SEARCH · Engineering Papers

Results for “ensemble 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.

SourceResearch connectionFrom yearThrough year

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

At least 433 records · Page 24

All sources NASA NTRS DOE OSTI

DOE OSTI2024

Selection of representative general circulation models under climatic uncertainty for Western North America

Abstract Climate change research uses an ensemble of general circulation model runs (GCMs-runs) to predict future climate under uncertainties. To reduce computational costs, this study selects representative GCM-runs (RGCM-runs) for Western North America (WNA) based on their performance in replicating historical climate conditions from 1981 to 2005 and projecting future changes from 1981–2010 to 2071–2100. This evaluation is conducted under two representative concentration pathways (RCPs) scenarios, RCP4.5 and RCP8.5, from the Coupled Model Intercomparison Project 5. By using an envelope-based selection technique and a multi-objective distance-based approach, we identify four RGCM-runs per RCP representing diverse climatic conditions, including wet-warm, wet-cold, dry-warm, and dry-cold. Compared to the full-set, these selected runs show a decreased mean absolute error (MAE) between the reference and RGCM-runs concerning the monthly average mean air temperature (T̄) and precipitation (P̄). For RCP4.5, T̄ MAE is 0.45 (vs. 0.58 in the full-set) and P̄ MAE is 0.31 (vs. 0.42). For RCP8.5, T̄ MAE is 0.51 (vs. 0.75) and P̄ MAE is 0.25 (vs. 0.36). The lower MAE values in the RGCM-run set indicate closer alignment between predicted and reference values, making the RGCM-run suitable for climate impact assessments in the region.

Water Resources↗

NASA NTRS2012

Global Modeling and Assimilation Office Annual Report and Research Highlights 2011-2012

Over the last year, the Global Modeling and Assimilation Office (GMAO) has continued to advance our GEOS-5-based systems, updating products for both weather and climate applications. We contributed hindcasts and forecasts to the National Multi-Model Ensemble (NMME) of seasonal forecasts and the suite of decadal predictions to the Coupled Model Intercomparison Project (CMIP5).

Rienecker, Michele M.↗

NASA NTRS2019

MERRA-2 Ocean: The NASA Global Modeling and Assimilation Office's Weakly Coupled Atmosphere-Ocean Reanalysis Using GEOS-S2S Version 3

The NASA Modern Era Reanalysis for Research and Applications (MERRA2) has been a respected and widely used reanalysis that has so far been restricted to the atmosphere. Now a newly released version of the atmosphere/ocean coupled data assimilation system (AODAS) has been developed by the NASA/Goddard Global Modeling and Assimilation Office to perform a retrospective ocean reanalysis from 1982 to present. In addition to assimilating all available in situ data (e.g. Argo, mooring, XBT and CTD data) and altimetry information into the ocean, the new version (GEOS-S2S Version 3) model includes a higher resolution, eddy-permitting ocean model than previous versions, a more realistic implementation of the atmosphere-ocean interface layer, and an improved coupling between glacier and ocean (among other improvements). In addition, this ocean data assimilation was expanded to include the assimilation of satellite sea surface salinity. The MERRA-2 AODAS will be described, and preliminary results will be shown from the assimilation reanalysis and from retrospective forecasts issued using a new ensemble strategy. Following the Global Ocean Data Assimilation Experiment (GODAE) protocols, we will present Class 1 through Class 4 validation results from the ocean reanalysis. Results indicate an improved ocean mixed layer depth, improved salinity near Greenland, an improved diurnal cycle of the sea surface skin temperature, an improved estimate of ocean evaporation, and better representation of western boundary currents (e.g. Gulf Stream) from our new ocean reanalysis. One of the motivations of this project is to provide optimal initial states for ENSO forecasting. Therefore, we will also present some preliminary results of retrospective ENSO forecasts. After thorough testing, it is expected that the GEOS-S2S Version 3 will replace our contributions to North American Multi-Model Ensemble (NMME), WCRP Subseasonal to Seasonal (S2S), and IRI seasonal prediction forecast projects.

Molod, Andrea↗

DOE OSTI2020

Machine Learning-Based Prediction of Distribution Network Voltage and Sensors Allocation: Preprint

Increasing penetration of fast-varying energy resources may negatively affect power systems' operation. At the same time, sensor deployment throughout distribution networks improves system awareness and enables the development of new and advanced voltage control solutions. Such control techniques rely on accurate prediction in anticipation to voltage violation scenarios. This paper analyzes various approaches for voltage prediction in a distribution system; it is shown that combining multiple techniques into a single regressor improves its predictive power. Moreover, a two-step regressor is proposed, where initial predictions based on a global regressor are refined by local regressors; in this case, prediction errors decrease significantly. Additionally, a clustering approach is employed for performing sensor allocation, so that only the most influential buses are selected for monitoring without diminishing prediction accuracy.

24 POWER TRANSMISSION AND DISTRIBUTION↗

DOE OSTI2020

Machine Learning-Based Prediction of Distribution Network Voltage and Sensor Allocation

61 RADIATION PROTECTION AND DOSIMETRY↗

DOE OSTI2024

Computationally efficient and error aware surrogate construction for numerical solutions of subsurface flow through porous media

Limiting the injection rate to restrict the pressure below a threshold at a critical location can be an important goal of simulations that model the subsurface pressure between injection and extraction wells. The pressure is approximated by the solution of Darcy’s partial differential equation for a given permeability field. The subsurface permeability is modeled as a random field since it is known only up to statistical properties. This induces uncertainty in the computed pressure. Solving the partial differential equation for an ensemble of random permeability simulations enables estimating a probability distribution for the pressure at the critical location. These simulations are computationally expensive, and practitioners often need rapid online guidance for real-time pressure management. An ensemble of numerical partial differential equation solutions is used to construct a Gaussian process regression model that can quickly predict the pressure at the critical location as a function of the extraction rate and permeability realization. The Gaussian process surrogate analyzes the ensemble of numerical pressure solutions at the critical location as noisy observations of the true pressure solution, enabling robust inference using the conditional Gaussian process distribution. Our first novel contribution is to identify a sampling methodology for the random environment and matching kernel technology for which fitting the Gaussian process regression model scales as O ( n log n ) instead of the typical O ( n 3 ) rate in the number of samples n used to fit the surrogate. The surrogate model allows almost instantaneous predictions for the pressure at the critical location as a function of the extraction rate and permeability realization. Our second contribution is a novel algorithm to calibrate the uncertainty in the surrogate model to the discrepancy between the true pressure solution of Darcy’s equation and the numerical solution. Finally, although our method is derived for building a surrogate for the solution of Darcy’s equation with a random permeability field, the framework broadly applies to solutions of other partial differential equations with random coefficients.

54 ENVIRONMENTAL SCIENCES↗

NASA NTRS2001

Are US Great Plains Droughts Predictable on Seasonal and Longer Time Scales?

The United States has experienced numerous episodes of unusually dry conditions lasting anywhere from months to several years. In this talk, I will examine the predictability of such episodes and the physical mechanisms controlling the variability of the summer climate of the continental United States. The analysis is based on ensembles of multi-year simulations and seasonal hindcasts generated with the NASA Seasonal-to-Interannual Prediction Project (NSIPP-1) General Circulation Model.

Schubert, Siegfried↗

NASA NTRS2001

Are Droughts in the United States Great Plains Predictable on Seasonal and Longer Time Scales?

The United States Great Plains has experienced numerous episodes of unusually dry conditions lasting anywhere from months to several years, In this presentation, we will examine the predictability of such episodes and the physical mechanisms controlling the variability of the summer climate of the continental United States. The analysis is based on ensembles of multi-year simulations and seasonal hindcasts generated with the NASA Seasonal to-Interannual Prediction Project (NSIPP-1) General Circulation Model.

Schubert, Siegfried D.↗

NASA NTRS2018

Prediction and Predictability of the Madden Julian Oscillation in the NASA GEOS-5 Seasonal-to-Subseasonal System

In this study, we examine the prediction skill and predictability of the Madden Julian Oscillation (MJO) in a recent version of the NASA GEOS-5 atmosphere-ocean coupled model run at at 1/2 degree horizontal resolution. The results are based on a suite of hindcasts produced as part of the NOAA SubX project, consisting of seven ensemble members initialized every 5 days for the period 1999-2015. The atmospheric initial conditions were taken from the Modern-Era Retrospective analysis for Research and Applications, Version 2 (MERRA-2), and the ocean and the sea ice were taken from a GMAO ocean analysis. The land states were initialized from the MERRA-2 land output, which is based on observation-corrected precipitation fields. We investigated the MJO prediction skill in terms of the bivariate correlation coefficient for the real-time multivariate MJO (RMM) indices. The correlation coefficient stays at or above 0.5 out to forecast lead times of 26-36 days, with a pronounced increase in skill for forecasts initialized from phase 3, when the MJO convective anomaly is located in the central tropical Indian Ocean. A corresponding estimate of the upper limit of the predictability is calculated by considering a single ensemble member as the truth and verifying the ensemble mean of the remaining members against that. The predictability estimates fall between 35-37 days (taken as forecast lead when the correlation reaches 0.5) and are rather insensitive to the initial MJO phase. The model shows slightly higher skill when the initial conditions contain strong MJO events compared to weak events, although the difference in skill is evident only from lead 1 to 20. Similar to other models, the RMM-index-based skill arises mostly from the circulation components of the index. The skill of the convective component of the index drops to 0.5 by day 20 as opposed to day 30 for circulation fields. The propagation of the MJO anomalies over the Maritime Continent does not appear problematic in the GEOS-5 hindcasts implying that the Maritime Continent predictability barrier may not be a major concern in this model. Finally, the MJO prediction skill in this version of GEOS-5 is superior to that of the current seasonal prediction system at the GMAO; this could be partly attributed to a slightly better representation of the MJO in the free running version of this model and partly to the improved atmospheric initialization from MERRA-2.

Achuthavarier, Deepthi↗

DOE OSTI2021

A consensus-based ensemble approach to improve transcriptome assembly

Systems-level analyses, such as differential gene expression analysis, co-expression analysis, and metabolic pathway reconstruction, depend on the accuracy of the transcriptome. Multiple tools exist to perform transcriptome assembly from RNAseq data. However, assembling high quality transcriptomes is still not a trivial problem. This is especially the case for non-model organisms where adequate reference genomes are often not available. Different methods produce different transcriptome models and there is no easy way to determine which are more accurate. Furthermore, having alternative-splicing events exacerbates such difficult assembly problems. While benchmarking transcriptome assemblies is critical, this is also not trivial due to the general lack of true reference transcriptomes. In this study, we first provide a pipeline to generate a set of the simulated benchmark transcriptome and corresponding RNAseq data. Using the simulated benchmarking datasets, we compared the performance of various transcriptome assembly approaches including both de novo and genome-guided methods. The results showed that the assembly performance deteriorates significantly when alternative transcripts (isoforms) exist or for genome-guided methods when the reference is not available from the same genome. To improve the transcriptome assembly performance, leveraging the overlapping predictions between different assemblies, we present a new consensus-based ensemble transcriptome assembly approach, ConSemble. Without using a reference genome, ConSemble using four de novo assemblers achieved an accuracy up to twice as high as any de novo assemblers we compared. When a reference genome is available, ConSemble using four genome-guided assemblies removed many incorrectly assembled contigs with minimal impact on correctly assembled contigs, achieving higher precision and accuracy than individual genome-guided methods. Furthermore, ConSemble using de novo assemblers matched or exceeded the best performing genome-guided assemblers even when the transcriptomes included isoforms. We thus demonstrated that the ConSemble consensus strategy both for de novo and genome-guided assemblers can improve transcriptome assembly. The RNAseq simulation pipeline, the benchmark transcriptome datasets, and the script to perform the ConSemble assembly are all freely available from: http://bioinfolab.unl.edu/emlab/consemble/.

59 BASIC BIOLOGICAL SCIENCES↗

NASA NTRS2014

The Impact of Microphysics and Planetary Boundary Layer Physics on Model Simulation of U.S. Deep South Summer Convection

Inspection of output from various configurations of high-resolution, explicit convection forecast models such as the Weather Research and Forecasting (WRF) model indicates significant sensitivity to the choices of model physics pararneterizations employed. Some of the largest apparent sensitivities are related to the specifications of the cloud microphysics and planetary boundary layer physics packages. In addition, these sensitivities appear to be especially pronounced for the weakly-sheared, multicell modes of deep convection characteristic of the Deep South of the United States during the boreal summer. Possible ocean-land sensitivities also argue for further examination of the impacts of using unique ocean-land surface initialization datasets provided by the NASA Short-term Prediction Research and Transition (SPoRn Center to select NOAAlNWS weather forecast offices. To obtain better quantitative understanding of these sensitivities and also to determine the utility of the ocean-land initialization data, we have executed matrices of regional WRF forecasts for selected convective events near Mobile, AL (MOB), and Houston, TX (HGX). The matrices consist of identically initialized WRF 24-h forecasts using any of eight microphysics choices and any of three planetary boWldary layer choices. The resulting 24 simulations performed for each event within either the MOB or HGX regions are then compared to identify the sensitivities of various convective storm metrics to the physics choices. Particular emphasis is placed on sensitivities of precipitation timing, intensity, and coverage, as well as amount and coverage oflightuing activity diagnosed from storm kinematics and graupel in the mixed phase layer. The results confirm impressions gleaned from study of the behavior of variously configured WRF runs contained in the ensembles produced each spring at the Center for the Analysis and Prediction of Storms, but with the benefit of more straightforward control of the physics package choices. The design of the experiments thus allows for more direct interpretation of the sensitivities to each possible physics combination. The results should assist forecasters in their efforts to anticipate and correct for possible biases in simulated WRF convection patterns, and help the modeling community refine their model parameterizations.

McCaul, Eugene W., Jr.↗

NASA NTRS2014

The Impacts of Microphysics and Planetary Boundary Layer Physics on Model Simulations of U. S. Deep South Summer Convection

Inspection of output from various configurations of high-resolution, explicit convection forecast models such as the Weather Research and Forecasting (WRF) model indicates significant sensitivity to the choices of model physics parameterizations employed. Some of the largest apparent sensitivities are related to the specifications of the cloud microphysics and planetary boundary layer physics packages. In addition, these sensitivities appear to be especially pronounced for the weakly-sheared, multicell modes of deep convection characteristic of the Deep South of the United States during the boreal summer. Possible ocean-land sensitivities also argue for further examination of the impacts of using unique ocean-land surface initialization datasets provided by the NASA Short-term Prediction Research and Transition (SPoRT Center to select NOAA/NWS weather forecast offices. To obtain better quantitative understanding of these sensitivities and also to determine the utility of the ocean-land initialization data, we have executed matrices of regional WRF forecasts for selected convective events near Mobile, AL (MOB), and Houston, TX (HGX). The matrices consist of identically initialized WRF 24-h forecasts using any of eight microphysics choices and any of three planetary boundary layer choices. The resulting 24 simulations performed for each event within either the MOB or HGX regions are then compared to identify the sensitivities of various convective storm metrics to the physics choices. Particular emphasis is placed on sensitivities of precipitation timing, intensity, and coverage, as well as amount and coverage of lightning activity diagnosed from storm kinematics and graupel in the mixed phase layer. The results confirm impressions gleaned from study of the behavior of variously configured WRF runs contained in the ensembles produced each spring at the Center for the Analysis and Prediction of Storms, but with the benefit of more straightforward control of the physics package choices. The design of the experiments thus allows for more direct interpretation of the sensitivities to each possible physics combination. The results should assist forecasters in their efforts to anticipate and correct for possible biases in simulated WRF convection patterns, and help the modeling community refine their model parameterizations.

McCaul, E. W., Jr.↗

DOE OSTI2023

Review of machine learning and deep learning models for toxicity prediction

The ever-increasing number of chemicals has raised public concerns due to their adverse effects on human health and the environment. To protect public health and the environment, it is critical to assess the toxicity of these chemicals. Traditional in vitro and in vivo toxicity assays are complicated, costly, and time-consuming and may face ethical issues. These constraints raise the need for alternative methods for assessing the toxicity of chemicals. Recently, due to the advancement of machine learning algorithms and the increase in computational power, many toxicity prediction models have been developed using various machine learning and deep learning algorithms such as support vector machine, random forest, k-nearest neighbors, ensemble learning, and deep neural network. This review summarizes the machine learning- and deep learning-based toxicity prediction models developed in recent years. Support vector machine and random forest are the most popular machine learning algorithms, and hepatotoxicity, cardiotoxicity, and carcinogenicity are the frequently modeled toxicity endpoints in predictive toxicology. It is known that datasets impact model performance. The quality of datasets used in the development of toxicity prediction models using machine learning and deep learning is vital to the performance of the developed models. The different toxicity assignments for the same chemicals among different datasets of the same type of toxicity have been observed, indicating benchmarking datasets is needed for developing reliable toxicity prediction models using machine learning and deep learning algorithms. This review provides insights into current machine learning models in predictive toxicology, which are expected to promote the development and application of toxicity prediction models in the future.

Research & Experimental Medicine↗

DOE OSTI2023

Large Ensemble Diagnostic Evaluation of Hydrologic Parameter Uncertainty in the Community Land Model Version 5 (CLM5)

Abstract Land surface models such as the Community Land Model version 5 (CLM5) seek to enhance understanding of terrestrial hydrology and aid in the evaluation of anthropogenic and climate change impacts. However, the effects of parametric uncertainty on CLM5 hydrologic predictions across regions, timescales, and flow regimes have yet to be explored in detail. The common use of the default hydrologic model parameters in CLM5 risks generating streamflow predictions that may lead to incorrect inferences for important dynamics and/or extremes. In this study, we benchmark CLM5 streamflow predictions relative to the commonly employed default hydrologic parameters for 464 headwater basins over the conterminous United States (CONUS). We evaluate baseline CLM5 default parameter performance relative to a large (1,307) Latin Hypercube Sampling‐based diagnostic comparison of streamflow prediction skill using over 20 error measures. We provide a global sensitivity analysis that clarifies the significant spatial variations in parametric controls for CLM5 streamflow predictions across regions, temporal scales, and error metrics of interest. The baseline CLM5 shows relatively moderate to poor streamflow prediction skill in several CONUS regions, especially the arid Southwest and Central U.S. Hydrologic parameter uncertainty strongly affects CLM5 streamflow predictions, but its impacts vary in complex ways across U.S. regions, timescales, and flow regimes. Overall, CLM5's surface runoff and soil water parameters have the largest effects on simulated high flows, while canopy water and evaporation parameters have the most significant effects on the water balance.

54 ENVIRONMENTAL SCIENCES↗

NASA NTRS1992

Continuation of tailored composite structures of ordered staple thermoplastic material

The search for the cost effective composite structure has motivated the investigation of several approaches to develop composite structure from innovative material forms. Among the promising approaches is the conversion of a planar sheet to components of complex curvature through sheet forming or stretch forming. In both cases, the potential for material stretch in the fiber direction appears to offer a clear advantage in formability over continuous fiber systems. A framework was established which allows the simulation of the anisotropic mechanisms of deformation of long discontinuous fiber laminates wherein the matrix phase is a viscous fluid. Predictions for the effective viscosities of a hyper-anisotropic medium consisting of collimated, discontinuous fibers suspended in viscous matrix were extended to capture the characteristics of typical polymers including non-Newtonian behavior and temperature dependence. In addition, the influence of fiber misorientation was also modeled by compliance averaging to determine ensemble properties for a given orientation distribution. A design tool is presented for predicting the effect of material heterogeneity on the performance of curved composite beams such as those used in aircraft fuselage structures. Material heterogeneity can be induced during manufacturing processes such as sheet forming and stretch forming of thermoplastic composites. This heterogeneity can be introduced in the form of fiber realignment and spreading during the manufacturing process causing radial and tangential gradients in material properties. Two analysis procedures are used to solve the beam problems. The first method uses separate two-dimensional elasticity solutions for the stresses in the flange and web sections of the beam. The separate solutions are coupled by requiring that forces and displacements match section boundaries. The second method uses an approximate Rayleigh-Ritz technique to find the solutions for more complex beams. Analyses are performed for curved beams of various cross-sections loaded in pure bending and with a uniform distributed load. Preliminary results show that the geometry of the beam dictates the effect of heterogeneity on performance. The role of heterogeneity is larger in beams with a small average radius-to-depth ration, R/t, where R is the average radius of the beam and t is the difference between the inside and outside radii. Results of the anlysis are in the form of stresses and displacements and are compared to both mechanics of materials and numerical solutions obtained using finite element analysis.

Santare, Michael H.↗

DOE OSTI2021

Uncertainty quantification for deep learning in particle accelerator applications

With the advent of increased computational resources and improved algorithms, machine learning-based models are being increasingly applied to complex problems in particle accelerators. However, such data-driven models may provide overly confident predictions with unknown errors and uncertainties. For reliable deployment of machine learning models in high-regret and safety-critical systems such as particle accelerators, estimates of prediction uncertainty are needed along with accurate point predictions. In this investigation, we evaluate Bayesian neural networks (BNN) as an approach that can provide accurate predictions along with reliably quantified uncertainties for particle accelerator problems, and compare their performance with bootstrapped ensembles of neural networks. We select three accelerator setups for this evaluation: a storage ring, a photoinjector, and a linac. The problems span different data volumes and dimensionalities (e.g., scalar predictions as well as image outputs). It is found that BNN provide accurate predictions of the mean along with reliable estimates of predictive uncertainty across the test cases. In this vein, BNN may offer an attractive alternative to deterministic deep learning tools to generate accurate predictions with quantified uncertainties in particle accelerator applications.

43 PARTICLE ACCELERATORS↗

DOE OSTI2026

Active operator learning with predictive uncertainty quantification for partial differential equations

With the increased prevalence of neural operators being used to provide rapid solutions to partial differential equations (PDEs), understanding the accuracy of model predictions and the associated error levels is necessary for deploying reliable surrogate models in scientific applications. Existing uncertainty quantification (UQ) frameworks employ ensembles or Bayesian methods, which can incur substantial computational costs during both training and inference. Here, we propose a lightweight predictive UQ method tailored for Deep operator networks (DeepONets) that also generalizes to other operator networks. Numerical experiments on linear and nonlinear PDEs demonstrate that the framework’s uncertainty estimates are unbiased and provide accurate out-of-distribution uncertainty predictions with a sufficiently large training dataset. Our framework provides fast inference and uncertainty estimates that can efficiently drive outer-loop analyses that would be prohibitively expensive with conventional solvers. We demonstrate how predictive uncertainties can be used in the context of Bayesian optimization and active learning problems to yield improvements in accuracy and data-efficiency for outer-loop optimization procedures. In the active learning setup, we extend the framework to Fourier Neural Operators (FNO) and describe a generalized method for other operator networks. To enable real-time deployment, we introduce an inference strategy based on precomputed trunk outputs and a sparse placement matrix, reducing evaluation time by more than a factor of five. Our method provides a practical route to uncertainty-aware operator learning in time-sensitive settings.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2021

Identifying precursors of daily to seasonal hydrological extremes over the USA using deep learning techniques and climate model ensembles

Focal Area(s): We focus on two areas of crosscutting interest for DOE: 1) predictability of extreme precipitation and drought in the USA and 2) the integration of climate models with new AI tools, such as convolutional neural networks (CNN) and methods to understand their output (e.g. layer-wise relevance propagation; LRP). This project fits into focus area 3 of this call for white paper using AI to gain insight from complex data, including explainable AI tools. Science Challenge: Predicting hydrological extremes is important due to their impacts on people, agriculture and infrastructure. This prediction is difficult due to the infrequent occurrence of extremes and their complexity. However, extreme events can be related to more predictable conditions in the ocean, such as El Nino, long-term soil moisture or large scale modes of climate variability, such as the North Atlantic Oscillation (NAO).

54 ENVIRONMENTAL SCIENCES↗