Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “performance modeling and prediction”

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 163 records · Page 9
All sourcesNASA NTRSDOE OSTI
DOE OSTI2020

The Performance of a Spectral Wave Model at Predicting Wave Farm Impacts

For renewable ocean wave energy to support global energy demands, wave energy converters (WECs) will likely be deployed in large numbers (farms), which will necessarily change the nearshore environment. Wave farm induced changes can be both helpful (e.g., beneficial habitat and coastal protection) and potentially harmful (e.g., degraded habitat, recreational, and commercial use) to existing users of the coastal environment. It is essential to estimate this impact through modeling prior to the development of a farm, and to that end, many researchers have used spectral wave models, such as Simulating WAves Nearshore (SWAN), to assess wave farm impacts. However, the validity of the approaches used within SWAN have not been thoroughly verified or validated. Herein, a version of SWAN, called Sandia National Laboratories (SNL)-SWAN, which has a specialized WEC implementation, is verified by comparing its wave field outputs to those of linear wave interaction theory (LWIT), where LWIT is theoretically more appropriate for modeling wave-body interactions and wave field effects. The focus is on medium-sized arrays of 27 WECs, wave periods, and directional spreading representative of likely conditions, as well as the impact on the nearshore. A quantitative metric, the Mean Squared Skill Score, is used. Results show that the performance of SNL-SWAN as compared to LWIT is “Good” to “Excellent”.

environmental impacts↗
DOE OSTI2020

Prediction of Performance Variation Caused by Manufacturing Tolerances and Defects in Gas Diffusion Electrodes of Phosphoric Acid (PA)–Doped Polybenzimidazole (PBI)-Based High-Temperature Proton Exchange Membrane Fuel Cells

The automated process of coating catalyst layers on gas diffusion electrodes (GDEs) for high-temperature proton exchange membrane fuel cells results inherently into a number of defects. These defects consist of agglomerates in which the platinum sites cannot be accessed by phosphoric acid and which are the consequence of an inconsistent coating, uncoated regions, scratches, knots, blemishes, folds, or attached fine particles—all ranging from μm to mm size. These electrochemically inactive spots cause a reduction of the effective catalyst area per unit volume (cm2/cm3) and determine a drop in fuel cell performance. A computational fluid dynamics (CFD) model is presented that predicts performance variation caused by manufacturing tolerances and defects of the GDE and which enables the creation of a six-sigma product specification for Advent phosphoric acid (PA)-doped polybenzimidazole (PBI)-based membrane electrode assemblies (MEAs). The model was used to predict the total volume of defects that would cause a 10% drop in performance. It was found that a 10% performance drop at the nominal operating regime would be caused by uniformly distributed defects totaling 39% of the catalyst layer volume (~0.5 defects/μm2). The study provides an upper bound for the estimation of the impact of the defect location on performance drop. It was found that the impact on the local current density is higher when the defect is located closer to the interface with the membrane. The local current density decays less than 2% in the presence of an isolated defect, regardless of its location along the active area of the catalyst layer.

Gurau, Vladimir (ORCID:0000000327429061)↗
DOE OSTI2022

Development and validation of a mathematical model of heart rate response to fluid perturbation

Physiological closed-loop controlled (PCLC) medical devices monitor and automatically adjust the patient’s condition by using physiological variables as feedback, ideally with minimal human intervention to achieve the target levels set by a clinician. PCLC devices present a challenge when it comes to evaluating their performance, where conducting large clinical trials can be expensive. Virtual physiological patients simulated by validated mathematical models can be utilized to obtain preclinical evidence of safety and assess the performance of the PCLC medical device during normal and worst-case conditions that are unlikely to happen in a limited clinical trial. A physiological variable that plays a major role during fluid resuscitation is heart rate (HR). For in silico assessment of PCLC medical devices regarding fluid perturbation, there is currently no mathematical model of HR validated in terms of its predictive capability performance. This paper develops and validates a mathematical model of HR response using data collected from sheep subjects undergoing hemorrhage and fluid infusion. The model proved to be accurate in estimating the HR response to fluid perturbation, where averaged between 21 calibration datasets, the fitting performance showed a normalized root mean square error (NRMSE) of 7.41 ± 2.8%. The model was also evaluated in terms of model predictive capability performance via a leave-one-out procedure (21 subjects) and an independent validation dataset (6 subjects). Two different virtual cohort generation tools were used in each validation analysis. The generated envelope of virtual subjects robustly met the defined acceptance criteria, in which 95% of the testing datasets presented simulated HR patterns that were within a deviation of 50% from the observed data. In addition, out of 16000 and 18522 simulated subjects for the leave one-out and independent datasets, the model was able to generate at least one virtual subject that was close to the real subject within an error margin of 9.56 ± 3.15% and 11.1 ± 1.22% NRMSE, respectively. In conclusion, the model can generate valid virtual HR physiological responses to fluid perturbation and be incorporated into future non-clinical simulated testing setups for assessing PCLC devices intended for fluid resuscitation.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2021

Design Space Exploration of Emerging Memory Technologies for Machine Learning Applications

Memory design space exploration methods study memory systems’ performances and limitations before implementation. The computer memory design space has grown exponentially because of the enormous growth of memory types, memory controllers, and application software. Computer simulators are commonly used for memory design space exploration. However, complex memory simulations take an enormous amount of time. Hence, in this paper, we proposed a machine learning-based design space exploration method for dynamic random-access memory and non-volatile memory systems. We applied our method to the CosmoGAN and LeNet applications to predict the following six memory response parameters: (i) bandwidth, (ii) power, (iii) average latency, (iv) average total latency, (v) memory reads, and (vi) memory writes. Our experimental results show that machine learning models can predict memory response parameter values faster than simulations. We used support vector machine, random forest, and gradient boosting machine learning models. We observed that the support vector machine provides better performance for bandwidth, average latency, and average total latency. The random forest model works better for memory reads and writes. The gradient boosting model provides superior prediction performance for power. We provide a detailed discussion on learning curve characteristics, error analysis, and memory type recommendation.

Hasan, S M Shamimul↗
DOE OSTI2021

Coupled Charge and Radiation Transport Processes in Thermophotovoltaic and Thermoradiative Cells

Accurate modeling of charge transport and both thermal and luminescent radiation is crucial to the understanding and design of radiative thermal energy converters. Charge-carrier dynamics in semiconductors are well-described by the Poisson-drift-diffusion equations, and thermal radiation in emitter-absorber structures can be computed using multilayer fluctuational electrodynamics. These two types of energy flows interact through radiation absorption or luminescence and charge-carrier generation or recombination. However, past research has typically only assumed limited interaction, with thermal radiation absorption as an input for charge-carrier models to predict device performance. To examine this assumption, we develop a fully coupled iterative model of charge and radiation transport in semiconductor devices, and we use our model to analyze near-field and far-field GaSb thermophotovoltaic and thermoradiative systems. By comparing our results with past methods that do not consider cross-influences between charge and radiation transport, we find that a fully coupled approach is necessary to accurately model photon recycling and near-field enhancement of external luminescence. As these effects can substantially alter device performance, our modeling approach can aid in the design of efficient thermophotovoltaic and thermoradiative systems.

30 DIRECT ENERGY CONVERSION↗
DOE OSTI2024

ML-based Dimension Reduction Strategies

Deep learning (DL)--based surrogate models have achieved success in various applications in carbon capture and storage (CCS). However, the model training on high-dimensional spaces is computationally expensive and impractical for large-scale and complex geological models, because the models usually contain hundreds of thousands to millions of grid cells, each with a set of parameters. Furthermore, the high cost of generating training data with sufficient variation is another limitation of model training on high-dimensional spaces, which may result in overfitting and reduce the model efficiency and prediction performance. We proposed the workflow incorporating dimension reduction methods and deep learning models, which aim to extract the latent variables of input parameters and output state variables, and then build the mapping function at the latent spaces. The proposed workflow can significantly reduce the computational complexity in solving both forward and inverse problems compared to models trained on high-dimensional spaces. Dimensionality reduction models showed great potential in workflows for fast reservoir simulation, history matching, prior model generation, visualization, and more, ultimately enhancing DL model performance in related SMART Work Packages.

Hosseini, Seyyed↗
DOE OSTI2023

Evaluation of BISON metallic fuel performance modeling against experimental measurements within FIPD and IMIS databases

Simulations were conducted using the BISON fuel performance code on an automated process to read initial and operating conditions from two databases—the Fuels Irradiation and Physics Database (FIPD) and Integral Fast Reactor Materials Information System (IMIS) database. These databases contain metallic fuel data from the Experimental Breeder Reactor-II (EBR-II) and the Fast Flux Test Facility (FFTF). The work demonstrates use of an integrated framework to access EBR-II fuel pin data for evaluating fuel performance models contained within BISON to predict fuel performance of next-generation metallic fuel systems. Between IMIS and FIPD, there is enough information to conduct 1,977 unique EBR-II metallic fuel pin histories from 29 different experiments, and 338 pins from FFTF MFF-3 and MFF-5 with varying levels of details between the two databases. Each of these fuel performance histories includes a high-resolution power history, flux history, coolant channel flow rates, and coolant channel temperatures, and new model developments in BISON since the initial demonstration of this integrated framework. Fission gas release (FGR), cumulative damage fraction, fuel axial swelling, FCCI wastage thickness, cladding profilometry, and burnup were all simulated in BISON and compared to post-irradiation examination (PIE) results to evaluate BISON fuel performance modeling. Implementation of new fuel performance models into a generic BISON input file coupled with IMIS and FIPD yielded results with a better representation of physics than the initial evaluation of the integrated framework. Cladding profilometry, FGR, and fuel axial swelling were found to be in good agreement with PIE measurements for most of the pins simulated. The chosen mechanical contact solver was found to significantly impact the axial fuel swelling and cladding strain predictions when used in conjunction with the U-Pu-Zr hot-pressing model since it bound the fuel to prevent further swelling and increased hydrostatic stresses. This work suggests that fuel performance modeling in BISON under steady-state conditions represents the PIE data well and should be reassessed when new PIE data become available in IMIS and FIPD databases and when improved physical models to better capture fuel performance are added to BISON.

Paaren, Kyle M.↗
DOE OSTI2023

A multiscale recurrent neural network model for predicting energy production from geothermal reservoirs

Optimization of energy production from geothermal reservoirs requires reliable prediction of energy production performance under alternative operation and development scenarios. Traditionally, reservoir simulation models are used for the evaluation and screening of alternative production and development plans. However, simulation models require extensive data collection and modeling efforts and are time-consuming to build, run, and update. Data-driven predictive models, on the other hand, can serve as efficient prediction tools that can be used for decision support and management of daily operations and surveillance activities. Data-driven models become particularly attractive when a reservoir simulation model for a field does not exist and/or is difficult to build. Machine learning (ML)-based data-driven models that have recently become popular in several fields exploit statistical patterns and relations in training data to generate predictions. As such, they tend to perform better in interpolation problems (that is, prediction within the training data range) than when they are used to extrapolate beyond the training data. Production data from geothermal reservoirs tend to exhibit short-term variabilities as well as long-term trends, such as monotonically declining production temperatures. Capturing both short-term features and long-term trends with ML-based models is not trivial. We evaluate the use of recurrent neural networks (RNN) for the prediction of energy production from geothermal reservoirs. RNN is a class of ML architectures that are used to represent and predict sequential/dynamic data. Thus, it can be challenging to apply RNN to problems where long-term trends must be captured and extrapolation beyond the training data range is needed. We introduce the multiscale RNN architecture to extend the application of RNN to detect and predict both short-term variabilities and long-term trends in geothermal data. The developed architecture consists of a long-term component to only capture low-frequency data patterns, and a short-term component to detect features with higher frequency and more nonlinearity. The final prediction is obtained by combining the long-term and short-term predictions. Both synthetic and field data are used to evaluate the presented multiscale RNN model. The prediction performance of the multiscale RNN is compared against those obtained from the regular RNN and the autoregressive (AR) model. The results suggest that the multiscale architecture improves the long-term prediction performance of the regular RNN and enhances its robustness against noise.

15 GEOTHERMAL ENERGY↗
DOE OSTI2024

Prognostic analysis of high-flow nasal cannula therapy and non-invasive ventilation in mild to moderate hypoxemia patients and construction of a machine learning model for 48-h intubation prediction—a retrospective analysis of the MIMIC database

Background This study aims to investigate the clinical outcome between high-flow nasal cannula (HFNC) and non-invasive ventilation (NIV) therapy in mild to moderate hypoxemic patients on the first ICU day and to develop a predictive model of 48-h intubation. Methods The study included adult patients from the MIMIC III and IV databases who first initiated HFNC or NIV therapy due to mild to moderate hypoxemia (100 < PaO2/FiO2 ≤ 300). The 48-h and 30-day intubation rates were compared using cross-sectional and survival analysis. Nine machine learning and six ensemble algorithms were deployed to construct the 48-h intubation predictive models, of which the optimal model was determined by its prediction accuracy. The top 10 risk and protective factors were identified using the Shapley interpretation algorithm. Result A total of 123,042 patients were screened, of which, 673 were from the MIMIC IV database for ventilation therapy comparison (HFNC n = 363, NIV n = 310) and 48-h intubation predictive model construction (training dataset n = 471, internal validation set n = 202) and 408 were from the MIMIC III database for external validation. The NIV group had a lower intubation rate (23.1% vs. 16.1%, p = 0.001), ICU 28-day mortality (18.5% vs. 11.6%, p = 0.014), and in-hospital mortality (19.6% vs. 11.9%, p = 0.007) compared to the HFNC group. Survival analysis showed that the total and 48-h intubation rates were not significantly different. The ensemble AdaBoost decision tree model (internal and external validation set AUROC 0.878, 0.726) had the best predictive accuracy performance. The model Shapley algorithm showed Sequential Organ Failure Assessment (SOFA), acute physiology scores (APSIII), the minimum and maximum lactate value as risk factors for early failure and age, the maximum PaCO 2 and PH value, Glasgow Coma Scale (GCS), the minimum PaO 2 /FiO 2 ratio, and PaO 2 value as protective factors. Conclusion NIV was associated with lower intubation rate and ICU 28-day and in-hospital mortality. Further survival analysis reinforced that the effect of NIV on the intubation rate might partly be attributed to the other impact factors. The ensemble AdaBoost decision tree model may assist clinicians in making clinical decisions, and early organ function support to improve patients’ SOFA, APSIII, GCS, PaCO 2 , PaO 2 , PH, PaO 2 /FiO 2 ratio, and lactate values can reduce the early failure rate and improve patient prognosis.

Fu, Wei↗
DOE OSTI2023

Current data are consistent with flat spatial hypersurfaces in the Λ CDM cosmological model but favor more lensing than the model predicts

Here, we study the performance of three pairs of tilted, and a pair of untilited, ΛCDM cosmological models, with three of these four pairs allowing for non-flat spatial hypersurfaces, against cosmic microwave background (CMB) temperature and polarization power spectrum data (P18), measurements of the Planck 2018 lensing potential power spectrum (lensing), and a large compilation of non-CMB data (non-CMB). For the eight models, we measure cosmological parameters and study whether or not pairs of the data sets (as well as subsets of them) are mutually consistent in these models. Half of these models allow the lensing consistency parameter A L , which re-scales the gravitational potential power spectrum, to be an additional free parameter to be determined from data, while the other three have A L = 1 which is the theoretically expected value. The pair of untilted non-flat ΛCDM models are incompatible with P18 data. The tilted spatially-flat models assume the usual primordial spatial inhomogeneity power spectrum that is a power law in wave number. The tilted non-flat models assume either the primordial power spectrum used in the Planck group anal yses [Planck P(q)], that has recently been numerically shown to be a good approximation to what is quantum-mechanically generated from a particular choice of closed inflation model initial conditions, or a recently computed power spectrum [new P(q)] that quantum-mechanically follows from a different set of non-flat inflation model initial conditions. In the tilted non-flat models with A L = 1 we find differences between P18 data and non-CMB data cosmological parameter constraints, which are large enough to rule out the Planck P(q) model at 3σ but not the new P(q) model. No significant differences are found when cosmological parameter constraints obtained with two different data sets are compared within the standard tilted flat ΛCDM model. While both P18 data and non-CMB data separately favor a closed geometry, with spatial curvature density parameter Ω k < 0, when P18+non-CMB data are jointly analyzed the evidence in favor of non-flat hypersurfaces subsides. Differences between P18 data and non-CMB data cosmological constraints subside when A L is allowed to vary. From the most restrictive P18+lensing+non-CMB data combination we get almost model-independent constraints on the cosmological parameters and find that the A L > 1 option is preferred over the Ω k < 0 one, with the A L parameter, for all models, being larger than unity by ~ 2.5σ. According to the deviance information criterion, in the P18+lensing+non-CMB analysis, the varying A L option is on the verge of being strongly favored over the A L = 1 one, which could indicate a problem for the standard tilted flat ΛCDM model. These data are consistent with flat spatial hypersurfaces but more and better data could improve the constraints on Ω k and might alter this conclusion. Error bars on some cosmological parameters are significantly reduced when non-CMB data are used jointly with P18+lensing data. For example, in the tilted flat ΛCDM model for P18+lensing+non-CMB data the Hubble constant H 0 = 68.09 ± 0.38 km s -1 Mpc -1 , which is consistent with that from a median statistics analysis of a large compilation of H 0 measurements as well as with a number of local measurements of the cosmological expansion rate. This H 0 error bar is 31% smaller than that from P18+lensing data alone.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2023

Machine Learning for First Principles Calculations of Material Properties for Ferromagnetic Materials

The investigation of finite temperature properties using Monte-Carlo (MC) methods requires a large number of evaluations of the system’s Hamiltonian to sample the phase space needed to obtain physical observables as function of temperature. DFT calculations can provide accurate evaluations of the energies, but they are too computationally expensive for routine simulations. To circumvent this problem, machine-learning (ML) based surrogate models have been developed and implemented on high-performance computing (HPC) architectures. In this paper, we describe two ML methods (linear mixing model and HydraGNN) as surrogates for first principles density functional theory (DFT) calculations with classical MC simulations. These two surrogate models are used to learn the dependence of target physical properties from complex compositions and interactions of their constituents. We present the predictive performance of these two surrogate models with respect to their complexity while avoiding the danger of overfitting the model. An important aspect of our approach is the periodic retraining with newly generated first principles data based on the progressive exploration of the system’s phase space by the MC simulation. The numerical results show that HydraGNN model attains superior predictive performance compared to the linear mixing model for magnetic alloy materials.

Eisenbach, Markus↗
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 OSTI2021

Improving protein tertiary structure prediction by deep learning and distance prediction in CASP14

Abstract Substantial progresses in protein structure prediction have been made by utilizing deep‐learning and residue‐residue distance prediction since CASP13. Inspired by the advances, we improve our CASP14 MULTICOM protein structure prediction system by incorporating three new components: (a) a new deep learning‐based protein inter‐residue distance predictor to improve template‐free (ab initio) tertiary structure prediction, (b) an enhanced template‐based tertiary structure prediction method, and (c) distance‐based model quality assessment methods empowered by deep learning. In the 2020 CASP14 experiment, MULTICOM predictor was ranked seventh out of 146 predictors in tertiary structure prediction and ranked third out of 136 predictors in inter‐domain structure prediction. The results demonstrate that the template‐free modeling based on deep learning and residue‐residue distance prediction can predict the correct topology for almost all template‐based modeling targets and a majority of hard targets (template‐free targets or targets whose templates cannot be recognized), which is a significant improvement over the CASP13 MULTICOM predictor. Moreover, the template‐free modeling performs better than the template‐based modeling on not only hard targets but also the targets that have homologous templates. The performance of the template‐free modeling largely depends on the accuracy of distance prediction closely related to the quality of multiple sequence alignments. The structural model quality assessment works well on targets for which enough good models can be predicted, but it may perform poorly when only a few good models are predicted for a hard target and the distribution of model quality scores is highly skewed. MULTICOM is available at https://github.com/jianlin-cheng/MULTICOM_Human_CASP14/tree/CASP14_DeepRank3 and https://github.com/multicom-toolbox/multicom/tree/multicom_v2.0 .

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2023

3D-equivariant graph neural networks for protein model quality assessment

Quality assessment (QA) of predicted protein tertiary structure models plays an important role in ranking and using them. With the recent development of deep learning end-to-end protein structure prediction techniques for generating highly confident tertiary structures for most proteins, it is important to explore corresponding QA strategies to evaluate and select the structural models predicted by them since these models have better quality and different properties than the models predicted by traditional tertiary structure prediction methods. We develop EnQA, a novel graph-based 3D-equivariant neural network method that is equivariant to rotation and translation of 3D objects to estimate the accuracy of protein structural models by leveraging the structural features acquired from the state-of-the-art tertiary structure prediction method—AlphaFold2. We train and test the method on both traditional model datasets (e.g. the datasets of the Critical Assessment of Techniques for Protein Structure Prediction) and a new dataset of high-quality structural models predicted only by AlphaFold2 for the proteins whose experimental structures were released recently. Our approach achieves state-of-the-art performance on protein structural models predicted by both traditional protein structure prediction methods and the latest end-to-end deep learning method—AlphaFold2. It performs even better than the model QA scores provided by AlphaFold2 itself. The results illustrate that the 3D-equivariant graph neural network is a promising approach to the evaluation of protein structural models. Integrating AlphaFold2 features with other complementary sequence and structural features is important for improving protein model QA.

59 BASIC BIOLOGICAL SCIENCES↗