Direct numerical simulation of laser powder bed fusion process on mesoscopic level can serve as a predictive tool for the construction quality of a part. The non-stationary fluid dynamics with surface tension and wetting effects, recoil pressure, heat transfer and powder deposition are included in the high-fidelity model formulated in the paper. The lattice Boltzmann method (LBM) is used for the simulation of the melt-pool dynamics and heat transfer while the heat equation is numerically solved on a large-scale adaptive grid. The discrete element method (DEM) is used for the simulation of powder deposition after each layer for the multilayer simulations. A hypothesis is proposed that the powder wetting is a key parameter defining the morphology of a multilayer thin-wall detail. A number of numerical experiments of the thin wall builds of Inconel 625 is carried out in comparison with the experimental data of Air Force Research Laboratory (AFRL) Additive Manufacturing (AM) Challenge Series. The model accurately predicts the wall shape and height profile for all thin wall cases available in the AFRL AM Challenge data. It is shown by the simulations, that the wall height variation and height of the hill at the beginning of the track is strongly dependent on the wetting condition of the powder by the melt. An explanation to the effect is proposed based on melt pool dynamics in the presence of powder. It is concluded that the powder wetting is necessary to be taken into account for the predictive modeling of the morphology of mesoscopic thin-walled structures.
Accurate fluid simulations require high computing cost. 3D modelling of fluid dynamic field evolution on a discrete mesh takes large amount of data storage, and data access becomes performance bottleneck. Our work is concerned with the task of mitigating the limitations that are caused by finite memory throughput in the parallel simulations. We use LRnLA algorithms for this issue, where localized tasks combine updates on several time layers. In this paper, the compact update for DiamondTorre LRnLA algorithm is constructed. It further improves localization of DiamondTorre algorithm, which improves arithmetic intensity for cross-stencil schemes. The ratio of loaded data to fully updated data approaches 1. The compact update is implemented with CUDA C++ for a numerical scheme for the advection-diffusion equation. 50 GLU/sec (billion lattice updates per second) performance is obtained on Nvidia RTX3090, and the maximal performance of almost 300 GLU/sec is obtained on an 8 GPU workstation. Note that the main data storage is in CPU RAM memory, but the host-device data exchange is concealed by temporal blocking: with appropriate the data transfers are concealed by the computing operations and do not affect the performance.
We demonstrate an efficient approach to numerical modeling of optical properties of large-scale structures with typical dimensions much greater than the wavelength of light. For this purpose, we use the finite-difference time-domain (FDTD) method enhanced with a memory efficient Locally Recursive non-Locally Asynchronous (LRnLA) algorithm called DiamondTorre and implemented for General Purpose Graphical Processing Units (GPGPU) architecture. We apply our approach to simulation of optical properties of organic light emitting diodes (OLEDs), which is an essential step in the process of designing OLEDs with improved efficiency. Specifically, we consider a problem of excitation and propagation of surface plasmon polaritons (SPPs) in a typical OLED, which is a challenging task given that SPP decay length can be about two orders of magnitude greater than the wavelength of excitation. We show that with our approach it is possible to extend the simulated volume size sufficiently so that SPP decay dynamics is accounted for. We further consider an OLED with periodically corrugated metallic cathode and show how the SPP decay length can be greatly reduced due to scattering off the corrugation. Ultimately, we compare the performance of our algorithm to the conventional FDTD and demonstrate that our approach can efficiently be used for large-scale FDTD simulations with the use of only a single GPGPU-powered workstation, which is not practically feasible with the conventional FDTD.
We present the results of 3D modeling of the laser and electron beam powder bed fusion process at the mesoscale with an in-house developed advanced multiphysical numerical tool. The hydrodynamics and thermal conductivity core of the tool is based on the lattice Boltzmann method. The numerical tool takes into account the random distributions of powder particles by size in a layer and the propagation of the laser (electron beam) with a full ray tracing (Monte Carlo) model that includes multiple reflections, phase transitions, thermal conductivity, and detailed liquid dynamics of the molten metal, influenced by evaporation of the metal and the recoil pressure. The model has been validated by a number of physical tests. We numerically demonstrate a strong dependence of the net energy absorption of the incoming heat source beam by the powder bed and melt pool on the beam power. We show the ability of our model to predict the measurable properties of a single track on a bare substrate as well as on a powder layer. We obtain good agreement with experimental data for the depth, width and shape of a track for a number of materials and a wide range of energy source parameters. We further apply our model to the simulation of the entire layer formation and demonstrate the strong dependence of the resulting layer morphology on the hatch spacing. The presented model could be very helpful for optimizing the additive process without carrying out a large number of experiments in a common trial-and-error method, developing process parameters for new materials, and assessing novel modalities of powder bed fusion additive manufacturing.
 scale(.21174)'%3e%3cpath d='m52.487 177.086 13.018 11.852 11.762-11.762 11.896 12.266 12.846-11.358 11.827 12.43 13.339-11.013 10.897 12.728 13.27-10.145 13.947 12.319'/%3e%3ccircle cx='89.425' cy='154.584' r='5.167'/%3e%3ccircle cx='111.237' cy='156.047' r='5.167'/%3e%3ccircle cx='30.05' cy='16.5' r='4'/%3e%3c/g%3e%3c/svg%3e)