inhibiting penetration of the lower hybrid wave to the plasma core when
operating at relatively high plasma densities [1.4,1.5]. These works have produced the guidelines used in the
FTU experiments, based on theoretical predictions that the PI–produced spectral broadening would diminish
under operations with higher electron temperatures at the periphery (T e_outer
) of the plasma column [1.3]. As
a result of operating in such a regime, the spectral broadening measured by radio–frequency (rf) probe has
been indeed observed to diminish. It has also been observed that the LH power penetration into a high density
core took place as expected.
The following options, available on the FTU device, have been exploited for producing different plasma edge
temperatures in high–density plasmas, and different plasma targets for LH power coupling: i) Toroidal or
poloidal limiter operations. With toroidal limiter operations, a stronger plasma–wall interaction occurs due to
the larger plasma–wall contact area. In these conditions lower recycling of particles from the vessel walls and