[Paper Review] Real-time Tomography-based Bayesian Inference from TCV Bolometry Data

D. Hamm, C. Theiler+4 more|arXiv (Cornell University)|Mar 12, 2026

Magnetic confinement fusion research0 citations

TL;DR

The paper introduces a real-time tomography-based Bayesian method to estimate radiated power from bolometry data on the TCV tokamak, using pre-computed coefficients and providing uncertainty quantification.

ABSTRACT

Radiated power information is crucial to diagnose and optimize the performance of fusion plasmas. Traditionally, at the TCV tokamak, radiated power analysis has only ever been possible following plasma discharge termination. However, recently, TCV bolometer data have become available in real-time. This offers the opportunity of integrating the radiated power information into the TCV plasma control system. In this work, we propose a novel real-time tomography-based Bayesian technique allowing estimation of the power radiated from user-defined regions of interest in the plasma. The real-time estimates are obtained as computationally cheap linear combinations of bolometer measurements, using pre-computed coefficients that are optimized for the specific discharge planned. This method is not, thus, trained on a set of synthetic or tomographically reconstructed emissivity profiles. We detail the derivation of the technique and show its equivalence to traditional tomographic estimates under suitable conditions. We then demonstrate that this technique enables accurate real-time estimation of the total, core, divertor and main chamber radiated power, by its application to a representative and heterogeneous set of TCV discharges. Finally, we discuss the robustness of the technique to faulty detectors, showing that simple precautions allow safe handling of many common issues. The computational routines implementing the described technique are provided as open-source code.

Motivation & Objective

• Provide real-time estimates of radiated power (total, core, divertor, main chamber) from TCV bolometry data.
• Incorporate magnetic equilibrium information via a diffusion prior to enhance spatial realism.
• Enable uncertainty quantification of radiated power through Bayesian posterior variance.
• Demonstrate robustness to faulty detectors and validate against post-discharge tomographic benchmarks.
• Release open-source implementation for reproducibility and reuse.

Proposed method

• Model the bolometry inversion as a Gaussian posterior using a diffusion prior that encodes magnetic equilibrium information.
• Derive that radiated power quantities of interest are linear functionals of the emissivity, allowing closed-form posterior moments.
• Pre-compute real-time coefficients beta_j from the planned discharge equilibrium (FBT) to express radiated power as a linear combination of bolometer measurements.
• Compute posterior mean as a linear map mu_post = A y and express radiated power estimates as P_rad_tot = b^T x, leading to real-time linear estimators with uncertainty.
• Provide real-time estimates for any region of interest by selecting pixels in that region and forming corresponding beta_j coefficients.
• Address practical issues like channel failures by evaluating multiple channel-selection strategies and demonstrating robustness.

Experimental results

Research questions

• RQ1 Can real-time radiated power estimates (total, core, divertor, main chamber) be obtained from real-time bolometer data without online tomographic inversion?
• RQ2 Does a tomography-based Bayesian approach with a pre-computed equilibrium-informed prior provide reliable uncertainty quantification for radiated power?
• RQ3 How robust is the real-time estimator to faulty or ill-behaved bolometer channels?
• RQ4 How close are real-time estimates to post-discharge tomographic benchmarks across diverse magnetic configurations?
• RQ5 Can the approach be deployed with pre-computed coefficients based on a planned discharge equilibrium (FBT) rather than real-time equilibrium reconstructions?

Key findings

• Real-time estimates of total, core, divertor, and main chamber radiated power closely track post-discharge tomographic estimates.
• The method provides meaningful uncertainty quantification via a Gaussian posterior, with computed variance for the radiated power.
• Coefficients pre-computed from planned equilibria (FBT) suffice for accurate real-time estimation across multiple magnetic configurations.
• The approach demonstrates robustness to faulty channels by testing various channel-selection strategies and maintaining accuracy.
• Open-source code and implementation are released for community use and reproducibility.
• Across 50 TCV discharges, the real-time estimates produced credible intervals that aligned with tomographic references.