EXPRESSION FOR THE THERMAL H-MODE ENERGY CONFINEMENT TIME UNDER ELM FREE CONDITIONS IN DEUTERIUM

The design of future tokamaks, which are supposed to reach ignition with the H mode, requires a reliable scaling expression for the H mode energy confinement time. In the present work, an H mode scaling expression for the thermal plasma energy confinement time has been developed by combining data from four existing divertor tokamaks, ASDEX, DIII-D, JET and PBX-M. The plasma conditions, which were as similar as possible to ensure a coherent set of data, were ELM free deuterium discharges heated by deuterium neutral beam injection. By combining four tokamaks, the dependence of the thermal energy confinement on the main plasma parameters, including the three main geometrical variables, was determined. Assuming a power law dependence, one obtains for the H mode in deuterium by linear regression: tau(th) = 0.066I(p) 1.11+/-0.04 P(L)-0.51+/-0.03 R1.75+/-0.08 a-0.30+/-0.07 kappa0.18+/-0.09 B(T)0.15+/-0.07. The density dependence is found to be close to zero. This expression is compared with the existing scaling relations and the differences are analysed. Predictions for ITER are given, and discussed with particular emphasis on the uncertainties. Correcting the data to take the closed divertor of ASDEX into account yields tau(th) = 0.048I(p) 1.10+/-0.04 P(L)-0.51+/-0.03 R1.83+/-0.08 a-0.34+/-0.07 kappa0.61+/-0.09 B(T)0.14+/-0.07 essentially characterized by a smaller numerical factor and a larger kappa dependence. It appears, however, that the predictions for ITER are only weakly changed when this expression is used. Finally, it is shown that taking the improved H mode confinement into account by means of a simple multiplier in front of an L mode confinement scaling expression is valid for ITER, but is not necessarily universally valid.