Topic/Type: 2.4 Gyro-kinetic and gyro-fluid methods, Poster

NBI heating simulations of fusion plasmas on the Grid (FAFNER2)

M. Rodr?guez-Pascual1, J. Guasp1, 2, F. Castej?n1, 2, A.J. Rubio-Montero1, I.M. Llorente3, R. Mayo1

1 CIEMAT
2 Spanish Laboratory of Fusion
3 UCM

FAFNER2 is a 3D code adapted to the TJ-II helical axis stellarator from the original one developed by Lister [1] that runs on Grid. It simulates the most common heating method for fusion devices, i.e. the Neutral Beam Injection (NBI) technology. To do so, Monte Carlo methods are applied to neutral atoms which are able to overcome the confining magnetic field of a tokamak or a stellarator and are ionised in the plasma via collisions with ions and electrons, thus heating and feeding the plasma.

Monte Carlo codes are suitable for running on distributed platforms such as Grid because of their easy parallelization. To the date, FAFNER2 has been usually run at CIEMAT by means of a batch mode on a shared memory of Cray architecture [2]. After the new implementation presented in this work, it can be executed on the current x86 distributed architectures either on cluster or on Grid by means of MPI library, but also based on DRMAA API with the aim of overcoming the circumstance of submitting jobs to non-MPI supporting sites. In addition, the metascheduler GridWay [3] has been also incorporated for maximizing the code performance.

According to this, FAFNER2 opens a new strategy for the fusion community since it is able to be coupled to many other codes, e.g. the ion kinetic transport code ISDEP [4] or the European Transport Solver (ETS) [5], and simulate a wider range of the phenomena that occurs in a fusion plasma. Into this framework, the FAFNER2 porting to the grid is the first step of a more ambitious work.

The performance of FAFNER2 on the Grid enhances their results because of the greater number of particles that can be calculated in a reasonable time, but is of most interest because it will allow the user not only to obtain the plasma parameters, but also the temporal evolution by creating dependencies in the different results obtained on the Grid. It runs adapted to the TJ-II stellarator, but can easily be modified to be adapted to new device geometries by changing the 3D lattice parameters, the inclusion of which is in an independent module.

This fact is of outmost importance since it will allow the fusion researchers to simulate, among others, the heating technique that will be used at ITER.

This work was made in part with support from the EC funded Project EGEE-III (http://www.eu-egee.org/).


[1] G.G. Lister, Technical Report 4/222, Max Planck IPP, Garching, Germany (1985)

[2] J. Guasp and M. Liniers, Fusion Technol. 24, 251 (1993)

[3] E. Huedo, R.S. Montero and I.M. Llorente, SCPE 6 (3), 1 (2005)

[4] F. Castej?n et al., Plasma Phys. Control. Fusion 49, 753 (2007)

[5] D. Kalupin et al., 35th EPS Conference on Plasma Phys. ECA Vol.32D, P-5.027 (2008)