Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Model selection”

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 865 records · Page 48
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

$\nu_μ$ Disappearance in MicroBooNE using the Deep Learning 1$\mu$1$p$ Selection

We test a 3+1 model with the MicroBooNE data using a 1µ1p selection developed using Deep-Learning-based reconstruction. In order to test this model we apply a muon neutrino disappearance effect to the selection, and search across a grid of oscillation model parameters using a Feldman Cousins technique. We determine MicroBooNE’s sensitivity across this model parameter space, and perform several validation studies to test this study’s robustness. Finally, we examine the allowed and excluded regions per MicroBooNE’s data at 90% confidence, using a data set corresponding to 6.67 × 10 20 protons on target. The null model remains allowed, and several of the high-disappearance models are excluded.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
NASA NTRS1995

Estimating the urban bias of surface shelter temperatures using upper-air and satellite data. Part 1: Development of models predicting surface shelter temperatures

Multiple regression techniques were used to predict surface shelter temperatures based on the time period 1986-89 using upper-air data from the European Centre for Medium-Range Weather Forecasts (ECMWF) to represent the background climate and site-specific data to represent the local landscape. Global monthly mean temperature models were developed using data from over 5000 stations available in the Global Historical Climate Network (GHCN). Monthly maximum, mean, and minimum temperature models for the United States were also developed using data from over 1000 stations available in the U.S. Cooperative (COOP) Network and comparative monthly mean temperature models were developed using over 1150 U.S. stations in the GHCN. Three-, six-, and full-variable models were developed for comparative purposes. Inferences about the variables selected for the various models were easier for the GHCN models, which displayed month-to-month consistency in which variables were selected, than for the COOP models, which were assigned a different list of variables for nearly every month. These and other results suggest that global calibration is preferred because data from the global spectrum of physical processes that control surface temperatures are incorporated in a global model. All of the models that were developed in this study validated relatively well, especially the global models. Recalibration of the models with validation data resulted in only slightly poorer regression statistics, indicating that the calibration list of variables was valid. Predictions using data from the validation dataset in the calibrated equation were better for the GHCN models, and the globally calibrated GHCN models generally provided better U.S. predictions than the U.S.-calibrated COOP models. Overall, the GHCN and COOP models explained approximately 64%-95% of the total variance of surface shelter temperatures, depending on the month and the number of model variables. In addition, root-mean-square errors (rmse's) were over 3 C for GHCN models and over 2 C for COOP models for winter months, and near 2 C for GHCN models and near 1.5 C for COOP models for summer months.

Epperson, David L.↗
DOE OSTI2020

Solvent selection for biphasic extraction of 5-hydroxymethylfurfural via multiscale modeling and experiments

We introduce a comprehensive conceptual framework for selecting solvents for reactive extraction in biphasic organic-water systems and demonstrate it for the separation of HMF (5-hydroxylmethylfurfural), a platform chemical produced in the acid-catalyzed dehydration of hexoses. We first perform in silico screening of ~2500 solvents, from the ADFCRS-2018 database using the ADF COSMO-RS implementation, and classification, based on the solvent partition coefficient. We then determine experimentally the partition coefficients for HMF, fructose, and products of HMF rehydration (levulinic acid (LA), and formic acid (FA)), the mutual water-organic solvent solubilities, and the separation factors in >50 select solvents spanning multiple homologous series at room temperature and a typical reaction temperature with in situ sampling. We find that COSMO-RS is excellent for screening purposes (typical error in most cases within a factor of ~2). Increased temperatures lead to significant reduction in partitioning, and room temperature measurements are clearly inadequate for solvent selection. Upon down selecting classes of solvents based on separation performance, we perform experimental thermal stability and reaction compatibility studies of a small set of solvents at relevant reactive-extraction temperatures. We discover that many substituted phenols exhibit an order-of-magnitude increase in partitioning compared to conventional solvents due chiefly to hydrogen bond interactions and show the necessary stability but retain a significant fraction of water and LA, factors that need to be considered in technoeconomic analysis. In contrast, anilines, aldehydes, and acids are good to excellent regarding separation but incompatible with this specific reaction medium. Furthermore, this multifaceted framework can be extended to other biomass-derived products and processes.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

An open-source framework for balancing computational speed and fidelity in production cost models

Studies of bulk power system operations need to incorporate uncertainty and sensitivity analyses, especially around exposure to weather and climate variability and extremes, but this remains a computational modeling challenge. Commercial production cost models (PCMs) have shorter runtimes, but also important limitations (opacity, license restrictions) that do not fully support stochastic simulation. Open-source PCMs represent a potential solution. They allow for multiple, simultaneous runs in high-performance computing environments and offer flexibility in model parameterization. Yet, developers must balance computational speed (i.e. runtime) with model fidelity (i.e. accuracy). In this paper, we present Grid Operations (GO), a framework for instantiating open-source, scale-adaptive PCMs. GO allows users to search across parameter spaces to identify model versions that appropriately balance computational speed and fidelity based on experimental needs and resource limits. Results provide generalizable insights on how to navigate the fidelity and computational speed tradeoff through parameter selection. We show that models with coarser network topologies can accurately mimic market operations, sometimes better than higher-resolution models. It is thus possible to conduct large simulation experiments that characterize operational risks related to climate and weather extremes while maintaining sufficient model accuracy.

42 ENGINEERING↗
DOE OSTI2022

Development of Short-Term Forecasting Models Using Plant Asset Data and Feature Selection

Nuclear power plants collect and store large volumes of heterogeneous data from various components and systems. With recent advances in machine learning (ML) techniques, these data can be leveraged to develop diagnostic and short-term forecasting models to better predict future equipment condition. Maintenance operations can then be planned in advance whenever degraded performance is predicted, thus resulting in fewer unplanned outages and the optimization of maintenance activities. This enables lower maintenance costs and improves the overall economics of nuclear power. This paper focuses on developing a short-term forecasting process that leverages a feature selection process to distill large volumes of heterogeneous data and predict specific equipment parameters. A variety of feature selection methods, including Shapley Additive Explanations (SHAP) and variance inflation factor (VIF), were used to select the optimal features as inputs for three ML methods: long short-term memory (LSTM) networks, support vector regression (SVR), and random forest (RF). Each combination of model and input features was used to predict a pump bearing temperature both 1 and 24 hours in advance, based on actual plant system data. The optimal inputs for the LSTM and SVR were selected using the SHAP values, while the optimal input for the RF consisted solely of the response variable itself. Each model produced similar 1-hour-ahead predictions, with root mean square errors (RMSEs) of roughly 0.006. For the 24-hour-ahead predictions, differences could be seen between LSTM, SVR, and RF, as reflected by model performances of 0.036 +- 0.014, 0.0026 +- 0, and 0.063 +- 0.004 RMSE, respectively. As big data and continuous online monitoring become more widely available, the proposed feature selection process can be used for many applications beyond the prediction of process parameters within nuclear infrastructure.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2025

Performance evaluation of automated data-driven feature extraction and selection methods for practical and scalable building energy consumption prediction models

Here, this study quantifies the impact of automated feature engineering methods (feature extraction and selection) on the quality and accuracy of machine learning models that predict building energy consumption. The case study compares model performance for three main scenarios: baseline (no feature extraction and selection), feature extraction only, and feature extraction combined with feature selection (filter and/or wrapper methods) for fully trained machine learning models for 200 metered/sub-metered energy measurements across 118 real buildings. For consistency, the same machine learning model architecture (a black box deep learning neural network with probabilistic forecast output) was used for all scenarios. Based on results, all feature engineering methods provided noticeable prediction accuracy improvements (e.g., 29%-68% median prediction improvement) compared to baseline scenarios. However, in this application, feature selection methods provide little practical value due to their limited performance gains and high computational cost. Smarter algorithm development supported by better computational environments will be needed before feature selection methods can reliably and efficiently improve predictive model performance.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS1984

Applications of model structure determination to flight test data

Several statistical and information criteria need to be considered when selecting an adequate model. Incorrect stability and control derivates result from inadequate aerodynamic model structure. Stepwise regression is used to determine the structure for an adequate model. Flight data which covers a nonlinear aerodynamic model range may be analyzed as a single data set or partitioned into several distinct sets. Stepwise regression for model structure detemination and parameter estimation was successfully applied to three aircraft types (single engine general aviation, unaugmented modern jet fighter, jet transport).

Batterson, J. G.↗
DOE OSTI2021

Modeling of transport in carrier-selective contacts in silicon heterojunction solar cells

High-performance silicon heterojunction (SHJ) solar cells use carrier-selective contact structures based on hydrogentated amorphous Si (a-Si:H) to maximize collection of photogenerated carriers. The high open circuit voltages observed experimentally in SHJ cells require that the carrier-selective contacts provide selectivity and passivation. However, a microscopic understanding of the dynamics of carrier transport through the a-Si layer is currently lacking. In this paper, we explicitly simulate the transport of holes across the a-Si:H(i) layer using a novel kinetic Monte Carlo approach. The hole-selective contact structure investigated in this paper uses p-type doped a-Si:H(p) and intrinsic a-Si:H(i) on an n-type crystalline silicon wafer, where the selectivity is provided by the a-Si:H(p) and the passivation is provided by the a-Si:H(i). However, in addition to the passivation provided by the a-Si:H(i), this layer also creates a potential barrier to the collection of photogenerated holes. There have been experimental studies in the literature that have suggested that multi-phonon processes are the main transport mechanism that assists in the transport of holes across the intrinsic a-Si:H barrier. Simulations presented here show that multi-phonon injection of holes into the a-Si:H(i) layer is the rate limiting step for transport across the a-Si:H(i) layer. Our results indicate that multi-phonon transport is strongly dependent on the electric field at the a-Si:H(i)/c-Si heterointerface as well. Transport simulations presented in this paper are consistent with experimental findings that multi-phonon processes limit transport across the a-Si:H(i) layer and are responsible for photocurrent suppression at the a-Si:H(i)/c-Si heterointerface when these processes are slower than the associated incident hole flux due to photo-excitation.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2023

Recent advances in integrated hydrologic models: Integration of new domains

Over the past several decades, hydrologic models have advanced from independent models of the surface and subsurface to integrated models that can capture the terrestrial hydrologic cycle within one framework. In recent years, these coupled frameworks have seen the inclusion of biogeochemical processes, ecohydrology, sedimentation and erosion, cold region hydrology, anthropogenic activities, and atmospheric processes. This expansion is the result of increased computational, data, and modeling capabilities and capacities, as well as improved understanding of the processes that drive these integrated systems. Here, in this study, we review these recent advances to integrate new processes and systems into existing terrestrial hydrologic models and highlight the significant challenges and opportunities that remain. We identify that with so many models currently available and in development, selecting the most appropriate model is difficult, and we suggest a path for new or novice modelers to find the most appropriate code based on their needs. In addition, data required to parameterize and calibrate these models can often constrain their applicability and usefulness. However, advances in environmental sensors and measurement technology, in addition to data assimilation of non-traditional data (e.g. remote sensing, qualitative data) are providing new ways of addressing this issue. As we expand hydrologic models to integrate more processes and systems, our computational demands also increase. Recent and emerging advances in computational platforms, including cloud and quantum computing, in addition to the use of machine learning to capture some processes, will continue to support the use of increasingly larger and more complex, process-based models. Finally, we highlight that it is critical to develop state-of-the-science models that are accessible to all model users, not just those applied for research and development. We encourage continued development of diverse modeling platforms, considering the user needs, data availability, and computational resources.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2022

Fast and fault-tolerant model predictive control of MMCs under selective harmonic elimination

Selective harmonic elimination (SHE) is a suitable pulse-width modulation (PWM) tech-nique for some applications of modular multilevel converters (MMC) with lower numberof semiconductor submodules (SM). However, SHE can make it difficult to control circu-lating currents in the MMC. Moreover, the MMC is vulnerable to SM switch failures thatreduce its output power magnitude and quality. This paper proposes a fault-tolerant modelpredictive control strategy for MMCs operating with SHE (FT SHE-MPC). The effects ofswitch failure on phase voltage waveforms are studied first. Then, a method is proposedto design the FT SHE waveform to preserve its harmonic performance. Next, the pro-posed FT SHE-MPC is designed to increase the capacitor voltages during an SM failure tomaintain the fundamental component of the faulty phase voltage and control the circulat-ing currents, maintaining balance among line voltages and the output power of the MMC.Additionally, the proposed method offers as much as 60% faster FT transition time and asmuch as 36% lower capacitor voltage ripple. The advantages of the proposed method areverified through offline and real-time simulation studies on a three-phase nine-level MMCin PSCAD/EMTDC software and RTDS.

42 ENGINEERING↗
DOE OSTI2020

Validation of Continuous-Energy ENDF/B-VIII.0 16 O, 56 Fe, and 63,65 Cu Cross Sections for Nuclear Criticality Safety Applications

Recently completed cross-section evaluations sponsored in part by the Nuclear Criticality Safety Program were incorporated into the 2018 release of the ENDF/B-VIII.0 cross-section library. Evaluated isotopes of interest to the nuclear data and criticality safety community include 16 O, 56 Fe, and 63,65 Cu. For performance validation, benchmark models defined in the International Criticality Safety Benchmark Evaluation Project Handbook were selected based on energy-integrated k eff sensitivities to total cross sections of interest and compared with experimental values. Of the 102 benchmark configurations that were utilized, 63 are sensitive to 16 O, 32 sensitive to 63,65 Cu, and 25 sensitive to 56 Fe. Selected benchmarks were modeled in SCALE 6.2.3 Criticality Safety Analysis Sequence (CSAS) continuous-energy Monte Carlo k eff calculations with ENDF/B-VII.1, with a hybrid ENDF/B-VII.1 with ENDF/B-VIII.0 data substituted for individual isotopes of interest, and with ENDF/B-VIII.0. ENDF/B-VIII.0 showed improved agreement with experimental k eff for 56 Fe, 63 Cu, elemental copper, and full library substitution while producing lessened agreement for 16 O and 65 Cu. Furthermore, with full library and isotope-specific ENDF/B-VIII.0 performance, a best-case ENDF library was formed by excluding underperforming isotopes’ ENDF/B-VIII.0 data, reverting 16 O and 65 Cu cross sections to ENDF/B-VII.1. This resulted in the average relative deviation between calculated and experimental data improving from 1.45σ for the ENDF/B-VIII.0 library to 1.32σ for the best-case library, relative to benchmark uncertainty.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2024

Ab initio leading order effective potential for elastic proton scattering based on the symmetry-adapted no-core shell model

Calculating microscopic optical potentials for elastic scattering at intermediate energies from light nuclei in an ab initio fashion within the Watson expansion has been established within the last few years. Based on the Watson expansion of the multiple scattering series, we employ a nonlocal translationally invariant nuclear density derived within the symmetry-adapted no-core shell model (SA-NCSM) framework from a chiral next-to-next-to-leading order (NNLO) nucleon-nucleon interaction and the very same interaction for a consistent full-folding calculation of the effective (optical) potential for nucleon-nucleus scattering for medium-heavy nuclei. The leading order effective (optical) folding potential is computed by integrating over a translationally invariant SA-NCSM one-body scalar density, spin-projected momentum distribution, and the Wolfenstein amplitudes 𝐴, 𝐶, and 𝑀. The resulting nonlocal potentials serve as input for a momentum space Lippmann-Schwinger equation. In the SA-NCSM, the model space is systematically up-selected using Sp⁡(3,ℝ) symmetry considerations. For the light nucleus of 6 He, we establish a systematic selection scheme in the SA-NCSM for scattering observables. Then, we apply this scheme to calculations of scattering observables, such as differential cross sections, analyzing powers, and spin rotation functions for elastic proton scattering from 20 Ne and 40 Ca in the energy regime between 65 and 200 MeV, and compare to available data. Furthermore, our calculations show that the leading order effective nucleon-nucleus potential in the Watson expansion of multiple scattering theory obtained from an up-selected SA-NCSM model space describes 40 Ca elastic scattering observables reasonably well to about 60 degrees in the center-of-mass frame, which coincides roughly with the validity of the NNLO chiral interaction used to calculate both the nucleon-nucleon amplitudes and the one-body scalar and spin nuclear densities.

Ab initio calculations↗
DOE OSTI2023

Machine learning-based inversion for acoustic impedance with large synthetic training data: Workflow and data characterization

Where wells are sparse or training data are difficult to label with high-quality wireline-derived impedance logs, machine learning (ML)-based inversion of acoustic impedance typically depends on small training data sets, leading to biased prediction. We have advanced a novel workflow that applies large synthetic seismic training data to reduce facies-related bias. Using a geologically realistic model as the truth model, we randomly select sparse seed wells to perform sequential Gaussian simulation (SGS) for impedance models of the same geometry and simulate facies variability. We implement random forest regression on 30 features extracted from the synthetic volume. We observe that more seed wells tend to reduce facies-induced bias by sampling more types of facies, resulting in a better prediction. We then focus on the responses of SGS models to facies changes, the number of seed wells necessary for a useful synthetic model, and how much a synthetic model can help ML-based inversion. Here, we observe that the SGS synthetic training model outperforms well-direct training in general. For modeled clastic shore-zone systems in Miocene Gulf of Mexico, two or more seed wells are necessary for a significant reduction of root-mean-square error and outliners, and improvement of facies imaging. In a field-data test, we apply a similar workflow to quantitatively predict acoustic impedance, which is then converted to a sand-volume map at a high-frequency sequence (10–100 m), revealing detailed facies and sandstone patterns. Such results are valuable in many geologic and engineering applications, such as hydrocarbon and CO 2 reservoir prospecting, reserve estimation, simulation, etc.

3D seismic↗
DOE OSTI2026

Geomatchd

Geomatchd is a software tool to assist non-experts in the rapid conceptual design of geothermal district heating systems. This tool comprises three main parts: user-side heating and cooling demand profiles, district heating and cooling (DHC) network models for sizing and optimal layout selection, and geothermal system models for heat extraction calculation and cost comparison. The architecture of Geomatchd is opensource and is intended to simplify efforts by other coders to improve on existing functions, to add functions and to leverage available existing code and data – to improve the tool for public use. Geomatchd and the material supporting its development may provide a more intuitive window on district heating and the use of local shallow geothermal resources.

AS↗
DOE OSTI2021

A Deep-Learning-Based Muon Neutrino CCQE Selection for Searches Beyond the Standard Model with MicroBooNE

The anomalous Low Energy Excess (LEE) of electron neutrinos and antineutrinos in MiniBooNE has inspired both theories and entire experiments to probe the heart of its mystery. One such experiment is MicroBooNE. This dissertation presents an important facet of its LEE investigation: how a powerful systematic can be levied on this signal through parallel study of a highly correlated channel in muon neutrinos. This constraint serves to strengthen MicroBooNE's ability to confirm or validate the cause of the LEE and will lay the groundwork for future oscillation experiments in Liquid Argon Time Projection Chamber (LArTPC) detector experiments like SBN and DUNE. In addition, this muon channel can be used to test oscillations directly, demonstrated through the world's first muon neutrino disappearance search with LArTPC data.

Cianci, Davio↗