Adaptive Remeshing Research Articles

Adaptively remeshed multiphysical modeling of resistance forge welding with experimental validation of residual stress fields and measurement processes

Welding processes used in the production of pressure vessels impart residual stresses in the manufactured component. Computational modeling is critical to predicting these residual stress fields and understanding how they interact with notches and flaws to impact pressure vessel durability. In this work, we present a finite element model for a resistance forge weld and validate it using laboratory measurements. Extensive microstructural changes, near-melt temperatures, and large localized deformations along the weld interface pose significant challenges to Lagrangian finite element modeling. The proposed modeling approach overcomes these roadblocks in order to provide a high-fidelity simulation that can predict the residual stress state in the manufactured pressure vessel; a rich microstructural constitutive model accounts for material recrystallization dynamics, a frictional-to-tied contact model is coordinated with the constitutive model to represent interfacial bonding, and adaptive remeshing is employed to alleviate severe mesh distortion. An interrupted-weld approach is applied to the simulation to facilitate comparison to displacement measures. Several approaches are employed for residual stress measurement in order to validate the finite element model: neutron diffraction, the contour method, and the slitting method. Model-measurement comparisons are supplemented with detailed simulations that reflect the configurations of the residual-stress measurement processes themselves. The model results show general agreement with experimental measurements, and we observe some similarities in the features around the weld region. Factors that contribute to model-measurement differences are identified. Finally, we conclude with some discussion of the model development and residual stress measurement strategies, including how to best leverage the efforts put forth here for other types of related weld problems.

An Investigation of Tensile, Fatigue, and Fracture Behavior of 3D-Printed Polymers

Abstract The progression of manufacturing technology has significantly benefited from the adoption of 3D printing techniques, which enable the production of parts with intricate geometries. However, it is important to acknowledge that components fabricated through this additive manufacturing method frequently manifest defects and are prone to failure under severe conditions. Therefore, a thorough examination of the mechanical properties of these parts is essential to effectively reduce the failure. This study aimed to explore the mechanical properties of two prevalently used 3D-printed polymers, specifically Onyx and acrylonitrile butadiene styrene (ABS), by integrating computational and experimental analyses. The experimental study utilized a material testing system and digital image correlation (DIC) technology, while the computational analysis covered the finite element (FE) modeling of the 3D-printed samples. The research focused on evaluating the tensile strength and fatigue resistance of the specimens printed in various orientations, alongside a detailed investigation of their fracture behavior. The crack propagation analysis was carried out using the DIC system and the separating morphing and adaptive re-meshing technology (SMART) scheme in ansys. It was observed that upright build orientation produced the weakest samples for axial loading and specimens with notches failed earlier than those without. Moreover, Onyx was found to have a higher resistance to fracture or failure compared to ABS. The FE modeling results demonstrated strong agreement with the experimental results, validating their accuracy and reliability in characterizing the critical mechanical response of 3D-printed parts rapidly and cost effectively.

Numerical investigation of flash formation and asymmetrical cooling in resistance projection welding

In resistance projection welding, the softened material is squeezed out from the joint area. The two workpieces being joined together often have different geometries, leading to asymmetrical cooling. The flash formation and asymmetrical cooling are not well quantified in the literature. This study developed a coupled electrical-thermal-mechanical model with adaptive remeshing and glued contact activation techniques for projection welding of steel fitting to a steel shell. The model was used to calculate the evolution of temperature distribution in the fitting and shell for welds made under two different hold times. The cooling rates were then inputted into a literature model to predict the joint hardness values, which correlated well with the experimental data. Additionally, the main factors for the accurate prediction of flash shape and size were the use of temperature-activated glued contact and high-temperature flow stress. Specifically, a combination of no glued contact and low flow stress can lead to an overprediction of the flash geometry.

A FEM/DEM adaptive remeshing strategy for brittle elastic failure initiation and propagation

AbstractThis article presents an adaptive remeshing strategy between the finite element method (FEM) and the discrete element method (DEM). To achieve this strategy, an edge‐to‐edge coupling method based on Lagrange multipliers has been set‐up to ensure the continuity of velocities at the interface. To switch from a computation initially purely FEM to a FEM‐DEM one, a field transfer method was required. In particular, a displacement field transfer method has been set‐up. The switching from a FEM subdomain to a DEM one is activated by a transition criterion. Each time a FEM subdomain is substituted by a DEM one, the DEM subdomain microscopic properties are set‐up with respect to the subdomain geometry and desired particle refinement. This is performed thanks to the linking to the so‐called “Cooker,” a tool distributed along with the GranOO Workbench. Two subdomain remeshing cases were dealt with: that of an initially FEM subdomain that is converted to DEM, and that of DEM subdomains which coalesce. A numerical test case shows that the dynamic remeshing method behaves as expected: FEM subdomains are substituted by DEM ones when the transition criterion is met, and DEM subdomains coalesce when required. The final numerical test case shows a good agreement with a crack propagation experiment of the literature, while a speedup of about 20 was observed when compared to pure DEM computation.

Effect of a subsurface void on the micromechanics of ductile metal indentation using remeshing

Near-surface voids and pores are generated in metal processing operations as diverse as additive manufacturing and powder processing, but their effect on indentation hardness has not been explored outside homogenized frameworks. Here we model the micromechanics of near-surface void deformation up to closure in a wedge-indentation field in ductile metal, and reveal the considerable softening effect of voids on the indentation hardness. Both symmetrically and eccentrically located voids, at various depths below the free surface, are studied. Notably, the extent of apparent reduction in the elastic modulus due to a void is much smaller than its effect on apparent hardness, e.g. 6.5% against 55% for the same void. Critical to the simulations is an adaptive remeshing finite element (FE) framework that allows accurate capture of processes like void closure and void-wall self-contact. The simulations reveal the subsurface plastic strain, strain-rate, and velocity fields with high fidelity, and their radical differences from the radial indentation field in a void-free specimen. These differences include the presence of localized pockets of high and low strain, and the initial accommodation of material displaced by the indenter by a corresponding reduction in void area. Unlike voids under uniform compression, the void-area evolution in indentation shows a characteristic sigmoidal pattern of reduction with indentation depth for all but the smallest voids. Interestingly, the indented surface profile can exhibit a one-sided pile-up feature which is diagnostic of the presence of an eccentrically located sub-surface void. Our remeshing scheme is versatile, as exemplified by its ability to model the extreme deformation associated with two nearly closed-spaced voids under an indenter. Our work shows that void-closure simulations in related applications like forming could benefit from the adoption of a remeshing framework.

Investigating the Influence of Holes as Crack Arrestors in Simulating Crack Growth Behavior Using Finite Element Method

The primary focus of this paper is to investigate the application of ANSYS Workbench 19.2 software’s advanced feature, known as Separating Morphing and Adaptive Remeshing Technology (SMART), in simulating the growth of cracks within structures that incorporate holes. Holes are strategically utilized as crack arrestors in engineering structures to prevent catastrophic failures. This technique redistributes stress concentrations and alters crack propagation paths, enhancing structural integrity and preventing crack propagation. This paper explores the concept of using holes as crack arrestors, highlighting their significance in increasing structural resilience and mitigating the risks associated with crack propagation. The crack growth path is estimated by applying the maximum circumferential stress criterion, while the calculation of the associated stress intensity factors is performed by applying the interaction integral technique. To analyze the impact of holes on the crack growth path and evaluate their effectiveness as crack arrestors, additional specimens with identical external dimensions but without any internal holes were tested. This comparison was conducted to provide a basis for assessing the role of holes in altering crack propagation behavior and their potential as effective crack arrestors. The results of this study demonstrated that the presence of a hole had a significant influence on the crack growth behavior. The crack was observed to be attracted towards the hole, leading to a deviation in its trajectory either towards the hole or deflecting around it. Conversely, in the absence of a hole, the crack propagated without any alteration in its path. To validate these findings, the computed crack growth paths and associated stress intensity factors were compared with experimental and numerical data available in the open literature. The remarkable consistency between the computational study results for crack growth path, stress intensity factors, and von Mises stress distribution, and the corresponding experimental and numerical data, is a testament to the accuracy and reliability of the computational simulations.

Simulation of fiber-induced melt pressure fluctuations within large scale polymer composite deposition beads

The process-structure-property relationship in Large Area Additive Manufacturing (LAAM) technology is an ongoing area of research as the inherent microstructural properties (chiefly fibers and voids) affect the performance of printed parts. Unfortunately, we currently lack adequate understanding of micro void nucleation and evolution during the LAAM and fused deposition modelling (FDM) processes. Modeling of the polymer melt flow during the extrusion process is important in understanding the underlying microstructural formation and associated properties of the print, that determines the part performance in service conditions. In this paper we compute fiber-induced local pressure fluctuations which may promote void formation in the bead’s microstructure. On a macro-scale, we determine flow fields of a purely viscous, Newtonian planar polymer deposition flow through a LAAM nozzle which are utilized on a micro-scale model where we simulate the evolution of a single ellipsoidal fiber along streamlines obtained from the macro-model. On the micro-scale, we determine instantaneous values of the translational and rotational velocities of the rigid ellipsoidal fiber that satisfies a balance of hydrodynamic forces and couples on the fiber’s surface based on a Newton Raphson algorithm and we track the fiber’s motion along the flow path via an explicit numerical iterative algorithm. Model verification is achieved by benchmarking results with solutions from well-known Jeffery’s equation of motion of a particle in homogeneous simple shear flow. We account for rotary diffusivity due to short-range fiber-fiber interaction in the FEA simulation by determining an effective fluid domain size representative of the interaction coefficient of the melt flow through a correlation analysis that yields an equivalent steady state orientation based on the Advani-Tucker equation. We also consider different possible motions of the fiber along individual LAAM flow paths from a given set of random initial fiber conditions to determine pressure bounds on the fiber surface along each streamline. For improved computational efficiency, calculations are carried out with respect to the fiber’s local coordinate axes to overcome the rigor of adaptive remeshing during the quasi-transient analysis. Results show low pressure extremes near the fiber’s surface which varies across the printed bead as well as through its thickness. Discussion is provided to gain insight into the effect of low-pressure extremes on micro void formation, particularly at the nozzle exit and during die swell/expansion.

Automated Surface Extraction: Adaptive Remeshing Meets Lagrangian Shrink-Wrapping

Automated Surface Extraction: Adaptive Remeshing Meets Lagrangian Shrink-Wrapping

Development of a GPGPU-parallelized FDEM based on extrinsic cohesive zone model with master-slave algorithm

This paper proposes a novel general-purpose graphic-processing-units (GPGPU) parallel computing approach to an extrinsic cohesive zone model (ECZM) - based combined finite-discrete element method (FDEM) for simulating rock fracturing. The proposed GPGPU-parallelized ECZM-FDEM incorporates a master–slave algorithm as an alternative to the complex adaptive remeshing process, which is usually used in ECZM but has prevented it from being parallelized using GPGPU. Numerical experiments of the Brazilian test and uniaxial compression test of rocks are conducted to compare the proposed ECZM-FDEM with a GPGPU-parallelized FDEM using the intrinsic cohesive zone model (ICZM-FDEM). Results show that the proposed method can not only overcome the accuracy degradation of calculated stresses and deformations that is inevitable in ICZM-FDEM but also reasonably simulate rock fracturing. Moreover, the proposed GPGPU-parallelized ECZM-FDEM achieves a maximum relative speed-up of 13 times over GPGPU-parallelized ICZM-FDEM due to efficient contact calculations and larger stable time steps. Thus, the proposed ECZM-FDEM is more physically sound and more computationally efficient compared with ICZM-FDEM, which may contribute to the further developments of FDEM.

On adaptive mesh coarsening procedures for the virtual element method for two-dimensional elastic problems

In the context of adaptive remeshing, the virtual element method provides significant advantages over the finite element method. The attractive features of the virtual element method, such as the permission of arbitrary element geometries, and the seamless permission of ‘hanging’ nodes, have inspired many works concerning error estimation and adaptivity. However, these works have primarily focused on adaptive refinement techniques with little attention paid to adaptive coarsening (i.e. de-refinement) techniques that are required for the development of fully adaptive remeshing procedures. In this work novel indicators are proposed for the identification of patches/clusters of elements to be coarsened, along with a novel procedure to perform the coarsening. The indicators are computed over prospective patches of elements rather than on individual elements to identify the most suitable combinations of elements to coarsen. The coarsening procedure is robust and suitable for meshes of structured and unstructured/Voronoi elements. Numerical results demonstrate the high degree of efficacy of the proposed coarsening procedures and sensible mesh evolution during the coarsening process. It is demonstrated that critical mesh geometries, such as non-convex corners and holes, are preserved during coarsening, and that meshes remain fine in regions of interest to engineers, such as near singularities.

Crack Growth Simulations With SMART Method in Adhesively Bonded Joints

Due to the important advantages of adhesive joints, such as their suitability for multi-material designs, their use has been increasing in the last decade. Determining the fracture behavior of structural adhesive bondings is essential for structural durability. In crack propagation analyses, adaptive meshing has drawn considerable attention because of its improvements in terms of complex preprocessing and time management. This paper presents a recently introduced separating morphing and adaptive remeshing technology (SMART) innovative crack growth simulation for adhesively bonded joints, considering static and cyclic cases. For the static case, an R-curve was obtained for the bonding joints of carbon steel and Araldite 2015. For the cyclic case, the Carbon Fiber Reinforced Polymer (CFRP) bonding joints were analyzed under constant-amplitude loading conditions. The crack-propagation rate and the number of cycles were estimated. Crack propagation simulations were validated using experimental data. Acceptable agreement was achieved between the experimental and estimated results.

NUMERICAL SIMULATION OF WAVE-INDUCED VEGETATION DYNAMICS USING A PARTITIONED COUPLING BETWEEN THE SPH METHOD AND AN FEA STRUCTURAL SOLVER

Sustainable coastal defense is promoted by wave-vegetation interaction that occurs in coastal zones and is able to dissipate wave energy and baffle currents. To study the complex fluid-structure interactions involved, numerical models present themselves as a suitable tool able to offer an in-depth view over the near field kinematics and dynamics. Taking into account the effect of vegetation dynamics is necessary, and often modelled through coupled fluid and structural solvers. This implies increased complexity, adaptive re-meshing tools, and numerically expensive simulations given the disparity between dimensions of the fluid domain and the very thin structural elements. The aim of this work is to present a novel numerical coupling tool based on the meshless Smoothed Particle Hydrodynamics (SPH) method, which can handle ultra-thin vegetation and resolve the vegetation kinematics without the use of any fitted parameters.

Development of a Numerical Approach for the CFD Simulation of a Gear Pump under Actual Operating Conditions

The geometric complexity and high-pressure gradients that characterize the design of the flow field of gear pumps make it very difficult to obtain an accurate CFD simulation of the component. Usually, assumptions are made both in terms of geometrical features and physics being included in the analysis. The contact between the teeth, which is a key factor for the correct functioning of these pumps, represents a critical challenge in 3D CFD simulations, mainly due to the intrinsic limits of the dynamic meshing techniques that can hardly effectively manage a zero or close to zero gap point forming during gear rotation. The geometric complexity and high-pressure gradients that characterize the gear pump flow field make a CFD analysis quite difficult, and the contact between the gear teeth is usually avoided, thus being an extremely important feature. In this paper, a gear pump composed of inlet and outlet pipes was considered, and the contact between the gear’s teeth was modeled in two different ways, one where it is effectively implemented and one where it is avoided using distancing and a proper casing modification. Herein, a new methodology is proposed for the application of the dynamic mesh method in the Simcenter STAR-CCM+ environment using an adaptive remeshing technique. The proposed methodology is compared with the alternative overset meshing method available in the software. The new meshing method is implemented using a user-routing that reproduces the real geometry of the gears while rotating during the pump operation, with teeth contact included. The routine is optimized in order to limit the additional computation and time needed for the remeshing process. The results that can be obtained using the two meshing approaches for the gear pump are compared in terms of computational effort and the accuracy of the results. The two methods showed opposite results in almost all the reported results, with the overset being more precise in the radial pressure evaluation and the dynamic being more reliable in the cavitation/aeration extension cloud.

Fatigue Crack Growth Studies under Mixed-Mode Loading in AISI 316 Stainless Steel

The objective of this study is to examine the behavior of fatigue crack growth (FCG) in the mixed mode (I/II) of the AISI 316 austenitic stainless steel alloy, considering mode mixity angles of 30°, 45°, and 60°. This particular alloy is widely used in the marine industry and various structural components because of its exceptional properties, such as high corrosion resistance, good formability, weldability, and high-temperature strength. By investigating the crack growth behavior, the study seeks to provide insights into the material’s durability and potential for long-term use in demanding applications. To analyze fatigue crack growth behavior using linear elastic fracture mechanics (LEFM), this study utilizes compact tension shear (CTS) specimens with varying loading angles. The CTS specimens provide an accurate simulation of real-world loading conditions by allowing for the application of various loading configurations, resulting in mixed-mode loading. The ANSYS Mechanical APDL 19.2 software, which includes advanced features such as separating, morphing, and adaptive remeshing technologies (SMART), was utilized in this study to precisely model the path of crack propagation, evaluate the associated fatigue life, and determine stress intensity factors. Through comparison with experimental data, it was confirmed that the loading angle had a significant impact on both the fatigue crack growth paths and the fatigue life cycles. The stress-intensity factor predictions from numerical models were compared to analytical data. Interestingly, it was observed that the maximum shear stress and von Mises stresses occurred when the loading angle was 45 degrees, which is considered a pure shear loading condition. The comparison shows consistent results, indicating that the simulation accurately captures the behavior of the AISI 316 austenitic stainless steel alloy under mixed-mode loading conditions.

An adaptive semi-implicit finite element solver for brain cancer progression modeling.

Glioblastoma is the most aggressive and infiltrative glioma, classified as Grade IV, with the poorest survival rate among patients. Accurate and rigorously tested mechanistic in silico modeling offers great value to understand and quantify the progression of primary brain tumors. This paper presents a continuum-based finite element framework that is built on high performance computing, open-source libraries to simulate glioblastoma progression. We adopt the established proliferation invasion hypoxia necrosis angiogenesis model in our framework to realize scalable simulations of cancer, and has demonstrated to produce accurate and efficient solutions in both two- and three-dimensional brain models. The in silico solver can successfully implement arbitrary order discretization schemes and adaptive remeshing algorithms. A model sensitivity analysis is conducted to test the impact of vascular density, cancer cell invasiveness and aggressiveness, the phenotypic transition potential, including that of necrosis, and the effect of tumor-induced angiogenesis in the evolution of glioblastoma. Additionally, individualized simulations of brain cancer progression are carried out using pertinent magnetic resonance imaging data, where the in silico model is used to investigate the complex dynamics of the disease. We conclude by arguing how the proposed framework can deliver patient-specific simulations of cancer prognosis and how it could bridge clinical imaging with modeling.

Numerical Analysis on Fatigue Crack Growth at Negative and Positive Stress Ratios.

The finite element method was used to investigate the effect of the stress ratio on fatigue crack propagation behavior within the framework of the linear elastic fracture mechanics theory. The numerical analysis was carried out using ANSYS Mechanical R19.2 with the unstructured mesh method-based separating, morphing, and adaptive remeshing technologies (SMART). Mixed mode fatigue simulations were performed on a modified four-point bending specimen with a non-central hole. A diverse set of stress ratios (R = 0.1, 0.2, 0.3, 0.4, 0.5, -0.1, -0.2, -0.3, -0.4, -0.5), including positive and negative values, is employed to examine the influence of the load ratio on the behavior of the fatigue crack propagation, with particular emphasis on negative R loadings that involve compressive excursions. A consistent decrease in the value of the equivalent stress intensity factor (ΔKeq) is observed as the stress ratio increases. The observation was made that the stress ratio significantly affects both the fatigue life and the distribution of von Mises stress. The results demonstrated a significant correlation between von Mises stress, ΔKeq, and fatigue life cycles. With an increase in the stress ratio, there was a significant decrease in the von Mises stress, accompanied by a rapid increase in the number of fatigue life cycles. The results obtained in this study have been validated by previously published literature on crack growth experiments and numerical simulations.

A Quasi-Optimal Shape Design Method for Electromagnetic Scatterers Based on NURBS Surfaces and Filter-Enhanced GWO

This paper presents a quasi-optimal and efficient method for designing the shape of electromagnetic (EM) scatterers. Optimizing the EM scatterer shape to achieve satisfactory scattering properties is a crucial, yet challenging problem due to the complex geometry involved and the high non-linearity in the optimization. In this paper, the geometric complexity is to be handled by modeling EM scatterer shapes as non-uniform rational basis spline (NURBS) surfaces, which enables an easy and intuitive shape control. Using the surfaces’ control points as optimization variables, the shape optimization problem is to be solved with the heuristic optimization strategy, which can effectively handle non-linearity and the associated local optima. Specifically, the recently developed grey wolf optimizer (GWO) is employed. Although GWO can effectively avoid local optima and achieve quasi-optimal results, it requires repeated calls to time-consuming EM simulations. To solve this efficiency issue, a filter-enhanced version of GWO is proposed to reduce the times of calling EM simulations, and an efficient, adaptive remeshing method is proposed to accelerate the numerical analysis process within each EM simulation call. Altogether, they lead to a quasi-optimal and efficient method for optimizing EM scatterers. Its effectiveness has also been validated by several examples.

Large Growth Deformations of Thin Tissue using Solid-Shells.

Simulating large scale expansion of thin structures, such as in growing leaves, is challenging. Solid-shells have a number of potential advantages over conventional thin-shell methods, but have thus far only been investigated for small plastic deformation cases. In response, we present a new general-purpose FEM growth framework for handling a wide range of challenging growth scenarios using the solid-shell element. Solid-shells are a middle-ground between traditional volume and thin-shell elements where volumetric characteristics are retained while being treatable as a 2D manifold much like thin-shells. These elements are adaptable to accommodate the many techniques that are required for simulating large and intricate plastic deformations, including morphogen diffusion, plastic embedding, strain-aware adaptive remeshing, and collision handling. We demonstrate the capabilities of growing solid-shells in reproducing buckling, rippling, curling, and collision deformations, relevant towards animating growing leaves, flowers, and other thin structures. Solid-shells are compared side-by-side with thin-shells to examine their bending behavior and runtime performance. The experiments demonstrate that solid-shells are a viable alternative to thin-shells for simulating large and intricate growth deformations.

Evaluation of computational homogenization methods for the prediction of mechanical properties of additively manufactured metal parts

It is well known that the strongly location-dependent microstructures observed in metal parts made using additive manufacturing (AM) processes differ from those found in components produced via traditional manufacturing processes. This is primarily because the microstructures in AM parts strongly depend on spatially-varying cooling rates that are caused by several factors including part geometry and proximity to the build plate. Consequently, it is necessary to take a fresh look into methods used for calculating resulting mechanical properties, as the increased spatial resolution that must be applied to the description of properties of AM parts calls for vast improvements in computational efficiency. Computational homogenization methods are employed for the calculation of these mechanical properties based on the polycrystalline microstructures typically encountered in metallic materials. Such calculations that enable the description of properties as a function of position increase the accuracy of a vital link in the structure–process–property–performance continuum that falls within the Integrated Computational Materials Engineering (ICME) paradigm. In this work we describe and critically review six of these methods and their application to AM microstructures. We find that the Fast Fourier Transform (FFT) method and the Finite Element Method (FEM) are the optimal choices based on several factors including accuracy, efficiency, and applicability to polycrystalline microstructures. Provided the microstructure is amenable to the use of uniform grids and the assumption of periodic boundary conditions, the FFT method has greater computational efficiency. Where unstructured meshes, adaptive remeshing and/or non-periodic boundary conditions are desired, the FEM is the method of choice. Our recommendations are equally applicable to functionally graded AM parts where site-specific microstructures are engineered using multiple materials or different process parameters. They can also be utilized for other material systems and circumstances where increased accuracy is achieved by describing mechanical properties as a function of location.

Enabling Part-Scale Scanwise process simulation for predicting melt pool variation in LPBF by combining GPU-based Matrix-free FEM and adaptive Remeshing

• A highly efficient GPU-based matrix-free finite element modeling with adaptive remeshing framework is developed to overcome the high computational expense of l -PBF thermal process modeling, making it feasible to do detailed scanwise thermal simulation for part-scale within days. • The effects of the scanning strategy, geometrical features, and proximity to the turning points on the melt pool size variability and lack of fusion porosity are investigated for two different parts. • The predicted variation in melt pool size and lack of fusion porosity has good correlation with experimental measurements. This work proposes to combine matrix-free finite element modeling (FEM), adaptive remeshing, and graphical processing unit (GPU) computing to enable, for the first time, scanwise process simulation of the Laser Powder Bed Fusion (L-PBF) process with temperature-dependent thermophysical properties at the part scale. Compared to the conventional FEM using the global stiffness approach and a uniform mesh running on 10 CPU cores, l -PBF process simulation based on the proposed methodology running on a GPU card with 5,120 Compute Unified Device Architecture (CUDA) cores enables a speedup of over 10,000x. This significant speedup facilitates detailed thermal history and melt pool geometry predictions at high resolution for centimeter-scale parts within days of computation time. Two parts consisting of various geometric features are simulated to reveal the effects of scan strategy and local geometry on melt pool size variation, which correlate well with melt pool and lack-of-fusion porosity measurements obtained via experiment.