arXiv subjects

A. John Hart

Publications and source records attributed to A. John Hart.

At least 19 recordsLinked to original sources

arXiv2026

Agentic AI in Engineering and Manufacturing: Industry Perspectives on Utility, Adoption, Challenges, and Opportunities

This work examines how AI, especially agentic systems, is being adopted in engineering and manufacturing workflows, what value it provides today, and what is needed for broader deployment. This is an exploratory and qualitative state-of-practice study grounded in over 30 interviews across four stakeholder groups (large enterprises, small/medium firms, AI developers, and CAD/CAM/CAE vendors). We find that near-term AI gains cluster around structured, repetitive work and data-intensive synthesis, while higher-value agentic gains come from orchestrating multi-step workflows across tools. Adoption is constrained less by model capability than by fragmented and machine-unfriendly data, stringent security and regulatory requirements, and limited API-accessible legacy toolchains. Reliability, verification, and auditability are central requirements for adoption, driving human-in-the-loop frameworks and governance aligned with existing engineering reviews. Beyond technical barriers there are also organizational ones: a persistent AI literacy gap, cultural heterogeneity, and governance structures that have not yet caught up with agentic capabilities. Together, the findings point to a staged progression of AI utility from low-consequence assistance toward higher-order automation, as trust, infrastructure, and verification mature. This highlights key breakthroughs needed, including integration with traditional engineering tools and data types, robust verification frameworks, and improved spatial and physical reasoning.

cs.CY↗

arXiv2025

GenCAD-3D: CAD Program Generation using Multimodal Latent Space Alignment and Synthetic Dataset Balancing

CAD programs, structured as parametric sequences of commands that compile into precise 3D geometries, are fundamental to accurate and efficient engineering design processes. Generating these programs from nonparametric data such as point clouds and meshes remains a crucial yet challenging task, typically requiring extensive manual intervention. Current deep generative models aimed at automating CAD generation are significantly limited by imbalanced and insufficiently large datasets, particularly those lacking representation for complex CAD programs. To address this, we introduce GenCAD-3D, a multimodal generative framework utilizing contrastive learning for aligning latent embeddings between CAD and geometric encoders, combined with latent diffusion models for CAD sequence generation and retrieval. Additionally, we present SynthBal, a synthetic data augmentation strategy specifically designed to balance and expand datasets, notably enhancing representation of complex CAD geometries. Our experiments show that SynthBal significantly boosts reconstruction accuracy, reduces the generation of invalid CAD models, and markedly improves performance on high-complexity geometries, surpassing existing benchmarks. These advancements hold substantial implications for streamlining reverse engineering and enhancing automation in engineering design. We will publicly release our datasets and code, including a set of 51 3D-printed and laser-scanned parts on our project site.

cs.GR↗

arXiv2025

Spreading of highly cohesive metal powders with transverse oscillation kinematics

Powder bed additive manufacturing processes such as laser powder bed fusion (LPBF) or binder jetting (BJ) benefit from using fine (D50 $\leq20~μm$) powders. However, the increasing level of cohesion with decreasing particle size makes spreading a uniform and continuous layer challenging. As a result, LPBF typically employs a coarser size distribution, and rotating roller mechanisms are used in BJ machines, that can create wave-like surface profiles due to roller run-out. In this work, a transverse oscillation kinematic for powder spreading is proposed, explored computationally, and validated experimentally. Simulations are performed using an integrated discrete element-finite element (DEM-FEM) framework and predict that transverse oscillation of a non-rotating roller facilitates the spreading of dense powder layers (beyond 50% packing fraction) with a high level of robustness to kinematic parameters. The experimental study utilizes a custom-built mechanized powder spreading testbed and X-ray transmission imaging for the analysis of spread powder layers. Experimental results generally validate the computational results, however, also exhibit parasitic layer cracking. For transverse oscillation frequencies above 200 Hz, powder layers of high packing fraction (between 50-60%) were formed, and for increased layer thicknesses, highly uniform and continuous layers were deposited. Statistical analysis of the experimental powder layer morphology as a function of kinematic spreading parameters revealed that an increasing transverse surface velocity improves layer uniformity and reduces cracking defects. This suggests that with minor improvements to the machine design, the proposed transverse oscillation kinematic has the potential to result in thin and consistently uniform powder layers of highly cohesive powder.

physics.app-ph↗

arXiv2025

Radiometric Temperature Measurement for Metal Additive Manufacturing via Temperature Emissivity Separation

Emission of blackbody emission from the meltpool and surrounding area in laser powder bed fusion (LPBF) makes this process visible to a range of optical monitoring instruments intended for online process and quality control. Yet, these instruments have yet to prove capable of reliably detecting the finest flaws that influence LPBF component mechanical performance, limiting their adoption. One hindrance lies in interpreting measurements of radiance as temperature, despite the physical link between these variables being readily understood as a combination of Plank's Law and spectral emissivity. Uncertainty in spectral emissivity arises as it is nearly impossible to predict and can be a strong function of wavelength; in turn, this manifests uncertainty in estimated temperatures and thereby obscures the LPBF process dynamics that indicate component defects. This paper presents temperature emissivity separation (TES) as a method for accurate retrieval of optically-measured temperatures in LPBF. TES simultaneously calculates both temperature and spectral emissivity from spectrally-resolved radiance measurements and, as the latter term is effectively measured, more accurate process temperatures result. Using a bespoke imaging spectrometer integrated with an LPBF testbed to evaluate this approach, three basic TES algorithms are compared in a validation experiment that demonstrates retrieval of temperatures accurate to $\pm 28$ K over a $1000$ K range. A second investigation proves industrial feasibility through fabrication of an LPBF test artifact. Temperature data are used to study the evolution of fusion process boundary conditions, including a decrease in cooling rate as layerwise printing proceeds. A provisional correlation of temperature fields to component porosity assessed by 3D computed tomography demonstrates in situ optical detection of micron-scale porous defects in LPBF.

physics.optics↗

arXiv2025

Additively manufacturable high-strength aluminum alloys with thermally stable microstructures enabled by hybrid machine learning-based design

Additively manufactured (AM) aluminum alloys with high strength and thermal stability have broad applications in turbine engines, vacuum pumps, heat exchangers, and many other industrial systems. Employing precipitates with an L1$_2$ structure to block dislocation motions is a widespread strategy to strengthen aluminum. However, to achieve high strength, a high volume fraction of small precipitates is required, and these characteristics are generally mutually exclusive. Here, we show that for certain compositions of Al alloys, L1$_2$ phases initially precipitate as sub-micron metastable ternary phases under the rapid solidification conditions of powder bed AM, yet the subsequent L1$_2$ phases that precipitate during heat treatment of the sample remain at the nanoscale, imparting high strength. For strength to be retained at elevated temperature, these nanoprecipitates must have low coarsening rates. To inversely design the composition of an alloy to have these target microstructural features, we used hybrid calculation of phase diagram (CALPHAD)-based integrated computational materials engineering (ICME) and Bayesian optimization techniques. We tested our approach by designing an Al-Er-Zr-Y-Yb-Ni model alloy, and the selected composition was manufactured in powder form as AM feedstock. The strength of specimens manufactured via laser powder bed fusion (LPBF) from the designed composition is comparable to that of wrought Al 7075, yet without cracking that occurs upon LPBF of Al 7075. After high-temperature (400$^\circ$C) aging the designed alloy is 50% stronger than the strongest known benchmark printable Al alloy.

cond-mat.mtrl-sci↗

arXiv2024

High-Fidelity Optical Monitoring of Laser Powder Bed Fusion via Aperture Division Multiplexing

Qualification of high-performance metal components produced by laser powder bed fusion (LPBF) must identify process-induced porous defects that reduce ductility and nucleate fatigue cracking. Detecting such defects via optical monitoring of LPBF provides a path towards in-process quality control without downstream testing such as by computed tomography. However, integration of in-process sensing with LPBF is hampered by geometric and optical complications and, as a result, it has yet to be proven that the finest pores that limit component fatigue life can be resolved via in situ data. We present aperture division multiplexing (ADM) as a method for simultaneously focusing the process laser and providing unobstructed optical access for high-fidelity process monitoring using a common optic. Construction of an ADM optic of achieving imaging at 50 micron spatial resolution in the mid-wave infrared is described, and this optic is demonstrated on a production-representative LPBF testbed. High-speed infrared video data are correlated to micro-CT measurement of pores as fine as 4.3 microns, through multiple process signatures, establishing the promise of ADM for qualification of LPBF component fatigue performance.

physics.optics↗

arXiv2024

Automated Segmentation and Analysis of Microscopy Images of Laser Powder Bed Fusion Melt Tracks

With the increasing adoption of metal additive manufacturing (AM), researchers and practitioners are turning to data-driven approaches to optimise printing conditions. Cross-sectional images of melt tracks provide valuable information for tuning process parameters, developing parameter scaling data, and identifying defects. Here we present an image segmentation neural network that automatically identifies and measures melt track dimensions from a cross-section image. We use a U-Net architecture to train on a data set of 62 pre-labelled images obtained from different labs, machines, and materials coupled with image augmentation. When neural network hyperparameters such as batch size and learning rate are properly tuned, the learned model shows an accuracy for classification of over 99% and an F1 score over 90%. The neural network exhibits robustness when tested on images captured by various users, printed on different machines, and acquired using different microscopes. A post-processing module extracts the height and width of the melt pool, and the wetting angles. We discuss opportunities to improve model performance and avenues for transfer learning, such as extension to other AM processes such as directed energy deposition.

cs.CV↗

arXiv2024

A rapid experimental workflow for studying melt track scaling in laser powder bed fusion using high-precision metal template substrates

Development and qualification of process parameters in laser powder bed fusion (LPBF) commonly involves many variables. At the outset of development, whether transferring known parameters to a new machine, or exploring a new material, single-track and single-layer experiments are a convenient means of down-selecting key variables and exploring parameter scaling behavior. We present an experimental workflow for single-layer LPBF experiments using etched high-precision metal template substrates, overcoming challenges with precision single-layer alignment in LPBF systems and enabling efficient processing and cross-sectional analysis. Templates are fabricated using chemical etching and machining, and are characterized using optical profilometry and X-ray transmission imaging of powder layers. Using the templates a single-track parameter study is performed in SS316 including three powder layer thicknesses, and spanning common laser melting modes (lack-of-fusion, conduction, and keyhole mode). Analysis of melt track geometries using automated image processing allows a scaling law to be applied to define the process window, quantifying the amount of material added with increasing powder layer thickness. Single-track results are verified with raster scanning experiments, showing the potential to transfer single-track results to full LPBF builds.

physics.app-ph↗

arXiv2024

Exploration of improved, roller-based spreading strategies for cohesive powders in additive manufacturing via coupled DEM-FEM simulations

Spreading of fine (D50 <=20um) powders into thin layers typically requires a mechanism such as a roller to overcome the cohesive forces between particles. Roller-based spreading requires careful optimization and can result in low density and/or inconsistent layers depending on the characteristics of the powder feedstock. Here, we explore improved, roller-based spreading strategies for highly cohesive powders using an integrated discrete element-finite element (DEM-FEM) framework. Powder characteristics are emulated using a self-similarity approach based on experimental calibration for a Ti-6Al-4V 0-20um powder. We find that optimal roller-based spreading relies on a combination of surface friction of the roller and roller kinematics that impart sufficient kinetic energy to break cohesive bonds between powder particles. However, excess rotation can impart excessive kinetic energy, causing ejection of particles and a non-uniform layer. Interestingly, the identified optimal surface velocities for counter-rotation as well as rotational oscillation are very similar, suggesting this quantity as the critical kinematic parameter. When these conditions are chosen appropriately, layers with packing fractions beyond 50% are predicted for layer thicknesses as small as ~2 times D90 of the exemplary powder, and the layer quality is robust with respect to substrate adhesion over a 10-fold range. The latter is an important consideration given the spatially varying substrate conditions in AM due to the combination of fused/bound and bare powder regions. As compared to counter-rotation, the proposed rotational oscillation is particularly attractive because it can overcome practical issues with mechanical runout of roller mechanisms. In particular, the application to rubber-coated rollers, which promises to reduce the risk of tool damage and particle streaking, is recommended for future investigation.

cs.CE↗

arXiv2022

Quantitative analysis of thin metal powder layers via transmission X-ray imaging and discrete element simulation: Roller-based spreading approaches

A variety of tools can be used for spreading metal, ceramic, and polymer feedstocks in powder bed additive manufacturing methods. Rollers are often employed when spreading powders with limited flowability, as arises in powders comprising fine particle sizes or high surface energy materials. Here, we study roller-based powder spreading for powder bed AM using the unique combination of a purpose-built powder spreading testbed with a proven method for X-ray mapping of powder layer depth. We focus on the density and uniformity of nominally 100 micrometer thick layers of roller-spread Ti-6Al-4V and Al-10Si-Mg powders. Our results indicate that when rotation is too rapid, roller-applied shear force impedes the creation of dense and uniform layers from powders of high innate flowability, or where inertial forces driven by particle density dominate cohesive forces. Roller counter-rotation augments the uniformity of cohesive powder layers, primarily though reducing the influence of particle clusters in the flowing powder, which are otherwise shown to cause deep, trench-like streaks. Companion discrete element method (DEM) simulations further contextualize the experiments through isolation of the effects of cohesion on layer attributes. Results suggest that roller motion parameters could apply a strategic level of additional shear force to the flowing powder, thereby mitigating the clumping behavior characteristic of highly cohesive feedstocks while maintaining high layer uniformity.

physics.app-ph↗

arXiv2022

Quantitative analysis of thin metal powder layers via transmission X-ray imaging and discrete element simulation: Blade-based spreading approaches

Spreading uniform and dense layers is of paramount importance to creating high-quality components using powder bed additive manufacturing (AM). Blade-like tools are often employed for spreading powder metal feedstocks, especially in laser powder bed fusion and electron beam melting, where powders are characterized by a D50 of 30 microns or greater. Along with variations in boundary conditions introduced by the layer-wise geometry and surface topography of the printed component, stochastic interactions between the spreading tool and powder result in spatial variations of layer quality that are still not well understood. Here, to study powder spreading under conditions representative of powder bed AM, we employ a modular, mechanized apparatus to create powder layers from moderately and highly cohesive powders with a selection of blade-like spreading tools. Powder layer effective depth is spatially mapped using transmission X-ray imaging, and uniformity is quantified via a statistical approach. We first compare layer density, or the effective depth of powder layer, and show that blade geometries with a curved profile lead to increased material deposition. Second, this approach enables quantification of local fluctuations, or layer defect severity. For example, we observe that the primary benefit of a V-shaped rubber blade, as compared to a 45 degree rigid blade, lies in enabling local deflection of the blade edge to eliminate streaking from large particles, while also increasing deposition. Additionally, we employ a custom DEM simulation to elucidate the opposing roles of particle density and surface energy with a pseudo-material approach, where the balance of inertial and cohesive forces determine macro-scale powder flowability. For specific alloy densities, we find a critical surface energy beyond which layer density is greatly impaired when powder spreading is performed using a blade.

physics.app-ph↗

arXiv2021

A spiral laser scanning routine for powder bed fusion inspired by natural predator-prey behavior

Additive manufacturing by laser powder bed fusion (LPBF) offers material versatility, capability for complex geometries, and control of mechanical properties and microstructure. However, achieving high-quality output requires process parameters that consider both local and global thermal gradients. Here, we propose a new scan pattern for mitigating part quality issues caused by non-uniform heating and cooling. A nature-inspired design method is employed to derive a spiral pattern stemming from the predator-prey behavior, and a power optimization routine is applied to maintain constant melt pool depth. Comparing simulated thermal histories for the spiral pattern to well-established zig-zag and helix scan patterns, we propose that the spiral pattern significantly reduces the spatial variation of temperature across the scan area, while a larger area remains above a specified threshold temperature at the end of the scan. Consequently, the spiral pattern is promising for LPBF of crack-prone materials, and for future optimization of LPBF overall.

cond-mat.mtrl-sci↗

arXiv2021

Printable, castable, nanocrystalline cellulose-epoxy composites exhibiting hierarchical nacre-like toughening

Due to their exceptional mechanical and chemical properties and their natural abundance, cellulose nanocrystals (CNCs) are promising building blocks of sustainable polymer composites. However, the rapid gelation of CNC dispersions has generally limited CNC-based composites to low CNC fractions, in which polymer remains the dominant phase. Here we report on the formulation and processing of crosslinked CNC-epoxy composites with a CNC fraction exceeding 50 wt.%. The microstructure comprises sub-micrometer aggregates of CNCs crosslinked to polymer, which are analogous to the lamellar structure of nacre and promotes toughening mechanisms associated with bulk ductile behavior, despite the brittle behavior of the aggregates at the nanoscale. At 63 wt.% CNCs, the composites exhibit a hardness of 0.66 GPa and a fracture toughness of 5.2 MPa.m$^{1/2}$. The hardness of this all-organic material is comparable to aluminum alloys, and the fracture toughness at the centimeter scale is comparable to that of wood cell wall. We show that CNC-epoxy composite objects can be shaped from the gel precursors by direct-write printing and by casting, while the cured composites can be machined into complex 3D shapes. The formulation, processing route, and the insights on toughening mechanisms gained from our multiscale approach can be applied broadly to highly loaded nanocomposites.

physics.app-ph↗

arXiv2021

Physics-Based Modeling and Predictive Simulation of Powder Bed Fusion Additive Manufacturing Across Length Scales

Powder bed fusion additive manufacturing (PBFAM) of metals has the potential to enable new paradigms of product design, manufacturing and supply chains while accelerating the realization of new technologies in the medical, aerospace, and other industries. Currently, wider adoption of PBFAM is held back by difficulty in part qualification, high production costs and low production rates, as extensive process tuning, post-processing, and inspection are required before a final part can be produced and deployed. Physics-based modeling and predictive simulation of PBFAM offers the potential to advance fundamental understanding of physical mechanisms that initiate process instabilities and cause defects. In turn, these insights can help link process and feedstock parameters with resulting part and material properties, thereby predicting optimal processing conditions and inspiring the development of improved processing hardware, strategies and materials. This work presents recent developments of our research team in the modeling of metal PBFAM processes spanning length scales, namely mesoscale powder modeling, mesoscale melt pool modeling, macroscale thermo-solid-mechanical modeling and microstructure modeling. Ongoing work in experimental validation of these models is also summarized. In conclusion, we discuss the interplay of these individual submodels within an integrated overall modeling approach, along with future research directions.

cs.CE↗

arXiv2021

A Testbed for Investigation of Selective Laser Melting at Elevated Atmospheric Pressure

Metal additive manufacturing (AM) by laser powder bed fusion (L-PBF) builds upon fundamentals established in the field of laser welding which include the influence of gas and plume dynamics on weld depth and quality. L-PBF demands a thorough investigation of the complex thermophysical phenomena that occur where the laser interacts with the metal powder bed. In particular, melt pool turbulence and evaporation are influenced by the ambient gas chemistry and pressure. This paper presents the design and validation of high pressure laser melting (HPLM) testbed; this accommodates bare metal plate samples as well as manually-coated single powder layers, and operates at up to 300 psig. The open architecture of this testbed allows for full control of all relevant laser parameters in addition to ambient gas pressure and gas flow over the build area. Representative melt tracks and rasters on bare plate and powder are examined in order to validate system performance, and preliminary analysis concludes that pressure has a significant impact on melt pool aspect ratio. The HPLM system thus enables careful study pressure effects on processing of common L-PBF materials, and can be applied in the future to materials that are challenging to process under ambient pressure, such as those with high vapor pressures.

physics.app-ph↗

arXiv2021

Substrate-Versatile Direct-Write Printing of Carbon Nanotube-Based Flexible Conductors, Circuits, and Sensors

Printed electronics rely on the deposition of conductive liquid inks, typically onto polymeric or paper substrates. Among available conductive fillers for use in electronic inks, carbon nanotubes (CNTs) have high conductivity, low density, processability at low temperatures, and intrinsic mechanical flexibility. However, the electrical conductivity of printed CNT structures has been limited by CNT quality and concentration, and by the need for nonconductive modifiers to make the ink stable and extrudable. This study introduces a polymer-free, printable aqueous CNT ink, and presents the relationships between printing resolution, ink rheology, and ink-substrate interactions. A model is constructed to predict printed feature sizes on impermeable substrates based on Wenzel wetting. Printed lines have conductivity up to 10,000 S/m. The lines are flexible, with < 5% change in DC resistance after 1,000 bending cycles, and <3% change in DC resistance with a bending radius down to 1 mm. Demonstrations focus on (i) conformality, via printing CNTs onto stickers that can be applied to curved surfaces, (ii) interactivity using a CNT-based button printed onto folded paper structure, and (iii) capacitive sensing of liquid wicking into the substrate itself. Facile integration of surface mount components on printed circuits is enabled by the intrinsic adhesion of the wet ink.

physics.app-ph↗

arXiv2021

A novel smoothed particle hydrodynamics formulation for thermo-capillary phase change problems with focus on metal additive manufacturing melt pool modeling

Laser-based metal processing including welding and three dimensional printing, involves localized melting of solid or granular raw material, surface tension-driven melt flow and significant evaporation of melt due to the applied very high energy densities. The present work proposes a weakly compressible smoothed particle hydrodynamics formulation for thermo-capillary phase change problems involving solid, liquid and gaseous phases with special focus on selective laser melting, an emerging metal additive manufacturing technique. Evaporation-induced recoil pressure, temperature-dependent surface tension and wetting forces are considered as mechanical interface fluxes, while a Gaussian laser beam heat source and evaporation-induced heat losses are considered as thermal interface fluxes. A novel interface stabilization scheme is proposed, which is shown to allow for a stable and smooth liquid-gas interface by effectively damping spurious interface flows as typically occurring in continuum surface force approaches. Moreover, discretization strategies for the tangential projection of the temperature gradient, as required for the discrete Marangoni forces, are critically reviewed. The proposed formulation is deemed especially suitable for modeling of the melt pool dynamics in metal additive manufacturing because the full range of relevant interface forces is considered and the explicit resolution of the atmospheric gas phase enables a consistent description of pore formation by gas inclusion. The accuracy and robustness of the individual model and method building blocks is verified by means of several selected examples in the context of the selective laser melting process.

cs.CE↗

arXiv2021

Spatial Mapping of Powder Layer Density for Metal Additive Manufacturing via X-ray Microscopy

Uniform powder spreading is a requisite for creating consistent, high-quality components via powder bed additive manufacturing (AM), wherein layer density and uniformity are complex functions of powder characteristics, spreading kinematics, and mechanical boundary conditions. High spatial variation in particle packing density, driven by the stochastic nature of the spreading process, impedes optical interrogation of these layer attributes. Thus, we present transmission X-ray imaging as a method for directly mapping the effective depth of powder layers at process-relevant scale and resolution. Specifically, we study layers of nominal 50-250 micrometer thickness, created by spreading a selection of commercially obtained Ti-6Al-4V, 316 SS, and Al-10Si-Mg powders into precision-depth templates. We find that powder layer packing fraction may be predicted from a combination of the relative thickness of the layer as compared to mean particle size, and flowability assessed by macroscale powder angle of repose. Power spectral density analysis is introduced as a tool for quantification of defect severity as a function of morphology, and enables separate consideration of layer uniformity and sparsity. Finally, spreading is studied using multi-layer templates, providing insight into how particles interact with both previously deposited material and abrupt changes in boundary condition. Experimental results are additionally compared to a purpose-built discrete element method (DEM) powder spreading simulation framework, clarifying the competing role of adhesive and gravitational forces in layer uniformity and density, as well as particle motion within the powder bed during spreading.

physics.app-ph↗