type-III Edge Localized Modes Research Articles

Integrated scenario with type-III ELMy H-mode edge: extrapolation to ITER

One of the most severe problems for fusion reactors is the power load on the plasma facing components. The challenge is to develop operation scenarios, which combine sufficient energy confinement with benign heat loads to the plasma facing components. The radiative type-III ELMy H-mode seems a possible solution for such an integrated ITER scenario. Nitrogen seeded type-III ELMy H-modes for the standard inductive scenario and the high beta stationary hybrid scenario are investigated with respect to their transient and steady-state power fluxes to the divertor, confinement properties, edge operational space, core operational space, plasma purity and MHD behaviour. A large database of highly radiative type-III ELMy H-modes on JET is used for extrapolations to ITER. On this basis the transient heat load should be acceptable for ITER. It was found that the scaling of the confinement time with respect to the ion gyroradius is close to the gyro-Bohm scaling. Scalings with respect to the plasma collisionality suggest that the confinement will be good enough for an ITER scenario at 17 MA with a power amplification factor (Q) of 10 and might be marginally good enough for a Q = 10 scenario at 15 MA. Those extrapolations are supported by simulations with an integrated core/edge model COREDIV. In addition the hybrid scenario with type-III edge localized modes has been proven to have improved edge conditions without any modification of the central plasma current profile, indicating it is compatible with a high beta operation for a steady-state ITER Q = 5 scenario.

Investigation of pellet-triggered MHD events in the ASDEX upgrade

In order to gain deeper insight into the process of MHD activity triggered by pellets we extended our previous analysis (standard type-I edge localized modes (ELMs)) to type-I ELMs in radiative edge scenarios, type-III ELMs, the quiescent H-mode regime, and core mode activity such as neoclassical tearing modes or snakes. Pellet triggering of mode activity has turned out to be a quite general feature, but only in case of the ELMs can it be unambiguously attributed to prompt local impact by the pellet. For edge plasma conditions characterized by higher resistivity, the growth time of spontaneous ELMs increases while the plasma changes from the type-I into the type-III regime. However, pellet-triggered ELMs retain fast rise times. In the quiescent Hmode, pellets still trigger ELM-like mode activity, but no longer accompanied by a significant release of energy from the plasma.

The poloidal distribution of type-III edge localized modes in the Mega-Ampère spherical tokamak (MAST)

This article describes the poloidal plasma particle distribution of type-III edge localized modes (ELMs) in the Mega-Ampère spherical tokamak [R.-J. Akers et al., Phys. Plasmas 9, 3919 (2002)]. A fast imaging camera with 10μs exposure time is used to record the Dα light coming from the entire poloidal cross section. Furthermore, three sets of probes, triggered at the same time, acquired at 1MHz, and located at different poloidal, radial, and toroidal locations in the tokamak are used. ELMs are observed to affect the Dα emission throughout the low-field scrape-off layer; on the high-field side, however, this effect is found to be small. The results obtained by imaging agree with the pointwise measurements using Langmuir probes. The radial propagation is shown to occur at a speed of 250m∕s, whereas the toroidal convection from the top to the bottom of the plasma is shown to be consistent with a transport at the local sound speed. Strong correlation amplitudes are reported among the probes that are poloidally and toroidally separated by several meters. The study of the cross-correlation coefficients as a function of the frequency indicates that this correlation is caused by the low-frequency component of the signal and that the high-frequency part is not correlated. Consequently, the filamentary structures are interpreted as caused by the onset of turbulence during an ELM and do not constitute the ELM itself.

Localized electromagnetic modes in MHD stable regime of the TJ-II Heliac

Edge localized mode (ELM) events are typically observed in a broad variety of tokamaks and stellarators. In the TJ-II Heliac, ELM-like events have been observed in experiments at low plasma parameters (Te ∼ Ti ∼ 100 eV, n ∼ 6 × 1012 cm−3, where Ti, Te and n are the ion and the electron temperatures and the plasma density, respectively) away from the electron cyclotron resonance heating power deposition region. Moreover, numerical calculations have shown that the experimental values for plasma pressure are more than an order of magnitude lower than the theoretical pressure threshold for both ideal and resistive magnetohydrodynamic modes. These transport events, now called localized electromagnetic modes, are found to be linked to the presence of rational magnetic surfaces in the vicinity of the steep density gradient region located at 0.5 ≤ r ≤ 0.7, r = reff/ap, where reff and ap are, respectively, the average radius of the magnetic surface and the plasma boundary; so, while the morphology of the time signals is similar to that of the type-III ELMs, the mode in TJ-II is localized at the edge of the confinement region. The characteristic frequency of the oscillations observed by the Mirnov coils is f = 15–50 kHz. The position of the rational magnetic surface is determined by the channel of the ECE diagnostic with the highest correlation to Mirnov probes and matches, within the experimental error, the position obtained from numerical calculations when the modification of the rotational transform due to toroidal currents is taken into account. These two coupled observations are attributed to the coupling of drift modes with Alfvèn and acoustic waves.

Integrated exhaust scenarios with actively controlled ELMs

An integrated radiative high performance scenario has been established at ASDEX Upgrade based on simultaneous feedback control of the average divertor neutral particle and power flux in combination with a high, pellet induced frequency of edge localized modes (ELMs). This approach is fully compatible with the present tungsten wall coating covering about 65% of the plasma facing components and is intended for application in the envisaged full-tungsten experiment. In these experiments, divertor recycling and effective divertor temperature (derived from thermoelectric currents) were tuned by acting on fuel gas puff and argon injection rates. The ELM frequency (fELM) was kept high by repetitive injection of small cryogenic deuterium pellets to avoid the radiative instabilities seen at low fELM and high radiated power, and to control the ELM energy. No confinement loss is observed in this radiative type-I ELMy scenario with relatively flat density profiles. In contrast, similar type-III ELM scenarios achieved in hydrogen show a confinement loss of 25% as compared to the type-I phase. In parallel to pellets, alternative ELM trigger techniques have been investigated as well. Fast vertical plasma oscillations are able to synchronize the ELM frequency to values higher and lower than the intrinsic fELM, but remain to be tested in the integrated scenario. Supersonic gas injection showed better fuelling efficiencies than usual gas puffing but instantaneous ELM release has not been achieved. A particular experimental challenge for AUG conditions is to obtain a high pace making frequency, to establish scalings of confinement and energy loss as a function of controlled ELM frequency.