Integrating Capsule Metrology Data into Bayesian Super Postshot Analysis for Enhanced Predictive Modeling of ICF Experiments

Accurate prediction of inertial confinement fusion (ICF) experiment outcomes is critical for advancing target design and achieving robust performance at the National Ignition Facility (NIF). The Bayesian Super Postshot (BSPS) framework has recently emerged as a powerful approach for combining sparse experimental data with large ensembles of radiation-hydrodynamics simulations, enabling rigorous inference of simulation input distributions and principled quantification of shot-to-shot variability. While BSPS has demonstrated improved predictive capabilities, current implementations do not explicitly incorporate capsule metrology data—such as measurements of pits, voids, and high-Z inclusions—which are known to influence implosion performance.

We will present a new methodology developed to integrate capsule quality data directly into the BSPS analysis pipeline. By augmenting the likelihood function and leveraging machine learning models to relate capsule defects to experimental observables, this approach aims to provide a more comprehensive and physically informed predictive model for ICF performance. Incorporating capsule metrology data into BSPS has the potential to significantly enhance the fidelity and interpretability of predictive modeling for future ICF experiments. This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.