Structure and shape of intermittent events

Blob theory and simulations predict the characteristic structure of blobs and holes in time and space which can be compared with experimental data. In order to understand the struc-ture of blobs, the aforementioned blob model needs to be extended with the effect of the background density. It can be assumed that a blob has monopole poloidal density distribu-tion with Gaussian distribudistribu-tion over a constant background density. The charge polarizadistribu-tion process of the blob results in a dipole charge distribution. In a toroidal device, the driving process for this effect is the curvature and the gradient of the magnetic field. The charge polarization speed is inverse proportional to the absolute plasma density of the blob, thus, the background density reduces the drive for the blob velocity and modifies the structure of the blob. The interaction of the blob with the background density has been studied in simulations [99,100,101,102,103]. This interaction produces a sheared flow pattern, which leads to the formation of a steep leading edge and a weak trailing edge. Experimentally the density and potential structure of blobs was verified using probe array in linear machines, stellarators, tokamaks and other small toroidal machines.

The steep leading edge and weak trailing edge structure was confirmed by conditional averaging analysis of measurements from Langmuir-probe data [82,83,95] (see Fig. 2.5b for TCV data). The results are qualitatively consistent with the simulated blob shapes, however, the exact blob shape depends on the density ratio between the blob density and the back-ground plasma density. This asymmetric pulse shape was seen in a number of tokamak experiments [79,83,85,96,104,105,106], stellarators [107] and linear machines [95,108].

Research showed that blobs are large positive bursts in the particle density fluctuations, while holes are negative density bursts on the background plasma. These bursts contribute to the negative or the positive tail of the density PDF. Therefore, the skewness of the mea-sured signal is going to be either positive or negative when blobs or holes are present, re-spectively. The radial skewness profile was measured to be negative in the edge plasma where holes are present and positive in the SOL where only blobs are present [74,80,85,84].

2.3. Recent experimental results on intermittent events 39

ESEL n_e¹⁹=4.4 n_e¹⁹=4.8 n_e¹⁹=6.5 n_e¹⁹=8.4 n_e¹⁹=11

-50 -25 0 25 50

<(n - n) / nrms | n - n >2.5nrms>

τ [µs]

Blob structure at different plasma densities b) Skewness and kurtosis profiles a)

-10 0 10 20 30

Δr [mm]

kurtosisskewness

0 1 2 3 -0.50 0.5 1 1.52

FIGURE 2.5: a) Conditionally averaged particle density fluctuations of blob pulses at the wall of the TCV tokamak at different line averaged core plasma

densities. The densityn_eis given in units of10¹⁹m⁻³[109].

b) Skewness and kurtosis profiles at the HL-2A tokamak. The blob birth zone is depicted with the green area [80].

The region of zero skewness is believed to be close to the birth-zone of the blobs and holes.

An example radial skewness profile and a kurtosis profile are plotted in Figure 2.5b from the HL-2A tokamak [80].

The data supports the physical picture, thus, inside a critial radius of the birth zone, negative density fluctuations (holes) dominate, while outside that radius, positive fluctua-tions (blobs) dominate. Theoretical considerafluctua-tions as well as simulafluctua-tions predict, that the birth-zone of blobs is located near the separatrix, which is indeed seen in the measurements [74,80,110,111]. The kurtosis also tends to be large in the far SOL and it also increases with radius. This is a result of the parabolic relation between the skewness and the kurtosis in such regimes [80,112].

41

Chapter 3

Aim of the doctoral thesis

The aim of the thesis is to contribute to the research of plasma turbulence, one of the most important transport mechanisms in fusion plasmas. One of the main goals was to develop a novel beam emission spectroscopy diagnostic which can measure electron density fluctua-tions resulting from turbulence by either utilizing a Deuterium or a Lithium beam. Another important goal was to utilize the measurement data gathered by this diagnostic and char-acterize plasma turbulence in the plasma edge and the scrape-off layer in different plasma scenarios. In order to achieve these goals, a combined beam emission spectroscopy diagnos-tic was built on the KSTAR tokamak (see Chapter 4), the main parameters and edge turbu-lence measurement capabilities of the system were investigated (see Chapter 5). At the end, scrape-off layer turbulence was characterized in different plasma scenarios (see Chapter 7).

43

Chapter 4

Development of the combined BES observation system

During the design of KSTAR a port was designed for beam emission diagnostics such as charge exchange spectroscopy, motional stark effect measurements and beam emission spec-troscopy. During 2011 a trial beam emission spectroscopy was designed and installed on KSTAR. The purpose of this trial diagnostic was to prove the feasibility of beam emission spectroscopy measurements on KSTAR. Although, the trial diagnostic utilized simple optics and non-optimized filtering, the measurements confirmed the expected capabilities of beam emission spectroscopy on KSTAR. The details of the trial system can be seen in [2]. Based on the experiences with the trial BES system, a final diagnostic was developed, designed and installed on KSTAR. The following sections describe the work regarding design, build and commissioning of the final KSTAR BES system. This section presents the effort carried out in the Korean-Hungarian Joint Laboratory for Fusion Diagnostics cooperation which was the main funding source of the project. The results presented in the first three sections of this chapter were a group effort by the authors of Ref. [1] including the author of this the-sis. The last section of this chapter presents results from solely the author’s work on testing and calibration of the KSTAR BES system. At the end of the chapter a method is presented, which was developed for subtracting spurious oscillations from BES data.

4.1 Observation geometry

During the initial diagnostic design of the KSTAR tokamak, a 6 inch diameter window was allocated for the BES diagnostic at the lowermost position of the M-port. The geometry is shown in Fig. 4.1. The window position allows imaging along field lines crossing the NBI beam line in a typical plasma equilibrium, and a sufficient Doppler-shift is also expected.

The Neutral Beam Injection (NBI) system on the L port of KSTAR consists of 3 ion sources.

The beam cross-section from one source is about 24×60 cm (width×height) while the beam energy is betweenE_b = 70−100keV. If more than one source is operated, the beam becomes stronger, but the cross-section increases causing some reduction in the spatial resolution of the BES diagnostic. Modeling with the comprehensive BES simulation tool (RENATE) [52]

indicated that in all cases, the radial spatial resolution of the DBES system is 1-3 cm, which is suitable for plasma turbulence measurements.

A suitable location was found for the Lithium beam on port K, where the beam could be injected into the same observation volume, where the NBI is observed with the BES optics [55]. Crossing the two beams raises the question whether a substantial excitation or ioniza-tion effect of the NBI on the Li-beam should be expected. The rates of these processes are R =σΦ, whereσ is the cross-section of the processes andΦis the flux of exciting/ionizing particles. In the plasma the dominant processes are via electron impact, therefore, the flux

Observation geometry for Deuterium and Lithium beam 3

2 1 0 -1 -2

-3

3 2 1 0 -1 -2 -3

x [m]

y [m]

1

2

4 3

0 1

-1

1.0 1.5 2.0 2.5

z [m]

R [m]

3 1 4

a) b)

M-port

K-por t

L-port

FIGURE4.1: KSTAR BES observation geometry (a) Toroidal view; b) Poloidal view); 1) M-port window, 2) Field of view, 3) Deuterium beam (NBI), 4)

Lithium beam.

of thermal electrons (Φe) in a typicalTe= 100eV,ne= 1·10¹⁹m⁻³edge plasma is compared to the flux of Deuterium atoms (ΦN BI) in the heating:

ΦN BI = Ib

Ae ≈5·10²⁰[s⁻¹m⁻²] (4.1)

Φ_e=n_e reT_e

me

≈4·10²⁵[s⁻¹m⁻²]. (4.2)

For the Deuterium beam, the expected dominant process is charge-exchange. As the charge-exchange cross sections do not differ from electron-impact cross section by more than 2 orders of magnitudes [113, 114], the heating beam has no significant effect on the Li-beam light emission profile.

According to modeling [52], the light intensity from the two beams is expected to be comparable, but the wavelength differ by about 14 nm. As the typical Doppler-shifts are only 2-5 nm for such beams, filtering of the LiBES and DBES light should not be a problem.

Closeness of the two line radiations simplifies the design of the optics, because they need to be designed only for a narrow wavelength range.