Nondimensional transport studies in TFTR

The machine parameters (Ip, Pheat, R) required for ignition in ITER have generally been extrapolated from power-law regression fits to global τE measurement on existing tokamaks. There remain important choices to be made in the form of the scaling relation which have not yet been resolved by theory. In particular, power flow Q(r) through a magnetic flux surface should scale as Q(r) = QBohmF, where F = F(ρ*, β, ν*, s, Te/Ti,...) is a function of local, nondimensional plasma parameters and QBohm is proportional to [neT2ea/eB]. Projections to ITER can be reduced to establishing the dependence of F on ρ* = ρi/a, because one can create plasmas in today's tokamaks which have similar values of the other non-dimensional parameters. Two common scalings suggested by theory are Bohm (F independent of ρ*) and gyroBohm (F is proportional to ρ*). Experiments have been carried out on TFTR to ascertain the dependence of F on ρ*, ν*, and β in L-mode plasmas, holding the other nondimensional parameters fixed. The observed variation of heat flow with ρ* was observed to be better described by Bohm scaling than gyroBohm. Comparisons with the critical temperature gradient transport model, which is gyroBohm in character, show that it overpredicts the temperature increase expected with increasing magnetic field. The ν* scan (remaining in the collisionless regime) revealed that the Bohm-normalized power flow is remarkably insensitive to collisionality, in agreement with ITER-P scaling. The β scan identified a deterioration of confinement with increasing β at fixed ρ* and ν*, of approximately the correct magnitude required to reconcile Bohm local transport scaling with ITER-P global scaling of τE. This may suggest a role for electromagnetic phenomena in governing tokamak transport even at very low beta. (author). 25 refs, 7 figs, 3 tabs