Abstract
The core components of nuclear reactors (e.g., fuel assemblies, spacer grids, control rods) encounter harsh environments due to high temperature, physical stress, and a tremendous level of radiation. The integrity of these elements is crucial for safe operation of nuclear power plants; post-irradiation examination (PIE) can reveal information about the integrity of these components. Neutron computed tomography (CT) is one important PIE measurement tool for nondestructively evaluating the structural integrity of these items. CT typically requires many projections to be acquired from different view angles, after which a mathematical algorithm is used for image reconstruction. However, when working with heavily irradiated materials and irradiated nuclear fuel, obtaining many projections is laborious and expensive. Image reconstruction from a smaller number of projections has been explored to achieve faster and more cost-efficient PIE. Classical reconstruction methods (e.g., filtered backprojection), unfortunately, do not typically offer stable reconstructions from a highly asymmetric, few-projection data set and often create severe streaking artifacts. We propose an iterative reconstruction technique to reconstruct curved, plate-type nuclear fuel assemblies using limited-angle CT. The performance of the proposed method is assessed using simulated data and validated through real projections. We also discuss the systematic strategy for establishing the conditions of reconstructions and finding the optimal imaging parameters for reconstructions of the fuel assemblies from few projections using limited-angle CT. Results show that a fuel assembly can be reconstructed using limited-angle CT if 36 or more projections are taken from a particular direction with 1° angular increment.