Search for: All records

Melt pool dynamics in metal additive manufacturing (AM) is critical to process stability, microstructure formation, and final properties of the printed materials. Physics-based simulation, including computational fluid dynamics (CFD), is the dominant approach to predict melt pool dynamics. However, the physics-based simulation approaches suffer from the inherent issue of very high computational cost. This paper provides a physics-informed machine learning method by integrating the conventional neural networks with the governing physical laws to predict the melt pool dynamics, such as temperature, velocity, and pressure, without using any training data on velocity and pressure. This approach avoids solving the nonlinear Navier–Stokes equation numerically, which significantly reduces the computational cost (if including the cost of velocity data generation). The difficult-to-determine parameters' values of the governing equations can also be inferred through data-driven discovery. In addition, the physics-informed neural network (PINN) architecture has been optimized for efficient model training. The data-efficient PINN model is attributed to the extra penalty by incorporating governing PDEs, initial conditions, and boundary conditions in the PINN model.

 

Data and R script repository for the manuscript to reproduce figs 5 and 6; Deglacial temperature and carbonate saturation state variability in the tropical Atlantic at Antarctic Intermediate Water Depths" (Oppo et al., 2023) 

This archived Paleoclimatology Study is available from the NOAA National Centers for Environmental Information (NCEI), under the World Data Service (WDS) for Paleoclimatology. The associated NCEI study type is Paleoceanography. The data include parameters of paleocean (oxygen isotopes) with a geographic location of North Atlantic Ocean. The time period coverage is from 22423 to 563 in calendar years before present (BP). See metadata information for parameter and study location details. Please cite this study when using the data. 

Time dependent observations of point-to-point correlations of the velocity vector field (structure functions) are necessary to model and understand fluid flow around complex objects. Using thermal gradients, we observed fluid flow by recording fluorescence of$${\text{He}}_{2}^{*}$$${\text{He}}_2^{\ast }$excimers produced by neutron capture throughout a ~ cm³volume. Because the photon emitted by an excited excimer is unlikely to be recorded by the camera, the techniques of particle tracking (PTV) and particle imaging (PIV) velocimetry cannot be applied to extract information from the fluorescence of individual excimers. Therefore, we applied an unsupervised machine learning algorithm to identify light from ensembles of excimers (clusters) and then tracked the centroids of the clusters using a particle displacement determination algorithm developed for PTV.