Upgrade of the Monte Carlo code FIDO

During rf heating on minority ions and current drive the velocity distribution is determined by the balance between the wave-particle interaction and the relaxation through Coulomb collision. This is simulated by the Monte Carlo code fido [1–3], which solves the Fokker-Planck equation in the low collisionality banana-transport regime, including finite orbit width and rf induced transport. The former version 1.2pl4 of fido is written in fortran assuming a circular cross section of the plasma. This limits the possibility of making comparisons with experiments. Further, the lack of elongation prohibits us from simulating scenarios with high currents. The new version 2.0, being developed, will be able to read a general Grad-Shafranov solution from the code chease [4], as well as to produce an internal equilibria including elongation, Shafranov-shift and triangularity [5]. Further, the development and maintainance time is reduced by using an object oriented design in c++. Monte Carlo methods are easy to parallelize. Version 2.0 is designed for running in parallel using a combination of mpi [6] and threading. Of course, the code is also runnable on a single node. Modern common available programming tools, such as yacc, lex and Doc++, are used. In order to maintain versions and consistency between the developers involved, the freely available cvs [7] version managing tool is used. 1 Fokker-Planck and Quasi linear rf-operator rf-wave fields can be used for auxiliary heating and for current drive. However it also induces transport. The velocity distribution during rf-scenarios is determined by a balance between • rf-induced diffusion in velocity space. • Coulomb collision relaxation; isotropization and thermalization. This can be described by a Fokker-Plank equation df dt ≡ ∂f ∂t + v ∂f ∂r + ∂v ∂t ∂f ∂v = C[f]+Q[f] (1) where C is the Coulomb collision operator and Q is the quasi linear rfoperator. The Fokker-Plank equation is an advection diffusion equation and can be written df dt = ∇ · (af +D · ∇f) a = arf + acol D = Drf +Dcol (2) where a is an advection vector and D is the diffusion matrix.