Sandia_Natl_Labs_FY19_LDRD_Annual_SAND2020-3752 R_2_S


Credible image meshing enables as-built computational analysis. A Sandia LDRD research team developed new technologies to automatically and credibly convert 3D images of materials or components arising from imaging techniques such as x-ray computed tomography (XCT) into high-quality computational meshes for multi-physics simulation. Deep Bayesian neural networks identify parts, materials, or phases of interest in the images. New facet-based meshing algorithms create high-quality and computationally efficient meshes. The uncertainty that is inherent in each of these processes is quantified and propagated, providing impact of geometric uncertainty on physics predictions. “Robust image-based meshing introduces a new paradigm for computational simulation, enabling enhanced surveillance of as-built components and direct feedback

on the impact of manufacturing variability on component performance, each of which enhances mission agility,” said PI Scott Roberts. The newly developed workflow has already been applied to a variety of Sandia mission applications, including thermal protection systems, battery materials, detonators, thermal sprays, and laser welds. Image-based simulation of a woven composite material showing greyscale XCT (top left), a high- quality tetrahedral mesh (bottom), and a thermal- mechanical simulation (top right). (Images courtesy of Lincoln Collins)

Forecasting marine sediment properties on and near the Arctic shelf with geospatial machine learning. This LDRD team proposed a combination of geospatial machine learning prediction and sediment thermodynamic/physical modeling to create probabilistic maps of geoacoustic and geomechanical sediment properties. This new technique for producing reliable estimates of Arctic seafloor properties will better support naval operations relying on sonar performance and seabed strength and can constrain models of shallow tomographic structure that are important for nuclear treaty compliance monitoring/ detection. By gaining more complete awareness of the battlespace environment through the development of seafloor forecasting capability, it will provide an assessment tool to support decision- making by warfare commanders. Such a tool can be made to automatically ingest data to produce a battlespace-assessment calculation that will improve blue force status awareness. It also benefits Sandia’s energy security mission areas, because it will provide a means to estimate the resource potential of seafloor petroleum systems. Two of the most important geologic parameters that determine seafloor acoustic properties are free gas and methane gas hydrate (natural gas). The project results were presented at the National

Geospatial-Intelligence Agency’s Maritime Community of Practice Meeting in November 2019, the American Geophysical Union Annual Fall Meeting in December 2019, and the Gordon Research Conference on Natural Gas Hydrate Systems in February 2020.



Made with FlippingBook Learn more on our blog