Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “process loads”

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 595 records · Page 33
All sourcesNASA NTRSDOE OSTI
NASA NTRS1993

Establishment of computerized numerical databases on thermophysical and other properties of molten as well as solid materials and data evaluation and validation for generating recommended reliable reference data

The Center for Information and Numerical Data Analysis and Synthesis, (CINDAS), measures and maintains databases on thermophysical, thermoradiative, mechanical, optical, electronic, ablation, and physical properties of materials. Emphasis is on aerospace structural materials especially composites and on infrared detector/sensor materials. Within CINDAS, the Department of Defense sponsors at Purdue several centers: the High Temperature Material Information Analysis Center (HTMIAC), the Ceramics Information Analysis Center (CIAC) and the Metals Information Analysis Center (MIAC). The responsibilities of CINDAS are extremely broad encompassing basic and applied research, measurement of the properties of thin wires and thin foils as well as bulk materials, acquisition and search of world-wide literature, critical evaluation of data, generation of estimated values to fill data voids, investigation of constitutive, structural, processing, environmental, and rapid heating and loading effects, and dissemination of data. Liquids, gases, molten materials and solids are all considered. The responsibility of maintaining widely used databases includes data evaluation, analysis, correlation, and synthesis. Material property data recorded on the literature are often conflicting, diverging, and subject to large uncertainties. It is admittedly difficult to accurately measure materials properties. Systematic and random errors both enter. Some errors result from lack of characterization of the material itself (impurity effects). In some cases assumed boundary conditions corresponding to a theoretical model are not obtained in the experiments. Stray heat flows and losses must be accounted for. Some experimental methods are inappropriate and in other cases appropriate methods are carried out with poor technique. Conflicts in data may be resolved by curve fitting of the data to theoretical or empirical models or correlation in terms of various affecting parameters. Reasons (e.g. phase transitions) must be found for unusual dependence or any anomaly. Such critical evaluation involves knowledge of theory, experience in measurement, familiarity with metallurgy (microstructural behavior) and not inconsiderable judgment. An examination of typical data compiled and analyzed by CINDAS shows that the thermal conductivity of a material reported in the literature may vary by a factor of two of more; the range of reported values increases as temperature increases reflecting the difficulty of high temperature measurements. Often only estimates of melt behavior are available, despite the importance of melt properties in modeling, welding, or other solidification processes. There may be only a few measurements available for properties such as kinematic viscosity, even for widely used materials such as stainless steel. In the face of such a paucity of existing data and in a national environment where too few new data are being generated it is nonetheless the responsibility of CINDAS to select and disseminate recommended values of a wide variety of thermophysical properties.

Ho, C. Y.↗
DOE OSTI2024

HydraGNN_Predictive_GFM_2024 - Ensemble of predictive graph foundation models for ground state atomistic materials modeling

We provide the ensemble of fifteen pre-trained graph foundation models (GFMs) for atomistic materials modeling applications. Each one of the fifteen GFMs has been trained on five open-source datasets that (once aggregated) amount to over 154 million atomistic structures, which cover over two-thirds of the natural elements of the periodic table and that comprises a broad set of organic and inorganic compounds. This vast set of atomistic structures comprises ground state configurations that are dynamically stable (i.e., equilibrated structures with atomic forces approximately close to zero values) as well as dynamically unstable structures (i.e., non-equilibrium structures with non-negligible non-zero values of atomic forces). The ensemble of datasets aggregated does NOT include excited states. The datasets have been curated to remove atomistic structures with spectral norm of the force tensor above 100 eV/angstrom. Moreover, a linear term of the energy was computed for each dataset using a linear regression model that uses the chemical concentration of each natural element as regressor. The linear term predicted by the linear regression model has been subtracted from each original energy value to perform a re-alignment of the energy values across different electronic structures approximation theories performed to generate the diverse multi-source, multi-fidelity datasets. The folder "ADIOS_files" contains the set of pre-processed datasets in Adaptable I/O System (ADIOS) format (https://www.exascaleproject.org/research-project/adios/) that have been used for the development and training of GFMs in this work. The "ADIOS_files" directory contains 6 sub-directories named as follows: - ANI1x-v3.bp - MPTrj-v3.bp - OC2020-20M-v3.bp - OC2020-v3.bp - OC2022-v3.bp - qm7x-v3.bp Each sub-directory contains the pre-processed datasets converted in Adaptable I/O System (ADIOS) format (https://www.exascaleproject.org/research-project/adios/) that have been used to the development, training, and performance testing of the ensemble go predictive graph foundation models. Each GFM was developed using HydraGNN (https://github.com/ORNL/HydraGNN) as underlying graph neural network (GNN) architecture. The multi-task learning (MTL) capability of HydraGNN was used to simultaneously train the GFMs on labeled values for direct predictions of energy (a total system property of an atomistic structure that measures the chemical stability) and atomic forces (an atomic level property of an atomistic structure that measures the dynamical stability). The hyper parameters of the GFM have been tuned using scalable hyperparameter optimization (HPO) algorithms implemented in the software DeepHyper (https://github.com/deephyper/deephyper). The pre-training of each HPO trial was performed using distributed data parallelism (DDP) to scale the training across 128 compute nodes of the exascale OLCF supercomputer Frontier. Each HPO trial was trained only for 10 epochs and an early stopping was performed to avoid wasting significant computational resources on GNN architectures that were clearly underperforming. For each HPO trial, the 'omnistat' tool developed by (AMD Research - Advanced Micro Device) was used to measure the total energy consumption in kWh. The ensemble of GFMs was obtained by selecting the fifteen best performing HPO trials. Four models have been selected for their clear advantage in accuracy, and these are the GFMs with IDs 229, 156, 147, 260. Additional eleven models have been selected based on judicious balance between accuracy and energy consumption needed for training, and these are the GFMs with IDs 165, 78, 137, 1, 175, 171, 181, 67, 179, 167, 351. Each selected GFM of the ensemble was continued to cumulate a total of at most 30 epochs. In some cases, the total number of epochs actually performed was les than 30 due to two combined factors: (1) the size of the GFM (i.e., the number of model parameters to train) and (2) the total wall-clock time for which the computational resources could be allocated on OLCF-Frontier. The "Ensemble_of_models" directory contains 15 sub-directories named as follows: - gfm_0.229 - gfm_0.156 - gfm_0.147 - gfm_0.260 - gfm_0.165 - gfm_0.78 - gfm_0.137 - gfm_0.1 - gfm_0.175 - gfm_0.171 - gfm_0.181 - gfm_0.67 - gfm_0.179 - gfm_0.167 - gfm_0.351 Each one of these sub-directories refers to one of the fifteen HPO trials that have been selected to continue the pre-training with at most 30 epochs. With each sub-directory associated with a specific HPO trial, the following files can be found: - config.json: file for argument parsing to develop and train an HydraGNN architecture - gfm_0.ID_epoch_N.pk: file with model parameters for HPO ID trial after N epochs of training The ensemble of fifteen GFM architectures was used for (1) ensemble averaging to stabilize the predictions of energy and atomic forces after pre-training for post-processing analysis and (2) ensemble uncertainty quantification (UQ). The code used to develop, pre-train, and load the pre-trained models for post-processing analysis is available on the ORNL-GitHub at the following link: https://github.com/ORNL/HydraGNN/tree/Predictive_GFM_2024

36 MATERIALS SCIENCE↗
DOE OSTI2021

OPFLearn.jl [SWR-21-109]

OPFLearn.jl is a Julia package for creating datasets for machine learning approaches to solving AC optimal power flow (AC OPF). It was developed to provide researchers with a standardized way to efficiently create AC OPF datasets that are representative of more of the AC OPF feasible load space compared to typical dataset creation methods. The OPFLearn dataset creation method uses a relaxed AC OPF formulation to reduce the volume of the unclassified input space throughout the dataset creation process. Over time this input space tightens around the relaxed AC OPF feasible region to increase the percentage of feasible load profiles found while uniformly sampling the input space. Load samples are processed using AC OPF formulations from PowerModels.jl. More information on the dataset creation method can be found in our publication, "OPF-Learn: An Open-Source Framework for Creating Representative AC Optimal Power Flow Datasets". To use OPFLearn.jl a PowerModels network data dictionary is required (can be loaded from Matpower ".m" files) to define the network the dataset is being created for.

Joswig-Jones, Trager↗
DOE OSTI2021

OPFLearn.jl v0.1.2 5/18/2023 [SWR-21-109]

OPFLearn.jl is a Julia package for creating datasets for machine learning approaches to solving AC optimal power flow (AC OPF). It was developed to provide researchers with a standardized way to efficiently create AC OPF datasets that are representative of more of the AC OPF feasible load space compared to typical dataset creation methods. The OPFLearn dataset creation method uses a relaxed AC OPF formulation to reduce the volume of the unclassified input space throughout the dataset creation process. Over time this input space tightens around the relaxed AC OPF feasible region to increase the percentage of feasible load profiles found while uniformly sampling the input space. Load samples are processed using AC OPF formulations from PowerModels.jl. More information on the dataset creation method can be found in our publication, "OPF-Learn: An Open-Source Framework for Creating Representative AC Optimal Power Flow Datasets". To use OPFLearn.jl a PowerModels network data dictionary is required (can be loaded from Matpower ".m" files) to define the network the dataset is being created for.

Joswig-Jones, Trager↗
NASA NTRS2007

System and method of designing a load bearing layer of an inflatable vessel

A computer-implemented method is provided for designing a restraint layer of an inflatable vessel. The restraint layer is inflatable from an initial uninflated configuration to an inflated configuration and is constructed from a plurality of interfacing longitudinal straps and hoop straps. The method involves providing computer processing means (e.g., to receive user inputs, perform calculations, and output results) and utilizing this computer processing means to implement a plurality of subsequent design steps. The computer processing means is utilized to input the load requirements of the inflated restraint layer and to specify an inflated configuration of the restraint layer. This includes specifying a desired design gap between pairs of adjacent longitudinal or hoop straps, whereby the adjacent straps interface with a plurality of transversely extending hoop or longitudinal straps at a plurality of intersections. Furthermore, an initial uninflated configuration of the restraint layer that is inflatable to achieve the specified inflated configuration is determined. This includes calculating a manufacturing gap between pairs of adjacent longitudinal or hoop straps that correspond to the specified desired gap in the inflated configuration of the restraint layer.

Spexarth, Gary R.↗
NASA NTRS1992

Environmental fatigue of an Al-Li-Cu alloy. Part 3: Modeling of crack tip hydrogen damage

Environmental fatigue crack propagation rates and microscopic damage modes in Al-Li-Cu alloy 2090 (Parts 1 and 2) are described by a crack tip process zone model based on hydrogen embrittlement. Da/dN sub ENV equates to discontinuous crack advance over a distance, delta a, determined by dislocation transport of dissolved hydrogen at plastic strains above a critical value; and to the number of load cycles, delta N, required to hydrogenate process zone trap sites that fracture according to a local hydrogen concentration-tensile stress criterion. Transgranular (100) cracking occurs for process zones smaller than the subgrain size, and due to lattice decohesion or hydride formation. Intersubgranular cracking dominates when the process zone encompasses one or more subgrains so that dislocation transport provides hydrogen to strong boundary trapping sites. Multi-sloped log da/dN-log delta K behavior is produced by process zone plastic strain-hydrogen-microstructure interactions, and is determined by the DK dependent rates and proportions of each parallel cracking mode. Absolute values of the exponents and the preexponential coefficients are not predictable; however, fractographic measurements theta sub i coupled with fatigue crack propagation data for alloy 2090 established that the process zone model correctly describes fatigue crack propagation kinetics. Crack surface films hinder hydrogen uptake and reduce da/dN and alter the proportions of each fatigue crack propagation mode.

Piascik, Robert S.↗
DOE OSTI2025

Identifying Challenges in Safeguards for Metallic Fuel Fabrication Facilities

As new advanced reactors gain popularity, there is an increasing interest in metallic fuel fabrication for fast reactors. While metallic fuels themselves are not a new idea, as many of the first reactors employed metallic fuels, new designs, compositions, and fabrication methods are appearing throughout the nuclear community. As the interest grows and facilities are constructed, both domestic and international safeguards will need to be heavily involved to support safeguards-by-design (SBD) measures from the start. This work compiles a review of historical and modern fuel types and fabrication methods, fabrication processes, safeguards gaps, and potential safeguards solutions. Metallic nuclear fuel types have been around for many decades and were included in some of the first reactors including the Experimental Breeder Reactor (EBR)-I and -II, the Fermi 1 reactor, the Integral Fast Reactor (IFR), and the Dounreay Fast Reactor (DFR). These reactors used various compositions including pure uranium (U) metal, U-zirconium (Zr) alloys, plutonium (Pu)-aluminum (Al) alloys, U-fissium (Fs) alloys, U-Pu-Zr alloys, and U-molybdenum (Mo) alloys [1, 2, 3, 4, 5]. These small alloying additions are included to improve the material properties of the pure U metal. The alpha-phase U (stable below 661C) suffers elongation in one direction causing grain boundary cracking and increasing creep rate due to irradiation growth, thermal cycling, and preferential crystal orientation. It is ideal to utilize the gamma-phase U (typically stable above 769C) by adding small amounts of alloying elements such as Zr or Mo to stabilize this phase down to room temperature [3]. Additionally, some research has been focused on U with transuranic (TRU) elements present, typically coming from the used fuel recycling process. Including these elements in fast reactor fuel can aid in the reduction of nuclear waste by burning minor long-lived actinides. However, the additions of TRU elements can cause concerns to arise when trying to fabrication or safeguard metallic fuels. A typical metallic fuel element is shown in Figure 1. Sodium is added into the cladding to create a thermal bond between the fuel slug and cladding wall. The fuel slug is then inserted and the end plug is welded on to the top of the fuel element. A gas plenum is left to create a headspace for gaseous fission products to escape rather than continue to build in the fuel itself [1, 5]. Other fuel element geometries exist as well, such as the Lightbridge twisted cruciform geometry shown in Figure 2 [6]. This design allows for better cooling performance and provides room for fuel rod swelling without impacting the fuel rod diameter. There are many different fabrication methods for metallic fuels, which is one of the many benefits of these fuel types. Many of these fabrication methods are relatively easy and cost-efficient. The most popular fabrication method is injection casting, sometimes called vacuum induction melting (VIM), shown in Figure 3 [4, 8, 9, 7, 10]. This method was largely used for EBR-II fuel fabrication. The injection casting system is contained inside of a vessel consisting of a Y2O3-coated graphite crucible surrounded by an induction coil with ZrO2-coated quartz molds suspended above the crucible. The fuel feedstock is placed inside of the graphite crucible and melted using the induction furnace. The induction furnace utilizes a dual frequency with the high frequency melting the feedstock and the low frequency causing stirring of the melted feedstock to form a homogeneous mixture. The mixture is heated to approximately 1600C in an argon environment. The vessel is evacuated and then the quartz molds are lowered into the graphite crucible containing the molten metal and the vessel is repressurized to inject the metal fuel upwards into the molds. The molds are removed and then shattered to release the fuel slugs. This fabrication method was used to fabricate 39,000 metallic fuel pins for EBR-II. While injection casting has been the most common metallic fuel fabrication method throughout the decades, many other methods have been explored including low-pressure gravity casting, microwave casting, continuous casting, centrifugal casting, coextrusion, and many others [11, 12, 8, 13, 14, 15]. Some of these methods aim to mitigate challenges that arise with americium (Am) volatilization during the casting process for TRU-containing fuel feedstocks, an issue with injection casting. Coextrusion is one of the methods explored at the Idaho National Laboratory (INL) and has been utilized for the initial fabrication tests of Lightbridge's unique fuels, as well as other metallic fuels with cladding coextruded. In this process, large billets are formed and machined and then inserted into a molten salt bath for approximately 30 minutes. The billets are then loaded into the extrusion press and extruded. This process can be seen in Figure 4 [15].

11 - NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2020

Experimental investigation of multi-scale strain-field heterogeneity in rocks

Rocks have inherently heterogeneous micro-structures. The mechanical response of these heterogeneous micro-structures primarily governs the overall macroscopic behavior of intact rocks, as these features cause heterogeneity in the distribution of stress and strain within a rock volume. These attributes have generally been studied through numerical tools, but in this study, the stress-induced multi-scale spatial heterogeneity in the strain-field of three types of rocks was analyzed experimentally. The full-field non-contact strain measurement approach of 2D Digital Image Correlation (DIC) was employed for real-time evaluation of the progression of strain-field heterogeneity across the surface of the rock specimens when subjected to uniaxial compression. The results show that the strain-field heterogeneity is a characteristic of the inherent heterogeneity of the rock specimens, and that it amplifies with the magnitude of stress applied on the rock specimens. The changes in the strain-field heterogeneity with loading were consistent with established microcrack evolution processes in the rock specimens. The statistical fluctuation in the strain-field heterogeneity of the three rocks at varying gauge lengths (GL) was then analyzed for the determination of the Representative Volume Element (RVE) length-scale at increasing load levels. The evolution of the RVE length-scale with loading facilitates an objective way of determination of the stress at which the strain-field heterogeneity increases and begins to influence the overall global response of the material. The RVE length-scale variation with loading showed that at the RVE length-scale reaches the specimen-scale at load levels well below the peak strength.

2D-digital image correlation↗
DOE OSTI2023

Unveiling the Fatigue Behavior of 2D Hybrid Organic–Inorganic Perovskites: Insights for Long-Term Durability

2D hybrid organic–inorganic perovskites (HOIPs) are commonly found under subcritical cyclic stresses and suffer from fatigue issues during device operation. However, their fatigue properties remain unknown. Here, the fatigue behavior of (C 4 H 9 -NH 3 ) 2 (CH 3 NH 3 ) 2 Pb 3 I 10 , the archetype 2D HOIP, is systematically investigated by atomic force microscopy (AFM). It is found that 2D HOIPs are much more fatigue resilient than polymers and can survive over 1 billion cycles. 2D HOIPs tend to exhibit brittle failure at high mean stress levels, but behave as ductile materials at low mean stress levels. These results suggest the presence of a plastic deformation mechanism in these ionic 2D HOIPs at low mean stress levels, which may contribute to the long fatigue lifetime, but is inhibited at higher mean stresses. The stiffness and strength of 2D HOIPs are gradually weakened under subcritical loading, potentially as a result of stress-induced defect nucleation and accumulation. The cyclic loading component can further accelerate this process. The fatigue lifetime of 2D HOIPs can be extended by reducing the mean stress, stress amplitude, or increasing the thickness. These results can provide indispensable insights into designing and engineering 2D HOIPs and other hybrid organic–inorganic materials for long-term mechanical durability.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2023

Techno-economic assessment and life cycle assessment of three potential pathways for biomass liquefaction

Biomass liquefaction can enhance corn-stover slurry transportation by reducing yield stress and overall viscosity, easing biomass handling, especially for slurries at high solids loading (≥30 % (w/v). Implementing a liquefaction process at the front-end of the biorefinery could reduce downtime caused by equipment fouling. Furthermore, liquefaction reduces process complexity by transforming high fiber mass into a pumpable slurry. Three scenarios for liquefaction of corn stover pellets: enzyme process, enzyme mimetic process, and a combined process of enzyme and enzyme mimetic were assessed. The ideal process will have the best performance-to-cost ratio, with minimal impact on the environment. Techno-economic and life-cycle analyses were used to understand the economic and environmental effects of the three liquefaction scenarios and to determine if one or more can achieve cost parity. Here, results show that the enzyme process had the best performance in terms of cost and emissions (gCO 2 -equivalent/kg slurry)compared to the other scenarios.

09 BIOMASS FUELS↗
DOE OSTI2025

Biogenic carbon capture at pulp mills via sodium spiking and oxy-fuel calcination

Over 13 million metric tons of biogenic carbon dioxide (CO 2 ) are mineralized yearly in United States (US) pulp mill recovery boilers as molten sodium carbonate. The mineralized CO 2 is released downstream in the rotary lime kiln where it can be captured at a relatively low cost for permanent sequestration. The use of biocarbon and bioenergy at pulp mills enables the possibility for atmospheric carbon removal when biogenic CO2 is captured and sequestered. We demonstrate the feasibility of capturing biogenic CO 2 via sodium spiking coupled with oxy-fuel calcination in the rotary lime kiln at existing kraft pulp mills. Sodium (Na) spiking elevates the Na ion loading in the kraft chemical looping process by replacing chlorine-based bleaching with a highly alkaline bleaching sequence. Chemical pulping processes are modeled and simulated to understand the technical limits of implementing sodium spiking in existing pulp mills. Each 1 % of sodium added to the bleaching operations increases the rate of CO 2 mineralization by 4 %, and the maximum increase in sodium content in the kraft process is 9 %. Without sodium spiking, estimated CO 2 capture costs are $131 and $107/mt CO 2 for air and oxy-fuel combustion, respectively. With the implementation of sodium spiking, the cost of CO 2 captured decreases by 27 % and 31 %, showing costs of $96 and $74/mt CO 2 for air and oxy-fuel combustion, respectively. Oxy-fuel combustion reduces the costs of CO 2 capture compared to air combustion, but merged deployment with sodium spiking seems to be the most cost-effective pathway to integrate carbon capture.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2022

Global Patterns and Controls of Nutrient Immobilization on Decomposing Cellulose in Riverine Ecosystems

Microbes play a critical role in plant litter decomposition and influence the fate of carbon in rivers and riparian zones. When decomposing low-nutrient plant litter, microbes acquire nitrogen (N) and phosphorus (P) from the environment (i.e., nutrient immobilization), and this process is potentially sensitive to nutrient loading and changing climate. Nonetheless, environmental controls on immobilization are poorly understood because rates are also influenced by plant litter chemistry, which is coupled to the same environmental factors. Here we used a standardized, low-nutrient organic matter substrate (cotton strips) to quantify nutrient immobilization at 100 paired stream and riparian sites representing 11 biomes worldwide. Immobilization rates varied by three orders of magnitude, were greater in rivers than riparian zones, and were strongly correlated to decomposition rates. In rivers, P immobilization rates were controlled by surface water phosphate concentrations, but N immobilization rates were not related to inorganic N. The N:P of immobilized nutrients was tightly constrained to a molar ratio of 10:1 despite wide variation in surface water N:P. Immobilization rates were temperature-dependent in riparian zones but not related to temperature in rivers. However, in rivers nutrient supply ultimately controlled whether microbes could achieve the maximum expected decomposition rate at a given temperature. Collectively, we demonstrated that exogenous nutrient supply and immobilization are critical control points for decomposition of organic matter.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Deep Reinforcement Learning for Distribution System Cyber Attack Defense with DERs

The use of smart inverter capabilities of distributed energy resources (DERs) enhances the grid reliability but in the meanwhile exhibits more vulnerabilities to cyber-attacks. This paper proposes a deep reinforcement learning (DRL)-based defense approach. The defense problem is reformulated as a Markov decision making process to control DERs and minimizing load shedding to address the voltage violations caused by cyber-attacks. The original soft actor-critic (SAC) method for continuous actions has been extended to handle discrete and continuous actions for controlling DERs' setpoints and loadshedding scenarios. Numerical comparison results with other control approaches, such as Volt-VAR and Volt-Watt on the modified IEEE 33-node, show that the proposed method can achieve better voltage regulation and have less power losses in the presence of cyber-attacks.

active distribution systems↗
DOE OSTI2020

Optimal Energy Storage Schedules for Load Leveling and Ramp Rate Control in Distribution Systems

Continued integration of distributed energy resources (DERs) into the grid, such as solar PVs, at a large-scale, contributes into the famous Duck Curve. New DER management algorithms are therefore deemed necessary to alleviate rapid variations within net load profiles of distribution systems. Herein this paper proposes a process to determine the optimal energy storage schedules for leveling the distribution circuit feederhead net load. A series of sensitivity analyses shows how the proposed method can be used to determine the optimal energy storage schedules with different capacities, state of charge requirements, and net load ramp rate limitations.

25 ENERGY STORAGE↗
DOE OSTI2022

Indirect Tool Condition Monitoring Using Ensemble Machine Learning Techniques

Abstract Tool condition monitoring (TCM) has become a research area of interest due to its potential to significantly reduce manufacturing costs while increasing process visibility and efficiency. Machine learning (ML) is one analysis technique which has demonstrated advantages for TCM applications. However, the commonly studied individual ML models lack generalizability to new machining and environmental conditions, as well as robustness to the unbalanced datasets which are common in TCM. Ensemble ML models have demonstrated superior performance in other fields, but have only begun to be evaluated for TCM. As a result, it is not well understood how their TCM performance compares to that of individual models, or how homogeneous and heterogeneous ensemble models’ performances compare to one another. To fill in these research gaps, milling experiments were conducted using various cutting conditions, and the model groups were compared across several performance metrics. Statistical t-tests were also used to evaluate the significance of model performance differences. Through the analysis of four individual ML models and five ensemble models, all based on the processes’ sound, spindle power, and axial load signals, it was found that on average, the ensemble models performed better than the individual models, and that the homogeneous ensembles outperformed the heterogeneous ensembles.

Engineering↗
DOE OSTI2022

Mechanical Properties of High-Temperature Fiber-Reinforced Thermoset Composites with Plain Weave and Unidirectional Carbon Fiber Fillers

Fiber-reinforced thermoset composites are a class of materials that address the arising needs from the aerospace and hypersonic industries for high specific strength, temperature-resistant structural materials. Among the high-temperature resistant thermoset categories, phenolic triazine (PT) cyanate esters stand out thanks to their inherent high degradation temperature, glass transition temperature, and mechanical strength. Despite the outstanding properties of these thermosets, the performance of carbon fiber composites using PT cyanate esters as matrices has not been thoroughly characterized. This work evaluated PT and carbon fiber composites’ compressive properties and failure mechanisms with different fiber arrangements. A PT resin with both plain weave (PW) and non-crimped unidirectional (UD) carbon fiber mats was analyzed in this research. Highly loaded thermoset composites were obtained using process temperatures not exceeding 260 °C, and the composites proved to retain compressive strength at temperatures beyond 300 °C. Compressive testing revealed that PT composites retained compressive strength values of 50.4% of room temperature for UD composites and 61.4% for PW composites. Post-compressive failure observations of the gage section revealed that the mechanisms for failure evolved with temperature from brittle, delamination-dominant failure to shear-like failure promoted by the plastic failure of the matrix. This study demonstrated that PT composites are a good candidate for structural applications in harsh environments.

Hall, Samuel Ernesto↗
NASA NTRS1972

Stress and buckling analysis of a double-bubble tank structure

An analytical study of a two-lobed circular arc pressure vessel (double-bubble tank) with a local perturbation from the cylindrical shape was made. It can be concluded that NASTRAN provides a good linear buckling analysis capability for structures that cannot be handled by other analysis techniques. However, solutions are relatively expensive from a computational standpoint. Two worthwhile additions to the NASTRAN program would be the inclusion of a new beam element that adequately represents the shear transfer when modeling stiffeners on a shell and the capability, in a rigid format, to keep certain prescribed loads constant during the eigenvalue extraction process.

Robinson, J. C.↗
NASA NTRS1975

Growth of surface and corner cracks in beta-processed and mill-annealed Ti-6Al-4V

Empirical stress-intensity expressions were developed to relate the growth of cracks from corner flaws to the growth of cracks from surface flaws. An experimental program using beta-processed Ti-6Al-4V verified these expressions for stress ratios, R greater than or equal to 0. An empirical crack growth-rate expression which included stress-ratio and stress-level effects was also developed. Cracks grew approximately 10 percent faster in transverse-grain material than in longitudinal-grain material and at approximately the same rate in longitudinal-grain mill-annealed Ti-6Al-4V. Specimens having surface and corner cracks and made of longitudinal-grain, beta-processed material were tested with block loads, and increasing the stresses in a block did not significantly change the crack growth rates. Truncation of the basic ascending stress sequence within a block caused more rapid crack growth, whereas both the descending and low-to-high stress sequences slowed crack growth.

Bell, P. D.↗