SEARCH · Engineering Papers

Results for “sequential”

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.

SourceResearch connectionFrom yearThrough year

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

At least 1,171 records · Page 65

All sources NASA NTRS DOE OSTI

DOE OSTI2025

Online thermal profile prediction for large format additive manufacturing: A hybrid CNN-LSTM based approach

Large format additive manufacturing (LFAM) is an advanced 3D printing technique that efficiently fabricates large-scale components through a layer-by-layer extrusion and deposition process. Accurate surface layer temperature monitoring is essential to prevent manufacturing failures and ensure final product quality. Traditional physics-based offline approaches for simulating thermal behavior are often inefficient and complex, posing challenges on real-time, in-situ monitoring. Here, to address this, we propose a data-driven hybrid CNN-LSTM model to predict sequential thermal images of arbitrary length using real-time infrared thermal imaging. In this approach, a Convolutional Neural Networks (CNN) is trained offline to capture spatial features, reduce dimensional complexity, and enhance time efficiency, while a stacked Long Short-Term Memory (LSTM) is applied online to capture temporal information for improved prediction of future thermal behavior in subsequent printing layers. Model performance is evaluated using MSE, SSIM, and PSNR metrics and is benchmarked against stacked LSTM and convolutional LSTM models, demonstrating superior accuracy and applicability. Additionally, to mitigate noise from moving extruders and gantry backgrounds in thermal images, a fine-tuned semantic segmentation model is implemented offline to extract printing geometry, enabling precise temperature tracking along the tool path for further thermal analysis. The frameworks developed in this study significantly advance temperature monitoring, thermal analysis, and in-situ manufacturing control for LFAM, bridging the gap between theoretical modeling and practical application.

Geometry extraction↗

DOE OSTI2025

A staged deep learning approach to spatial refinement in 3D temporal atmospheric transport

High-resolution spatiotemporal simulations effectively capture the complexities of atmospheric plume dispersion in complex terrain. However, their high computational cost makes them impractical for applications requiring rapid responses or iterative processes, such as optimization, uncertainty quantification, or inverse modeling. To address this challenge, this work introduces the Dual-Stage Temporal Three-dimensional UNet Super-resolution (DST3D-UNet-SR) model, a highly efficient deep learning model for plume dispersion predictions. DST3D-UNet-SR is composed of two sequential modules: the temporal module (TM), which predicts the transient evolution of a plume in complex terrain from low-resolution temporal data, and the spatial refinement module (SRM), which subsequently enhances the spatial resolution of the TM predictions. We train DST3D-UNet-SR using a comprehensive dataset derived from high-resolution large eddy simulations (LES) of plume transport. We propose the DST3D-UNet-SR model to significantly accelerate LES of three-dimensional (3D) plume dispersion by three orders of magnitude. Additionally, the model demonstrates the ability to dynamically adapt to evolving conditions through the incorporation of new observational data, substantially improving prediction accuracy in high-concentration regions near the source.

3D temporal sequences↗

DOE OSTI2023

Deep reinforcement learning for class imbalance fault diagnosis of equipment in nuclear power plants

In equipment fault diagnosis in nuclear power plants, there may be far more samples in one class (e.g., a health state) than in another class (e.g., a fault state). The distribution of data in each class is highly skewed. Most machine learning algorithms are suitable for balanced training datasets. When faced with imbalanced samples, these algorithms tend to provide good identification for the majority classes and bias for the minority classes. However, the misclassification of minority classes can lead to high costs. To address the above problem, this paper develops a deep reinforcement learning-based diagnosis method that models fault diagnosis as a sequential decision-making process. At each time step, the agent receives the state of the environment represented by the training samples and then takes a diagnosis action guided by a policy. If the action is correct/incorrect, the agent receives a positive/negative reward. The reward for minority classes is higher than that for majority classes. The agent’s goal is to obtain as many cumulative rewards as possible in the process, i.e., to identify the sample as correctly as possible. Six demonstration scenarios are constructed, depending on the selected fault datasets and the designed model structures. Experiments show that the proposed method achieves a higher weighted-averaged F1 score than the classical supervised learning method in most cases of class imbalance. Finally, the proposed method has potential applications in the field of class imbalance fault diagnosis of equipment in nuclear power plants.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

DOE OSTI2023

BISON validation to in situ cladding burst test and high-burnup LOCA experiments

The process for developing and qualifying nuclear fuels for commercial nuclear application requires fundamental material development, characterization, and design; out-of-pile testing on unirradiated materials; integral fuel rod irradiations, testing, and postirradiation examinations; and transient analyses. The historical approach depends on the generation of large empirical datasets and series of integral fuel rod irradiations, and this approach ultimately takes ~20 years—or sometimes longer—to acquire data through extensive sequential testing. Thus, the qualification and eventual deployment of new fuel systems constitute a long process. However, recent technological advancements have provided researchers the opportunity to perform out-of-cell, in situ measurements to assess material performance for the duration of the experiment. One such example of this capability is the use of digital image coordination and thermal imaging to assess Zircaloy cladding performance under a simulated loss-of-coolant accident (LOCA) transient condition. In situ measurements generally provide high-fidelity strain, strain rates, and temperature surface maps. This is critical for the US nuclear industry, which is actively developing a technical basis to support extending the peak rod average burnup from 62 to ~75 GWd/tU and the deployment of accident-tolerant fuel. However, the US Nuclear Regulatory Commission (NRC) outlined in its research information letter several technical issues that the industry must address before extending burnup. One topic of specific interest is understanding the cladding balloon and rupture geometry during the LOCA heat-up phase. By leveraging these advanced in situ capabilities, this work used in situ data generated from a simulated LOCA to better understand high-temperature creep and its effect on Zircaloy balloon and rupture performance. Here, this work used the BISON fuel performance code to assess the high-temperature creep model predictions with in situ data.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

DOE OSTI2020

Engineering the Composition of Bimetallic Nanocrystals to Improve Hydrodeoxygenation Selectivity for 2-Acetylfuran

Hydrodeoxygenation (HDO) of 2-acetylfuran (2-AF) was examined at 473 K and 33 bar over carbon-supported, bimetallic catalysts. Monodispersed, Pt-based (PtNi, PtCo, and PtCu) and Ni-based (NiCu) nanocrystals (NCs) with well-defined sizes and compositions were prepared via colloidal synthesis. Steady-state, flow-reactor measurements demonstrate that selectivities for the HDO reaction to 2-ethylfuran (2-EF) are strongly dependent on the alloy compositions. The HDO of 2-AF is shown to be sequential, where 2-EF can react further to form over-hydrogenated products. The NiCu 3 /C nanocatalyst was superior to the other bimetallics; with the optimal residence time, it was possible to achieve a selectivity of 75% 2-EF with 80% conversion of 2-AF, with good stability for up to at least 5 hours time-on-stream. Deactivation due to coking became an issue when the NCs were deposited on oxide supports with weak acidity (SiO 2 and γ-Al 2 O 3 ). Furthermore, this work provides insights into which bimetallic catalytic systems might be useful for HDO of 2-AF reaction.

2-acetylfuran↗

DOE OSTI2020

Toward hydrogen economy: Selective guaiacol hydrogenolysis under ambient hydrogen pressure

The conversion of guaiacol into the nylon precursors, cyclohexanol and cyclohexanone, is catalyzed by palladium supported on high surface area ceria (Pd/HS CeO 2 ) under mild conditions (100 °C, ≤1 bar H 2 ) via sequential hydrodemethoxylation and hydrogenation. In contrast, the 2-methoxycyclohexanol side product is generated by direct guaiacol hydrogenation. Reaction selectivity is determined by competing CO bond hydrogenolysis versus arene hydrogenation. Hydrogenolysis selectivity increases as H 2 pressure decreases, and over 80 % H 2 is incorporated into the products at 1 bar H 2 . Higher reactivity of guaiacol than anisole implies that the hydroxyl group is essential in Pd/HS CeO 2 catalysis. Participation of water in the conversion is evidenced by deuterium incorporation at non-exchangeable positions when the reaction is performed in D 2 O. Here, the combination of near quantitative mass balance of H 2 , high recyclability, and use of water as a solvent offers a simple, green and efficient conversion of lignin-derived aromatics into commercial products.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2022

Effect of framework Al pairing on NO storage properties of Pd-CHA passive NOx adsorbers

For this study, three Pd/H-CHA samples were prepared containing 53.0%, 10.8% and 6.5% paired Al sites at near fixed Si/Al ratio and similar Pd loading. According to H 2 temperature-programmed reduction, Pd was present almost exclusively as isolated cations in the two samples containing the higher concentrations of paired Al sites, whereas in the other sample PdO was also present. Simulated lean cold start tests on the fresh samples conducted in a microflow reactor showed that the sample containing PdO stored the lowest amount of NOx. When tested with CO/H 2 , the sample containing 53.0% paired Al sites showed significantly better storage capacity than the other samples and deactivated less rapidly upon sequential tests. Experiments using lean gasoline engine exhaust revealed similar trends. This study showed that a high concentration of paired Al sites in Pd/H-CHA is beneficial for NOx storage capacity, thermal durability, and minimizing deactivation in the presence of CO/H 2 .

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2023

Efficient and selective dual-pathway polyolefin hydro-conversion over unexpectedly bifunctional M/TiO 2 -anatase catalysts

The huge scale of plastic waste generation and its environmental consequences drive the demands for catalytic plastic upcycling processes. Here, we report efficient polypropylene (PP) hydro-conversion over Ru and Ni supported on a sol-gel anatase TiO 2 (TiO 2 -A-SG). A small number of Brønsted acid sites on TiO 2 -A-SG enables hydrocracking, improving efficiency, steering the selectivity towards high-valued C 4-20 , and allowing more isomerization, compared to Ru-based monofunctional hydrogenolysis catalysts. The relative contribution of hydrocracking increases with lower hydrogen partial pressure, more branched substrates, and lower Ru loading. Ni/TiO 2 -A-SG exhibits superior hydrocracking activity and selectivity to Ni on conventional Brønsted-acidic supports, particularly zeolites. Further, mechanistic studies shows fast isomerization and sequential ß-scissions caused by strong polymer-catalyst interaction and fast cracking with abundant 3 C. This favors C 4-12 formation and prevents secondary reactions. This work demonstrated a novel, highly efficient noble-metal-free catalyst for plastic waste upcycling, while advancing the mechanistic understanding of polyolefin hydro-conversion.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2025

Effect of CO on alcohol oxidation over commercial oxidation catalysts for control of emissions from lean-burn engines

For commercialization, alcohol-based net-zero carbon fuels in diesel-driven internal combustion engines must meet emissions standards set by the U.S. Environmental Protection Agency. Here, a series of aged commercial oxidation catalysts, Pt diesel oxidation catalyst (DOC), Pd+Pt DOC, and Pt oxidation catalyst (Pt OC), were studied using an automated flow reactor and in situ DRIFTS techniques to investigate the impact of CO on alcohol oxidation. Surface formate and acetate intermediates were formed due to partial oxidation of methanol and ethanol, respectively. Sequential introduction of alcohol and CO experiments using in situ DRIFTS on all three catalysts showed that CO inhibited low temperature methanol oxidation. Mutual inhibition effects between CO and ethanol were observed during co-oxidation. Acetate species shifted CO oxidation to higher temperatures on Pt DOC and Pt OC, whereas CO pre-exposure inhibited acetate formation. CO oxidation occurred at lower temperatures during ethanol and CO co-oxidation on Pd+Pt DOC.

33 ADVANCED PROPULSION SYSTEMS↗

DOE OSTI2020

Optimal carbon storage reservoir management through deep reinforcement learning

Model-based optimization plays a central role in energy system design and management. The complexity and high-dimensionality of many process-level models, especially those used for geosystem energy exploration and utilization, often lead to formidable computational costs when the dimension of decision space is also large. This work adopts elements of recently advanced deep learning techniques to solve a sequential decision-making problem in applied geosystem management. Specifically, a deep reinforcement learning framework was formed for optimal multiperiod planning, in which a deep Q-learning network (DQN) agent was trained to maximize rewards by learning from high-dimensional inputs and from exploitation of its past experiences. To expedite computation, deep multitask learning was used to approximate high-dimensional, multistate transition functions. Both DQN and deep multitask learning are pattern based. As a demonstration, the framework was applied to optimal carbon sequestration reservoir planning using two different types of management strategies: monitoring only and brine extraction. Both strategies are designed to mitigate potential risks due to pressure buildup. Results show that the DQN agent can identify the optimal policies to maximize the reward for given risk and cost constraints. Finally, experiments also show that knowledge the agent gained from interacting with one environment is largely preserved when deploying the same agent in other similar environments.

15 GEOTHERMAL ENERGY↗

DOE OSTI2026

Consumer safety-oriented scheduling of rotating power outages during heat waves

Extreme heat events have widespread effects on power systems, reducing available generation capacity, limiting transmission capabilities, and causing unusual demand patterns on the consumer side. As these combined effects expose bulk transmission systems to potential large-scale blackouts, utilities may be required to schedule and apply rotating outages, by temporarily and alternately disconnecting distribution substations to reduce overload. However, utilities lack mechanisms to inform these events, exacerbating the negative effects of heat waves on affected communities. This paper introduces a novel framework for scheduling rotating outages during heat waves while considering impacts on consumers’ safety. Instead of random sequential load shedding, we propose a methodology to rotate power outages considering a metric that quantifies the indoor overheating risk of groups of consumers during a power outage. The overheating risk is derived from a detailed building simulation using CityBES, where the buildings are modeled based on available data—use type, year built, floor area, number of stories, location—while presence of air conditioning and occupancy are calibrated from smart meter data. Based on the metric, an algorithm to schedule the rotating outages is applied to prioritize feeders for disconnection at each hour according to their overheating risk to meet a utility load reduction target. Applied to two substations and seven feeders in the Portland General Electric territory, the results show that this approach effectively leads to the lowest overheating risk during the resulting outage schedules, with an average 10.1% lower overheating compared to uninformed schedules.

Building thermal simulation↗

DOE OSTI2021

Systematic investigations on iron cycling in phosphorus/siderophore systems: Synergism or antagonism?

Synergisms between microbial exudates on Fe (hydr)oxide dissolution as an effective Fe acquisition pathway have been recently addressed and vigorously debated. However, Fe liberation mechanisms and where siderophores and phosphorus (P) coexist received little attentions. Current study systematically investigated ferrihydrite dissolution in the presence of desferrioxamine B (DFOB) (a kind of fungally-derived siderophores) and inorganic/organic phosphorus (orthophosphate, Pi; myo-inositol hexaphosphate, IHP), as a function of solute pH, reaction time and reagent content. Reacted solids were characterized by N 2 -BET adsorption, zeta (ζ) potential analysis, sequential extraction analysis (SEDEX), field emission scanning electron microscopy (FESEM), high-resolution transmission electron microscopy (HRTEM), X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), micro Raman spectroscopy and attenuated total reflectance-Fourier transform infrared spectroscopy (ATR-FTIR). Our results indicate that upon reaction with P-only or (DFOB + P) systems, interfacial complexation partially switched from monolayer bidentate-binuclear surface complexes to ternary complexes, or laterally transformed into amorphous Fe–P precipitates. The Fe-Pi complex precipitated more readily under acidic conditions, and Fe-IHP complex preferentially nucleated in neutral-alkaline environments. Phosphorus slightly promoted Fe release from minerals and fixation to the leached layer or interfacial liquid zone initially, but subsequently prevented further attacks from protons and DFOB. The co-effects of P and DFOB likely correspond to two successive scenarios: 1) DFOB is preferentially attracted toward ferrihydrite surfaces by negative electrical fields induced by adsorbed phosphorus and can act synergistically with labile P–Fe complexes, resulting in intensive temporal dissolution of Fh; 2) subsequent Fe shuttling to DFOB can be prohibited by stabilized, passive P/Fe–P layers. Finally, our results emphasize the antagonism between P compounds and siderophores (i.e., DFOB here) on ferrihydrite dissolution to improve understanding of the biologically-mediated Fe cycling in natural systems.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2020

Elucidating the Influence of Residual Polymer and Gas Environment on the Electronic Structure of a Graphene Layer Using in Situ APXPS

We use in situ ambient pressure X-ray photoelectron spectroscopy (APXPS) and ultraviolet photoelectron spectroscopy (UPS) to develop an effective method for studying changes in the graphene (Gr) electronic structure according to certain circumstance. The amount of polymethyl methacrylate (PMMA) residual polymer (RP), inevitably generated during the Gr transfer process, is significantly reduced from Gr surface by thermal annealing. This processed Gr is then sequentially exposed to specific gas environments (Ar, N 2 , O 2 , and CO 2 ), and APXPS or UPS is carried out to investigate the variations in the Gr electronic structure including the work function. When the amount of PMMA RP on Gr is reduced, the position of the main carbon peak shifts by more than 0.4 eV to a higher binding energy (in XPS spectra), and the secondary electron cutoff moves by about 0.2 eV to a lower binding energy (in UPS spectra). These changes are generally caused by a decrease in the Gr work function. On the other hand, exposure to the gas environments at different temperatures that we investigated did not produce significant changes in the work function and chemical states of Gr. These results confirm that the material in contact with Gr should be considered to achieve the desired Gr performance in electronics.

36 MATERIALS SCIENCE↗

DOE OSTI2022

Reinforcement Learning for feedback-enabled cyber resilience

The rapid growth in the number of devices and their connectivity has enlarged the attack surface and made cyber systems more vulnerable. As attackers become increasingly sophisticated and resourceful, mere reliance on traditional cyber protection, such as intrusion detection, firewalls, and encryption, is insufficient to secure the cyber systems. Cyber resilience provides a new security paradigm that complements inadequate protection with resilience mechanisms. A Cyber-Resilient Mechanism (CRM) adapts to the known or zero-day threats and uncertainties in real-time and strategically responds to them to maintain the critical functions of the cyber systems in the event of successful attacks. Feedback architectures play a pivotal role in enabling the online sensing, reasoning, and actuation process of the CRM. Reinforcement Learning (RL) is an important gathering of algorithms that epitomize the feedback architectures for cyber resilience. It allows the CRM to provide dynamic and sequential responses to attacks with limited or without prior knowledge of the environment and the attacker. In this work, we review the literature on RL for cyber resilience and discuss the cyber-resilient defenses against three major types of vulnerabilities, i.e., posture-related, information-related, and human-related vulnerabilities. Here we introduce moving target defense, defensive cyber deception, and assistive human security technologies as three application domains of CRMs to elaborate on their designs. The RL algorithms also have vulnerabilities themselves. We explain the major vulnerabilities of RL and present develop several attack models where the attacker target the information exchanged between the environment and the agent: the rewards, the state observations, and the action commands. We show that the attacker can trick the RL agent into learning a nefarious policy with minimum attacking effort. The paper introduces several defense methods to secure the RL-enabled systems from these attacks. However, there is still a lack of works that focuses on the defensive mechanisms for RL-enabled systems. Last but not least, we discuss the future challenges of RL for cyber security and resilience and emerging applications of RL-based CRMs.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2022

The role of information structures in game-theoretic multi-agent learning

Multi-agent learning (MAL) studies how agents learn to behave optimally and adaptively from their experience when interacting with other agents in dynamic environments. The outcome of a MAL process is jointly determined by all agents’ decision-making. Hence, each agent needs to think strategically about others’ sequential moves, when planning future actions. The strategic interactions among agents makes MAL go beyond the direct extension of single-agent learning to multiple agents. With the strategic thinking, each agent aims to build a subjective model of others decision-making using its observations. Such modeling is directly influenced by agents’ perception during the learning process, which is called the information structure of the agent’s learning. As it determines the input to MAL processes, information structures play a significant role in the learning mechanisms of the agents. This review creates a taxonomy of MAL and establishes a unified and systematic way to understand MAL from the perspective of information structures. Here we define three fundamental components of MAL: the information structure (i.e., what the agent can observe), the belief generation (i.e., how the agent forms a belief about others based on the observations), as well as the policy generation (i.e., how the agent generates its policy based on its belief). In addition, this taxonomy enables the classification of a wide range of state-of-the-art algorithms into four categories based on the belief-generation mechanisms of the opponents, including stationary, conjectured, calibrated, and sophisticated opponents. We introduce Value of Information (VoI) as a metric to quantify the impact of different information structures on MAL. Finally, we discuss the strengths and limitations of algorithms from different categories and point to promising avenues of future research.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2025

Integrated thermal and biological conversion of microalgal proteins to lipids

Microalgal composition varies with cultivation strategy, and low-cost approaches often produce high-protein biomass. This presents challenges for biorefineries designed around static, lipid-rich feedstocks. In particular, hydrolysates from high-protein algae are nitrogen-rich and sugar-poor, limiting microbial conversion and reducing product yields. This study develops a sequential thermal conditioning and biological upgrading strategy to integrate high-protein hydrolysate processing within conventional lipid extraction and upgrading designs. Oxidative deconstruction was used to break down proteins into ammonium and short-chain carboxylates. Ammonium was subsequently removed to yield a nitrogen-depleted, carboxylate-rich medium suitable for microbial lipid production. Bioconversion trials with Cutaneotrichosporon oleaginosum showed lipid accumulation only from hydrolysates treated with both oxidative deconstruction and nitrogen removal, reaching 1.2 g/L lipids at 30 % intracellular content. This integrated approach enables protein-to-lipid conversion and improves flexibility to process variable algal feedstocks, advancing fuel-oriented microalgal biorefineries.

09 BIOMASS FUELS↗

DOE OSTI2021

The Stellar decomposition: A compact representation for simplicial complexes and beyond

Here, we introduce the Stellar decomposition, a model for efficient topological data structures over a broad range of simplicial and cell complexes. A Stellar decomposition of a complex is a collection of regions indexing the complex’s vertices and cells such that each region has sufficient information to locally reconstruct the star of its vertices, i.e., the cells incident in the region’s vertices. Stellar decompositions are general in that they can compactly represent and efficiently traverse arbitrary complexes with a manifold or non-manifold domain. They are scalable to complexes in high dimension and of large size, and they enable users to easily construct tailored application-dependent data structures using a fraction of the memory required by a corresponding global topological data structure on the complex. As a concrete realization of this model for spatially embedded complexes, we introduce the Stellar tree, which combines a nested spatial tree with a simple tuning parameter to control the number of vertices in a region. Stellar trees exploit the complex’s spatial locality by reordering vertex and cell indices according to the spatial decomposition and by compressing sequential ranges of indices. Stellar trees are competitive with state-of-the-art topological data structures for manifold simplicial complexes and offer significant improvements for cell complexes and non-manifold simplicial complexes. We conclude with a high-level description of several mesh processing and analysis applications that utilize Stellar trees to process large datasets.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2022

Additive manufacturing of nanotube-loaded thermosets via direct ink writing and radio-frequency heating and curing

Direct Ink Writing (DIW) is an extrusion-based additive manufacturing method where the print medium is a liquid-phase ‘ink’ dispensed out of nozzles and deposited along digitally defined paths. Conventional DIW of thermosetting resins relies on viscosity modifying agents, novel crosslinking chemistries, and/or long curing schedules in an oven. Here we demonstrate the use of a co-planar radio frequency applicator to generate an electric field, which can be used to rapidly heat and cure nano-filled composite resins as they are printed. This method avoids the need for an oven or post- curing step. This process consists of a sequential print-and-cure cycle which allows for printing of high-resolution, multi-layered structures. Every extruded layer is partially cured using RF before depositing the next layer; this allows the printed part to maintain structural integrity. The process enables both increased throughput and decreased touch time relative to traditional manufacturing. Commercial epoxy resin with varied nano-filler loadings were examined as DIW candidates. After printing, the thermo-mechanical properties, surface finish, and shape retention of RF-cured samples were comparable to conventionally cured samples. Furthermore, this method of manufacturing establishes RF heating as a suitable alternative to conventional methods, facilitating rapid, free-form processing of thermosetting resins without a mold.

36 MATERIALS SCIENCE↗