Using Bayesian analysis and Gaussian processes to infer electron temperature and density profiles on the Mega-Ampere...

Area of Science:

• Plasma physics
• Fusion energy research
• Computational physics

Background:

• Accurate measurement of plasma profiles is crucial for understanding and controlling fusion devices.
• Traditional analysis methods for Thomson scattering (TS) data may have limitations in profile reconstruction.

Purpose of the Study:

• To develop and present a unified, Bayesian inference method for determining midplane electron temperature and density profiles.
• To incorporate novel techniques, including Gaussian process priors and Gauss-Laguerre quadratures, to enhance profile inference accuracy.

Main Methods:

• A Bayesian inference framework is employed, integrating data from Thomson scattering (TS) and interferometry.
• A Gaussian process prior is utilized to infer the mollification length-scale of the electron density and temperature profiles.
• Gauss-Laguerre quadratures are used for direct calculation of the depolarization term in the TS forward model.

Main Results:

• The unified Bayesian inference method was applied to high-resolution TS data from the Mega-Ampere Spherical Tokamak (MAST).
• The inferred profiles demonstrate improved accuracy and robustness compared to those obtained from standard analysis techniques.
• The novel features of the inference method contribute to a more reliable reconstruction of plasma profiles.

Conclusions:

• The developed unified Bayesian inference provides a powerful tool for accurate plasma profile reconstruction in fusion devices.
• The integration of Gaussian processes and advanced quadrature methods enhances the fidelity of Thomson scattering data analysis.
• This approach offers a significant advancement in the analysis of experimental data for magnetic confinement fusion research.