Experimental determination of the correlation properties of plasma turbulence using 2D BES systems

Abstract

Abstract. A procedure is presented to map from the spatial correlation parameters of a turbulent density field (the radial and binormal correlation lengths and wavenumbers, and the fluctuation amplitude) to correlation parameters that would be measured by a Beam Emission Spectroscopy (BES) diagnostic. The inverse mapping is also derived, which results in resolution criteria for recovering correct correlation parameters, depending on the spatial response of the instrument quantified in terms of Point-Spread Functions (PSFs). Thus, a procedure is presented that allows for a systematic comparison between theoretical predictions and experimental observations. This procedure is illustrated using the MAST BES system and the validity of the underlying assumptions is tested on fluctuating density fields generated by direct numerical simulations using the gyrokinetic code GS2. The measurement of the correlation time, by means of the cross-correlation time-delay (CCTD) method, is also investigated and is shown to be sensitive to the fluctuating radial component of velocity, as well as to small variations in the spatial properties of the PSFs.Keywords: Beam-emission spectroscopy, point-spread functions, synthetic diagnostics, plasma turbulence, plasma diagnostics, tokamaks. Inner Outer Figure 1. Example e −1 amplitude contours of PSFs [1] for five specific cases taken from MAST shots and described in Section 4.3. The dots mark the locations of the focal points of the detector channels. The characterisation of the PSFs is described in Section 4.2.

Ion-scale turbulence in MAST: anomalous transport, subcritical transitions, and comparison to BES measurements

“…Finally, the application of the synthetic diagnostic leads to a reduction of roughly 50% of the RMS fluctuation amplitude, i.e., from (δn i /n i ) GS2 rms ∼ 0.01-0.08 for the raw density fluctuations to (δn i /n i ) SYNTH rms ∼ 0.005-0.03. This observation is consistent with a recent detailed analysis [49] of the effect of PSFs on the measurement of MAST turbulence using a subset of GS2 simulations reported here.…”

“…Comparison between results of analysis with and without the synthetic diagnostic shows that the synthetic diagnostic has a measurable effect on several turbulence characteristics, including the poloidal correlation length and the fluctuation amplitude, consistent with the conclusions of Ref. [49]. Finally, we present the correlation lengths and times as functions of the ion heat flux and again show that the structure of the turbulence in our simulations is effectively only a function of this parameter, which measures the distance to the turbulence threshold.…”

“…Density fluctuations are inferred from D α emission produced by the NBI beam as it penetrates the plasma. The measured fluctuating intensity of the D α emission is proportional to the local plasma density at the corresponding viewing location, and the two quantities are related via point-spread functions (PSFs) [47,51,49]. The PSFs depend on the magnetic equilibrium, beam parameters, viewing location, and plasma profiles and as a result, have to be calculated explicitly for each measurement [47].…”

“…[42,43,47] and, for a detailed study of the effect of PSFs on the measurement of turbulent structures, to Ref. [49].…”

“…Following [47], we can use this to calculate the correlation time τ c by tracking turbulent structures as they move poloidally and measuring their temporal decorrelation. This "cross-correlation time delay" technique [44,47,49] is as follows. We calculate the correlation function C ∆Z (∆t) = C(∆R = 0, ∆Z, λ(θ = 0), ∆t) for several poloidal separations ∆Z, as shown in figure 33.…”