MHD analysis of edge instabilities in the JET tokamak

Abstract

The aim of nuclear fusion energy research is to demonstrate the feasibility of nuclear fusion reactors as a future energy source. The tokamak is the most advanced fusion machine to date, and is most likely the first system to be converted into a reactor. An important subject of nuclear fusion research is the study of the equilibrium and stability of a plasma with respect to large scale displacements. In a tokamak, several instabilities can occur. A class of edge instabilities that occur in the high confinement regime, H-mode, have been called Edge Localised Modes (ELMs). ELMs are relaxation oscillations that cause quasiperiodic energy and particle losses out of the confined plasma into the scrape-off layer. These losses are of concern for future burning fusion plasmas, such as ITER, due to the large transient heat loads expected on plasma facing components in contact with the scrape-off layer. These heat loads may reduce the target lifetime below tolerable levels. Although the existence of ELMs has been known for many years, their physics is not well understood yet. Much effort has been spent world-wide in an attempt to improve the understanding of these instabilities. A review of the present state of ELM research is given. Empirically, at least three types of ELMs have been identified, which are normally classified as type-I, type-II and type-III ELMs. From the point of view of plasma stability, research has increasingly focussed on the role of certain MHD instabilities, namely (finite-n) ballooning and kink (peeling) modes, as well as coupled ballooning-kink modes, leading to the proposition of a theoretical model called the peeling-ballooning cycle. This thesis presents new insight into ELMs obtained from the analysis of experimental data in the JET tokamak, and compares the observations with present theoretical ELM models. Low frequency coherent type-I ELM precursor modes have been identified. Their properties are studied in detail. Precursors with low toroidal mode numbers are known to be external kink instabilities, while experimental findings and their comparison with stability calculations suggest that the precursor modes with higher toroidal mode numbers are not pure external kinks but coupled ballooning-kink modes. In spite of their regular occurrence, there is no evidence that the precursor mode growth rate accelerates rapidly before the ELM. This is regarded as an indication that external kink modes or coupled peeling-ballooning modes do not trigger the ELMs, which would contradict present theoretical ELM models. Another type of instability are the Washboard (WB) modes, a very common edge instability regularly observed in the H-mode regime in JET. Since their discovery in the late 90s, rather little attention has been paid to them, either theoretically or experimentally. So far they have not been regarded as an ELM-relevant instability. However, evidence for their involvement in the pedestal and ELM dynamics is presented.