«Association EURATOM / IPP.CR I N S T I T U T E O F P L A S M A P H Y S I C S, v.v. i. ACADEMY OF SCIENCES OF THE CZECH REPUBLIC ANNUAL REPORT ...»
To evaluate the transmission factor of the RFA slit, information about the transverse electron energy was needed. To this purpose, a modeling of fast electron generation including the electron transverse energy development was performed. It was found that the gain of the transverse energy of the electrons in the fast particle beam obtained from the LH field or from the transformation of the longitudinal velocity of the fast particles to the transverse velocity by collisions is negligible in the RFA transmission factor calculations.
The fast electron energy distribution of accelerated particles in the fast electron beam was determined: Also for as low LH power as 0.35 MW, fast electrons with energy larger than 400 eV are generated. Even for UG1= -1000V, fast electrons with energy higher than 1 keV remain to be present in the “2nd” beam near the LCFS. The fast electron beam generated by the parasitic wave ANNUAL REPORT 2007 ASSOCIATION EURATOM/IPP.CR absorption extends for high PLH up to the LCFS and is up to 5 cm or even more radially wide, with possible corresponding radially broader heat loads in case that the beam impacts on the wall.
The observed existence of the wide beam needs to be explained by theory, as the current theory predicts the radial width of the beam of several mm. The fast electron beam intensity very near to the LCFS decreases more quickly with decreasing PLH than the fast beam intensity near the LH grill mouth. As the comparison of the 1st and 2nd beam shows, the mean collector current (and therefore similar unwanted heat loads) in the two beams is comparable at higher PLH. The 2nd beam generation might perhaps be connected with the nonlinear production of higher spatial harmonics, studied in exploration of the so called spectral gap problem. No locally produced fast ions by the LH wave were found.
A more sophisticated modeling of fast particle generation shows that the locally LH produced gain of the transverse energy of the fast particles can be neglected in the calculations of the transmission factor of the RFA slit. The knowledge of the transmission factor is needed in parasitic absorption heat load calculations from RFA measurements. No clear indications of fast particles emanating from the region of ICRF antenna were found by the RFA. The flux tubes connected to ICRH Q5 antenna.were explored by the RFA to determine sheath potential in presence of RF as a function of SOL density.
All results mentioned above were obtained in a team effort.
 V. Petrzilka et al., presented at the Tarragona EPS05 Conference, paper P-2.095.
 V. Petrzilka, Plasma Physics Contr. Fusion 33 (no 4) (1991) 365.
Part II - PHYSICS Modeling of LH wave ionization, development of EDGE2D code to 3D V. Petržílka
In collaboration with:
J. Mailloux, G. Corrigan, V. Parail, K. Erents, P. Jacquet, K. Kirov, J. Spence, Assoc.
M. Goniche, A. Ekedahl, Assoc. EURATOM-CEA-Cadarache;
K. Rantamäki, Association Euratom – Tekes;
P. Belo, Association Euratom-IST, Lisboa;
J. Ongena, Ass. EURATOM – Belgian State We performed numerical modeling of near Lower Hybrid (LH) grill Scrape-off-Layer (SOL) plasma density variations ne,SOL as a function of gas puff and LH power ionization with the fluid code EDGE2D. Novel 3D limiter-like features in the EDGE2D are introduced. The modeling shows that both puff and heating/ionization are important in raising the density in the far SOL, what is important for LH wave coupling in ITER like discharges.
The modeling of gas puff ionization is a 3D problem, but an amended 2D poloidal – radial model can provide useful insights. The code includes direct SOL ionization by the LH wave . It is assumed that the ionization by the LH wave is produced due to the local SOL electron heating by the wave. A fraction of power lost in SOL increases the SOL Te, which in turn increases ionisation in the SOL. Locally generated fast particles can also participate in the heating process.
EDGE2D  is modified so that it can model a SOL up to 10 cm or more. Modelling of the private space in front of the grill needs to introduce 3D limiter-like features into EDGE2D. This is attempted in the present contribution. The poloidal limiters are modeled as spatially localized sinks, where the recombination is artificially strongly enhanced. This quasi 3D modeling with poloidal limiters acting as sinks allows to distinguish the private space of the grill between the poloidal limiters in the modeling.
Fig. 1. Schematic representation of the 2D and 3D configuration of the LH grill and RCP.
In this way, it is possible to investigate the impact of parameters such as gas puffing rate, gas puffing poloidal location, LH power dissipated in the SOL and radial location of the LH dissipated power (e.g., radially near the grill mouth, or further away from the grill mouth), separately on the ne,SOL in the private launcher SOL and in the private RCP (Reciprocating Probe) SOL space. However, the gas puff location in the 2d modeling is always magnetically connected to the LH grill private space, Fig. 1. It is therefore impossible to distinguish between magnetically connected and not-connected gas puff. Similarly, the RCP location is also always connected to the LH grill private space. Obviously, there is a need for a fully 3D modeling code, to obtain a more quantitative description of the trends obtained now by the quasi-3D model.
0 0.02 0.04 0.06 0.08 0.1 Fig. 2. Left figure: Effects of the limiter boundary location. x- axis: distance from separatrix in m. Right figure: Profile “D” of the LH field dissipation [a.u.]. Series 1 (blue diamonds): dLW =4.75 cm (grill ~ 3 cm behind the limiter), series 2 (magenta rectangles): dLW= 2.75 cm (grill ~ 1 cm behind the limiter), and series 3 (yellow triangles): dLW= 1.25 cm (grill is ~ flush with the limiter).
Figure 2 then shows ne,SOL in LH private SOL as a function of the limiter boundary location (changing the distance limiter – wall dLW is similar to changing the distance launcher-limiter) Gas puff is 1e22 el/s near the outer mid-plane (OMP), i.e., by “GIM6”. Heating in front of the grill is 150 kW. The assumed profile of the LH SOL dissipation used in this contribution is shown as profile “D” on the bottom figure. The upper figure shows the ne,SOL in the OMP (between limiters acting as a sink) in the launcher private SOL. The next Figure 3 shows ne,SOL in the limiter sink (top figure) and at the RCP location (bottom figure) as a function of the limiter boundary location. Gas puffing and heating is the same as in Fig.2.
Part II - PHYSICS Fig. 3. Profiles of ne,SOL near limiters and in the RCP private SOL.
Fig. 4. Upper figure: ne,SOL as a function of the heating and puff rates, at the bottom figure there are corresponding neutrals profiles in the OMP. Blue diamonds: 0 heating, 0 puff, magenta rectangulars: 0 heating, puff = 1e22 el/s, yellow triangles: heating = 150 kW, 0 puff, Cyan diamonds: heating = 150 kW, puff = 1e22 el/s.
Fig. 5. Profile of ne,SOL in the OMP in grill private SOL, as a function of the gas puff location.
Blue diamonds: 0 heating, 0 puff, magenta rectangulars: gas puff 1e22 el/s near OMP, yellow triangles: gas puff near RCP, cyan diamonds: gas puff at the top.
Figure 4 shows ne,SOL in the OMP (upper figure) as a function of the heating and puff rates, with dLW = 4.75 cm; on the bottom figure there are neutrals profiles in the OMP. Let us note that joined heating and gas puff tend to flatten the ne,SOL profile, cf. the cyan curve. It can be demonstrated that the flatness of the ne,SOL profile depends also on the assumed profile of the LH wave dissipation. The more near to the grill the LH power is dissipated, the more flat is the ne,SOL profile. Figure 5 shows ne,SOL in the OMP in grill private SOL, as a function of the gas puff location, dLW = 1.25 cm, with heating in front of the grill = 150 kW.
Main results: The modeling shows that both puff and heating/ionization are important in raising the density in the far SOL, what is important for LH wave coupling in ITER like discharges.
Although OMP seems to be the best location for gas puffing, the other two puff locations (near RCP, at the top) also give an increase in ne,SOL with heating. This is again an important information for ITER, where top gas puffing is assumed. The modeling shows the flattening of the far ne,SOL profile, which is observed in experiments .
All results mentioned above were obtained in a team effort.
 V. Petrzilka et al., 34th EPS07 Conference, Warszaw, July 2-6th, 2007, paper P4.100, preprint EFDA–JET–CP(07)03/60, to be submitted to Nuclear Fusion.
 M.Goniche et al., 34th EPS07 Conference, Warszaw, July 2-6th, 2007 paper P1. 152, preprint EFDA–JET–CP(07)03/40.
Part II - PHYSICS
Cherenkov detectors for fast electron measurements V. Weinzettl, J. Stöckel, M. Vácha, M. Peterka
In collaboration with:
L. Jakubowski, M.J. Sadowski, J. Stanislawski, K. Malinowski, J. Zebrowski, M. Jakubowski, Association EURATOM IPPLM, The Andrzej Soltan Institute for Nuclear Studies, OtwockSwierk, Poland Fast electrons generated by the ohmic heating in the CASTOR tokamak plasma were recorded by means of the Cherenkov detection system described in . The experimental data were collected in about 500 discharges. The shots were 25 ms long and performed at the toroidal magnetic field BT ranging from 0.8 T to 1.4 T, the plasma current Ip varied from 5 kA to 15 kA. The measurements were limited to relatively low plasma densities ne reaching 0.5-1.5×1019 m-3 and a relatively high acceleration voltage VLOOP, typically higher than 2 V.
Measurements indicate that introduced modifications of the Cherenkov detection system enabled an electromagnetic interference to be reduced and a direct hard radiation to be eliminated completely. Using a radiation attenuation in lead shielding, an energy of the direct radiation was estimated as E = (3,9 ± 1,0) MeV that gives the same energy as for a test water shielding, E = (3,8 ± 0,1) MeV. A question on an unclear origin of this hard radiation reaching the detection system only during the tokamak discharges stays still unanswered.
The entrance window of the detector, by default oriented in the direction opposite to the plasma current allowing fast electron measurements, was turned around its axis by 180 degrees in agreement with the plasma current in some shots of the campaign. The recorded signals were in this case very low, see Fig.1, what confirmed that the recorded Cherenkov signals were induced just by fast electrons (50 keV).
On the CASTOR tokamak, the dependences of the fast electron signals on the radial position of the Cherenkov detector, as well as on the plasma density, see Fig.1-2, plasma current and toroidal magnetic field were investigated [2,3]. A character of the signals depended very strongly on the radial position of the detector as well as on the plasma density. With an increase of the observation radius, the Cherenkov signal decreased and it Fig. 1. Integrated Cherenkov emission as a function of the appeared later. It was found that after plasma density. The blue curve corresponds to the 25 ms, when the transformer primary orientation of the red detector allowing fast electron the measurements, curve shows electron signal at the winding (inducing the ohmic heating) opposite detector orientation.
was short-circuited and the stabilization system was turned off, a strong increase in the signal intensity appeared as a result of the destruction of the plasma column. A considerable influence of the plasma density was also observed. In the high-density discharges, the initial increase in the Cherenkov signal was followed by a stationary phase. It can be directly interpreted as a constant loss rate of the fast electrons. On the contrary, the low-density discharges showed almost an exponential growth of
the Cherenkov signals lasting until the end of the discharge. The described effect might be connected with a magnitude of the acceleration electric field reaching a critical value in the lowdensity discharges.
Fig. 2. Evolution of the Cherenkov signals as a function of the position of the detector head for highdensity (left) and low-density (right) discharges.
A connection between the Cherenkov emission and hard X-rays (HXR), which were measured at a close vicinity of the Cherenkov detector, was also studied, see Fig.3. As it can be clearly seen, a global behaviour of the hard X-ray radiation differs from the Cherenkov emission; however, there are some common peaks at certain moments confirming an interaction of fast electrons with the Cherenkov detector head. But a correlation of the Cherenkov and local HXR signals can reach up to 70 % and its magnitude practically does not depend on the detector position, plasma density or toroidal magnetic field. A correlation peak FWHM varies from 6 to 11 µs and slowly increases towards the plasma center. It should be noticed that a position of the correlation peak indicates a small but nonzero delay between the Cherenkov and hard X-ray emission in the range of 1-3 µs (the hard X-rays outruns the Cherenkov emission).
The results of the statistical approach using a single-count analysis of the time-resolved Fig. 3. Example of the temporal evolution of the Cherenkov data indicates possible transport Cherenkov emission measured at r = 60 mm, the mechanisms of fast electrons – the fast burst local hard X-ray, and the total hard X-ray losses in combination with a slow diffusion. intensity.
 L. Jakubowski, et al., 34th EPS Conference on Plasma Physics, Warsaw, Poland, July 2-6, 2007, ECA Vol.31F, P-5.097 (2007)  J. Zebrowski, et al., PLASMA 2007 – International Conference on Research and Applications of Plasmas, Greifswald, Germany, October 16-19, 2007, WeP3  L. Jakubowski, et al., 17th IAEA Technical Meeting on Research Using Small Fusion Devices, Lisbon, Portugal, October 22-24, 2007, P17
Part II - PHYSICS
JET neutron data analyses via inversion algorithms based on Minimum Fisher Regularisation J. Mlynář, V. Svoboda
In collaboration with: