Explore the vast range of titles available.

Metal additive manufacturing (AM) presents advantages such as increased complexity for a lower part cost and part consolidation compared to traditional manufacturing. The multiscale, multiphase AM processes have been shown to produce parts with non-homogeneous microstructures, leading to variability in the mechanical properties based on complex process–structure–property (p-s-p) relationships. However, the wide range of processing parameters in additive machines presents a challenge in solely experimentally understanding these relationships and calls for the use of digital twins that allow to survey a larger set of parameters using physics-driven methods. Even though physics-driven methods advance the understanding of the p-s-p relationships, they still face challenges of high computing cost and the need for calibration of input parameters. Therefore, data-driven methods have emerged as a new paradigm in the exploration of the p-s-p relationships in metal AM. Data-driven methods are capable of predicting complex phenomena without the need for traditional calibration but also present drawbacks of lack of interpretability and complicated validation. This review article presents a collection of physics- and data-driven methods and examples of their application for understanding the linkages in the p-s-p relationships (in any of the links) in widely used metal AM techniques. The review also contains a discussion of the advantages and disadvantages of the use of each type of model, as well as a vision for the future role of both physics-driven and data-driven models in metal AM.

The overarching aim of this paper is to explore the use of machine learning (ML) to predict the microstructure-sensitive evolution of a three-dimensional (3D) crack surface in a polycrystalline alloy. A convolutional neural network (CNN)-based methodology is developed to establish spatial relationships between micromechanical/microstructural features in a cyclically loaded, uncracked microstructure and the 3D crack path, the latter quantified by the vertical deviation (i.e., z -offset) of the crack along a specified axis. The proposed methodology consists of (i) a feature selection and reduction scheme to identify a lower-dimensional representation of the experimentally measured microstructure and computed micromechanical fields, which allows for computational feasibility in predicting the z -offsets; (ii) a CNN model to compute the z -offset as a function of the local, lower-dimensional feature data; and (iii) a radial basis function smoothing spline to ensure spatial continuity between the independently predicted z -offsets. The proposed CNN-based methodology is shown to improve on the accuracies obtained using existing ML models such as XGBoost and to provide a definitive way of quantifying model uncertainty associated with CNN predictions. To further investigate the applicability of ML models, multiple prediction strategies with which to deploy ML algorithms are proposed and the relative performance of ML algorithms corresponding to each prediction strategy are analyzed. The presented work thus provides a framework to find an encoded representation of 3D microstructure and micromechanical data and develop methods to predict microstructure-sensitive crack evolution based on this encoded representation, while quantifying associated prediction uncertainties.

Porosity, a commonly occurring void defect in casting and additive manufacturing, is known to affect the mechanical response of metals, making it difficult or impossible to predict response variability. We introduce a new method of uniquely characterizing pore networks using a void descriptor function (VDF), which can be used to predict ductile-metal failure properties, namely, toughness modulus, ultimate strength, elongation, and fracture location. The VDF quantifies the inter-relationships of pores by accounting for pore location, size, and distance to free surface. Using a finite-element-modeling framework, 120 tensile specimens with statistically similar pore networks were simulated (virtually tested) to failure. The pore networks were characterized by the proposed VDF, which was then compared to the nominal location of fracture (defined as the fracture-initiation location corresponding to the dominant crack responsible for final rupture). The location of maximum VDF accurately predicted the fracture location (within ± 0.2 mm) for 91 (76%) of the 120 samples and proved to be a more reliable indicator than the location of maximum reduced cross-section area and the location of largest pore diameter for predicting fracture location. Furthermore, the maximum VDF value was found to be more highly correlated than fraction porosity, pore size, reduced-cross section area, and total number of pores to the ultimate tensile strength, elongation, and toughness modulus.

Conventional fracture mechanics asserts that the relevant physics governing small crack growth occurs near the crack front. However, for fatigue, computing these physics for each crack-growth increment over the entire microstructurally small crack regime is computationally intractable. Properly trained deep-learning surrogate models can massively accelerate fatigue crack-growth predictions by virtually propagating an initial crack using micromechanical fields corresponding to just the initially cracked microstructure. As the predicted crack front advances, however, the fields no longer reflect relevant near-crack-front physics, leading to error and uncertainty accumulation. To address this, we present an interleaved physics-based deep-learning (PBDL) framework, where updates to the crack representation in the physics-based model are triggered intermittently using model uncertainty, thereby updating micromechanical fields passed to the deep-learning model. We show that this framework, representing a novel cycle-jumping approach, effectively limits error accumulation in history-dependent fatigue crack evolution and forms a template for other time-series applications in materials.

Advances in additive manufacturing enable the production of tailored lattice structures and thus, in principle, coronary stents. This study investigates the effects of process-related irregularities, heat and surface treatment on the morphology, mechanical response, and expansion behavior of 316L stainless steel stents produced by laser powder bed fusion and provides a methodological approach for their numerical evaluation. A combined experimental and computational framework is used, based on both actual and computationally reconstructed laser powder bed fused stents. Process-related morphological deviations between the as-designed and actual laser powder bed fused stents were observed, resulting in a diameter increase by a factor of 2-2.6 for the stents without surface treatment and 1.3-2 for the electropolished stent compared to the as-designed stent. Thus, due to the increased geometrically induced stiffness, the laser powder bed fused stents in the as-built (7.11 ± 0.63 N) or the heat treated condition (5.87 ± 0.49 N) showed increased radial forces when compressed between two plates. After electropolishing, the heat treated stents exhibited radial forces (2.38 ± 0.23 N) comparable to conventional metallic stents. The laser powder bed fused stents were further affected by the size effect, resulting in a reduced yield strength by 41% in the as-built and by 59% in the heat treated condition compared to the bulk material obtained from tensile tests. The presented numerical approach was successful in predicting the macroscopic mechanical response of the stents under compression. During deformation, increased stiffness and local stress concentration were observed within the laser powder bed fused stents. Subsequent numerical expansion analysis of the derived stent models within a previously verified numerical model of stent expansion showed that electropolished and heat treated laser powder bed fused stents can exhibit comparable expansion behavior to conventional stents. The findings from this work motivate future experimental/numerical studies to quantify threshold values of critical geometric irregularities, which could be used to establish design guidelines for laser powder bed fused stents/lattice structures.

In their seminal work, Hey et al. identify four distinct paradigms that represent the historical evolution of science and technology: (1) empirical science, (2) theoretical-model science, (3) computational science (simulations), and (4) the emerging paradigm of data-driven science. A recent perspective from Agrawal and Choudhary points out that this sequence of paradigms appears throughout the materials science community specifically and corresponds to advancements over time in the ability to observe, interpret, and represent material behavior. In the first two paradigms, the representation of physical or mechanical behavior is typically of low dimension and is limited to those behaviors with relatively simple governing physics. Prime examples include elasticity and yield in structural metals, where the governing physics and the corresponding salient features of the material structure have been identified to a high level of confidence.

Systematic correlation analysis was performed between simulated micromechanical fields in an uncracked polycrystal and the known path of an eventual fatigue-crack surface based on experimental observation. Concurrent multiscale finite-element simulation of cyclic loading was performed using a high-fidelity representation of grain structure obtained from near-field high-energy x-ray diffraction microscopy measurements. An algorithm was developed to parameterize and systematically correlate the three-dimensional (3D) micromechanical fields from simulation with the 3D fatigue-failure surface from experiment. For comparison, correlation coefficients were also computed between the micromechanical fields and hypothetical, alternative surfaces. The correlation of the fields with hypothetical surfaces was found to be consistently weaker than that with the known crack surface, suggesting that the micromechanical fields of the cyclically loaded, uncracked microstructure might provide some degree of predictiveness for microstructurally small fatigue-crack paths, although the extent of such predictiveness remains to be tested. In general, gradients of the field variables exhibit stronger correlations with crack path than the field variables themselves. Results from the data-driven approach implemented here can be leveraged in future model development for prediction of fatigue-failure surfaces (for example, to facilitate univariate feature selection required by convolution-based models).

Three computational methods for modeling fracture are compared in the context of a class’ participation in the Third Sandia Fracture Challenge (SFC3). The SFC3 was issued to assess blind predictions of ductile fracture in a complex specimen geometry produced via additive manufacturing of stainless steel 316L powder. In this work, three finite-element-based methods are investigated: (1) adaptive remeshing, with or without material-state mapping; (2) element deletion; and (3) the extended finite element method. Each student team was tasked with learning about its respective method, calibrating model parameters, and performing blind prediction(s) of fracture/failure in the challenge-geometry specimen. Out of 21 teams who participated in the SFC3, three of the seven student teams from this class project ranked among the top five based on either global force-displacement or local strain predictions. Advantages and disadvantages of the three modeling approaches are identified in terms of mesh dependency, user-friendliness, and accuracy compared to experimental results. Recommendations regarding project management and organization are offered to facilitate future classroom participation in the Sandia Fracture Challenge or similar blind round-robin exercises.

The Sandia Fracture Challenges provide a forum for the mechanics community to assess its ability to predict ductile fracture through a blind, round-robin format where mechanicians are challenged to predict the deformation and failure of an arbitrary geometry given experimental calibration data. The Third Challenge (SFC3) required participants to predict fracture in an additively manufactured (AM) 316L stainless steel bar containing through holes and internal cavities that could not have been conventionally machined. The volunteer participants were provided extensive data including tension and notched tensions tests of 316L specimens built on the same build-plate as the Challenge geometry, micro-CT scans of the Challenge specimens and geometric measurements of the feature based on the scans, electron backscatter diffraction (EBSD) information on grain texture, and post-test fractography of the calibration specimens. Surprisingly, the global behavior of the SFC3 geometry specimens had modest variability despite being made of AM metal, with all of the SFC3 geometry specimens failing under the same failure mode. This is attributed to the large stress concentrations from the holes overwhelming the stochastic local influence of the AM voids and surface roughness. The teams were asked to predict a number of quantities of interest in the response based on global and local measures that were compared to experimental data, based partly on Digital Image Correlation (DIC) measurements of surface displacements and strains, including predictions of variability in the resulting fracture response, as the basis for assessment of the predictive capabilities of the modeling and simulation strategies. Twenty-one teams submitted predictions obtained from a variety of methods: the finite element method (FEM) or the mesh-free, peridynamic method; solvers with explicit time integration, implicit time integration, or quasi-statics; fracture methods including element deletion, peridynamics with bond damage, XFEM, damage (stiffness degradation), and adaptive remeshing. These predictions utilized many different material models: plasticity models including J2 plasticity or Hill yield with isotropic hardening, mixed Swift-Voce hardening, kinematic hardening, or custom hardening curves; fracture criteria including GTN model, Hosford-Coulomb, triaxiality-dependent strain, critical fracture energy, damage-based model, critical void volume fraction, and Johnson-Cook model; and damage evolution models including damage accumulation and evolution, crack band model, fracture energy, displacement value threshold, incremental stress triaxiality, Cocks-Ashby void growth, and void nucleation, growth, and coalescence. Teams used various combinations of calibration data from tensile specimens, the notched tensile specimens, and literature data. A detailed comparison of results based of these different methods is presented in this paper to suggest a set of best practices for modeling ductile fracture in situations like the SFC3 AM-material problem. All blind predictions identified the nominal crack path and initiation location correctly. The SFC3 participants generally fared better in their global predictions of deformation and failure than the participants in the previous Challenges, suggesting the relative maturity of the models used and adoption of best practices from previous Challenges. This paper provides detailed analyses of the results, including discussion of the utility of the provided data, challenges of the experimental-numerical comparison, defects in the AM material, and human factors.

Aim: Maternal corticosteroid administration in pregnancy is known to enhance fetal lung maturity in at risk fetuses. The aim of this study was to test the hypothesis that corticosteroid therapy alters fetal pulmonary blood flow in pregnancies at risk for preterm birth (PTB). Material and methods: We prospectively evaluated main fetal pulmonary artery (MPA) blood flow in pregnant women at risk for PTB and treated with corticosteroids (betamethasone), compared to an uncomplicated cohort without steroid therapy. The Doppler indices of interest included Peak Systolic Velocity (PSV), Resistive Index (RI), Pulsatility Index (PI), Systolic/Diastolic ratio (S/D ratio), Acceleration Time (AT), and Acceleration Time/Ejection Time Ratio (AT/ ET ratio), with the latter serving as the primary outcomes due to its stability irrespective of gestational age. Results: When compared with controls, fetuses treated with corticosteroids demonstrated significantly decreased pulmonary artery acceleration time (median: 28.89 (22.22-51.11) vs. 33.33 (22.20-57.00), p=0.006), while all other indices remained similar. We found no difference in pulmonary blood flow between fetuses who developed respiratory distress syndrome (RDS) and those that did not (31.56 ± 6.842 vs. 32.36 ± 7.265, p= 0.76). Conclusion: Our data demonstrate altered fetal pulmonary blood flow with corticosteroid therapy, possibly due to increased arterial elastance brought on by medication effect, which leads to the decreased acceleration time or possible gestational age affect. Contrary to a recent report, we did not observe any Doppler differences in fetuses with RDS, which underscores the need for further examination of this proposed association.