Bayesian Activity Estimation and Uncertainty Quantification of Spent Nuclear Fuel Using Passive Gamma Emission...

Area of Science:

• Nuclear Engineering
• Applied Mathematics
• Computational Physics

Background:

• Passive gamma emission tomography (PGET) is crucial for spent nuclear fuel characterization.
• Accurate activity estimation is vital for safe storage and reprocessing.
• Existing methods face challenges with noise and complex geometries.

Purpose of the Study:

• To develop and validate a robust Bayesian method for activity estimation in PGET.
• To compare the performance of isotropic Gaussian and Poisson noise models.
• To assess the method's robustness under model mis-specification with non-linear forward models.

Main Methods:

• Formulation as a Bayesian linear inverse problem.
• Application of a Bernoulli-truncated Gaussian prior for sparse configurations.
• Utilizing a split and augmented Gibbs sampler (Markov chain Monte Carlo - MCMC).
• Validation with synthetic and simulated realistic data.

Main Results:

• The proposed MCMC algorithm demonstrates superior performance in estimating pin activities.
• The method accurately quantifies uncertainty measures.
• Robustness analysis shows effectiveness even with mis-specified linear models for non-linear problems.

Conclusions:

• The developed Bayesian MCMC approach offers enhanced accuracy and uncertainty quantification for PGET activity estimation.
• This method provides a significant improvement over existing techniques for spent nuclear fuel analysis.
• The approach is valuable for complex scenarios, including non-linear forward models.