A robust reconstruction of three-dimensional asymmetries in ICF implosions at the NIF, using a physics model and neural networks applied to multiple heterogeneous data sources

Three-dimensional (3D) asymmetries represent major performance-degradation mechanisms in inertial-confinement fusion (ICF) implosions at the National Ignition Facility (NIF). These asymmetries can be diagnosed with the three neutron imaging systems (NIS) fielded with orthogonal views to the implosion. Conventional tomographic reconstructions are typically used to reconstruct the 3D morphology of the hot-spot and surrounding high-density fuel-shell in an implosion using NIS data, but the reconstruction problem is ill-posed with only three imaging lines of sight. Under certain assumptions, relative low-mode asymmetries of the surrounding high-density fuel-shell can also be diagnosed with the suite of real-time neutron activation diagnostics (RTNADs) and the neutron time-of-flight (nTOF) detectors. In this work, a machine-learning based 3D reconstruction technique has been developed and used to overcome these limitations by combining information from NIS, RTNAD and nTOF data. In this reconstruction technique, a physics model combined with a group of neural networks is used to analyze the NIS, RTNAD and nTOF data to robustly reconstruct the 3D morphology of the hot-spot and surrounding high-density fuel-shell in a NIF implosion. This technique provides new insights about the origins of the 3D asymmetries in an implosion, information that cannot be obtained from the different data sources individually. In this presentation, development and use of this technique will be discussed, and how it will be used to guide the program toward minimizing asymmetries in ICF implosions and achieving higher performance.