Implementation and validation of swept density reflectometry for integrated data analysis at ASDEX Upgrade

Area of Science:

• Plasma physics
• Fusion energy research
• Diagnostic techniques

Background:

• Integrated Data Analysis (IDA) in tokamaks like ASDEX Upgrade infers plasma properties using diverse data.
• Accurate forward modeling is crucial for probabilistic evaluation and Bayesian inference in IDA.
• Existing methods for density profile determination have limitations, particularly in the outer core region.

Purpose of the Study:

• To introduce a novel, efficient forward model for O-mode profile reflectometry suitable for Bayesian inference within IDA.
• To overcome limitations of traditional methods like Abel inversion for determining plasma density profiles.
• To improve the accuracy and uncertainty quantification of electron density profiles and their gradients.

Main Methods:

• Development of an efficient forward model based on the analytic solution for a piece-wise linear density description.
• Integration of O-mode profile reflectometry data into the IDA framework.
• Utilizing complementary diagnostics (e.g., lithium beam) for initial profile shape analysis.
• Probabilistic evaluation incorporating error propagation from measurements and uncertain sources.

Main Results:

• The new forward model enables overcoming issues with established cut-off location determination methods.
• The model allows for analysis of initial density profile shapes using diagnostics beyond hard-coded initializations.
• Joint analysis of complementary diagnostics, including the new reflectometry model, provides density profiles with uncertainties.
• The method successfully closes a data gap in the outer core region.

Conclusions:

• The new O-mode profile reflectometry forward model is a valuable addition to ASDEX Upgrade's IDA.
• The probabilistic approach intrinsically handles error propagation, providing reliable uncertainty quantification.
• The method supports robust density profile evaluation even with missing or ambiguous data.
• This approach is suitable for real-time analysis, offering density profiles with error bars.