Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Laterality”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 163 records · Page 9
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

Autonomous direct freeform fabrication strategy for multi-axis additive manufacturing

Multi-axis additive manufacturing (M-AM) enables precise material deposition along both planar and curved layers, eliminating the need for support structures through a continuous material deposition approach. In contrast to conventional 2-dimensional planar layers constrained to a fixed building orientation, the deposition on freeform layers demands the specification of guided curves to determine material deposition directions which are no longer to be fixed to a build direction. There are challenges that arise when fabricating components with multiple “build” directions, necessitating the decomposition of geometries and the specification of guided curves for the resulting volumes. Furthermore, multi-axis systems introduce heightened challenges due to an increased degree of freedom in motion. Consequently, the potential risks of collision between the deposited geometry and the motion platform become a notable concern. This research proposes a freeform layering algorithm to address the challenge of seamless transitions between planar and curved layers in the process planning of M-AM. The algorithm computes 3D “printable” layers by leveraging topological information derived from the geometry to be built and integrates collision-free manufacturability considerations into the computational process. These accumulated volumes serve as a “substrate” and support volumes for subsequent deposition, allowing later layers to be built without the need for additional support material. In conclusion, the proposed method successfully devises a freeform layering approach suitable for intricate models that demand substantial support, thus enabling the fabrication of diverse geometries in a manner previously unachievable.

36 MATERIALS SCIENCE↗
DOE OSTI2026

Process-driven roadmap for depositing super duplex stainless steel via wire Arc additive manufacturing

Here, this study systematically investigates the effects of shielding gas, bead spacing, weld mode, and travel speed on the phase balance, porosity, and hardness of wire arc additively manufactured (WAAM) super duplex stainless steel (ER2594). Robotic WAAM was employed to fabricate multilayer walls under systematically varied process conditions, followed by phase transformation simulations, X-ray computed tomography (XCT), electron backscatter diffraction (EBSD), and microhardness evaluation. Thermodynamic simulations predicted rapid cooling of AM process can suppress the potential formation of deleterious precipitates which was later validated via cross-sectional microstructure analyses of printed samples. XCT revealed porosity levels below 0.2% for all deposits, with reduced travel speed significantly lowering defect volume. Microstructural analyses revealed the evolution of various austenite precipitates, including grain boundary austenite (GBA), Widmanstätten austenite (WA), and intergranular austenite (IGA), sequentially upon cooling of the ferrite phase. Among all process parameters, weld transfer mode exhibited the strongest influence on phase balance; pulsed mode promoted higher ferrite retention (~ 36%) compared to RapidX mode. No consistent relationship between stabilized phase fraction and captured microhardness was observed. This work provides critical insights for optimizing WAAM parameters to control phase balance and mechanical performance, which is essential for producing high-integrity super duplex stainless-steel components for nuclear and marine applications.

Grain orientation↗
DOE OSTI2025

Main bearing response in a waked 15-MW floating wind turbine in below-rated conditions

Increased wind turbine size raises unknowns related to structural flexibility. Moreover, moving to deeper waters, component reliability becomes more critical. This work investigates main bearing response dependence on drivetrain flexibility and wake impingement in a two-turbine wind farm. A 15-MW floating direct-drive turbine is considered. Large eddy simulations (LES) are employed to model neutral, stable and unstable atmospheric conditions at below-rated mean wind speed, while the engineering codes OpenFAST and FAST.Farm simulate turbine and wake behavior. Results indicate significant sensitivities in fatigue estimates to lateral distance between the upstream and downstream turbine. The trends are most substantial in stable conditions, where the waked downwind main bearing sees twice the fatigue damage estimates of the upstream turbine for one position and 50% for another. Main bearing fatigue sensitivity to drivetrain flexibility is minor, while properly including generator rotor inertia loads is important for the axial forces of the locating (axially fixed) bearing, especially in stable conditions.

17 WIND ENERGY↗
DOE OSTI2026

Entropy Stable Conservative Flux Form Neural Networks

We propose an entropy-stable conservative flux form neural network (CFN) to predict the dynamics of unknown governing conservation laws. The design of the network is based on the entropy-stable, second-order, and non-oscillatory Kurganov-Tadmor (KT) scheme. The proposed entropy-stable CFN, hereafter referred to as ESCFN, uses slope limiting as a denoising mechanism, ensuring accurate predictions in both noisy and sparse observation environments, as well as in both smooth and discontinuous regions. Importantly, our method is designed to predict long term dynamics of the unknown conservation law exclusively from a short temporal window of observed data, that is, without oracle knowledge of the PDE or later-time solution profiles. Numerical experiments demonstrate that the ESCFN achieves both stability and conservation while maintaining accuracy over extended time domains, and successfully predicts shock propagation speeds in long-term simulations. Furthermore, it is also robust to both noisy and sparse data environments.

Hyperbolic conservation laws↗
DOE OSTI2024

Effects of Plume Targeted Cooling on Residual Stress in Controlled Atmosphere Plasma Sprayed Coatings

Thermal spray processes can benefit from cooling to maintain substrate temper, reduce processing times, and manage thermally induced residual stresses. “Plume quenching” is a plume-targeted cooling technique which has been shown to reduce substrate temperatures by redirection of hot plume gases using a lateral argon curtain injected into the plume, while limiting interaction with the substrate or affecting coating properties. Here, this study explores the use of this technique for residual stress management by reducing the thermally driven component in nickel and tantalum coatings on titanium and aluminum substrates. The in-situ residual stress profiles were measured for all substrate and coating pairings during spraying and cooling, and the deposition and thermal stresses recorded. For substrate and coating pairings where the predominant component of residual stress was thermal (driven by a large difference in coefficient of thermal expansion, Δα, between coating and substrate), plume quenching reduced both the thermal stress and the final stress state of the coating. This was seen primarily in tantalum on aluminum coatings where the Δα was -17 × 10 -6 /°C, and thermal stress was reduced by 7.5% and 22.4% for the plume quenching rates of 50 and 100 slpm, respectively.

42 ENGINEERING↗
DOE OSTI2025

Effects of an External Magnetic Field on Keyhole Mode Laser Melting of 316 Stainless Steel

Keyhole-mode laser melting is an efficient method for joining or cutting large, thick components, but controlling keyhole depth and fluctuations has remained challenging. Applying an external magnetic field can control melt pool flows and indirectly influence keyhole morphology and dynamics. The induced Lorentz force, comprising Seebeck and damping components, plays a crucial role in the melt pool dynamics, depending on temperature gradient, flow rates, and magnetic field orientation and magnitude. This research investigates the effects of an external magnetic field on keyhole behavior during laser spot melting of 316 stainless steel using synchronized high-speed synchrotron X-ray and thermal imaging. Findings revealed that a longitudinal magnetic field (120 mT) increased keyhole depth but exacerbated lateral fluctuations, resulted in a 20% increase in the melt pool temperature gradient and a 27% decrease in cooling rate. Conversely, a transverse magnetic field (760 mT) reduced keyhole depth and improved porosity formation. The findings suggest that a decrease in keyhole depth correlates with a decrease in fluctuations, and vice versa. These insights enhance understanding of external magnetic fields’ impact on laser melting, with implications for improving part quality.

Alamdari, Aslan Bafahm [The Ohio State Univ., Colu↗
DOE OSTI2025

New electric force and charge exchange modules in discrete element model enables particle dynamics simulation in electric field

Discrete element modeling (DEM) is an important technique for particle dynamics simulation. The field of metal additive manufacturing often utilizes DEM to simulate the rheological behaviors of powder. Standard contact and short-range interactions are sufficient in most cases but insufficient to describe the particle dynamics with the influence of an electric field. Modeling such a system requires additional physics to describe the particle–field interactions. The relevant physics has been experimentally understood but is not yet available in DEM. Here, we develop a charge exchange and an electric force module. The electric force module governs particle response to the electric field, while the charge exchange module enables particles to acquire proper charge during contact with charged geometries. We validate the modules against analytical calculations and high-speed videos of electrostatic powder deposition experiments. Notably, the model struggles to capture the initial particle levitation. We later deploy a modified electric field, as supported by static electric field simulation, to better approximate the electric field penetration into the powder layer. This modification improves the model’s capability of simulating realistic particle levitation. The results highlight the challenges of modeling particle behaviors in the electric field while demonstrating the feasibility of obtaining quantitative results, which are difficult to measure experimentally.

charge exchange↗
DOE OSTI2025

Direct observation of strain-enhanced hydrogen segregation and failure at high-angle grain boundaries in nickel

Understanding the mechanisms underlying hydrogen embrittlement remains difficult, even in single-element metals. Both microstructure and stress state influence hydrogen distribution in metals and alloys, which impacts deformation and failure. Here, in this work, we use in-situ Kelvin probe force microscopy (KPFM) to monitor the hydrogen distribution in pure nickel over time at 1.1 % and 3.5 % strain. The sample strained to 3.5 % results in preferential hydrogen segregation to high-angle grain boundaries whereas the sample strained to 1.1 % does not exhibit preferential hydrogen segregation. Optical digital image correlation (DIC) shows that hydrogen charging results in both localized and reduced strains during tensile testing of a notched sample. Later stages of deformation and failure (e.g. microcracking) are studied by using in-situ transmission X-ray microscopy (TXM). TXM reveals nanoscale structural changes to a propagating crack in a hydrogen environment. Localized void growth and secondary cracking occur at grain boundaries near the primary crack front. Correlative electron backscattering diffraction (EBSD) is used to relate the cracking at grain boundaries to the hydrogen segregation observed in KPFM. These findings are unified in a proposed hydrogen embrittlement mechanism that describes the interaction of hydrogen with grain boundaries, and the role of grain boundaries in hydrogen embrittlement.

Grain boundary fracture↗
DOE OSTI2025

Scaling microstructural processes in the sintering of ionic ceramics

A multi-scale framework, combining a multiphase field formulation and large deformation mechanics, was developed as a stepping stone to perform the data analytics of the microstructural level kinetics of a sintering solid. Relevant microstructural information from this framework, such as grain, stress, and porosity statistics, was scaled up to describe the macroscopic level sintering kinetics. Here, the developed formulation was applied to describe the electric field assisted sintering of Y 2 O 3 . Microstructural inhomogeneities in a multi-granular solid result in the formation of a field of compressive stress networks, which interleave with low compression and weakly tensile regions, defining a scaffolding for sintering concentration regions to develop. A Poisson effect-induced lateral stress network is also naturally self-induced as a result of the mechanical constraints imposed by the sintering apparatus. For long sintering times, localized shear stresses enhancing mass flow along grain boundaries and internal surfaces develop. Three-sided pores are removed by either vacancy transport to the surrounding pores, or move towards the external surfaces through grain boundary diffusion. Four- and higher order-sided pores stabilize because an equal amount of vacancies are gained and lost through the connecting grain boundaries. Grain dewetting contributes to pore coalescence, suggesting that pore kinetics and grain growth are coupled and should be analyzed in concert. The combined sintering and grain growth kinetics define six regimes of sintering behavior: (1) T, the transient regime; (2) E$_Υ$, the surface energy dominated, early sintering regime, where the grain growth exponent, p = 1, and the stress concentration factor, $f$ ~ $1/\hat{ρ}^{4.6}$; (3) E S , the stress dominated, early sintering regime, where p = 1 and $f$ ~ $1/\hat{ρ}^{4}$; (4) I$_Υ$, the surface energy dominated, intermediate sintering regime, where p = 2 and $f$ ~ $1/\hat{ρ}^{4.6}$; (5) I S , the stress dominated, intermediate sintering regime, where p = 2 and $f$ ~ $1/\hat{ρ}^{4}$; and (6) L, the late sintering regime, where p = 3 and $f$ ~ 1. At the macroscopic level, the rapid densification and suppression of grain growth observed in the electric field assisted sintering process is a consequence of the compounding effects of the underlying stress-, transport-, and interfacial-energy-induced energy minimization kinetics, as predicted by the multi-scale framework.

11 - NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2025

A general kinetic framework for dislocation mobility derived from probabilistic cellular automaton simulations of the kink-pair mechanism

Dislocation mobility laws are essential components of dislocation-density-based crystal plasticity models. For dislocations governed by the kink-pair mechanism, however, existing formulations are typically limited to specific regimes due to the com plex interplay between stochastic kink-pair nucleation and lateral kink migration. In this work, we develop a general kinetic framework that expresses the average dislocation velocity as a function of mechanism-level variables: positive/negative kink pair nucleation rates, kink migration velocity, dislocation segment length, critical kink-pair width, and kink height. Probabilis tic cellular automaton simulations are used to capture the behavior of conceptual dislocation segments between the limiting conditions of migration outpacing nucleation on the one end and nucleation outpacing migration on the other. An elemen tary functional form that captures the system dynamics is suggested and fitted against the simulation results. This framework remains valid for arbitrary combinations of the six variables and is, therefore, compatible with any admissible constitutive re lations that describe their stress and temperature dependence. Comparisons with established approaches and experimental results confirm the robustness and physical consistency of the formulation, making it broadly applicable to material systems in which dislocation motion is governed by the kink-pair mechanism.

36 MATERIALS SCIENCE↗
DOE OSTI2026

Paleotribological models using preserved fossil tissue properties reveal functional significance of Eurasian mammoth dental evolution

Eurasian mammoths (Mammuthus) underwent substantial modifications in molar morphology as later-diverging species evolved progressively thinner enamel and increased enamel crest complexity. These features have been hypothesized to reduce whole-tooth wear and extend dental longevity as increasingly graze-dominated diets evolved within the lineage. This hypothesis has yet to be directly tested. Here, in this study, we developed an in-silico wear model using experimentally derived wear rates from fossil and extant proboscidean dental tissues. The models revealed that shifts in tissue topology do not affect whole-tooth wear rate, as inverse trends in lamellar frequency and enamel thickness preserve a consistent surface enamel area fraction; the determining factor of wear. Rather, topological shifts produce a wear-emergent secondary occlusal surface with greater numbers of triturating crests that create a regular, low-relief, file-like shearing pavement. These changes in occlusal architecture likely directly impacted the mastication capacity of Mammuthus dentitions, facilitating their dietary expansion to incorporate fibrous, lower-nutrient graze.

Dental wear↗
DOE OSTI2024

A study on the impact of using a subchannel resolution for modeling of large break loss of coolant accidents

The nuclear industry is investigating the feasibility of transitioning from 18- to 24-month fuel cycles because of the positive impact it would have on the operational costs for the current fleet of light-water reactors. A challenge to making this change is the increased risk of fuel fragmentation, relocation, and dispersal (FFRD) due to the known potential for ceramic fuel to pulverize into fine particles at the higher discharge burnups. Previous work has been performed by the Nuclear Energy Advanced Modeling and Simulation program to assess FFRD risk in high-burnup cores using the BISON fuel performance code and a coarse mesh thermal hydraulics (T/H) solution for a loss-of-coolant accident (LOCA) using the TRACE system T/H code. Because of the importance of the T/H solution for FFRD assessment, this study seeks to investigate the impact of using higher-fidelity subchannel techniques for modeling of the LOCA transient. CTF was used to model a subregion of a high-burnup core that was depleted by the Virtual Environment for Reactor Applications (VERA) multiphysics core simulator. Both coarse-mesh and pin-resolved models were created in CTF, and a consistent coarse-mesh TRACE model was also developed to allow for benchmarking the code results. Further, a large-break loss-of-coolant accident (LBLOCA) reflood transient was simulated using these three models, and results were compared. Results showed some consistent differences between the CTF and TRACE coarse models, including a higher peak cladding temperature (PCT) prediction in CTF and later quenching in CTF; however, the transient clad temperature behavior was similar, and these differences are likely due to post-critical heat flux heat transfer modeling differences and minimum film boiling temperature model differences. The pin-resolved results indicate that the PCT in the lumped model is often under-predicted by as much as 70 °C and that PCT occurs at a different location than the high-power pin in the assembly. The lumped model predicts a difference of 10 °C or less between the average and hot pins in the assembly, whereas the pin-resolved model predicts a range of over 100 °C. These results indicate that higher-fidelity T/H results may have an impact on predicted core behavior during LOCA, which may be important to consider when assessing FFRD risk.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗
DOE OSTI2024

Development of a Griffin model of the advanced test reactor

In the pursuit of a higher fidelity deterministic simulation capability of the Advanced Test Reactor, it is important to have a fast yet accurate deterministic neutronics model. Here, to achieve this, we employed an advanced two-step method. The first step involves generating homogenized cross sections using OpenMC, a cutting-edge Monte Carlo neutron transport code. OpenMC offers excellent modular capabilities, allowing for easy component integration and flexibility in incorporating new designs into the model. The second step involves deterministic transport calculations, which are performed using Griffin, a reactor physics application based on the Multiphysics Object-Oriented Simulation Environment (MOOSE). To ensure the accurate spatial resolution and assignment of material cross sections, a Cubit-generated mesh for the Advanced Test Reactor is utilized as an intermediate step between the OpenMC and Griffin models; Griffin utilizes the mesh for its finite element solution, while OpenMC material identifications are written to the mesh file to be used in Griffin material assignments. Additionally, a Python-based script converts the cross sections generated by OpenMC into the ISOXML format required by Griffin. Initial comparisons using the Griffin diffusion solver indicated good agreement between the neutron multiplication factors obtained from the standalone OpenMC model and the Griffin model, with differences of less than 10 pcm in the 2D geometry configuration; it was later determined that this agreement was likely due to compensating effect and was more likely on the order of –700 pcm relative to the OpenMC solution. However, in three-dimensional calculations, an unacceptably large error (almost 8,000 pcm) was found in the Griffin solution with the diffusion solver. Subsequent calculations using Griffin’s discrete ordinates solver demonstrated substantially improved agreement, within 116 pcm of the OpenMC solution used to generate the cross sections for Griffin. Building on this capability, future work will seek to perform more detailed validation calculations. The ultimate goal is to evaluate both transient and multiphysics simulations of the reactor.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗
DOE OSTI2025

MURR LEU structural and thermal hydraulics analyses: Part I – Preliminary irradiation thermo-mechanical behavior

The University of Missouri Research Reactor (MURR) is expected to be converted from highly enriched uranium (HEU, ≥ 20 wt% U-235) U-Al x dispersion fuel to low-enriched uranium (LEU, < 20 wt% U-235) with U-10Mo monolithic fuel. This work introduces high-fidelity irradiation thermo–mechanical (T-M) analysis of the MURR LEU focusing on changes in coolant channel gap thickness. Three-dimensional (3D) finite element (FE) models were developed to simulate the irradiation T-M behavior of the MURR LEU element with all 23 curved fuel plates, the two side plates, and the combs. It was shown that channel gap thickness changes were influenced not only by plate thickness variations due to fuel swelling and creep but also by the radial displacement of consecutive MURR LEU plates. Modeling the fuel element assembly captured side plate displacements, which were shown to reduce radial fuel plate displacements towards the convex side. The maximum local radial displacement in the element was predicted at the end of life (EOL) as 23.7 mil (602.0 µm) on the lateral centerline of plate 23 towards the convex side. The maximum stripe-averaged reduction in channel gap thickness, particularly relevant for thermal hydraulics (TH) safety analysis, was calculated as 15.9 mil (403.9 µm) in single-side heated channel 24 (the outermost channel). These results account for the thermal resistance from the oxide build-up on cladding surfaces which was shown to be up to 0.82 mil (20.8 µm) thick. It was demonstrated that accounting for oxide layer thermal resistance led to a 10 °C higher peak fuel temperature and a 4.4 mil (111.8 µm) greater maximum local radial displacement. The impact of the calculated channel gap thickness changes on the MURR LEU TH safety analysis is evaluated in Part II.

Irradiation thermo-mechanical analysis↗
DOE OSTI2024

Agent-based modeling for multimodal transportation of CO 2 for carbon capture, utilization, and storage: CCUS-agent

Here, to understand the system-level interactions between the entities in Carbon Capture, Utilization, and Storage (CCUS), an agent-based foundational modeling tool, CCUS-Agent, is developed for a large-scale study of transportation flows and infrastructure in the United States. Key features of the tool include (i) modular design, (ii) multiple transportation modes, (iii) capabilities for extension, and (iv) testing against various system components and networks of small and large sizes. Five matching algorithms for CO 2 supply agents (e.g., powerplants and industrial facilities) and demand agents (e.g., storage and utilization sites) are explored: Most Profitable First Year (MPFY), Most Profitable All Years (MPAY), Shortest Total Distance First Year (SDFY), Shortest Total Distance All Years (SDAY), and Shortest distance to long-haul transport All Years (ACAY). Before matching, the supply agent, demand agent, and route must be available, and the connection must be profitable. A profitable connection means the supply agent portion of revenue from the 45Q tax credit must cover the supply agent costs and all transportation costs, while the demand agent revenue portion must cover all demand agent costs. A case study employing over 5500 supply and demand agents and multimodal CCUS transportation infrastructure in the contiguous United States is conducted. The results suggest that it is possible to capture over 9 billion tonnes (GT) of CO 2 from 2025 to 2043, which will increase significantly to 22 GT if the capture costs are reduced by 40 %. The MPFY and SDFY algorithms capture more CO 2 earlier in the time horizon, while the MPAY and SDAY algorithms capture more later in the time horizon.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2024

Wellbore cement alteration and roles of CO 2 and shale during underground hydrogen storage

To mitigate climate change and adopt renewable energy, energy storage is crucial and can be done in the form of hydrogen gas (H 2 ). Subsurface geologic reservoirs are positioned to store H 2 on the largest scales for the longest terms of all potential options. However, H 2 injection may boost reactions that consume hydrogen, generate undesired gases, and alter pore structures of geomedia. To explore the extent of H 2 -associated biotic reactions at a near wellbore location, four experiments were conducted under underground storage conditions with wellbore cement cores and, in most instances, shale samples submerged in synthetic formation brine. Post-reaction gas, aqueous, and solid phase samples were analyzed using olfactory screening and, later, gas chromatography (GC-MS), inductively coupled plasma optical emission spectroscopy (ICP-OES), scanning electron microscopy (SEM), and synchrotron micro-scale x-ray fluorescence (μ-XRF). Within a period of 16 weeks, hydrogen sulfide (H 2 S) was generated in systems containing both H 2 and shale. XRF mapping identified a zone enriched in iron(II) and reduced sulfur along the rim of cement cross sections that was largely associated with CO 2 -induced cement carbonation. Shale did not show noticeable alteration, but there is evidence it contributed to the initial inoculation of the system and provided nutrients for microbes via water-rock interactions. Here, this study considers both rock formations and wellbore cement not previously evaluated concurrently. Findings support understanding and modeling of H 2 -associated biogeochemical reactions during underground hydrogen storage.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Experimental investigation on phase change material–based finned tube heat exchanger for thermal energy storage and building envelope thermal management

Phase change materials (PCMs) are attractive solutions for thermal energy storage (TES) applications by absorbing and releasing large amounts of latent heat during solid–liquid phase transitions. However, their relatively low thermal conductivity requires novel heat exchanger–based solutions to improve the power density and overall energy storage efficiency of the TES system. This work presents the design and experimental results of a finned tube heat exchanger to store collected natural thermal energy from a building envelope in a latent-based TES and to release it later for building heating/cooling applications. We experimentally evaluate the finned tube heat exchanger and evaluate the performance of TES in reducing building heating and cooling loads over 3–4 h of desired time of operation (e.g., peak load). The optimized design allows for maximum energy density by minimizing the heat exchanger volume, and the system is evaluated experimentally using commercially available heat exchanger materials and an organic PCM. Here, the experimental results reveal that the TES system is able to charge and discharge stored latent energy within 3–4 h, matching peak building electricity demand duration under an average fluid flow rate of 0.136 kg/s and temperature difference of 5.55 °C. Importantly, such optimized designs illuminate a path toward TES designs that are low-cost, scalable, and optimized for thermal energy and power availability under the desired time of operation.

25 ENERGY STORAGE↗
DOE OSTI2025

Deformable phrase level attention: A flexible approach for improving AI based medical coding

Objective: Improving the AI-driven automated medical encoding of clinical text plays a vital role in gathering information on the occurrence of diseases to improve population-level health. This work presents a novel attention mechanism designed to enhance text classification models and ensure appropriate classification of medical concepts in unstructured electronic health records. Materials and Methods: We developed a deformable, phrase-level attention mechanism to identify important lexical word-level and contextual phrase-level information from clinical text documents. We evaluated conventional and transformer-based deep learning models that we extended with our attention mechanism on the extraction of critical cancer information (e.g., site, subsite, laterality, histology, behavior) from 629,908 electronic pathology reports and on the automated medical encoding of 52,722 hospital discharge summaries. Results: Transformer-based models with the deformable, phrase-level attention mechanism achieved the best performance on the extraction of critical cancer information from pathology reports. Conventional- and transformer-based models show similar or better performance than their baseline counterparts on the automated medical encoding of clinical documents. Discussion: The addition of phrase-level information allowed models extended with our proposed method to outperform standard word-level attention. Our method showed favorable properties for the real-world application in terms of model robustness and phenotyping. These results indicate that our method is promising for automated data harmonization for common data models. Conclusion: This work proposes a novel deformable, phrase-level attention mechanism that enhances text classification models in the extraction of medical concepts from clinical text documents. We demonstrate strong performances on two clinical text datasets and showcase real-world deployability of our method.

Automated medical encoding↗