FT/P2-29 The Physics of Design and Operation of High Power Neutral Beam Injection Ducts – Extrapolating JET Experience to ITER Situations

Neutral beam injection systems have proven to be the single most effective form of heating for tokamak plasmas. Typical beam pulse lengths are of the order of ten seconds and the major limitation to increased pulse length in multi-megawatt beamlines is the effect of re-ionised neutral particles in the restricted drift space, or “duct”, connecting the beamline to the tokamak vessel. These particles are deflected and frequently focused by the stray magnetic field of the tokamak and can produce significant power density on the walls of the duct. In JET the power density due to re-ionisation can reach ten megawatt per square metre and is the main limitation to beam pulse length. The effect of the re-ionised power is to cause local heating of the duct wall and evolution of gas trapped within the wall material. This raises the pressure in the duct, causing further reionisation of the beam and hence increased wall heating. Unchecked, this process can lead to complete reionisation of the beam and possible structural failure of the duct wall. A new model is presented that describes an effective source rate of excess gas evolved from the wall in terms of the surface temperature and area subjected to heating. This approach reduces the predicted dependency of duct pressure on beam flux relative to conventional models, parametrised by an ion-induced desorption coefficient, and is validated by comparison with measurements from the 80keV and 130keV JET beamlines over similar power ranges. In conjunction with a particle trajectory re-ionisation code to determine the size and power loading of the affected area, a selfconsistent description of the duct pressure balance may be determined for a given heat-transfer characteristic at the wall. This can be directly applied to the design of systems for ITER such as the duct liner and the electrostatic residual ion dump panels. The time response of the duct pressure can be used to establish the mechanism by which gas is released. It is shown that only the percolation of occluded gas within the structure of the wall can account for the timescale over which the pressure is observed to rise and the quantity of gas released. These occlusions occur as a result of localised damage within the wall material and hence it follows that gas evolution will be a function of the ageing process of future systems.