Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Prediction Accuracy”

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 1,135 records · Page 63
All sourcesNASA NTRSDOE OSTI
DOE OSTI2020

Optimizing Simulation Parameters for Weak Lensing Analyses Involving Non-Gaussian Observables

We performed a series of numerical experiments to quantify the sensitivity of the predictions for weak lensing statistics obtained in ray-tracing dark matter (DM)-only simulations, to two hyper-parameters that influence the accuracy as well as the computational cost of the predictions: the thickness of the lens planes used to build past light cones and the mass resolution of the underlying DM simulation. The statistics considered are the power spectrum (PS) and a series of non-Gaussian observables, including the one-point probability density function, lensing peaks, and Minkowski functionals. Counterintuitively, we find that using thin lens planes (< 60 h {sup −1} Mpc on a 240 h {sup −1} Mpc simulation box) suppresses the PS over a broad range of scales beyond what would be acceptable for a survey comparable to the Large Synoptic Survey Telescope (LSST). A mass resolution of 7.2 × 10{sup 11} h {sup −1} M {sub ⊙} per DM particle (or 256{sup 3} particles in a (240 h {sup −1} Mpc){sup 3} box) is sufficient to extract information using the PS and non-Gaussian statistics from weak lensing data at angular scales down to 1′ with LSST-like levels of shape noise.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2021

An initial framework for the rapid qualification of long-term creep rupture strength via microstructural modeling

This report describes the development and testing of a new method for extrapolating short-term creep rupture test data to predict long-term rupture strength. The goal of this work is to reduce the time required to qualify new materials for nuclear service by reducing the lead time required for dedicated, long-term material testing to establish key long-term material properties. The new approach described here uses a physics-based model to predict the long-term creep rupture strength of 316H stainless steel using only short-term test data. The key idea is to use Bayesian inference to find the statistical distribution of the model parameters that best explain the short-term rupture data. Because the model is physics-based these parameters are all microstructural quantities that can be measured through detailed material characterization experiments. The Bayesian prior distributions provide a means for incorporating this characterization data into the final model to improve the accuracy of the long-term model predictions. However, where such data is not available the process still produces an accurate model based on an uniformed prior. Our hypothesis is that this approach more accurately extrapolates the short-term test data when compared to current, empirical methods. The report proves this hypothesis using actual long-term rupture data available for 316H, including tests with rupture times greater than 200,000 hours. The general approach developed here could be applied to other materials and other time-dependent material properties. Applying this new technique to develop long-term qualified material properties, potentially in conjunction with other accelerated qualification approaches like staggered qualification test programs, could greatly reduce the time required to qualify new materials for nuclear service.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2025

Developing a robust strength model using physically-informed genetic programming

The strength of materials is influenced by a range of external conditions, such as temperature and deformation rate. Consequently, materials that demonstrate substantial variations in their mechanical behavior due to fluctuations in temperature and strain rate require complex strength models to accurately predict material performance in real-world applications. To predict such complex behavior, a robust and flexible strength model is necessary. In this work, we utilize genetic programming-based symbolic regression (GPSR) to develop data-driven strength models that accurately represent the measured stress–strain responses of tin across a wide range of strain, strain rate and temperature regimes. The GPSR models are constrained by physically-informed conditions, which leads to significant improvement in extrapolation. The best model is integrated into a multi-physics code to perform Taylor impact simulations, validating the model’s accuracy and robustness. In conclusion, the model predictions showed excellent agreement with experimental results, particularly when compared to predictions using traditional strength models.

Genetic programming↗
DOE OSTI2020

Comparison of linear regression, k-nearest neighbour and random forest methods in airborne laser-scanning-based prediction of growing stock

Abstract In this study, for five sites around the world, we look at the effects of different model types and variable selection approaches on forest yield modelling performances in an area-based approach (ABA). We compared ordinary least squares regression (OLS), k-nearest neighbours (kNN) and random forest (RF). Our objective was to test if there are systematic differences in accuracy between OLS, kNN and RF in ABA predictions of growing stock volume. The analyses are based on a 5-fold cross-validation at five study sites: an eucalyptus plantation, a temperate forest and three different boreal forests. Two completely independent validation datasets were also available for two of the boreal sites. For the kNN, we evaluated multiple measures of distance including Euclidean, Mahalanobis, most similar neighbour (MSN) and an RF-based distance metric. The variable selection approaches we examined included a heuristic approach (for OLS, kNN and RF), exhaustive search among all combinations (OLS only) and all variables together (RF only). Performances varied by model type and variable selection approaches among sites. OLS and RF had similar accuracies and were more efficient than any of the kNN variants. Variable selection did not affect RF performance. Heuristic and exhaustive variable selection performed similarly for OLS. kNN fared the poorest amongst model types, and kNN with RF distance was prone to overfitting when compared with a validation dataset. Additional caution is therefore required when building kNN models for volume prediction though ABA, being preferable instead to opt for models based on OLS with some variable selection, or RF with all variables together.

Cosenza, Diogo N.↗
DOE OSTI2021

Deep-Learning-Derived Evaluation Metrics Enable Effective Benchmarking of Computational Tools for Phosphopeptide Identification

Tandem mass spectrometry (MS/MS)-based phosphoproteomics is a powerful technology for global phosphorylation analysis. However, applying four computational pipelines to a typical mass spectrometry (MS)-based phosphoproteomic dataset from a human cancer study, we observed a large discrepancy among the reported phosphopeptide identification and phosphosite localization results, underscoring a critical need for benchmarking. While efforts have been made to compare performance of computational pipelines using data from synthetic phosphopeptides, evaluations involving real application data have been largely limited to comparing the numbers of phosphopeptide identifications due to the lack of appropriate evaluation metrics. We investigated three deep learning-derived features as potential evaluation metrics: phosphosite probability, Delta RT and spectral similarity. Predicted phosphosite probability is computed by MusiteDeep, which provides high accuracy as previously reported; Delta RT is defined as the absolute retention time (RT) difference between RTs observed and predicted by AutoRT; and spectral similarity is defined as the Pearson’s correlation coefficient between spectra observed and predicted by pDeep2. Using a synthetic peptide dataset, we found that both Delta RT and spectral similarity provided excellent discrimination between correct and incorrect peptide-spectrum matches (PSMs) both when incorrect PSMs involved wrong peptide sequences and even when incorrect PSMs were caused by only incorrect phosphosite localization. Based on these results, we used all the three deep learning-derived features as evaluation metrics to compare different computational pipelines on diverse set of phosphoproteomic datasets and showed their utility in benchmarking performance of the pipelines. The benchmark metrics demonstrated in this study will enable users to select computational pipelines and parameters for routine analysis of phosphoproteomics data and will offer guidance for developers to improve computational methods.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2021

Autonomous nondestructive evaluation of resistance spot welded joints

The application of non-destructive evaluation approaches has attracted strong interests in modern automotive industries. Here, we present an autonomous deep-computing framework to analyze raw videos from infrared systems and to predict weld nugget shape and size with unprecedented accuracy and speed. In a comprehensive training and testing experiment with 90 videos (seven sets of welding material stack-ups), a new method was developed to assemble sufficient datasets for neural network training. Our framework successfully predicts all the nugget shapes with F1 scores that range from 0.84 to 0.92. The total training time on Nvidia DGX station takes less than 10 min for each set of welding material stack-up. The real inference time of an individual dataset (with 30 video frames) takes about 0.005 s. The procedure and methods developed in the study can be applied to other image-based weld property prediction, as well as other manufacturing processes. Furthermore, our well-trained neural networks take limited memory resources (2.3 MB) and are suitable for embedded microprocessors for in-situ welding quality control as edge computing within an intelligent welding framework.

42 ENGINEERING↗
DOE OSTI2021

Modeling and simulation of transitional Rayleigh–Taylor flow with partially averaged Navier–Stokes equations

In this work, the partially averaged Navier–Stokes (PANS) equations are used to predict the variable-density Rayleigh–Taylor (RT) flow at Atwood number 0.5 and maximum Reynolds number 500. This is a prototypical problem of material mixing, featuring laminar, transitional, and turbulent flow, instabilities and coherent structures, density fluctuations, and production of turbulence kinetic energy by both shear and buoyancy mechanisms. These features pose numerous challenges to modeling and simulation, making the RT flow ideal to develop the validation space of the recently proposed PANS Besnard–Harlow–Rauenzahn-linear eddy viscosity model closure. The numerical simulations are conducted at different levels of physical resolution and test three approaches to set the parameters $f_\phi$ defining the range of physically resolved scales. The computations demonstrate the efficiency (accuracy vs cost) of the PANS model predicting the spatiotemporal development of the RT flow. Results comparable to large-eddy simulations and direct numerical simulations are obtained at significantly lower physical resolution without the limitations of the Reynolds-averaged Navier–Stokes equations in these transitional flows. The data also illustrate the importance of appropriate selection of the physical resolution and the resolved fraction of each dependent quantity $\phi$ of the turbulent closure, $f_\phi$. These two aspects determine the ability of the model to resolve the flow phenomena not amenable to modeling by the closure and, as such, the computations’ fidelity.

Navier Stokes equations↗
DOE OSTI2021

Cellular, molecular, and therapeutic characterization of pilocarpine-induced temporal lobe epilepsy

Animal models have expanded our understanding of temporal lobe epilepsy (TLE). However, translating these to cell-specific druggable hypotheses is not explored. Herein, we conducted an integrative insilico-analysis of an available transcriptomics dataset obtained from animals with pilocarpine-induced-TLE. A set of 119 genes with subtle-to-moderate impact predicted most forms of epilepsy with ~ 97% accuracy and characteristically mapped to upregulated homeostatic and downregulated synaptic pathways. The deconvolution of cellular proportions revealed opposing changes in diverse cell types. The proportion of nonneuronal cells increased whereas that of interneurons, except for those expressing vasoactive intestinal peptide (Vip), decreased, and pyramidal neurons of the cornu-ammonis (CA) subfields showed the highest variation in proportion. A probabilistic Bayesian-network demonstrated an aberrant and oscillating physiological interaction between nonneuronal cells involved in the blood–brain-barrier and Vip interneurons in driving seizures, and their role was evaluated insilico using transcriptomic changes induced by valproic-acid, which showed opposing effects in the two cell-types. Additionally, we revealed novel epileptic and antiepileptic mechanisms and predicted drugs using causal inference, outperforming the present drug repurposing approaches. These well-powered findings not only expand the understanding of TLE and seizure oscillation, but also provide predictive biomarkers of epilepsy, cellular and causal micro-circuitry changes associated with it, and a drug-discovery method focusing on these events.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2023

Deep neural network uncertainty quantification for LArTPC reconstruction

We evaluate uncertainty quantification (UQ) methods for deep learning applied to liquid argon time projection chamber (LArTPC) physics analysis tasks. As deep learning applications enter widespread usage among physics data analysis, neural networks with reliable estimates of prediction uncertainty and robust performance against overconfidence and out-of-distribution (OOD) samples are critical for their full deployment in analyzing experimental data. While numerous UQ methods have been tested on simple datasets, performance evaluations for more complex tasks and datasets are scarce. Here we assess the application of selected deep learning UQ methods on the task of particle classification using the PiLArNet monte carlo 3D LArTPC point cloud dataset. We observe that UQ methods not only allow for better rejection of prediction mistakes and OOD detection, but also generally achieve higher overall accuracy across different task settings. We assess the precision of uncertainty quantification using different evaluation metrics, such as distributional separation of prediction entropy across correctly and incorrectly identified samples, receiver operating characteristic curves (ROCs), and expected calibration error from observed empirical accuracy. We conclude that ensembling methods can obtain well calibrated classification probabilities and generally perform better than other existing methods in deep learning UQ literature.

47 OTHER INSTRUMENTATION↗
DOE OSTI2022

Confronting a set of Skyrme and $\chi _{EFT}$ predictions for the crust of neutron stars

With the improved accuracy of neutron star observational data, it is necessary to derive new equation of state where the crust and the core are consistently calculated within a unified approach. For this purpose we describe non-uniform matter in the crust of neutron stars employing a compressible liquid-drop model, where the bulk and the neutron fluid terms are given from the same model as the one describing uniform matter present in the core. We then generate a set of fifteen unified equations of state for cold catalyzed neutron stars built on realistic modelings of the nuclear interaction, which belongs to two main groups: the first one derives from the phenomenological Skyrme interaction and the second one from χEFT Hamiltonians. The confrontation of these model predictions allows us to investigate the model dependence for the crust properties, and in particular the effect of neutron matter at low density. Here, the new set of unified equations of state is available at the CompOSE repository.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2022

Improving Residential Building Simulations Through Large-Scale Empirical Validation

Residential building energy simulations are increasingly used for energy-efficient building design, codes and standards analysis, home certifications and ratings, utility programs, and technology assessments. Various software tools exist to perform residential building simulations, and these tools often use different models, inputs, and assumptions. This leads to inconsistencies that can undermine confidence in the predicted results. Validation of these tools can increase confidence by ensuring their accuracy and consistency. One way to validate simulation tools is through empirical testing, which compares predicted energy usage to measured utility billing data. This paper describes the process of data collection, data standardization, and empirical validation, and illustrates its use with our residential EnergyPlus (R)-based software. The data and process can be extended to other simulation tools and contribute to improving residential building simulations more broadly.

empirical validation↗
DOE OSTI2020

On The Use of Sectional Techniques for the Solution of Depolymerization Population Balances: Results on a Discrete-Continuous Mesh

To study the discrete bond-breaking phenomena of depolymerization, the use of a fully continuous Population Balance Equation (PBE) is inadequate to embody all the inherent characteristics of the process, thus resulting in the need for a discrete-continuous mesh. In this work, the performance of the three most state-of-the-art sectional techniques, i.e. the fixed pivot technique (FPT), cell average technique (CAT) and finite volume scheme (FVS) in approximating discrete depolymerization using discrete-continuous PBEs was extensively compared and evaluated. The solutions from these three methods show different accuracy depending on the breakage mechanisms. For chain-end scission, the FPT and the CAT satisfactorily predict the population densities and moments whereas the FVS fails to predict the population densities but preserves the zeroth and the first moments. In the application of a discrete-continuous model, we identified a previously-not-reported issue of a precipitous drop in the number density at the boundary of discrete and continuous region specifically for chain-end scission. We successfully fixed this problem by employing the alterations proposed in this paper, to the particle allocation functions at the boundary points. For random scission, all three sectional techniques predict the population densities and moments to a high degree of accuracy, even at a very coarse mesh, through the use of our new stoichiometric kernel which is able to closely approximate the inherently discrete bond-breaking depolymerization process. The assessments in this present work intends to provide a clear-cut direction to efficient and economical modelling of depolymerization processes.

Ahamed, Firnaaz↗
DOE OSTI2025

Data-Driven Kinetic Reaction Networks for Separation Chemistry

Understanding complex, multistep chemical reactions at the molecular level is a major challenge whose solution would greatly benefit the design and optimization of numerous chemical processes. The separation of rare-earth (4f) and actinide (5f) elements is an example where improving our chemical understanding is important for designing and optimizing new chemistries, even with a limited number of observations. Here, in this work, we leverage data-driven artificial intelligence and machine-learning approaches to develop kinetic reaction networks that describe the liquid–liquid extraction mechanism of uranium using N,N-di-2-ethylhexyl-isobutyramide (DEHiBA). Specifically, we compare and contrast the properties of two classes of models: (1) purely data-driven models that are regularized using chemistry-agnostic, L1 regression and (2) chemistry-informed models that are regularized using relative reaction energies provided by quantum mechanical calculations. We observe that purely data-driven models are unbiased, simple, and accurate in their predictions of experimental measurements when provided with sufficient data but are difficult to fully constrain and interpret. In contrast, chemistry-informed models exhibit significantly improved chemical interpretability and consistency, providing a detailed description of the separation process while achieving high accuracy through ensemble averaging. Overall, the dominant species predicted to be extracted into the organic phase is UO 2 (NO 3 ) 2 (DEHiBA) 2 , agreeing with experimental slope analysis, thermodynamic modeling, EXAFS, and crystal structures. This work demonstrates that leveraging the fundamental structure of the problem can lead to efficient learning schemes that provide both accurate predictions and chemical insights at a low computational cost.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2022

Prospects and Limitations of Predicting Fuel Ignition Properties from Low-Temperature Speciation Data

Using chemical kinetic modeling and statistical analysis, we investigate the possibility of correlating key chemical “markers”–typically small molecules–formed during very lean (φ ~ 0.001) oxidation experiments with near-stoichiometric (φ ~ 1) fuel ignition properties. One goal of this work is to evaluate the feasibility of designing a fuel-screening platform, based on small laboratory reactors that operate at low temperatures and use minimal fuel volume. Buras et al. [Combust. Flame2020,216, 472–484] have shown that convolutional neural net (CNN) fitting can be used to correlate first-stage ignition delay times (IDTs) with OH/HO2 measurements during very lean oxidation in low-T flow reactors with better than factor-of-2 accuracy. In this work, we test the limits of applying this correlation-based approach to predict the low-temperature heat release (LTHR) and total IDT, including the sensitivity of total IDT to the equivalence ratio, φ. We demonstrate that first-stage IDT can be reliably correlated with very lean oxidation measurements using compressed sensing (CS), which is simpler to implement than CNN fitting. LTHR can also be predicted via CS analysis, although the correlation quality is somewhat lower than for first-stage IDT. In contrast, the accuracy of total IDT prediction at φ = 1 is significantly lower (within a factor of 4 or worse). Furthermore, these results can be rationalized by the fact that the first-stage IDT and LTHR are primarily determined by low-temperature chemistry, whereas total IDT depends on low-, intermediate-, and high-temperature chemistry. Oxidation reactions are most important at low temperatures, and therefore, measurements of universal molecular markers of oxidation do not capture the full chemical complexity required to accurately predict the total IDT even at a single equivalence ratio. As a result, we find that φ-sensitivity of ignition delay cannot be predicted at all using solely correlation with lean low-T chemical speciation measurements.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2025

Physics-informed latent neural operator for real-time predictions of time-dependent parametric PDEs

Deep operator network (DeepONet) has shown significant promise as surrogate models for systems governed by partial differential equations (PDEs), enabling accurate mappings between infinite-dimensional function spaces. However, when applied to systems with high-dimensional input-output mappings arising from large numbers of spatial and temporal collocation points, these models often require heavily overparameterized networks, leading to long training times. Latent DeepONet addresses some of these challenges by introducing a two-step approach: first learning a reduced latent space using a separate model, followed by operator learning within this latent space. While efficient, this method is inherently data-driven and lacks mechanisms for incorporating physical laws, limiting its robustness and generalizability in data-scarce settings. Here, in this work, we propose PI-Latent-NO, a physics-informed latent neural operator framework that integrates governing physics directly into the learning process. Our architecture features two coupled DeepONets trained end-to-end: a Latent-DeepONet that learns a low-dimensional representation of the solution, and a Reconstruction-DeepONet that maps this latent representation back to the physical space. By embedding PDE constraints into the training via automatic differentiation, our method eliminates the need for labeled training data and ensures physics-consistent predictions. The proposed framework is both memory and compute-efficient, exhibiting near-constant scaling with problem size and demonstrating significant speedups over traditional physics-informed operator models. We validate our approach on a range of parametric PDEs, showcasing its accuracy, scalability, and suitability for real-time prediction in complex physical systems.

Latent representations↗
DOE OSTI2022

Solvation Thermodynamics of Solutes in Water and Ionic Liquids Using the Multiscale Solvation-Layer Interface Condition Continuum Model

We report molecular assembly processes are generally driven by thermodynamic properties in solutions. Atomistic modeling can be very helpful in designing and understanding complex systems, except that bulk solvent is very inefficient to treat explicitly as discrete molecules. The SLIC/CDC multiscale model combines continuum solvent electrostatics based on the solvent layer interface condition (SLIC) with statistical thermodynamic models for hydrogen bonding and nonpolar modes: cavity formation, dispersion interactions, combinatorial mixing (CDC). The SLIC/CDC model predicts Gibbs energies of solvation for a database of 500 solutes in water with average accuracy better than 1 kcal/mol both for experimental measurements and for explicit-solvent molecular dynamics simulations. The separate SLIC/CDC energy mode values agree quantitatively with those computed from explicit-solvent molecular dynamics. The SLIC/SASA multiscale model combines the SLIC continuum electrostatic model with the solvent accessible surface area (SASA) nonpolar energy mode. The SLIC/SASA model predicts Gibbs energies of solvation with better than 1.4 kcal/mol average accuracy in aqueous systems and better than 1.6 kcal/mol average accuracy in ionic liquids. Both models predict solvation entropies, and are the first implicit-solvation models capable of predicting solvation heat capacities.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2021

Data‐Driven Safety Risk Prediction of Lithium‐Ion Battery

Abstract Inevitable safety issues have pushed battery engineers to become more conservative in battery system design; however, battery‐involved accidents still frequently are reported in headlines. Identifying, understanding, and predicting safety risks have become priorities to further accelerate technology and industry development. However, diverse loading scenarios, significantly varied stress‐induced short circuit mechanisms, and highly coupled mechanical–electrochemical safety behaviors have remained grand challenges. Herein, the safety risk is termed as the probability of the mechanical triggering of an internal short circuit, to reflect the safety related behaviors of lithium‐ion batteries. Based on a mechanical model and experimental results, a sufficient dataset is generated consisting of strain states and their corresponding safety risks, covering both cylindrical and pouch cells, various states of charges, and loading conditions. Machine‐learning tools combined with the established finite element mechanical model are applied to predict the safety risks of the cells. The results achieve a high level of accuracy on the test data (the relative error of the average short circuit prediction deviation is less than 6.2%.). This work underpins the safety risk concept and highlights the promise of physics combined with data‐driven modeling methodology to predict the safety behaviors of energy storage systems.

Jia, Yikai↗
DOE OSTI2021

Thermal loading analysis of the ring injection dump for the Spallation Neutron Source facility

The ring injection dump (RID) is the largest beam dump in the Spallation Neutron Source facility at Oak Ridge National Laboratory and accepts a fraction of the beam from the ring that is not captured in the ring during injection. Thermo-fluid modeling of the RID components was performed during the initial design by using the software ANSYS CFX to characterize the power rating for the RID. From the calculations, the power rating was lowered from 200 kW to 150 kW, at 1.3 GeV beam energy due to concerns regarding the heating of the concrete structure. The Proton Power Upgrade project scope included reevaluating the power limit to the RID in anticipation that increasing it would provide more operational flexibility. The focus was on the shielding because it was the limiting factor, and data collected from thermocouples installed on the structure were used to benchmark the analysis. A three-stage process was adopted in the latest study. First, the already existing steady-state calculations were validated and extended with the latest Monte Carlo N-Particle source term calculations. Second, transient calculations were conducted to capture the dynamic state of the system in response to the energy dumped to the RID during the following operational period of the beam from 2005–2019 (14 years), and sensitivity analyses of crucial parameters were performed to benchmark the model with reasonable accuracy. Finally, an idealized transient cycle analysis was conducted with realistic duty factors to predict the temperature distribution at higher beam powers. Overall, the results confirmed the accuracy of the original steady state model but allowed the development of a new validated transient model with higher accuracy for future analyses.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗