Geometrical scheme of the NBI lines and line of sight of the CNPA in TJ-II (top view). In TJ-II, the plasma column has a vertical excursion of every period ( ), so the CNPA LOS does not look straight to NBI .
CNPA channels and energy width (error bars in the plot, usually smaller than the symbols) and efficiency. Measurements in the low energy channels are not precise because these channels overlap in energy.
Plasma parameters (effective radius, inclination of the field lines, and neutral density) as a function of the distance x along the LOS (sector B2, tangential port).
Left plot: plasma profiles used in ISDEP for the comparison with the CNPA: density, electron and ion temperatures, electric potential, and neutral density. Right plot: Spitzer slowing down time, 4 showing that the typical timescale of this system is a few milliseconds.
Left plot: cross sections for the atomic processes in the fast ion detection. 15 The charge exchange ( ), ionization protons impact ( ) and electron impact ( ) cross sections are expressed in terms of the fast particle energy (E). The ionization by electron impact rate ( ) depends on the background electron temperature Te . In this case, it is plotted for . Right plot: dependency of the electron impact ionization rate coefficient with the electron temperature. It is approximately constant for the TJ-II electron temperatures in NBI discharges ( ).
Charge exchange time in TJ-II, according to Eq. (4) , in the conditions considered. It turns out that CX processes are not negligible only in the external part of the plasma and for low energies, where is short enough. Since ISDEP deals with orbits in the SOL, where is calculated in the SOL assuming that the profiles of plasma parameters are constant in the SOL and equal to their value at .
Left plot: line density, X-rays and NBI traces for shot #18982. Right plot: CNPA raw data for selected channels (see Fig. 2 for the energy associated to each channel and its width). For this shot, a 30 ms plateau is found in the signals. We can find a low error flux spectrum averaging each channel in time and taking the mean with similar shots.
Steady state distribution function for TJ-II with the profiles shown in Fig. 4 for (top, left); (top, right); (bottom, left), and (bottom, right). Traces of the injected ions at 10, 15, and 30 keV can be seen. The distribution function is normalized so . The statistical error in f depends on the value of f, see Fig. 9 .
Relative statistical errors in f (see Fig. 8 ), in %.
Comparison of the flux spectrum measured by the CNPA and calculated with ISDEP. Both flux spectra are normalized to one for . In the CNPA spectrum, the error bars are the standard deviation of the measurements of the different discharges. In the ISDEP curve, the error bars are the Monte Carlo statistical errors. The error bars are smaller than the symbol size for many points in the spectrum. The persistence of fast ions P(t) is shown in the small chart indicating the existence of two timescales: small NBI direct losses around and CX or hits to the vacuum vessel at larger times.
Plasma profiles for a low density NBI shot (left). CNPA spectra and ISDEP simulation (right).
Plasma volume, average minor radius, and iota range (from the axis to the edge) for the two magnetic configurations considered.
Comparison of absolute values of the fast neutral spectra at 10 keV. Errors are indicated in parentheses.
Article metrics loading...
Full text loading...