Observation of the intensity of the recycling particle flux at the main plasma edge for various limiter and divertor discharges indicates that the gross energy confinement of beam-heated discharges is closely related to the intensity of the edge particle flux. In limiter discharges, the global particle confinement time and the energy confinement time τE show many similarities: 1) linear Ip dependence at Ip < 600 kA, 2) no BT dependence, and 3) deterioration against injection power. Improvement of τE by increasing Ip, for example, is associated with high temperatures at the plasma edge region accompanied by reduced particle recycling. – Divertor discharges with low particle recycling around the main plasma show better energy confinement than limiter discharges at high plasma densities. The improvement of τE is primarily originated in the reduction of heat transport at the main plasma edge region, which is associated with the reduction of recycling particle flux at the main plasma edge. Under certain operation condition, for example, excessive cold-gas puffing, the discharge shows relatively high scrape-off plasma density and strong particle recycling between the main plasma and the limiter. The energy confinement time of these discharges degrades somewhat or reduces completely to that of the limiter discharge. – In low-recycling divertor discharges, the central electron and ion temperature is proportional to the injection power, and the plasma stored energy is proportional to n̄ePabs (scales as INTOR scaling). With ≈ 4 MW beam injection, high-temperature and high-density plasmas were obtained (stored energy up to 280 kJ, Te(0) ≈ Ti(0) ≈ 2.5–3.0 keV at n̄e ≈ (6–7) × 1013 cm−3, τE* ≈ 70 ms).

The empirical scaling of the electron thermal diffusivity, Xe, is investigated for more than 100 beam-heated discharges. These discharges include two major features: 1) high-temperature and highdensity plasma (T e (O)«Tj(O)« 2.5-3 keV at n ^ (5-7) X 10 13 cm" 3 , P NB i £ 4 MW), which will be the basis for the breakeven experiments in the next-generation tokamaks, 2) three types of discharges, i.e. good and poor confinement divertor discharges and limiter discharges. -All kinds of discharges (good heating and poor heating divertor discharges, limiter discharges) have fhe same functional form in x e within ~ 40% at 0.25 a < r <0.65 a, where Xe in the simplest expression scales as Xe a Vn e -The * on t n e r m a l diffusivity, Xi, is consistent with the assumption that the neoclassical Xi can be applicable to our three kinds of discharges. For discharges with different heating efficiency, there is no systematic difference, in the adjustable multiplier, to the neoclassical theory by Hinton-Hazeltine for an ion collisionality of v*= 0.02-0.5.

No abstract

A two-dimensional time dependent fluid code has been developed for transport processes in the edge plasma of a tokamak, coupled with a Monte Carlo method for neutral gas behaviour. The code employs a particle-incell method for the numerical solution of fluid equations. A simulation of the Doublet III divertor experiment has been performed with this code. It has been confirmed that the radial profiles of temperature and density in the scrape-off and divertor region can be simulated fairly well

A Langmuir probe array in the divertor plate has been used to investigate the dense, cold divertor plasma associated with remote radiative cooling in neutral-beam-heated, single-null open-divertor discharges in Doublet-Ill. With injected powers of up to 1.2 MW, the divertor plasma becomes denser and colder as the main plasma line-averaged density n̄e increases, reaching ned= 2.8 X 1014 cm−3. Since the electron temperature drops to Ted = 3.5 eV under these conditions, this cold, dense plasma can provide a solution to the problem of wall erosion.