Search NASA⌕ Search

SEARCH · Search NASA

Results for “data driven model”

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 433 records · Page 24

Domain knowledge-informed, process-mapping AI graph for designing Fe-based alloys

<span style="font-family: Calibri, sans-serif; font-size: 12pt;">Continuous improvement in efficiency of a power plant relies on designing materials for use at increasingly higher temperature and/or pressure, for 100,000s hours of operation. Due to complexity, non-linearity and high-dimensionality of the problem, traditional Machine Learning (ML) approaches require unreasonably large datasets for the data-driven model development. Science-based material and process engineering complements hard data with, sometimes soft and intuitive, empirical domain knowledge. Artificial Intelligence (AI) was used in this study to incorporate such knowledge into computational graph architecture (process-mimicking artificial neuron design, causal layer and graph structures, ensemble modeling of latent states) and learning procedures (variable transformation, fuzzy physics pre-training and freezing of deep layers, virtual microstructure representation, and adversarial multi-objective optimization). The first alloys design pathways suggested by the AI tool (pyroMind) passed a preliminary engineering review on soundness and transparency.</span>

Romanov, Vyacheslav↗

Model-Data Fusion Approaches for Retrospective and Predictive Assessment of the Pan-Arctic Scale Permafrost Carbon Feedback to Global Climate

Policy decisions regarding climate change mitigation and adaptation rely on having reliable projections of future climate change. Coupled climate-carbon models predict that the northern high latitudes will serve as a substantial land carbon sink during the 21stcentury, but these models lack many of the key processes governing high-latitude ecosystem processes. Arctic tundra and boreal forests are unique from other biomes particularly because of the presence of snow, ice and frozen ground, and it is important to understand the mechanisms that drive interconnected processes in these landscapes. Increasing our confidence in climate projections requires an integrated approach where existing, on-going and planned observational and experimental studies of high-latitude ecosystem processes are organized and analyzed in a framework designed to target improvements in understanding key ecosystem-climate feedbacks. For this project we designed an approach to improve our scientific understanding of the permafrost carbon feedback through broad-scale assessment of the drivers and responses of the Arctic system carbon cycle using data-driven models of simple-to-intermediate complexity. The made a valuable contribution to the first of its kind, data-driven quantitative assessment of the impact of permafrost thaw on pan-Arctic scale carbon cycle feedbacks to the global climate system. The approach drew on a conceptual framework for integrating the key system components and their underlying data within a “book-keeping” framework for assessment of the key processes driving the permafrost carbon feedback. The results provided quantitative model-data benchmarking to constrain future projections of carbon cycle responses to a warming Arctic. The framework used environmental proxy data sets examining recent-era geomorphologic land-form transitions and vegetation dynamics, through remotely-sensed data, to guide the extrapolation of site-based data on the key system components to larger regions. These products contribute to the development of finer-scale, more complex representation of permafrost carbon processes in terrestrial biogeochemistry models, to operate within coupled Earth system modeling frameworks.

54 ENVIRONMENTAL SCIENCES↗

Search for R-parity-violating supersymmetry in a final state containing leptons and many jets with the ATLAS experiment using $\sqrt{s}=13\hbox{TeV}$ proton–proton collision data

A search for R-parity-violating supersymmetry in final states characterized by high jet multiplicity, at least one isolated light lepton and either zero or at least three b-tagged jets is presented. The search uses 139fb –1 of $\sqrt{s}=13\hbox{TeV}$ proton–proton collision data collected by the ATLAS experiment during Run 2 of the Large Hadron Collider. The results are interpreted in the context of R-parity-violating supersymmetry models that feature gluino production, top-squark production, or electroweakino production. The dominant sources of background are estimated using a data-driven model, based on observables at medium jet multiplicity, to predict the b-tagged jet multiplicity distribution at the higher jet multiplicities used in the search. Machine-learning techniques are used to reach sensitivity to electroweakino production, extending the data-driven background estimation to the shape of the machine-learning discriminant. No significant excess over the Standard Model expectation is observed and exclusion limits at the 95% confidence level are extracted, reaching as high as 2.4 TeV in gluino mass, 1.35 TeV in top-squark mass, and 320 (365) GeV in higgsino (wino) mass.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Machine Learning for a-posteriori model-observed data fusion to enhance predictive value of ESM output

The proposed research is consistent with focal areas 2 and 3: when successful, this effort will enhance the predictive value of Earth System Model (ESM) output by identifying and quantifying its systematic deviations from available observations and by correcting and recalibrating to existing data. The product is a hybrid process & data-driven modeling tool whose estimation can also bring insights through pattern discovery recognition on model deficiencies on the one hand and data gaps on the other.

54 ENVIRONMENTAL SCIENCES↗

Parametric and Nonparametric Models of U.S. Cost Overruns for Nuclear Power Plants

This study presents new data-driven models to estimate the effect of capacity on the percentage of cost overruns in the United States for nuclear power plant construction projects before and after the Three Mile Island accident. Parametric and nonparametric models have been developed that describe the significant shifts in nuclear energy costs during the dynamic environment. Employing a contemporary descriptive methodology and a quantitative analysis, we furnish a comprehensive overview of the alterations in cost overrun distribution and show the changes observed in other pivotal metrics alongside cost overruns. Our emphasis lies in documenting the fluctuations in cost overruns alongside nuclear reactor capacity levels and the increase of the overnight capital costs to build nuclear reactors. Our results show that increasing the size of nuclear reactors is not a factor statistically significant to decrease the percentage of cost overruns, and the probit model results provide evidence that an increase in size increases the probability of having cost overruns larger than 100% (double the estimated cost). We also compare our findings to two other regions: Asia and Europe.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Invertible Temper Modeling using Normalizing Flows and the Effects of Structure Preserving Loss

Advanced manufacturing research and development is typically small-scale, owing to costly experiments associated with these novel processes. Deep learning techniques could help accelerate this development cycle but frequently struggle in small-data regimes like the advanced manufacturing space. While prior work has applied deep learning to modeling visually plausible advanced manufacturing microstructures, little work has been done on data-driven modeling of how microstructures are affected by heat treatment, or assessing the degree to which synthetic microstructures are able to support existing workflows. We propose to address this gap by using invertible neural networks (normalizing flows) to model the effects of heat treatment, e.g., tempering. The model is developed using scanning electron microscope imagery from samples produced using shear-assisted processing and extrusion (ShAPE) manufacturing. This approach not only produces visually and topologically plausible samples, but also captures information related to a sample’s material properties or experimental process parameters. We also demonstrate that topological data analysis, used in prior work to characterize microstructures, can also be used to stabilize model training, preserve structure, and improve downstream results. We assess directions for future work and identify our approach as an important step towards end-to-end deep learning system for accelerating advanced manufacturing research and development.

Howland, Sylvia↗

Opportunities and data requirements for data-driven prognostics and health management in liquid hydrogen storage systems

During the past decade, Prognostics and Health Management (PHM) has become an important set of tools in various areas of industry and academic reliability engineering. PHM consists of a variety of mathematical and computational methods used to support data-driven decision-making to increase the safety, availability, and reliability of complex engineering systems. In particular, PHM can provide crucial insight into reliability and safety design improvements for developing technologies where historical performance and failure data are limited. This is the case of hydrogen fueling and storage technologies. This work presents a high-level approach for designing data-driven PHM applications for bulk liquid hydrogen (LH 2 ) storage systems for hydrogen fueling stations. This paper addresses core aspects of the design, development, and implementation of data-driven PHM applications that can improve the reliability assessment of hydrogen components. The analysis focuses on the relationship between data availability and diagnostic/prognostic capabilities; potential challenges; and integration schemes for current risk mitigation measures. We identify potential condition-monitoring data sources for key components in an LH 2 storage system, including storage tanks, piping, and pumps. Further, we determine that the short-term goals for the implementation of data-driven models in PHM frameworks in hydrogen systems should focus on developing adequate data collection and analysis strategies, as well as exploring the effect on reliability, safety, and regulations for hydrogen systems.

08 HYDROGEN↗

Deep Learning-Based Failure Prognostic Model for PV Inverter Using Field Measurements

Here, this study presents a novel approach for the precise monitoring and prognosis of photovoltaic (PV) inverter status, which is crucial for the proactive maintenance of PV systems. It addresses the gaps in traditional model-based methods, which tend to neglect the overall reliability of inverters, and the limitations of data-driven approaches that largely depend on simulated data. This research presents a robust solution applicable to real-world scenarios. The proposed data-driven model for PV inverter failure prognosis employs actual inverter measurements, integrating various operational and weather-related factors based on domain knowledge. This approach effectively represents inverter stressors and operational status. Utilizing an Enhanced Siamese Convolutional Neural Network (ESCNN), the model merges operational data with domain knowledge features, redefining the prognosis challenge as a classification task. Furthermore, the paper discusses an ESCNN-based real-time inverter failure monitoring method developed on the well-trained model. The proposed models are rigorously trained and tested with real inverter data and a novel filtering method is included to address accidental failures in practical scenarios. The results validate the model's efficacy, and the directions for future research are also outlined.

42 ENGINEERING↗

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↗

Data-Driven Wind Farm Optimization Incorporating Effects of Turbulence Intensity

We demonstrate assimilation of field data from a nacelle-mounted lidar and meteorological tower into a medium-fidelity Reynolds-Averaged Navier Stokes wind farm flow model to better predict the effects of atmospheric stability. Increased predictability under a variety of atmospheric conditions can lead to more effective control design and optimization of a wind farm. In particular, atmospheric stability affects wind turbine wake propagation and, therefore, aspects of wind farm control and performance, such as active wind farm control, layout optimization, and power output. Accurately modeling wakes in different stability conditions remains a persistent challenge. This paper presents an optimization framework that leverages high-fidelity field or simulation data to correct a lower-fidelity flow model. Optimal model corrections are found by solving a regularized, high-dimensional, gradient-based optimization problem using adjoint flowfield information. We validate the trained model against large eddy simulation results, and perform separate gradient-based layout optimizations of a simulated utility-scale wind farm to maximize power. Using the data-driven model corrections, we find that atmospheric stability significantly impacts layout optimization and power production: the optimal layout for stable conditions produced 9.1% more power than the optimal layout for unstable conditions, and the optimal layout for neutral conditions underperformed by 8.5% in unstable conditions.

17 WIND ENERGY↗

Low‐dimensional manifold learning for uncertainty quantification in complex multi‐scale stochastic systems

Broadly speaking, the goals of the project are to develop techniques to use manifold learning to develop reduced‐order and surrogate models for "hyper‐reduction" of very high‐dimensional complex multi‐scale systems. This is being achieved by employing a newly proposed form of manifold projection and learning that leverages recent advancements in computational geometry and data‐driven modeling. In particular, we are applying a manifold projection technique to project the solutions of very high‐dimensional systems onto the so‐called Grassmannmanifold, a Reimannian manifold comprised of orthonormal matrices. We then apply data‐driven machine learning techniques to classify the solutions on the manifold (e.g. clustering techniques) according to their proximity on the manifold and leverage a further nonlinear dimension reduction to organize the structured data on the manifold. Finally, we are developing novel techniques that enable us to directly interpolate the hyper‐reduced data such that we can predict the solution of the complex, high‐ dimensional system without need to call the full expensive computational model. Given their adherence to the underlying structure of the solution of the physical system, it is expected that these approximate solutions will be sufficiently constrained so as to (approximately) adhere to physical principles.

97 MATHEMATICS AND COMPUTING↗

Using Particle Image Velocimetry to Determine Turbulence Model Parameters

The primary parameter of a standard k-ϵ model, C μ , was calculated from stereoscopic particle image velocimetry (PIV) data for a supersonic jet exhausting into a transonic crossflow. This required the determination of turbulent kinetic energy, turbulent eddy viscosity, and turbulent energy dissipation rate. Image interrogation was optimized, with different procedures used for mean strain rates and Reynolds stresses, to produce useful turbulent eddy viscosity fields. The eddy viscosity was calculated by a least-squares fit to all components of the three-dimensional strain-rate tensor that were available from the PIV data. This eliminated artifacts and noise observed when using a single strain component. Local dissipation rates were determined via Kolmogorov’s similarity hypotheses and the second-order structure function. The eddy viscosity and dissipation rates were then combined to determine C μ . Considerable spatial variation was observed in C μ , with the highest values found in regions where turbulent kinetic energy was relatively low but where turbulent mixing was important, e.g., along the high-strain jet edges and in the wake region. This suggests that use of a constant C μ in modeling may lead to poor Reynolds stress predictions at mixing interfaces. A data-driven modeling approach that can predict this spatial variation of C μ based on known state variables may lead to improved simulation results without the need for calibration.

42 ENGINEERING↗

Perspectives on Machine Learning-assisted Plasma Medicine: Towards Automated Plasma Treatment

Cold atmospheric plasmas (CAPs) have shown great promise for medical applications through their synergistic chemical, electrical, and thermal effects, which can induce therapeutic outcomes. However, safe and reproducible plasma treatment of complex biological surfaces poses a major hurdle to the widespread adoption of CAPs for medical applications. Predictive modeling of the mutual interactions between the plasma and biological surfaces and, thus, systematic approaches to quantify and predict plasma treatment outcomes remain largely elusive due to the lack of mechanistic understanding of plasma-surface interactions that can span across vastly different length-and time-scales. In addition, real-time sensing capabilities in biomedical CAP devices are often limited, which can be detrimental to plasma treatment due to the intrinsic plasma and surface variability during the treatment, as well as sensitivity to external perturbations. All of these challenges can make reproducible and effective plasma treatment of biological surfaces difficult to realize, which is further compounded by errors due to human operation of hand-held CAP devices. Machine learning and data-driven approaches can be particularly useful in addressing these challenges in three major ways: (i) data-driven modeling of hard-to-model plasma-surface interactions and plasma treatment outcomes; (ii) learning data analytics for plasma and surface diagnostics in real-time; and (iii) developing predictive controllers that enable reliable and effective CAP treatments. Furthermore, this paper discusses the promise of machine learning to accelerate plasma medicine research in these areas, toward machine learning-assisted and automated CAP treatment of complex biological surfaces.

60 APPLIED LIFE SCIENCES↗

Data-driven global ocean modeling for seasonal to decadal prediction

Accurate modeling of ocean dynamics is crucial for enhancing our understanding of complex ocean circulation processes, predicting climate variability, and tackling challenges posed by climate change. Although great efforts have been made to improve traditional numerical models, predicting global ocean variability over multiyear scales remains challenging. Here, we propose ORCA-DL (Oceanic Reliable foreCAst via Deep Learning), a data-driven three-dimensional ocean model for seasonal to decadal prediction of global ocean dynamics. ORCA-DL accurately simulates the three-dimensional structure of global ocean dynamics with high physical consistency and outperforms state-of-the-art numerical models in capturing extreme events, including El Niño–Southern Oscillation and upper ocean heat waves. Moreover, ORCA-DL stably emulates ocean dynamics at decadal timescales, demonstrating its potential even for skillful decadal predictions and climate projections. Our results demonstrate the high potential of data-driven models for providing efficient and accurate global ocean modeling and prediction.

Science & Technology - Other Topics↗

Physics-constrained, low-dimensional models for magnetohydrodynamics: First-principles and data-driven approaches

Plasmas are highly nonlinear and multiscale, motivating a hierarchy of models to understand and describe their behavior. However, there is a scarcity of plasma models of lower fidelity than magnetohydrodynamics (MHD), although these reduced models hold promise for understanding key physical mechanisms, efficient computation, and real-time optimization and control. Galerkin models, obtained by projection of the MHD equations onto a truncated modal basis, and data-driven models, obtained by modern machine learning and system identification, can furnish this gap in the lower levels of the model hierarchy. This work develops a reduced-order modeling framework for compressible plasmas, leveraging decades of progress in projection-based and data-driven modeling of fluids. We begin by formalizing projection-based model reduction for nonlinear MHD systems. To avoid separate modal decompositions for the magnetic, velocity, and pressure fields, we introduce an energy inner product to synthesize all of the fields into a dimensionally consistent, reduced-order basis. Next, we obtain an analytic model by Galerkin projection of the Hall-MHD equations onto these modes. We illustrate how global conservation laws constrain the model parameters, revealing symmetries that can be enforced in data-driven models, directly connecting these models to the underlying physics. We demonstrate the effectiveness of this approach on data from high-fidelity numerical simulations of a three-dimensional spheromak experiment. Finally, this manuscript builds a bridge to the extensive Galerkin literature in fluid mechanics and facilitates future principled development of projection-based and data-driven models for plasmas.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

When not to use machine learning: A perspective on potential and limitations

Abstract The unparalleled success of artificial intelligence (AI) in the technology sector has catalyzed an enormous amount of research in the scientific community. It has proven to be a powerful tool, but as with any rapidly developing field, the deluge of information can be overwhelming, confusing, and sometimes misleading. This can make it easy to become lost in the same hype cycles that have historically ended in the periods of scarce funding and depleted expectations known as AI winters. Furthermore, although the importance of innovative, high-risk research cannot be overstated, it is also imperative to understand the fundamental limits of available techniques, especially in young fields where the rules appear to be constantly rewritten and as the likelihood of application to high-stakes scenarios increases. In this article, we highlight the guiding principles of data-driven modeling, how these principles imbue models with almost magical predictive power, and how they also impose limitations on the scope of problems they can address. Particularly, understanding when not to use data-driven techniques, such as machine learning, is not something commonly explored, but is just as important as knowing how to apply the techniques properly. We hope that the discussion to follow provides researchers throughout the sciences with a better understanding of when said techniques are appropriate, the pitfalls to watch for, and most importantly, the confidence to leverage the power they can provide. Graphical abstract

36 MATERIALS SCIENCE↗

Improving Deployment Availability of Energy Storage With Data-Driven AGC Signal Models

Energy Storage (ES) provides great flexibility and large benefits to power system operations and control. When providing ancillary services (e.g., regulation, reserve, etc.), the real-time (RT) deployment of ES is uncertain, and it is important to manage state of charge accordingly. Aiming to improve the ES performance for providing energy and regulation service in the electricity market, we propose two data-driven Automatic Generation Control (AGC) signal models. The first one is a historical-data-driven AGC signal model, which is based on the analysis of the historical AGC signals, and is designed for ES participation in the day-ahead (DA) market. The second one is a prediction-data-driven AGC signal model, which is based on the prediction of the AGC signals, and is designed for ES participation in the RT market. We also develop a deployment availability check model and solution algorithm. The proposed framework is applied to an ES bidding problem in the DA and RT markets. The results indicate that deployment availability and operational performance of the ES are improved with the proposed data-driven AGC models compared to traditional benchmarks.

AGC signal model↗

Toward an understanding of foreground emission in the BICEP2 region

BICEP2 has reported the detection of a degree-scale B-mode polarization pattern in the Cosmic Microwave Background (CMB) and has interpreted the measurement as evidence for primordial gravitational waves. Motivated by the profound importance of the discovery of gravitational waves from the early Universe, we examine to what extent a combination of Galactic foregrounds and lensed E-modes could be responsible for the signal. We reanalyze the BICEP2 results and show that the 100 ×150 GHz and 150 ×150 GHz data are consistent with a cosmology with r=0.2 and negligible foregrounds, but also with a cosmology with r=0 and a significant dust polarization signal. We give independent estimates of the dust polarization signal in the BICEP2 region using a number of different approaches: (1) data-driven models based on Planck 353 GHz intensity, polarization fractions inferred from the same Planck data used by the BICEP2 team but corrected for CMB and CIB contributions, and polarization angles from starlight polarization data or the Planck sky model; (2) the same set of pre-Planck models used by the BICEP2 team but taking into account the higher polarization fractions observed in the CMB- and CIB-corrected map; (3) a measurement of neutral hydrogen gas column density N{sub HI} in the BICEP2 region combined with an extrapolation of a relation between HI column density and dust polarization derived by Planck; and (4) a dust polarization map based on digitized Planck data, which we only use as a final cross-check. While these approaches are consistent with each other, the expected amplitude of the dust polarization power spectrum remains uncertain by about a factor of three. The lower end of the prediction leaves room for a primordial contribution, but at the higher end the dust in combination with the standard CMB lensing signal could account for the BICEP2 observations, without requiring the existence of primordial gravitational waves. By measuring the cross-correlations between the pre-Planck templates used in the BICEP2 analysis and between different versions of a data-based template, we emphasize that cross-correlations between models are very sensitive to noise in the polarization angles and that measured cross-correlations are likely underestimates of the contribution of foregrounds to the map. These results suggest that BICEP1 and BICEP2 data alone cannot distinguish between foregrounds and a primordial gravitational wave signal, and that future Keck Array observations at 100 GHz and Planck observations at higher frequencies will be crucial to determine whether the signal is of primordial origin.

79 ASTRONOMY AND ASTROPHYSICS↗