This paper describes the updates to and analysis of the International Tokamak Physics Activity (ITPA) Global H-Mode Confinement Database version 3 (DB3) over the period 1994-2004. Data have now been collected from 18 machines of different sizes and shapes: ASDEX, ASDEX

The condition of the latest version of the ELMy H-mode database has been re-examined. It is shown that there is bias in the ordinary least squares regression for some of the variables. To address these shortcomings three different techniques are employed: (a) principal component regression, (b) an error in variables technique and (c) the selection of a better conditioned dataset with fewer variables. Scalings in terms of the dimensionless physics variables, as well as the standard set of engineering variables, are also derived. The new scalings give a very similar performance for existing scalings for ITER at the standard βn of 1.6, but a much improved performance at higher βn.

Density profiles in pedestal region (H-mode) are measured in HL-2A and the characteristics of the density pedestal are described. Cold particle deposition by Supersonic Molecular Beam Injection (SMBI) within the pedestal is verified. ELM mitigation by SMBI into the H-mode pedestal is demonstrated and the relevant physics is elucidated. The sensitivity of the effect to SMBI pressure and duration are studied. Following SMBI, the ELM frequency increases and ELM amplitude decreases for a finite duration period. Increases in ELM frequency of SMBI ELM f / 0 ELM f 2-3.5 are achieved. This experiment argues that the ELM mitigation results from an increase in Page 2 higher frequency fluctuations and transport events in the pedestal, which are caused by SMBI. These inhibit the occurrence of large transport events which span the entire pedestal width. The observed change in the density pedestal profiles and edge particle flux spectrum with and without SMBI supports this interpretation. An analysis of the experiment and a model shows that ELMs can be mitigated by SMBI with shallow particle penetration into the pedestal.

Typical ELMy H-mode discharges have been achieved on the HL-2A tokamak with combined auxiliary heating of NBI and ECRH. The minimum power required is about 1.1 MW at a density of 1.6 × 10 19 m −3 and increases with a decrease in density, almost independent of the launching order of the ECRH and NBI heating. The energy loss by each edge localized mode (ELM) burst is estimated to be lower than 3% of the total stored energy. At a frequency of typically 400 Hz, the energy confinement time is only marginally reduced by the ELMs. The supersonic molecular beam injection fuelling is found to be beneficial for triggering an L-H transition due to less induced recycling and higher fuelling efficiency. The dwell time of the L-H transition is 20-200 ms, and tends to decrease as the power increases. The delay time of the H-L transition is 10-30 ms for most discharges and is comparable to the energy confinement time. The ELMs with a period of 1-3 ms are sustained for more than ten times the energy confinement time with enhanced confinement factor H 89 > 1.5, which tends to decrease with the total heating power. The confinement time in the H-mode discharges increases with plasma current approximately linearly.

Two groups of frequency sweeping modes are observed and interpreted in the HL-2 A plasmas with qmin ∼ 1. The tokamak simulation code calculations indicate the presence of a reversed shear q-profile during the existence of these modes. The mode frequencies lie in between TAE and BAE frequencies, i.e. ωBAE < ω < ωTAE, and these modes are highly localized near qmin, i.e. r/a ∼ 0.25. A group of modes characterized by down-sweeping frequency with qmin decrease due to qmin > 1 and nqmin − m > 0, and another group of modes characterized by up-sweeping frequency with qmin drop, owing to qmin < 1 and nqmin − m < 0 before sawtooth crash. The kinetic Alfvén eigenmode code analysis supports that the down-sweeping modes are kinetic reverse shear Alfvén eigenmodes (KRSAEs), and the up-sweeping modes are RSAEs, which exist in the ideal or kinetic MHD limit. In addition, the down- and up-sweeping RSAEs both have fast nonlinear frequency behaviour in the process of slow frequency sweeping, i.e. producing pitch-fork phenomena. These studies provide valuable constraint conditions for the q-profile measurements.