Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Error Metrics”

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 235 records · Page 13
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

CrossCheck: Rapid, Reproducible, and Interpretable Model Evaluation

Evaluation beyond aggregate performance metrics, e.g. F1-score, is crucial to both establish an appropriate level of trust in machine learning models and identify future model improvements. In this paper we demonstrate CrossCheck, an interactive visualization tool for rapid crossmodel comparison and reproducible error analysis. We describe the tool and discuss design and implementation details. We then present three use cases (named entity recognition, reading comprehension, and clickbait detection) that show the benefits of using the tool for model evaluation. CrossCheck allows data scientists to make informed decisions to choose between multiple models, identify when the models are correct and for which examples, investigate whether the models are making the same mistakes as humans, evaluate models’ generalizability and highlight models’ limitations, strengths and weaknesses. Furthermore, CrossCheck is implemented as a Jupyter widget, which allows rapid and convenient integration into data scientists’ model development workflows.

Arendt, Dustin L.↗
DOE OSTI2023

Stochastic Framework for Optimal Control of Planetary Reentry Trajectories Under Multilevel Uncertainties

We present a novel stochastic optimal control framework that accounts for various types of uncertainties, with application to reentry trajectory planning. The formulation of the optimal trajectory control problem is presented in the context of an indirect method where a functional objective associated with the terminal vehicle speed is to be minimized. Uncertain input parameters in the optimal trajectory control model, including aerodynamic parameters and initial and terminal conditions, are modeled as aleatory random variables, while the statistical parameters of these aleatory distributions are themselves random variables. The parametric and model uncertainties are simultaneously propagated through an extended polynomial chaos expansion (EPCE) formalism. Several metrics are described to evaluate response statistics and presented as insightful tools for robust decision making. Specifically, the response probability density function (PDF) reflecting influence of both epistemic and aleatory uncertainties is obtained. By sampling over the random variables representing model error, an ensemble of response PDFs is generated and the associated failure probability is estimated as a random variable with its own polynomial chaos expansion. Besides, the sensitivity index functions of response PDF with respect to the statistical parameters are evaluated. Coupling parametric and model uncertainties within the EPCE framework leads to a robust and efficient paradigm for multilevel uncertainty propagation and PDF characterization in general optimal control problems.

Engineering↗
DOE OSTI2023

Stochastic multiscale modeling for quantifying statistical and model errors with application to composite materials

This paper provides a coherent and efficient computational framework for stochastic multiscale analysis of material systems in the presence of parametric uncertainties and modeling errors. Uncertainty in those model parameters that are not deduced as upscaled quantities is attributed to an uncertainty “germ”. While such parameters can appear at any scale, they are predominant at the finest analysis scale. Additional uncertainties stemming from statistical estimation, attributed to lack of data and model error, are associated with each submodel contributing to the multiscale system. Here, a robust and efficient framework based on a generalized extended polynomial chaos expansion (gEPCE) is proposed to simultaneously propagate all these uncertainties in order to provide a probabilistic representation of specific quantities of interest (QoI). We characterize the full probability distribution of the QoI and the uncertainty in the failure probability pertaining to its tails. By combining gEPCE with kernel density estimation (KDE) and directional derivatives, we construct sensitivity measures that connect these statistical metrics of QoI to the various sources of uncertainty to assess their individual and combined impacts. An illustrative problem featuring three-point bending of a composite beam is investigated to demonstrate the presented approach.

36 MATERIALS SCIENCE↗
DOE OSTI2020

Three dimensional cluster analysis for atom probe tomography using Ripley’s K-function and machine learning

The size and structure of spatial molecular and atomic clustering can significantly impact material properties and is therefore important to accurately quantify. Ripley’s K-function (K(r)), a measure of spatial correlation, can be used to perform such quantification when the material system of interest can be represented as a marked point pattern. This work demonstrates how machine learning models based on K (r)-derived metrics can accurately estimate cluster size and intra-cluster density in simulated three dimensional (3D) point patterns containing spherical clusters of varying size; over 90% of model estimates for cluster size and intra-cluster density fall within 11% and 18% error of the true values, respectively. These K (r)-based size and density estimates are then applied to an experimental APT reconstruction to characterize MgZn clusters in a 7000 series aluminum alloy. Here we find that the estimates are more accurate, consistent, and robust to user interaction than estimates from the popular maximum separation algorithm. Using K (r) and machine learning to measure clustering is an accurate and repeatable way to quantify this important material attribute.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Intracardiac Electrical Imaging using the 12-lead ECG: A Machine Learning Approach using Synthetic Data

Current state-of-the-art techniques for non-invasive imaging of cardiac electrical phenomena require voltage recordings from dozens of different torso locations and anatomical models built from expensive medical diagnostic imaging procedures. Here this study aimed to assess if recent machine learning advances could alternatively reconstruct electroanatomical maps at clinically relevant resolutions using only the standard 12-lead electrocardiogram (ECG) as input. To that end, a computational study was conducted to generate a dataset of over 16000 detailed cardiac simulations, which was then used to train neural network (NN) architectures designed to exploit both spatial and temporal correlations in the ECG signal. Analysis over a validation set showed average errors in activation map reconstruction below 1.7 msec over 75 intracardiac locations. Furthermore, phenotypical patterns of activation and the morphology of the activation potential were correctly reconstructed. The approach offers opportunities to stratify patients non-invasively, both retrospectively and prospectively, using metrics otherwise only available through invasive clinical procedures.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2021

Management of the Three Mile Island, Unit 2, Accident Corium and Severely Damaged Fuel Debris

The Three Mile Island, Unit Two (TMI-2) pressurized water reactor core melted down in 1979 due to an untimely combination of maintenance problems that led to a loss of feedwater, followed by a series of operational misunderstandings and errors. The melted core recovery process required development of a wide array of tools. After approximately three years of water management and other cleanup actions, the first views of the core revealed a much higher degree of damage than previously expected. Approximately 62 metric tons of the core had melted, leaving only 42 of the 177 fuel assemblies standing with a majority of rods intact. The core to be recovered was composed of loose debris and a solidified mass of formerly molten fuel. Robotic tools that had been designed for the task were of limited effectiveness due to the range of material types and phases. Over a period of several years, the central melt was broken up, primarily by use of a core-boring drill originally designed to acquire samples through the depth of the debris field. The broken pieces were loaded into specially designed debris canisters by use of long-handled pick-and-place tools and suctioned into baffled knockout canisters using an airlift vacuum system. A remotely operable underwater plasma arc torch was developed for removal of the lower core support structure to be able to retrieve the fuel pieces and secondary melt that had accumulated on the bottom reactor vessel head. Canister design played a central role in defining the initial retrieval process and later affected transportation, interim wet and its current interim dry storage. Due to concerns about the potential for radiolysis of residual water in the debris and other canister material, the canisters were fitted with hydrogen recombiner catalyst units to prevent a buildup of flammable gas and potential pressurization. To confirm the effectiveness of the recombiners during shipping, eight dewatered canisters were kept sealed for up to 205 days with periodic sampling of the headspace gas. The highest hydrogen value (9 vol%) was observed in a canister held for 147 days while the longest stored canister resulted in a 5% hydrogen concentration. The oxygen concentration never exceeded 0.5 %, and the primary backfill was >80% argon, meaning there was no flammability risk. Radiolytic hydrogen was also observed in wet pool storage, where the vented canisters discharged a portion of the water backfill as a result of gas production. The 344 canisters (268 fuel debris, 62 filter, and 12 knockout type) were dewatered, loaded in groups of seven into a double barrier shipping cask, and transported to the U.S. Department of Energy site in Idaho for 10 years of pool storage.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2023

Scaling Optimal Allocation of Cloud Resources Using Lagrange Relaxation

The rapid growth of Cloud Computing (CC) has increased the variety of computing resources, storage, and communication services that pose significant new challenges for the efficient use of cloud resources. The cost-efficient allocation of cloud resources has become a decisive premise for the adoption of CC services. The cost-efficient selection and scheduling of these resources to meet the demands of a scientific workflow is a challenging problem that is exacerbated by the inclusion of multiple CC providers. In this paper, we present a novel strategy for the cost-efficient selection of CC resources using Lagrange relaxation. Our approach is based on preselection of resources and demand decomposition to create a series of smaller sub-problems, which allow the estimation of the best cost-structures and selection of CC service providers for a subset of the time period of the planning horizon. Decomposition of the demand is achieved through the boundary analysis of a continuous relaxation of the problem. Using the metrics defined in terms of the cost and time of completion, we demonstrate excellent performance in relation to optimal solutions. Our method reduced the computational time from hours to seconds for a representative 36-month problem and provided high-quality solutions (< 0.05% relative error). Given the importance of selecting resources and scheduling complex scientific workflows, we believe that this novel strategy will be beneficial for many researchers and users of cloud computing resources.

De La Torre, Luis↗
DOE OSTI2025

Analyzing and Exploring Training Recipes for Large-Scale Transformer-Based Weather Prediction

Abstract The rapid rise of deep learning (DL) in numerical weather prediction (NWP) has led to a proliferation of models which forecast atmospheric variables with comparable or superior skill than traditional physics-based NWP. However, among these leading DL models, there is a wide variance in both the training settings and architecture used. Further, the lack of thorough ablation studies makes it hard to discern which components are most critical to success. In this work, we show that it is possible to attain high forecast skill even with relatively off-the-shelf architectures, simple training procedures, and moderate compute budgets. Specifically, we train a minimally modified Swin Transformer V2 (SwinV2) on ERA5 data and find that it attains superior skill in terms of mean-square errors of deterministic forecasts when compared against the European Centre for Medium-Range Weather Forecasts’ Integrated Forecasting System (IFS). Almost all DL–NWP systems share a core set of hyperparameters and design decisions. To aid and expedite future DL–NWP research, we present an in-depth, systematic exploration of different loss functions, model sizes and depths, patch sizes, and multistep training objectives. We also examine the model performance with metrics beyond the typical accuracy (ACC) and RMSE and investigate how the performance scales with model size. Through our open-source code, scoring pipelines, and models, we share our findings on key aspects of the training pipeline. These ablations reduce the necessity for expensive hyperparameter tuning and lower the barrier to entry for future DL–NWP research. Significance Statement This study investigates the potential of using large-scale transformer-based models for weather prediction, showing that it is possible to achieve high forecast accuracy with simpler, off-the-shelf architectures. By training a minimally modified SwinV2 transformer on ERA5 data, we show that the model achieves competitive forecast skill in terms of mean-square error for key variables, outperforming the European Centre for Medium-Range Weather Forecasts’ Integrated Forecasting System (IFS) at all lead times. Our findings suggest that effective training strategies, such as multistep fine-tuning and channel-weighted losses, significantly enhance the model’s performance. However, we also highlight that these improvements come with trade-offs in other areas, such as ensemble spread and high-frequency spatial detail. This work highlights the promise of deep learning in improving weather forecasts, which could lead to better preparedness and response to weather events, ultimately benefiting society by providing more reliable weather predictions.

Willard, Jared D. [Lawrence Berkeley National Labo↗
DOE OSTI2020

Assessing uncertainty in the timing of energy use during cost-optimal distributed energy technology selection and sizing

This paper empirically derives uncertainty ranges in cost-optimal solar PV and storage sizing by comparing results from the REopt Lite optimization platform from metered data and a set of simulated Department of Energy Commercial Reference Building (CRB) profiles at 65 sites. We find load profile shape alone does not explain a site’s optimal configurations (i.e., PV, Storage, PV and Storage, No System). Still, load profile shape does introduce uncertainty to optimal PV and storage capacities. Across all cases where PV is part of an optimal configuration, we find the average ratio of power capacities derived from metered loads to capacities derived from CRB profiles to be 0.97 (and as high as 1463), where 1 would be a perfect match in system size. For storage, the ratio is 1.6 (and as high as 42). We also assess how, in the absence of complete metered data, a CRB profile can be selected that would be expected to yield the most similar solar PV and storage capacities. From those metrics that can be available from billing data (i.e., peak demand, monthly load totals), we find that uncertainty is most reduced by selecting the CRB’s with an annual peak occurring at the most similar time, or those with the lowest average root mean square error (RMSE) among monthly peak loads. This research can help improve the implementation and interpretation of results derived from simulated load profiles and is an important next step in advancing smart grid solutions.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗
DOE OSTI2020

Dynamic Undervolting to Improve Energy Efficiency on Multicore X86 CPUs

Chip manufacturers introduce redundancy at various levels of CPU design to guarantee correct operation, even for worst-case combinations of non-idealities in process variation and system operating conditions. This redundancy is implemented partly in the form of voltage margins. However, for a wide range of real-world execution scenarios these margins are excessive and merely translate to increased power consumption, hindering the effort towards higher-energy efficiency in both HPC and general purpose computing. Our study on the x86-64 Haswell and Skylake multicore microarchitectures reveals-wide voltage margins, which vary across different microarchitectures, different chip parts of the same microarchitecture, and across different workloads. We find that it is necessary to quantify-voltage margins using multi-threaded and multi-instance workloads, as characterization with single-threaded and single-instance workloads that do not stress the CPU to its full capacity typically identifies overly optimistic margins that lead to errors when applied in realistic program execution scenarios. In addition, we introduce, deploy and evaluate a run-time governor that dynamically reduces the supply voltage of modern multicore x86-64 CPUs. Our governor employs a model that takes as input a set of performance metrics which are directly measurable via performance monitoring counters and have high predictive value for the minimum tolerable supply voltage (V min ), to predict and apply the appropriate reduction for the workload at hand. Compared with the conventional DVFS governor, our approach in this study achieves up to 42 percent energy savings for the Skylake family and 34 percent for the Haswell family for complex, real-world applications.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Day-ahead continuous double auction-based peer-to-peer energy trading platform incorporating trading losses and network utilisation fee

Integration of distributed energy resources, such as photovoltaic solar (PV), introduces new opportunities to establish local energy market frameworks to improve renewable energy utilisation in residential sectors. Such peer-to-peer (P2P) energy trading refers to a local market structure where customers (and prosumers) interact to share excess PV generation to enhance the individual and community social welfare. In this work, a day-ahead continuous double auction (CDA)-based P2P market structure considering network losses and network utilisation fees was designed. Day-ahead PV energy is modelled using fractional integral polynomials and the output is forecasted using an autoregressive integrated moving average model for each market interval. Based on the customer load and excess PV energy, the CDA market is cleared using a bid/ask matching mechanism. The performance of the P2P market was evaluated by computing different welfare metrics while analysing the effect of network constraints. The results show that the designed CDA-based P2P market structure increases the social welfare of all participants by an average of 17.75% compared to the baseline for the presented cases. Moreover, the impact of the forecasting error between the day-ahead and real-time market was also quantified.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2024

Velocity reconstruction in the era of DESI and Rubin/LSST. I. Exploring spectroscopic, photometric, and hybrid samples

Peculiar velocities of galaxies and halos can be reconstructed from their spatial distribution alone. This technique is analogous to the baryon acoustic oscillations reconstruction, using the continuity equation to connect density and velocity fields. The resulting reconstructed velocities can be used to measure imprints of galaxy velocities on the cosmic microwave background like the kinematic Sunyaev-Zel’dovich effect or the moving lens effect. As the precision of these measurements increases, characterizing the performance of the velocity reconstruction becomes crucial to allow unbiased and statistically optimal inference. In this paper, we quantify the relevant performance metrics: the variance of the reconstructed velocities and their correlation coefficient with the true velocities. We show that the relevant velocities to reconstruct for kSZ and moving lens are actually the halo—rather than galaxy—velocities. We quantify the impact of redshift-space distortions, photometric redshift errors, satellite galaxy fraction, incorrect cosmological parameter assumptions and smoothing scale on the reconstruction performance. Here, we also investigate hybrid reconstruction methods, where velocities inferred from spectroscopic samples are evaluated at the positions of denser photometric samples. We find that using exclusively the photometric sample is better than performing a hybrid analysis. The 2 Gpc/ℎ length simulations from abacussummit with realistic galaxy samples for DESI and Rubin LSST allow us to perform this analysis in a controlled setting. In the companion paper [B. Hadzhiyska, S. Ferraro, B. Ried Guachalla, and E. Schaan, companion paper, Phys. Rev. D 109, 103534 (2024).], we further include the effects of evolution along the light cone and give realistic performance estimates for DESI luminous red galaxies, emission line galaxies, and Rubin LSST-like samples.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2023

Explainable machine learning for hydrogen diffusion in metals and random binary alloys

Hydrogen diffusion in metals and alloys plays an important role in the discovery of new materials for fuel cell and energy storage technology. While analytic models use hand-selected features that have clear physical ties to hydrogen diffusion, they often lack accuracy when making quantitative predictions. Machine learning models are capable of making accurate predictions, but their inner workings are obscured, rendering it unclear which physical features are truly important. To develop interpretable machine learning models to predict the activation energies of hydrogen diffusion in metals and random binary alloys, we create a database for physical and chemical properties of the species and use it to fit six machine learning models. Our models achieve root-mean-squared errors between 98–119 meV on the testing data and accurately predict that elemental Ru has a large activation energy, while elemental Cr and Fe have small activation energies. By analyzing the feature importances of these fitted models, we identify relevant physical properties for predicting hydrogen diffusivity. While metrics for measuring the individual feature importances for machine learning models exist, correlations between the features lead to disagreement between models and limit the conclusions that can be drawn. Instead grouped feature importance, formed by combining the features via their correlations, agree across the six models and reveal that the two groups containing the packing factor and electronic specific heat are particularly significant for predicting hydrogen diffusion in metals and random binary alloys. In conclusion, this framework allows us to interpret machine learning models and enables rapid screening of new materials with the desired rates of hydrogen diffusion.

36 MATERIALS SCIENCE↗
DOE OSTI2026

LDRD FY25 Program Overview

As Lawrence Livermore National Laboratory’s (LLNL’s) Laboratory Directed Research and Development (LDRD) program enters its fifth decade of leading-edge research and development, its impact and importance have never been stronger. The program continues to advance strategic investments in pioneering science, technology, and engineering, ensuring LLNL will be ready to deliver on our mission as it evolves over the coming decades. Investing in LDRD research, and the people who perform this critical work, gives LLNL the ability to sustain our role as a leader in the Department of Energy and National Nuclear Security Administration enterprise. The LDRD program enables high-risk, high-payoff research that anticipates emerging threats and future mission needs. By nurturing the ingenuity of the Lab’s greatest asset, its people, LDRD funding advances not only our research but also grows and nurtures our workforce: engaging future innovators with student mentoring, challenging postdoctoral researchers to apply their skills to support national security, and strengthening the leadership skills of early career staff. This annual report documents how LDRD investments advance LLNL’s science, technology, and engineering across our mission space. To assess LDRD’s impact we track both short and long-term metrics such as peer-reviewed publications, number of students, or professional fellows. In addition to reviewing these metrics, I encourage you to delve deeper into the breadth of science and technology that illustrate the strategic value of this research portfolio. For instance, a recent exploratory research project used advanced manufacturing to construct miniaturized three-dimensional ion traps for a quantum computer with reduced quantum error rates to enable applications that address national security missions and support basic science. Another project has delved into studying detonation by examining deflagration to enhance the safety and security of the nuclear weapons stockpile. LDRD researchers are also deploying AI agents on two of the world’s most powerful supercomputers to automate and accelerate inertial confinement fusion experiments. Other teams are delivering more accurate optical constants to enable improved validation for aluminum to advance atomic and molecular physics models. LDRD-driven discoveries of how metals deform under extreme conditions strengthen our ability to model and design materials for demanding national security environments. National security challenges are increasingly complex and continuously evolving. LDRD focuses our most innovative science and technology on these challenges, ensuring the Laboratory is developing creative, forward-leaning solutions for our nation and the world. The following pages feature highlights of published scientific advances, patents, and honors that stem from LDRD investments. As you read this report, I hope you will understand how these investments position the Laboratory, and our partners, to meet the demands of the decades ahead.

36 MATERIALS SCIENCE↗
DOE OSTI2025

Examining the Impact of Local Constraint Violations on Energy Computations in DFT

ABSTRACT This work examines the impact of locally imposed constraints in Density Functional Theory (DFT). Using a metric referred to as the extent of violation index (EVI), we quantify how well exchange‐correlation functionals adhere to local constraints. Applying EVIs to a diverse set of molecules for GGA functionals reveals constraint violations, particularly for semi‐empirical functionals. We leverage EVIs to explore potential connections between these violations and errors in chemical properties. While no correlation is observed for atomization energies, a significant statistical correlation emerges between EVIs and total energies. Similarly, the analysis of reaction energies suggests weak positive correlations for specific constraints. However, definitive conclusions about error cancellation mechanisms cannot be made at this time. These observations revealed by EVIs may be useful for consideration when designing future generations of semilocal functionals.

Khanna, Vaibhav [Department of Chemistry Universit↗
DOE OSTI2024

Multivariate prognosis of battery advanced state of health via transformers

De-risking energy storage investments necessary to meet CO 2 reduction targets requires a deep understanding of the connections between battery health, design, and use. The historical definition of the battery state of health (SOH) as the percentage of current versus initial capacity is inadequate for this purpose, motivating an expanded SOH consisting of an interrelated set of descriptors including capacity, energy, ionic and electronic impedances, open-circuit voltages, and microstructure metrics. In this work, we introduce deep transformer networks for the simultaneous prognosis of 28 battery SOH descriptors using two cycling datasets representing six lithium-ion cathode chemistries, multiple electrolyte/anode compositions, and different charge-discharge scenarios. The accuracy of these predictions for battery life (with an unprecedented mean absolute error of 19 cycles in predicting end of life for a lithium-iron-phosphate fast-charging dataset) illustrates the promise of deep learning toward providing enhanced understanding and control of battery health.

25 ENERGY STORAGE↗
DOE OSTI2022

Putting AlphaFold models to work with phenix.process_predicted_model and ISOLDE

AlphaFold has recently become an important tool in providing models for experimental structure determination by X-ray crystallography and cryo-EM. Large parts of the predicted models typically approach the accuracy of experimentally determined structures, although there are frequently local errors and errors in the relative orientations of domains. Importantly, residues in the model of a protein predicted by AlphaFold are tagged with a predicted local distance difference test score, informing users about which regions of the structure are predicted with less confidence. AlphaFold also produces a predicted aligned error matrix indicating its confidence in the relative positions of each pair of residues in the predicted model. The phenix.process_predicted_model tool downweights or removes low-confidence residues and can break a model into confidently predicted domains in preparation for molecular replacement or cryo-EM docking. These confidence metrics are further used in ISOLDE to weight torsion and atom–atom distance restraints, allowing the complete AlphaFold model to be interactively rearranged to match the docked fragments and reducing the need for the rebuilding of connecting regions.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2023

Comparison of Polarized Radiative Transfer Codes Used by the EHT Collaboration

Interpretation of resolved polarized images of black holes by the Event Horizon Telescope (EHT) requires predictions of the polarized emission observable by an Earth-based instrument for a particular model of the black hole accretion system. Such predictions are generated by general relativistic radiative transfer (GRRT) codes, which integrate the equations of polarized radiative transfer in curved spacetime. A selection of ray-tracing GRRT codes used within the EHT Collaboration is evaluated for accuracy and consistency in producing a selection of test images, demonstrating that the various methods and implementations of radiative transfer calculations are highly consistent. When imaging an analytic accretion model, we find that all codes produce images similar within a pixel-wise normalized mean squared error (NMSE) of 0.012 in the worst case. When imaging a snapshot from a cell-based magnetohydrodynamic simulation, we find all test images to be similar within NMSEs of 0.02, 0.04, 0.04, and 0.12 in Stokes I , Q , U , and V , respectively. We additionally find the values of several image metrics relevant to published EHT results to be in agreement to much better precision than measurement uncertainties.

79 ASTRONOMY AND ASTROPHYSICS↗