Bayesian inference of radial impurity transport in the pedestal of ASDEX Upgrade discharges using charge-exchange spectroscopy

Abstract

We present a novel framework for quantifying radial impurity transport in the pedestal of ASDEX Upgrade (AUG) discharges. Our method is based on charge-exchange recombination spectroscopy measurements of line radiation from multiple impurity charge states, each along a radially distributed line-of-sight array in steady-state plasmas. Inverse inference based on the diffusive-convective transport solver Aurora combined with a synthetic diagnostic enables us to separate diffusive and convective transport contributions and to derive the impurity density and charge state distribution profiles. Robust uncertainty quantification is provided as the full probability distribution of the parameters is obtained according to Bayesian statistics with the use of a nested sampling algorithm. The approach allows for a high radial resolution and data quality due to the steady-state plasma, but requires data from multiple impurity charge states. It is, therefore, particularly suitable for impurity transport studies in the region of steep edge gradients. In this paper, we present thorough tests of the method based on synthetic data. Furthermore, we show an application to AUG measurement data, inferring the pedestal neon transport in the quasi-continuous exhaust (QCE) regime without large edge-localized modes. The comparison of the transport result with neoclassical simulations shows a clear contribution of turbulent diffusion in the QCE pedestal. This supports the hypothesis of additional transport associated with the predicted high-n ballooning-unstable region and the observed quasi- coherent mode.

Keywords

• ASDEX Upgrade
• impurity transport
• pedestal
• charge-exchange spectroscopy
• Bayesian inference
• nested sampling