Search NASA⌕ Search

Engineering topics

Shin, Dongil

Publications and source records attributed to Shin, Dongil.

7 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Paper #50 Preliminary Sizing of Active Magnetic Bearings for sCO2 Waste-Heat Recovery Application

The team of Southwest Research Institute (SwRI) and GE Vernova (GE) is executing a project developing conceptual designs of high-temperature active magnetic bearings (AMBs) for sCO2 machinery applications. The goals of the project are to develop conceptual designs for radial and thrust AMBs capable of operating in sCO2 environments up to 1000°F. A waste-heat recovery (WHR) application with hermetic machinery is used as a reference case to apply AMBs. This application was previously investigated to compare hermetic machinery enabled by process-lubricated bearings and high-speed motors/generators with traditional machinery configurations using oil-lubricated bearings, gearboxes, and grid-frequency electric machines. In this paper, AMBs are sized for the WHR application machinery. Load capacity estimates are made using first-principle-derived formulas that consider operation at elevated temperatures. Estimates of linearized AMB system coefficients are also made to aid future controller development. Rotor models are developed that account for AMB rotor geometry and housing space requirements. Rotordynamics studies are performed to identify reasonable target values for closed-loop AMB stiffness and closed-loop AMB damping. These target values will be used in future control studies.

Lipham, Robert↗
DOE OSTI2023

Deep material network via a quilting strategy: visualization for explainability and recursive training for improved accuracy

Recent developments integrating micromechanics and neural networks offer promising paths for rapid predictions of the response of heterogeneous materials with similar accuracy as direct numerical simulations. The deep material network is one such approaches, featuring a multi-layer network and micromechanics building blocks trained on anisotropic linear elastic properties. Once trained, the network acts as a reduced-order model, which can extrapolate the material’s behavior to more general constitutive laws, including nonlinear behaviors, without the need to be retrained. However, current training methods initialize network parameters randomly, incurring inevitable training and calibration errors. Here, we introduce a way to visualize the network parameters as an analogous unit cell and use this visualization to “quilt” patches of shallower networks to initialize deeper networks for a recursive training strategy. The result is an improvement in the accuracy and calibration performance of the network and an intuitive visual representation of the network for better explainability.

97 MATHEMATICS AND COMPUTING↗