Recent simulations by Putvinskij et al. (PSI Conference, 1996) have shown that introducing impurities into the plasma in order to mitigate adverse disruption effects in ITER may actually be deleterious because of a potentially unwelcome phenomenon: generation of multi-MeV runaway electrons by the collisional avalanche mechanism (Rosenbluth, M.N., et al., in Fusion Energy 1996 (Proc. 16th Int. Conf. Montreal, 1996), Vol. 2, IAEA, Vienna (in press) Paper FP-26). The injection of a liquid hydrogen jet to deliver a massive density increase is proposed as a means of avoiding runaways, while providing the same beneficial effects as impurities. A discussion of many jet related topics, such as ablation/penetration, jet breakup time and stability, is presented. Owing to an ablation pressure instability, it is predicted that the jet will quickly break up into a regular chain of droplets with dimensions of approximately the size of the jet radius. It is found that while deep penetration in the plasma can easily be achieved, bubble growth and disruptive boiling (hashing) during the propagation in the vacuum gap between the nozzle exit and the plasma are the main processes limiting the jet survival time. Calculations indicate that for ITER reference parameters, the jet can remain coherent in vacuum for a distance ~1 m before disintegrating. On the basis of this present understanding, the prospect for the safe termination of ITER discharges by high density liquid jet injection appears promising

The bounce average Fokker-Planck equation describing fast particles in a tokamak with toroidal magnetic field ripple has been solved numerically by a Monte Carlo approach. The essential element is that the ripple effect is treated as a diffusion of banana trapped particles. The diffusion coefficient is precalculated from a semi-analytical expression for a given plasma geometry. The kinetic equation is solved for the processes of fast particle slowing down, pitch angle scattering, ripple diffusion and acceleration by ion cyclotron resonance heating. This approach was found to be particularly useful for the study of fast particle behaviour in non-stationary conditions, including sawtooth effects. A ripple loss code has been developed on the basis of this principle. The code has been benchmarked against a full orbit following Monte Carlo code. Applications of the code to the experimental results from the JET experiments with 16 and 32 toroidal field coils are given

It is shown that the kink distortion of the magnetic axis and, as a consequence, the variation of the magnetic field along magnetic surfaces in the toroidal direction lead to a new type of 'superbanana' particle orbits and can drastically change the classical particle and heat transport in the central core. The phenomena discussed have a time-scale of approximately=10-3 s and seem to be too slow to affect the sawtooth crash process (which is an order of magnitude faster). They can, however, influence the evolution of the plasma parameters between crashes, the formation of 'snakes' and other phenomena in the central plasma inside the q=1 magnetic surface

The inverse problem for drift losses of fusion alpha particles in a tokamak -the problem of calculating plasma parameters from known distributions of the alpha particle flux over the wall surface -is analysed. As the PLT experiments have shown, the energy spectrum of fusion charged particles escaping from the plasma can give an idea on the plasma temperature. The paper demonstrates that from a known spatial distribution of the particle flux it is possible also to determine the distributions of plasma density, ion temperature and plasma current. -The results of numerical modelling are given for some particular cases in which the aim was to establish the current density and the fusion alpha particle source profiles.