SEARCH · Engineering Papers

Results for “Learning with errors”

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

SourceResearch connectionFrom yearThrough year

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

At least 1,243 records · Page 69

All sources NASA NTRS DOE OSTI

DOE OSTI2025

A Neural Optimizer With Decision-Focused Learning for Optimal Energy Storage Operation

Here, this article introduces a neural optimizer-based framework for optimizing battery energy storage system (BESS) control for grid services, including demand charge and energy cost reduction. By leveraging decision-focused learning (DFL), the proposed framework ensures seamless integration and adaptation, significantly enhancing control performance. A patch time-series transformer is employed for peak load forecasting, incorporating aleatoric uncertainty quantification to account for forecasting uncertainties within the decision-making process. The framework utilizes a solver-in-the-loop approach to generate optimal BESS actions, which are then used to train the neural optimizer-based agent. By co-optimizing both BESS operational modes and output power within the NN, the system achieves improved performance and robustness. After initial training, the forecasting and control models are jointly fine-tuned to account for forecasting errors, further improving decision precision and efficiency through DFL. Case studies are performed to validate the performance of the framework using multiple real-world datasets, demonstrating superior performance in monthly peak load forecasting compared to state-of-the-art models. In addition, the results are compared against existing decision-making approaches. The results demonstrate a reduction in monthly peak forecasting error by approximately 15% across various performance measures and achieve an optimization gap for BESS operation that is about three times smaller compared to existing methods.

Kim, Hyeonjin [Pacific Northwest National Laborato↗

DOE OSTI2024

Surrogate modelling for urban building energy simulation based on the bidirectional long short-term memory model

Here, the urban microclimate is essential for accurate simulation-based urban building energy modelling (UBEM). However, a high spatial-resolution microclimate can increase the computational resources demands of UBEM. Surrogate modelling is one of the promising approaches for fast UBEM. This study proposes a bidirectional Long Short-Term Memory (LSTM)-based approach for simulation-based UBEM surrogate modelling. The estimations are aggregated into census tracts using total building floor area. A case study using UBEM to estimate annual hourly building energy use and anthropogenic heat from all existing buildings in Los Angeles County found that most of the surrogate models can complete the annual hourly simulation within 90 minutes with a normalized mean absolute error lower than 10%, and that the bidirectional LSTM outperforms the standard LSTM in accuracy. This study demonstrates the advantages of bidirectional RNN architecture in building energy surrogate modelling and is expected to promote long-term and high-resolution UBEM with detailed microclimates.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

DOE OSTI2025

Data-Driven State of Health Estimation for Second-Life Batteries Using Interpolated Synthetic Data and Feature Selection

Accurate estimation of the State of Health (SOH) for second-life batteries (SLBs) is crucial given their increasing use in energy storage applications. Precise SOH prediction is essential for safe operation and robust battery management systems. A major challenge is the limited availability of datasets for building reliable degradation models. To address this, synthetic data generation through linear interpolation is performed to extend the available data, making it more representative of real-world battery operating conditions. By analyzing feature correlation with SOH, the most relevant features are selected for the model. The proposed approach employs a convolutional neural network (CNN) model trained on this interpolated, feature-selected dataset, using time series data of voltage, temperature, and current over a cycle. By focusing on highly correlated features, the model achieves over 95% accuracy, with mean absolute error and root mean squared error up to 2.27% and 2.64%, respectively, in SOH estimation for two battery datasets tested. These results highlight the potential of combining synthetic data generation and feature selection to enhance SOH predictions, showcasing the superior performance of the proposed CNN model for both new batteries and SLBs.

feature selection↗

DOE OSTI2021

A fast multi-fidelity method with uncertainty quantification for complex data correlations: Application to vortex-induced vibrations of marine risers

Here we develop a fast multi-fidelity modeling method for very complex correlations between high- and low-fidelity data by working in modal space to extract the proper correlation function. We apply this method to infer the amplitude of motion of a flexible marine riser in cross-flow, subject to vortex-induced vibrations (VIV). VIV are driven by an absolute instability in the flow, which imposes a frequency (Strouhal) law that requires a matching with the impedance of the structure; this matching is easily achieved because of the rapid parametric variation of the added mass force. As a result, the wavenumber of the riser spatial response is within narrow bands of uncertainty. Hence, an error in wavenumber prediction can cause significant phase-related errors in the shape of the amplitude of response along the riser, rendering correlation between low- and high-fidelity data very complex. Working in modal space as outlined herein, dense data from low-fidelity data, provided by the semi-empirical computer code VIVA, can correlate in modal space with few high-fidelity data, obtained from experiments or fully-resolved CFD simulations, to correct both phase and amplitude and provide predictions that agree very well overall with the correct shape of the amplitude response. We also quantify the uncertainty in the prediction using Bayesian modeling and exploit this uncertainty to formulate an active learning strategy for the best possible location of the sensors providing the high fidelity measurements.

42 ENGINEERING↗

DOE OSTI2025

Navigating the Noise: Bringing Clarity to ML Parameterization Design With O $\boldsymbol{\mathcal{O}}$(100) Ensembles

Abstract Machine‐learning (ML) parameterizations of subgrid processes (here of turbulence, convection, and radiation) may one day replace conventional parameterizations by emulating high‐resolution physics without the cost of explicit simulation. However, uncertainty about the relationship between offline and online performance (i.e., when integrated with a large‐scale general circulation model) hinders their development. Much of this uncertainty stems from limited sampling of the noisy, emergent effects of upstream ML design decisions on downstream online hybrid simulation. Our work rectifies the sampling issue via the construction of a semi‐automated, end‐to‐end pipeline for size ensembles of hybrid simulations, revealing important nuances in how systematic reductions in offline error manifest in changes to online error and online stability. For example, removing dropout and switching from a Mean Squared Error to a Mean Absolute Error loss both reduce offline error, but they have opposite effects on online error and online stability. Other design decisions, like incorporating memory, converting moisture input from specific humidity to relative humidity, using batch normalization, and training on multiple climates do not come with any such compromises. Finally, we show that ensemble sizes of may be necessary to reliably detect causally relevant differences online. By enabling rapid online experimentation at scale, we can empirically settle debates regarding subgrid ML parameterization design that would have otherwise remained unresolved in the noise.

Lin, Jerry [Department of Earth System Sciences Un↗

DOE OSTI2024

Learning model combining convolutional deep neural network with a self-attention mechanism for AC optimal power flow

Alternating current optimal power flow (OPF) analysis is critical for efficient and reliable operation of power systems. For large systems or repetitive computations, the traditional methods such as the direct and gradient methods, or non-traditional methods, such as the genetic algorithm and simulating annealing, are time-consuming and unsuitable for real-time computing. The work in this paper proposes a novel framework to obtain the optimal solution of power flow in real-time using a combination of convolutional neural networks and a self-attention mechanism. All parameters of the power networks are rearranged in an image-like shape of a multi-channel image where each channel is a two-dimensional matrix. The proposed approach is adaptive with every input size of power systems as well as frequent variations of network topologies without intervention to the framework core. The encompassment of all power system contexts in which all parameters of internal elements, generation costs, and topology information are included, contributes to the higher accuracy of inference compared to other current machine-learning-based OPF-solving methods. Besides, the proposed framework established on ubiquitous platforms is effortlessly integrated into current infrastructures of power systems, and the great efficiency along with the computation speed may serve as a critical point for practical implications, such as enabling faster decision-making during real-time operations, predicting system contingencies, and remedial actions based on an offline pre-trained model. Furthermore, this supervised learning process is applied to the dataset of four case studies of meshed power systems: the IEEE 5-bus system (IEEE-5), the IEEE 30-bus system (IEEE-30), the IEEE 39-bus system (IEEE-39), and the IEEE 57-bus system (IEEE-57) to prove the efficacy of the proposed method.

42 ENGINEERING↗

DOE OSTI2024

Validation of new and existing methods for time-domain simulations of turbulence and loads

We seek to obtain a second-by-second match between the simulated and measured structural loads of a utility-scale wind turbine. To obtain the one-to-one load simulations, we start with the furthest upstream component of the modeling chain: the turbulent inflow. We consider new and existing methods to generate constrained-turbulence flow fields. The new method is based on large-eddy simulations (LES) and machine learning (ML). The existing methods include Kaimal-based TurbSim and the superstatistical wind field model. The inflow measurements used to constrain these simulations are obtained with a nacelle-mounted scanning lidar. We compare the flow fields for the different inflow simulation approaches and validate their associated load predictions against measurements collected in the Rotor Aero-dynamics, Aeroelastics, and Wake (RAAW) field campaign. We find that the rotor-position control developed for this study is key in enabling the time match between measurements and simulations. When this control approach is used, the load simulation performance tracks with the inflow simulation fidelity, with LES+ML yielding errors ≤ 4% for the damage-equivalent loads of flapwise bending moment, and tower fore-aft bending moments.

17 WIND ENERGY↗

DOE OSTI2022

The stellar parameters and elemental abundances from low-resolution spectra – I. 1.2 million giants from LAMOST DR8

As a typical data-driven method, deep learning becomes a natural choice for analysing astronomical data. In this study, we built a deep convolutional neural network (NN) to estimate basic stellar parameters $T\rm {_{eff}}$, log g , metallicity ([M/H] and [Fe/H]) and [α/M] along with nine individual elemental abundances ([C/Fe], [N/Fe], [O/Fe], [Mg/Fe], [Al/Fe], [Si/Fe], [Ca/Fe], [Mn/Fe], and [Ni/Fe]). The NN is trained using common stars between the APOGEE survey and the LAMOST survey. We used low-resolution spectra from LAMOST survey as input, and measurements from APOGEE as labels. For stellar spectra with the signal-to-noise ratio in g band larger than 10 in the test set, the mean absolute error (MAE) is 29 K for $T\rm {_{eff}}$, 0.07 dex for log g , 0.03 dex for both [Fe/H] and [M/H], and 0.02 dex for [α/M]. The MAE of most elements is between 0.02 and 0.04 dex. The trained NN was applied to 1210 145 giants, including sub-giants, from LAMOST DR8 within the range of stellar parameters 3500 K < $T\rm {_{eff}}$ < 5500 K, 0.0 dex < log g < 4.0 dex, −2.5 dex < [Fe/H] < 0.5 dex. The distribution of our results in the chemical spaces is highly consistent with APOGEE labels and stellar parameters show consistency with external high-resolution measurements from GALAH. The results in this study allow us to further studies based on LAMOST data and deepen our understanding of the accretion and evolution history of the Milky Way. The electronic version of the value added catalog is available at http://www.lamost.org/dr8/v1.1/doc/vac.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2024

Neural network emulation of spontaneous fission

Large-scale computations of fission properties are an important ingredient for nuclear reaction network calculations simulating rapid neutron-capture process (the 𝑟 process) nucleosynthesis. Due to the large number of fissioning nuclei potentially contributing to the 𝑟 process, a microscopic description of fission based on nuclear density functional theory (DFT) is computationally challenging. Here, we explore the use of neural networks (NNs) to construct DFT emulators capable of predicting potential energy surfaces and collective inertia tensors across the whole nuclear chart, starting from a minimal set of DFT calculations. We use constrained Hartree-Fock-Bogoliubov (HFB) calculations to predict the potential energy and collective inertia tensor in the axial quadrupole and octupole collective coordinates, for a set of nuclei in the 𝑟-process region. We then employ NNs to emulate the HFB energy and collective inertia tensor across the considered region of the nuclear chart. Least-action pathways characterizing spontaneous fission half-lives and fragment yields are then obtained by means of the nudged elastic band method. The potential energy predicted by NNs agrees with the DFT value to within a root-mean-square error of 500 keV, and the collective inertia components agree to within an order of magnitude. These results are largely independent of the NN architecture. The exit points on the outer turning line are found to be well emulated. For the spontaneous fission half-lives the NN emulation provides values that are found to agree with the DFT predictions within a factor of 10 3 across more than 70 orders of magnitude. Neural networks are able to emulate the potential energy and collective inertia well enough to reasonably predict physical observables. Future directions of study, such as the inclusion of additional collective degrees of freedom and active learning, will improve the predictive power of microscopic theory and further enable large-scale fission studies.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

DOE OSTI2024

AXI4MLIR: User-Driven Automatic Host Code Generation for Custom AXI-Based Accelerators

Tensor algebra operations represent an important class of algorithms used across many applications, including machine learning, scientific computing, and data analytics. As a result, the efficient generation of custom accelerators for tensor operations has received increased attention. Previous efforts have produced automated tools enabling users to prototype and explore optimized accelerators. However, little effort has been focused on the host-accelerator interaction in these tools. Efficient use of hardware accelerators requires knowledge about the accelerator's capabilities (operations, data formats, and opcode support), the host CPU microarchitecture (e.g., memory hierarchy), the host-accelerator interface, and the application's features (which code regions should be mapped onto an accelerator). Manually rewriting the original applications to facilitate improved custom accelerator mapping is an error-prone and time-consuming endeavor. To cope with this, we propose AXI4MLIR, a new framework to automatically generate and optimize the communication between the host CPU and arbitrary accelerators that implement linear algebra algorithms. AXI4MLIR extends the MLIR compiler framework to automatically generate efficient host-accelerator driver code for accelerators with AXI-based interfaces. Our compiler extensions enable automatic driver code generation while carefully considering the host's memory hierarchy and target accelerator features. To demonstrate the flexibility and utility of AXI4MLIR, we test it with diverse use cases that include different types of accelerators, tiling scenarios, and dataflow schemes. We compare our experimental results to manual implementations of host-accelerator driver code and find that our approach can reduce CPU cache references by 56% and deliver up to a 1.65x speedup.

Bohm Agostini, Nicolas↗

DOE OSTI2023

High-Resolution PM2.5 Concentrations Estimation Based on Stacked Ensemble Learning Model Using Multi-Source Satellite TOA Data

Nepal has experienced severe fine particulate matter (PM2.5) pollution in recent years. However, few studies have focused on the distribution of PM2.5 and its variations in Nepal. Although many researchers have developed PM2.5 estimation models, these models have mainly focused on the kilometer scale, which cannot provide accurate spatial distribution of PM2.5 pollution. Based on Gaofen-1/6 and Landsat-8/9 satellite data, we developed a stacked ensemble learning model (named XGBLL) combined with meteorological data, ground PM2.5 concentrations, ground elevation, and population data. The model includes two layers: a XGBoost and Light GBM model in the first layer, and a linear regression model in the second layer. The accuracy of XGBLL model is better than that of a single model, and the fusion of multi-source satellite remote sensing data effectively improves the spatial coverage of PM2.5 concentrations. Besides, the spatial distribution of the daily mean PM2.5 concentrations in the Kathmandu region under different air conditions was analyzed. The validation results showed that the monthly averaged dataset was accurate (R2 = 0.80 and root mean square error = 7.07). In addition, compared to previous satellite PM2.5 datasets in Nepal, the dataset produced in this study achieved superior accuracy and spatial resolution.

Environmental Sciences & Ecology↗

DOE OSTI2022

WinnowML: Stable feature selection for maximizing prediction accuracy of time-based system modeling

Online deep learning (ODL) has become an important methodology for modeling time-based performance of computer systems. An open problem is the intelligent selection of features from raw workload traces of computer systems. The best methods are overly sensitive to noisy data, causing frequent feature changes and re-training. Using all available features inflates training time and introduces model artifacts if some features should have been dropped. We present WinnowML, a method for automatically determining the most relevant feature subset for a predictive time-series model. WinnowML combines existing feature ranking algorithms and a history of each feature's ranking to iteratively rank a feature set to lower prediction error and maximize long term relevance. From this ranked feature set, the most relevant and stable subset is selected to train a model. Experimentally, we show how WinnowML can lower a model's mean absolute relative error up to 42% on average compared to the closest performing approach. Additionally, we lower the fluctuation in feature ranking and selection up to 65%. We also demonstrate how to combine WinnowML and a model search tool to provide improvements in performance of up to 14.5% when compared to using all the feature available.

Bel, Oceane MS↗

DOE OSTI2024

Understanding GPU Memory Corruption at Extreme Scale: The Summit Case Study

GPU memory corruption and in particular double-bit errors (DBEs) remain one of the least understood aspects of HPC system reliability. Albeit rare, their occurrences always lead to job termination and can potentially cost thousands of node-hours, either from wasted computations or as the overhead from regular checkpointing needed to minimize the losses. As supercomputers and their components simultaneously grow in scale, density, failure rates, and environmental footprint, the efficiency of HPC operations becomes both an imperative and a challenge. We examine DBEs using system telemetry data and logs collected from the Summit supercomputer, equipped with 27,648 Tesla V100 GPUs with 2nd-generation high-bandwidth memory (HBM2). Using exploratory data analysis and statistical learning, we extract several insights about memory reliability in such GPUs. We find that GPUs with prior DBE occurrences are prone to experience them again due to otherwise harmless factors, correlate this phenomenon with GPU placement, and suggest manufacturing variability as a factor. On the general population of GPUs, we link DBEs to short- and long-term high power consumption modes while finding no significant correlation with higher temperatures. We also show that the workload type can be a factor in memory’s propensity to corruption.

Oles, Vlad↗

DOE OSTI2021

A daily, 250 m and real-time gross primary productivity product (2000–present) covering the contiguous United States

Abstract. Gross primary productivity (GPP) quantifies the amount of carbon dioxide (CO2) fixed by plants through photosynthesis. Although as a key quantity of terrestrial ecosystems, there is a lack of high-spatial-and-temporal-resolution, real-time and observation-based GPP products. To address this critical gap, here we leverage a state-of-the-art vegetation index, near-infrared reflectance of vegetation (NIRV), along with accurate photosynthetically active radiation (PAR), to produce a SatelLite Only Photosynthesis Estimation (SLOPE) GPP product for the contiguous United States (CONUS). Compared to existing GPP products, the proposed SLOPE product is advanced in its spatial resolution (250 m versus >500 m), temporal resolution (daily versus 8 d), instantaneity (latency of 1 d versus >2 weeks) and quantitative uncertainty (on a per-pixel and daily basis versus no uncertainty information available). These characteristics are achieved because of several technical innovations employed in this study: (1) SLOPE couples machine learning models with MODIS atmosphere and land products to accurately estimate PAR. (2) SLOPE couples highly efficient and pragmatic gap-filling and filtering algorithms with surface reflectance acquired by both Terra and Aqua MODIS satellites to derive a soil-adjusted NIRV (SANIRV) dataset. (3) SLOPE couples a temporal pattern recognition approach with a long-term Cropland Data Layer (CDL) product to predict dynamic C4 crop fraction. Through developing a parsimonious model with only two slope parameters, the proposed SLOPE product explains 85 % of the spatial and temporal variations in GPP acquired from 49 AmeriFlux eddy-covariance sites (324 site years), with a root-mean-square error (RMSE) of 1.63 gC m−2 d−1. The median R2 over C3 and C4 crop sites reaches 0.87 and 0.94, respectively, indicating great potentials for monitoring crops, in particular bioenergy crops, at the field level. With such a satisfactory performance and its distinct characteristics in spatiotemporal resolution and instantaneity, the proposed SLOPE GPP product is promising for biological and environmental research, carbon cycle research, and a broad range of real-time applications at the regional scale. The archived dataset is available at https://doi.org/10.3334/ORNLDAAC/1786 (download page: https://daac.ornl.gov/daacdata/cms/SLOPE_GPP_CONUS/data/, last access: 20 January 2021) (Jiang and Guan, 2020), and the real-time dataset is available upon request.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2021

Generalization Equations for Machine Learners Based on Physical and Abstract Laws

The physical and abstract laws derived from the first principles have been recently exploited to customize and sharpen machine learning (ML) methods and also derive their generalization equations. These laws often encapsulate knowledge that complements datasets and ML models. We present a generic framework that uses these laws to provide ML codes that are transferable across multiple areas, including data transport infrastructures and thermal hydraulics analytics of nuclear reactors. By anchoring on datasets from these areas and the statistical generalization theory, we present a rigorous approach to co-develop ML solutions and the generalization equations that characterize them, by exploiting the structure and constraints from the laws. We present illustrative examples using practical problems from existing literature on the performance characterization of data transport infrastructures, and the sensor error and power level estimation in nuclear reactor systems using sensor measurements of primary and secondary coolant systems, respectively.

Rao, Nageswara↗

DOE OSTI2025

Systematic softening in universal machine learning interatomic potentials

Machine learning interatomic potentials (MLIPs) have introduced a new paradigm for atomic simulations. Recent advancements have led to universal MLIPs (uMLIPs) that are pre-trained on diverse datasets, providing opportunities for universal force fields and foundational machine learning models. However, their performance in extrapolating to out-of-distribution complex atomic environments remains unclear. In this study, we highlight a consistent potential energy surface (PES) softening effect in three uMLIPs: M3GNet, CHGNet, and MACE-MP-0, which is characterized by energy and force underprediction in atomic-modeling benchmarks including surfaces, defects, solid-solution energetics, ion migration barriers, phonon vibration modes, and general high-energy states. The PES softening behavior originates primarily from the systematically underpredicted PES curvature, which derives from the biased sampling of near-equilibrium atomic arrangements in uMLIP pre-training datasets. Our findings suggest that a considerable fraction of uMLIP errors are highly systematic, and can therefore be efficiently corrected. We argue for the importance of a comprehensive materials dataset with improved PES sampling for next-generation foundational MLIPs.

36 MATERIALS SCIENCE↗

DOE OSTI2021

Rapid quantitative analysis of trace elements in plutonium alloys using a handheld laser-induced breakdown spectroscopy (LIBS) device coupled with chemometrics and machine learning

Here, we present the first reported quantification of trace elements in plutonium via a portable laser-induced breakdown spectroscopy (LIBS) device and demonstrate the use of chemometric analysis to enhance the handheld device's sensitivity and precision. Quantification of trace elements such as iron and nickel in plutonium metal via LIBS is a challenging problem due to the complex nature of the plutonium optical emission spectra. While rapid analysis of plutonium alloys has been demonstrated using portable LIBS devices, such as the SciAps Z300, their detection limits for trace elements are severely constrained by their achievable pulse power and length, light collection optics, and detectors. In this paper, analytical methods are evaluated as a means to circumvent the detection constraints. Three chemometric methods often used in analytical spectroscopy are evaluated; principal component regression, partial least-squares regression, and artificial neural networks. These models are evaluated based on goodness-of-fit metrics, root mean-squared error, and their achievable limits of detection (LoDs). Partial least squares proved superior for determining content of iron and nickel in plutonium metal, yielding LoDs of 15 and 20 ppm, respectively. These results of identifying the undesirable trace elements in plutonium components are critical for applications such as fabricating radioisotope thermoelectric generators or nuclear fuel.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

DOE OSTI2025

Variational AutoEncoders Reveal Intensifying GPP Extremes in Continental United States based on CESM2 Simulations

Climate extremes significantly impact terrestrial carbon cycle dynamics, necessitating robust methods for detecting and analyzing anomalous behavior in plant productivity. This study presents a novel application of variational autoencoders (VAE) for identifying extreme events in gross primary productivity (GPP) from Community Earth System Model version 2 simulations across four AR6 regions in the Continental United States. We compare VAE-based anomaly detection with traditional singular spectral analysis (SSA) methods across three time periods: 1850-80, 1950-80, and 2050-80 under SSP5-8.5 scenario. The VAE architecture employs three dense layers and a latent space with input sequence length of 12 months, training on normalized GPP time series to reconstruct the GPP and identify anomalies based on reconstruction errors. Extreme events are defined using 5th percentile thresholds applied to both VAE and SSA anomalies. Results demonstrate strong regional agreement between VAE and SSA methods in spatial patterns of extreme event frequencies, despite VAE consistently producing higher threshold values (179-756 GgC for VAE vs. 100-784 GgC for SSA across regions and periods). Both methods reveal increasing magnitudes and frequencies of negative carbon cycle extremes toward 2050-80, particularly in Western and Central North America. The VAE approach shows comparable performance to established SSA techniques while offering computational advantages and enhanced capability for capturing non-linear temporal dependencies in carbon cycle variability. This research demonstrates the potential of deep learning approaches for extremes detection and provides a foundation for improved understanding of future carbon cycle risks under future conditions.

Sharma, Bharat [ORNL] (ORCID:0000000266982487)↗