UCLA

This dissertation reports advancements in measuring density fluctuations and flow velocity in tokamak plasmas using the Doppler back-scattering (DBS) diagnostic. Two separate investigations are presented: measurements of the back-scattered power perpendicular wavenumber spectrum, and a thorough cross-diagnostic comparison of E × B rotation measurements. To measure the perpendicular wavenumber spectrum using DBS, the launch- angle(s) of the probing beam were scanned to probe different values of k⊥ at the cutoff layer. Measurements were made in RF-heated H-mode plasmas with low collisionality and dominant electron-heating – a regime expected to be relevant for future burning plasmas. The back-scattered power spectrum, Ps(k⊥), is found to exhibit exponential decay. In normalized wavenumbers, the exponential decay factor (Ps(k⊥) ~ e−ζ(k⊥ρs)) was found to be ζ ≈ 1.7 over the majority of the spectrum and ζ ≈ 3.9 at high-k⊥ρs. Novel synthetic diagnostic modeling was performed to test models of plasma turbulence against the measured Ps(k⊥). DBS instrumental effects and the plasma density fluctuations, δn, were each modeled with varying levels of physics-fidelity. Reduced model beam-tracing and quasi-linear gyro-fluid results are compared with higher-fidelity full-wave and nonlinear gyrokinetics simulations. Ultimately, mixed-agreement between measurements and simulations across the wavenumber spectrum is found. At the lowest measured wavenumbers k⊥ρs ≤ 0.4 and for k⊥ρs ≥ 1.2 simulations and measurements agree as to the shape of the wavenumber spectrum. Over the range 0.4 < k⊥ρs < 1.2, simulations and measurements disagree as to the rate of spectral decay. Transport modeling found that while the reduced quasi-linear model can match the experimental fluxes, it does not reproduce the higher-fidelity nonlinear gyrokinetic behavior in the electron energy channel. In a separate investigation, rigorous cross-diagnostic comparisons were made between DBS and the charge-exchange recombination spectroscopy (CER) diagnostics. Measurements were made in an L-mode plasma where neutral beam injection (NBI) applied a variable external torque. In analyzing the DBS measurements, novel spectral fitting methods and uncertainty propagation were used to construct the DBS-CER comparison dataset. Statistical analysis suggests a tendency for DBS and CER to agree within their respective uncertainties. Across the comparison dataset, the average difference in ωE×B was found to be |d| = 0.1 kRad/s. The 95% confidence interval for their agreement was found to be approximately 5 kRad/s in terms of ωE×B. The phase velocity of δn fluctuations was estimated using linear gyro-fluid analysis and found to be negligible relative to vE×B. The results of the rotation investigation provide a certain level of mutual-validation for DBS and CER in measuring E × B rotation.