Search NASA⌕ Search

SEARCH · Search NASA

Results for “Multivariate outputs”

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 37 records · Page 2

Ecological Forecasting in Chesapeake Bay: Using a Mechanistic-Empirical Modelling Approach

The Chesapeake Bay Ecological Prediction System (CBEPS) automatically generates daily nowcasts and three-day forecasts of several environmental variables, such as sea-surface temperature and salinity, the concentrations of chlorophyll, nitrate, and dissolved oxygen, and the likelihood of encountering several noxious species, including harmful algal blooms and water-borne pathogens, for the purpose of monitoring the Bay's ecosystem. While the physical and biogeochemical variables are forecast mechanistically using the Regional Ocean Modeling System configured for the Chesapeake Bay, the species predictions are generated using a novel mechanistic empirical approach, whereby real-time output from the coupled physical biogeochemical model drives multivariate empirical habitat models of the target species. The predictions, in the form of digital images, are available via the World Wide Web to interested groups to guide recreational, management, and research activities. Though full validation of the integrated forecasts for all species is still a work in progress, we argue that the mechanistic–empirical approach can be used to generate a wide variety of short-term ecological forecasts, and that it can be applied in any marine system where sufficient data exist to develop empirical habitat models. This paper provides an overview of this system, its predictions, and the approach taken.

Brown, C. W.↗

Reduced order models for assessing CO 2 impacts in shallow unconfined aquifers

Risk assessment studies of potential CO 2 sequestration projects consider many factors, including the possibility of brine and/or CO 2 leakage from the storage reservoir. Detailed multiphase reactive transport simulations have been developed to predict the impact of such leaks on shallow groundwater quality; however, these simulations are computationally expensive and thus difficult to directly embed in a probabilistic risk assessment analysis. Here we present a process for developing computationally fast reduced-order models which emulate key features of the more detailed reactive transport simulations. A large ensemble of simulations that take into account uncertainty in aquifer characteristics and CO 2 /brine leakage scenarios were performed. Twelve simulation outputs of interest were used to develop response surfaces (RSs) using a MARS (multivariate adaptive regression splines) algorithm (Milborrow, 2015). A key part of this study is to compare different measures of ROM accuracy. We then show that for some computed outputs, MARS performs very well in matching the simulation data. The capability of the RS to predict simulation outputs for parameter combinations not used in RS development was tested using cross-validation. Again, for some outputs, these results were quite good. For other outputs, however, the method performs relatively poorly. Performance was best for predicting the volume of depressed-pH-plumes, and was relatively poor for predicting organic and trace metal plume volumes. We believe several factors, including the non-linearity of the problem, complexity of the geochemistry, and granularity in the simulation results, contribute to this varied performance. The reduced order models were developed principally to be used in probabilistic performance analysis where a large range of scenarios are considered and ensemble performance is calculated. We demonstrate that they effectively predict the ensemble behavior. But, the performance of the RSs is much less accurate when used to predict time-varying outputs from a single simulation. If an analysis requires only a small number of scenarios to be investigated, computationally expensive physics-based simulations would likely provide more reliable results. Finally, if the aggregate behavior of a large number of realizations is the focus, as will be the case in probabilistic quantitative risk assessment, the methodology presented here is relatively robust.

54 ENVIRONMENTAL SCIENCES↗

Search for associated production of z and Higgs bosons in proton-antiproton collisions at 1.96 TeV

We present a search for associated production of Z and Higgs bosons in 4.2 fb -1 of $\bar{p}$p collisions at √s = 1.96 TeV, produced in RunII of the Tevatron and recorded by the D0 detector. The search is performed in events containing at least two muons and at least two jets. The ZH signal is distinguished from the expected backgrounds by means of multivariate classifiers known as random forests. Binned random forest output distributions are used in comparing the data to background-only and signal+background hypotheses. No excess is observed in the data, so we set upper limits on ZH production with a 95% confidence level.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

TU-FG-201-05: Varian MPC as a Statistical Process Control Tool

Purpose: Quality assurance in radiotherapy requires the measurement of various machine parameters to ensure they remain within permitted values over time. In Truebeam release 2.0 the Machine Performance Check (MPC) was released allowing beam output and machine axis movements to be assessed in a single test. We aim to evaluate the Varian Machine Performance Check (MPC) as a tool for Statistical Process Control (SPC). Methods: Varian’s MPC tool was used on three Truebeam and one EDGE linac for a period of approximately one year. MPC was commissioned against independent systems. After this period the data were reviewed to determine whether or not the MPC was useful as a process control tool. Analyses on individual tests were analysed using Shewhart control plots, using Matlab for analysis. Principal component analysis was used to determine if a multivariate model was of any benefit in analysing the data. Results: Control charts were found to be useful to detect beam output changes, worn T-nuts and jaw calibration issues. Upper and lower control limits were defined at the 95% level. Multivariate SPC was performed using Principal Component Analysis. We found little evidence of clustering beyond that which might be naively expected such as beam uniformity and beam output. Whilst this makes multivariate analysis of little use it suggests that each test is giving independent information. Conclusion: The variety of independent parameters tested in MPC makes it a sensitive tool for routine machine QA. We have determined that using control charts in our QA programme would rapidly detect changes in machine performance. The use of control charts allows large quantities of tests to be performed on all linacs without visual inspection of all results. The use of control limits alerts users when data are inconsistent with previous measurements before they become out of specification. A. Carver has received a speaker’s honorarium from Varian.

07 ISOTOPE AND RADIATION SOURCES↗

Learning likelihood ratios with neural network classifiers

The likelihood ratio is a crucial quantity for statistical inference in science that enables hypothesis testing, construction of confidence intervals, reweighting of distributions, and more. Many modern scientific applications, however, make use of data- or simulation-driven models for which computing the likelihood ratio can be very difficult or even impossible. By applying the so-called “likelihood ratio trick,” approximations of the likelihood ratio may be computed using clever parametrizations of neural network-based classifiers. A number of different neural network setups can be defined to satisfy this procedure, each with varying performance in approximating the likelihood ratio when using finite training data. We present a series of empirical studies detailing the performance of several common loss functionals and parametrizations of the classifier output in approximating the likelihood ratio of two univariate and multivariate Gaussian distributions as well as simulated high-energy particle physics datasets.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Inferring Atmospheric Release Characteristics in a Large Computer Experiment using Bayesian Adaptive Splines

An atmospheric release of hazardous material, whether accidental or intentional, can be catastrophic for those in the path of the plume. Predicting the path of a plume based on characteristics of the release (location, amount and duration) and meteorological conditions is an active research area highly relevant for emergency and long-term response to these releases. As a result, researchers have developed particle dispersion simulators to provide plume path predictions that incorporate release characteristics and meteorological conditions. However, since release characteristics and meteorological conditions are often unknown, the inverse problem is of great interest, that is, based on all the observations of the plume so far, what can be inferred about the release characteristics? This is the question we seek to answer using plume observations from a controlled release at the Diablo Canyon Nuclear Power Plant in Central California. With access to a large number of evaluations of a computationally expensive particle dispersion simulator that includes continuous and categorical inputs and spatio-temporal output, building a fast statistical surrogate model (or emulator) presents many statistical challenges, but is an essential tool for inverse modeling and sensitivity analysis. We achieve accurate emulation using Bayesian adaptive splines to model weights on empirical orthogonal functions. Here, we use this emulator as well as appropriately identifiable simulator discrepancy and observational error models to calibrate the simulator, thus finding a posterior distribution of characteristics of the release. Since the release was controlled, these characteristics are known, making it possible to compare our findings to the truth.

54 ENVIRONMENTAL SCIENCES↗

Sensitivity Analysis and Emulation for Functional Data using Bayesian Adaptive Splines

When a computer code is used to simulate a complex system, one of the fundamental tasks is to assess the sensitivity of the simulator to the different input parameters. In the case of computationally expensive simulators, this is often accomplished via a surrogate statistical model, a statistical output emulator. An effective emulator is one that provides good approximations to the computer code output for wide ranges of input values. In addition, an emulator should be able to handle large dimensional simulation output for a relevant number of inputs; it should flexibly capture heterogeneities in the variability of the response surface; it should be fast to evaluate for arbitrary combinations of input parameters, and it should provide an accurate quantification of the emulation uncertainty. In this paper we discuss the Bayesian approach to multivariate adaptive regression splines (BMARS) as an emulator for a computer model that outputs curves. We introduce modifications to traditional BMARS approaches that allow for fitting large amounts of data and allow for more efficient MCMC sampling. We emphasize the ease with which sensitivity analysis can be performed in this situation. We present a sensitivity analysis of a computer model of the deformation of a protective plate used in pressure-driven experiments. Our example serves as an illustration of the ability of BMARS emulators to fulfill all the necessities of computability, flexibility and reliable calculation on relevant measures of sensitivity.

97 MATHEMATICS AND COMPUTING↗

STSR-INR: Spatiotemporal super-resolution for multivariate time-varying volumetric data via implicit neural representation

Implicit neural representation (INR) has surfaced as a promising direction for solving different scientific visualization tasks due to its continuous representation and flexible input and output settings. We present STSR-INR, an INR solution for generating simultaneous spatiotemporal super-resolution for multivariate time-varying volumetric data. Inheriting the benefits of the INR-based approach, STSR-INR supports unsupervised learning and permits data upscaling with arbitrary spatial and temporal scale factors. Unlike existing GAN- or INR-based super-resolution methods, STSR-INR focuses on tackling variables or ensembles and enabling joint training across datasets of various spatiotemporal resolutions. Here we achieve this capability via a variable embedding scheme that learns latent vectors for different variables. In conjunction with a modulated structure in the network design, we employ a variational auto-decoder to optimize the learnable latent vectors to enable latent-space interpolation. To combat the slow training of INR, we leverage a multi-head strategy to improve training and inference speed with significant speedup. We demonstrate the effectiveness of STSR-INR with multiple scalar field datasets and compare it with conventional tricubic+linear interpolation and state-of-the-art deep-learning-based solutions (STNet and CoordNet).

97 MATHEMATICS AND COMPUTING↗

Organic Aerosol Component (OACOMP) Value-Added Product (Report)

Significantly improved returns in their aerosol chemistry data can be achieved via the development of a value-added product (VAP) of deriving OA components, called Organic Aerosol Components (OACOMP). OACOMP is primarily based on multivariate analysis of the measured organic mass spectral matrix. The key outputs of OACOMP are the concentration time series and the mass spectra of OA factors that are associated with distinct sources, formation and evolution processes, and physicochemical properties.

54 ENVIRONMENTAL SCIENCES↗

Stacked Low-Inertia Converter or Solid-State Transformer: Modeling and Model Predictive Priority-Shifting Control for Voltage Balance

This paper presents control challenges of stacked low-inertia converter (SLIC) or cascaded reduced dc-link solid-state transformer (SST) and proposes a novel model predictive priority-shifting (MPPS) control with implicit modulator and a discrete-time large-signal model for voltage balancing and dc-link regulation. Low-inertia converters, featuring small electrolytic capacitor-less dc links, dramatically reduce cost, size, and weight compared to conventional solutions. However, without a large dc-link buffer, the input and output are tightly coupled, leading to significant control challenges. The control becomes even more challenging with these converters stacked input-series output-parallel (ISOP) for medium-voltage (MV) grid, which causes coupling between the modules besides the coupling within each module. This paper analyzes the multi-objective, multi-degree of freedom control problem, using the modular soft-switching solid-state transformer (M-S4T) as an example of the SLIC. First, distribution of control efforts under controller saturation is critical because multiple control objectives can be conflicting, especially when the module voltages are unbalanced and are being restored. The MPPS can shift the priorities to address this issue. Second, due to the low inertia and high dc-link ripple, classic space vector pulse-width modulation (SVPWM), average model with small-ripple assumption, and control design based on small-signal model cannot accurately modulate, model, and control the nonlinear reduced dc link. Therefore, a discrete-time large-signal model of the M-S4T is established to derive the predictive control in the MPPS. The MPPS and the PI control are compared in MV simulations to show the issue of applying the PI to the SLIC and the effectiveness of the MPPS for voltage balancing and dc-link regulation in a deadbeat manner. Finally, the proposed control is tested on a 5 kV ISOP SiC SST prototype to verify priority shifting to address controller saturation issue and fast and robust voltage balancing.

14 SOLAR ENERGY↗

End-to-End Models for Effects of System Noise on LIMS Analysis of Igneous Rocks

The ChemCam instrument on the Mars Science Laboratory will be the first extraterrestial deployment of laser-induced breakdown spectroscopy (UBS) for remote geochemical analysis. LIBS instruments are also being proposed for future NASA missions. In quantitative LIBS applications using multivariate analysis techniques, it is essential to understand the effects of key instrument parameters and their variability on the elemental predictions. Baseline experiments were run on a laboratory instrument in conditions reproducing ChemCam performance on Mars. These experiments employed Nd:YAG laser producing 17 mJ/pulse on target and an with a 200 {micro}m FWHM spot size on the surface of a sample. The emission is collected by a telescope, imaged on a fiber optic and then interfaced to a demultiplexer capable of >40% transmission into each spectrometer. We report here on an integrated end-to-end system performance model that simulates the effects of output signal degradation that might result from the input signal chain and the impact on multivariate model predictions. There are two approaches to modifying signal to noise (SNR): degrade the signal and/or increase the noise. Ishibashi used a much smaller data set to show that the addition of noise had significant impact while degradation of spectral resolution had much less impact on accuracy and precision. Here, we specifically focus on aspects of remote LIBS instrument performance as they relate to various types of signal degradation. To assess the sensitivity of LIBS analysis to signal-to-noise ratio (SNR) and spectral resolution, the signal in each spectrum from a suite of 50 laboratory spectra of igneous rocks was variably degraded by increasing the peak widths (simulating misalignment) and decreasing the spectral amplitude (simulating decreases in SNR).

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Power Prediction of Airborne Wind Energy Systems Using Multivariate Machine Learning

Kites can be used to harvest wind energy at higher altitudes while using only a fraction of the material required for conventional wind turbines. In this work, we present the kite system of Kyushu University and demonstrate how experimental data can be used to train machine learning regression models. The system is designed for 7 kW traction power and comprises an inflatable wing with suspended kite control unit that is either tethered to a fixed ground anchor or to a towing vehicle to produce a controlled relative flow environment. A measurement unit was attached to the kite for data acquisition. To predict the generated tether force, we collected input–output samples from a set of well-designed experimental runs to act as our labeled training data in a supervised machine learning setting. We then identified a set of key input parameters which were found to be consistent with our sensitivity analysis using Pearson input–output correlation metrics. Finally, we designed and tested the accuracy of a neural network, among other multivariate regression models. The quality metrics of our models show great promise in accurately predicting the tether force for new input/feature combinations and potentially guide new designs for optimal power generation.

17 WIND ENERGY↗

Downscaled CMIP5 projections of physical fire risk understate historical trends

Reliable projections of wildfire risk are important for multi-sector impacts analysis. Statistically downscaled and bias-corrected Earth system model ensemble products are routinely used to analyze regional physical wildfire risk, but evaluations of historical observed trends and variability are lacking. Here, we evaluate physical fire risk over the western United States using the Canadian Forest Fire Weather Index (FWI) by comparing model outputs from the Coupled Model Intercomparison Project Phase 5 (CMIP5), statistically downscaled via the Multivariate Adaptive Constructed Analogs (MACA) approach, against the observational target dataset gridMET, a gridded high-resolution surface meteorological product. We analyze multidecadal trends and interannual variability in seasonal average FWI for the historical period and future projections under two emissions scenarios, and we compare MACA-CMIP5 ensemble results with a simple time series model that generates historical and future projections of seasonal FWI based on bootstrapping observed historical trends and variability. Our findings indicate that MACA-CMIP5 accurately captures the magnitude and spatial patterns of seasonally averaged FWI but tends to underestimate historical decadal trends. We show that future increases in fire risk may be underestimated relative to the simple time series model that projects historical variability into the future. We also highlight that model biases in relative humidity contribute significantly to model-data differences. Our results underscore the importance of historical hindcasting exercises for informing broader multi-sector applications.

FWI↗

Efficient Surrogate Model Development: Impact of Sample Size and Underlying Model Dimensions

This study compares eight surrogate-model construction approaches using computational experiments. Here, the construction approaches considered include: Artificial Neural Networks (ANNs), Automated Learning of Algebraic Models using Optimization (ALAMO), Radial Basis Networks (RBNs), Extreme Learning Machines (ELMs), Gaussian Progress Regression (GPR), Random Forests (RFs), Support Vector Regression (SVR), and Multivariate Adaptive Regression Splines. Each approach is used to construct surrogate models for predicting the outputs of thirty-four test functions. The data used to construct surrogate models were generated using Latin Hypercube Sampling (LHS), Halton and Sobol sampling methods. We studied the impact of sampling method, sample size, and challenge function characteristics on the accuracy of surrogate model predictions. Our results revealed that for smaller sample sizes, on average, surrogate models trained using the data points generated by Sobol sequences provided the best estimation for all surrogate model construction approaches. As the sample size increased, the impact of sampling methods diminished. The comparison of surrogate-modeling construction approaches at large sample sizes revealed that surrogate models trained using ANN, ALAMO, and ELM yielded smaller root mean squared error and higher adjusted R-squared values than the models trained using the rest of the approaches.

42 ENGINEERING↗

Search for the Higgs Boson Produced in Association with a Vector Boson Using Like-Sign Dilepton Events in 1.96-TeV Proton-Antiproton Collisions

This thesis presents a search for the standard model and fermiophobic Higgs bosons produced in association with a vector boson using like-sign dilepton events collected by Collider Detector at Fermilab (CDF II). This analysis uses the data corresponding to an integrated luminosity of 9.4 fb -1 . The search is characterized by minimally biased cuts which maximize the acceptance of the Higgs bosons. The numbers of expected signal events passing requirements for like-sign dilepton are 5.6 for the fermiophobic Higgs on the mass of 110 GeV/c 2 assuming the standard model production cross section, and 1.7 for the standard model on the mass of 160 GeV/c 2 . The expectation of background events are 696.1, while the number of observed events are 624. To discriminate signal events from backgrounds, the search employs a multivariate analysis based on boosted decision trees. There are no significant disagreements in the BDT outputs, then, upper limits at a 95% confidence level are set on σ(p$\bar{p}$→V h) × B(h→WW ) for both Higgs bosons scenarios. The expected and observed limits are 5.9 and 9.2, respectively, relative to the standard model expectation for the standard model Higgs boson mass of 160 GeV/c 2 . For the fermiophobic Higgs boson corresponding to the mass of 110 GeV/c 2 , the expected and observed limits are 2.6 and 4.4, respectively.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Bayesian Treed Multivariate Gaussian Process with Adaptive Design: Application to a Carbon Capture Unit

Computer experiments (numerical simulations) are widely used in scientific research to study and predict the behavior of complex systems, which usually have responses consisting of a set of distinct outputs. The computational cost of the simulations at high resolution are often expensive and become impractical for parametric studies at different input values. To overcome these difficulties we develop a Bayesian treed multivariate Gaussian process (BTMGP) as an extension of the Bayesian treed Gaussian process (BTGP) in order to model and evaluate a multivariate process. A suitable choice of covariance function and the prior distributions facilitates the different Markov chain Monte Carlo (MCMC) movements. We utilize this model to sequentially sample the input space for the most informative values, taking into account model uncertainty and expertise gained. A simulation study demonstrates the use of the proposed method and compares it with alternative approaches. We apply the sequential sampling technique and BTMGP to model the multiphase flow in a full scale regenerator of a carbon capture unit. The application presented in this paper is an important tool for research into carbon dioxide emissions from thermal power plants.

Konomi, Bledar A.↗

Explainable AI for Multivariate Time Series Pattern Exploration: Latent Space Visual Analytics With Temporal Fusion Transformer and Variational Autoencoders in Power Grid Event Diagnosis

Detecting and analyzing complex patterns in multivariate time-series data is crucial for decision-making in urban and environmental system operations. However, challenges arise from the high dimensionality, intricate complexity, and interconnected nature of complex patterns, which hinder the understanding of their underlying physical processes. Existing AI methods often face limitations in interpretability, computational efficiency, and scalability, reducing their applicability in real-world scenarios. This paper proposes a novel visual analytics framework that integrates two generative AI models, Temporal Fusion Transformer (TFT) and Variational Autoencoders (VAEs), to reduce complex patterns into lower-dimensional latent spaces and visualize them in 2D using dimensionality reduction techniques such as PCA, t-SNE, and UMAP with DBSCAN. These visualizations, presented through coordinated and interactive views and tailored glyphs, enable intuitive exploration of complex multivariate temporal patterns, identifying patterns’ similarities and uncover their potential correlations for a better interpretability of the AI outputs. The framework is demonstrated through a case study on power grid signal data, where it identifies multi-label grid event signatures, including faults and anomalies with diverse root causes. Additionally, novel metrics and visualizations are introduced to validate the models and assess the performance, efficiency, and consistency of latent maps generated by VAE, which have been utilized in prior studies for latent space cartography and used as a benchmark in this study, and the emerging TFT architecture under various configurations. These analyses provide actionable insights for model parameter tuning and reliability improvements. Comparative results highlight that TFT achieves shorter run times and superior scalability to diverse time-series data shapes compared to VAE. This work advances fault diagnosis in multivariate time series, fostering explainable AI to support critical system operations.

Explainable AI↗

Simulation of the output power of copper bromide lasers by the MARS method

The dependence of the output power of CuBr lasers (operating at wavelengths of 510.6 and 578.2 nm) on ten input physical parameters has been statistically analysed based on a large amount of experimental data accumulated for these lasers. Regression models have been built using the flexible nonparametric method of multivariate adaptive regression splines (MARS) to describe both linear and nonlinear local dependences. These models cover more than 97% initial data with an error comparable with the experimental error; they are applied to estimate and predict the output powers of both existing and future lasers. The advantage of the models constructed for estimating laser parameters over the standard parametric methods of multivariate factor and regression analysis is demonstrated.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗