Search NASA⌕ Search

SEARCH · Search NASA

Results for “hypothesis tests”

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 73 records · Page 4

Feature Selection and Inferential Procedures for Video Data [Slides]

With the rise of electronic and high-dimensional data, new and innovative feature detection and statistical methods are required to perform accurate and meaningful statistical analysis of these datasets that provide unique statistical challenges. In the area of feature detection, much of the recent feature detection research in the computer vision community has focused on deep learning methods, which require large amounts of labeled training data. However, in many application areas, training data is very limited and often difficult to obtain. We develop methods for fast, unsupervised, precise feature detection for video data based on optical flows, edge detection, and clustering methods. We also use pretrained neural networks and interpretable linear models to extract features using very limited training data. In the area of statistics, while high-dimensional data analysis has been a main focus of recent statistical methodological research, much focus has been on populations of high-dimensional vectors, rather than populations of high-dimensional tensors, which are three-dimensional arrays that can be used to model dependent images, such as images taken of the same person or ripped video frames. Our feature detection method is a non-model-based method that fusses information from dense optical flow, raw image pixels, and frame differences to generate detections. Our hypothesis testing methods are based on the assumption that dependent images are concatenated into a tensor that follows a tensor normal distribution, and from this assumption, we derive likelihood-ratio, score, and regression-based tests for one- and multiple-sample testing problems. Our methods will be illustrated on simulated and real datasets. We conclude this report with comments on the relationship between feature detection and hypothesis testing methods.

97 MATHEMATICS AND COMPUTING↗

Effects of Phasor Measurement Uncertainty on Power Line Outage Detection

Phasor measurement unit (PMU) technology provides an effective tool to enhance the wide-area monitoring systems (WAMSs) in power grids. Although extensive studies have been conducted to develop several PMU applications in power systems (e.g., state estimation, oscillation detection and control, voltage stability analysis, and line outage detection), the uncertainty aspects of PMUs have not been adequately investigated. This paper focuses on quantifying the impact of PMU uncertainty on power line outage detection and identification, in which a limited number of PMUs installed at a subset of buses are utilized to detect and identify the line outage events. Specifically, the line outage detection problem is formulated as a multi-hypothesis test, and a general Bayesian criterion is used for the detection procedure, in which the PMU uncertainty is analytically characterized. We further apply the minimum detection error criterion for the multi-hypothesis test and derive the expected detection error probability in terms of PMU uncertainty. The framework proposed provides fundamental guidance for quantifying the effects of PMU uncertainty on power line outage detection. Case studies are provided to validate our analysis and show how PMU uncertainty influences power line outage detection.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Annual Scientific Focus Area Report for the Energy Exascale Earth System Model Project: Phase II

The project's long-term goal is to assert and maintain an international scientific leadership position in the development of Earth system models that address the grand challenge of actionable predictions of Earth system variability and change, with an emphasis on the most critical scientific questions facing the nation and DOE. The grand challenge will be addressed by overcoming three major limitations: Pushing the high-resolution frontier of Earth system modeling through a combination of strategic model development and computational advances targeting exascale computers and beyond. Bridging the gap in scales and processes by developing a new class of models that couple the natural, managed, and man-made systems across scales to address energy sector vulnerability to variability and change. The project will achieve its goal through four intersecting project elements: A series of prediction and simulation experiments addressing scientific questions and mission needs; A well-documented and tested, continuously advancing, evolving, and improving system of model codes that comprise the E3SM Earth system model; The ability to use effectively leading (and “bleeding”) edge computational facilities soon after their deployment at DOE national laboratories; and An infrastructure to support code development, hypothesis testing, simulation execution, and analysis of results. To achieve the Grand Challenge of actionable Earth system predictions to address the most critical scientific questions facing our nation and DOE, the E3SM project defined its strategy in a fifteen-year roadmap with four intersecting project elements: A series of prediction and simulation experiments addressing scientific questions and mission needs; A well-documented and tested, continuously advancing, evolving, and improving system of model codes that comprise the E3SM Earth system model ; The ability to use effectively leading (and “bleeding”) edge computational facilities soon after their deployment at DOE national laboratories; and An infrastructure to support code development, hypothesis testing, simulation execution, and analysis of results. Figure 1 depicts the E3SM Project Roadmap, showing the relationships among the first three major project elements: the simulations, the modeling system to perform those simulations, and the machines on which they will be executed. Unlike the other three elements that have distinct but overlapping phases, the fourth element, the infrastructure, will evolve continuously based on the requirements imposed by project needs.

54 ENVIRONMENTAL SCIENCES↗

ASSESSMENT OF NORMALITY FOR CRITICALITY SAFETY BIAS AND BIAS UNCERTAINTY CALCULATION

The single-sided lower tolerance factors frequently used for non-trending assessment of the validation bias and bias uncertainty are sensitive to departures from normality. When used properly, the tolerance limits ensure that an appropriate fraction of the true population of applicable critical experiments lies above the calculated lower tolerance limit with the required statistical confidence level. One condition necessary to ensure that the appropriate proportion of the true population of keff values in the validation suite lies above the lower tolerance limit is that the assumption that the normality of the underlying population of critical experiments is valid or conservative.This paper discusses various methods used to assess whether the assumption that the validation suite may be treated as a random sample drawn from a normal distribution is acceptable. Techniques for assessing the validity of this underlying assumption include common omnibus hypothesis tests for normality, assessment of sample skewness and kurtosis of the validation suite, and graphical techniques. These techniques are used to assess the nature and potential conservatism/nonconservatism imparted by various departures from normality where possible. A review of hypothesis testing is also presented to frame the discussion of omnibus normality tests. Additionally, two cases are analysed with these techniques to provide an example of how they should be implemented.

Clarity, Justin↗

How to COAAD Images. II. A Coaddition Image that is Optimal for Any Purpose in the Background-dominated Noise Limit

Image coaddition is one of the most basic operations that astronomers perform. In Paper I, we presented the optimal ways to coadd images in order to detect faint sources and to perform flux measurements under the assumption that the noise is approximately Gaussian. Here, we build on these results and derive from first principles a coaddition technique that is optimal for any hypothesis testing and measurement (e.g., source detection, flux or shape measurements, and star/galaxy separation), in the background-noise-dominated case. This method has several important properties. The pixels of the resulting coadded image are uncorrelated. This image preserves all the information (from the original individual images) on all spatial frequencies. Any hypothesis testing or measurement that can be done on all the individual images simultaneously, can be done on the coadded image without any loss of information. The PSF of this image is typically as narrow, or narrower than the PSF of the best image in the ensemble. Moreover, this image is practically indistinguishable from a regular single image, meaning that any code that measures any property on a regular astronomical image can be applied to it unchanged. In particular, the optimal source detection statistic derived in Paper I is reproduced by matched filtering this image with its own PSF. This coaddition process, which we call proper coaddition, can be understood as the maximum signal-to-noise ratio measurement of the Fourier transform of the image, weighted in such a way that the noise in the entire Fourier domain is of equal variance. This method has important implications for multi-epoch seeing-limited deep surveys, weak lensing galaxy shape measurements, and diffraction-limited imaging via speckle observations. The last topic will be covered in depth in future papers. We provide an implementation of this algorithm in MATLAB.

79 ASTRONOMY AND ASTROPHYSICS↗

OTD Observations of Continental US Ground Flashes Detected by NLDN

Lightning optical flash parameters (e.g., radiance, area, duration, number of optical groups, and number of optical events) derived from almost 5 yrs of Optical Transient Detector (OTD) data are compared with peak current and multiplicity observations derived from the US National Lightning Detection Networkm (NLDN). Despite the relatively low lightning geolocation accuracy afforded by OTD, a total of 48,870 NLDN cloud-to-ground (CG) flashes were correlated with OTD flashes, or about 10,000 CGs per year. The median values of the above OTD flash parameters for the 48,870 CGs were, respectively: 0.137 J/square meters/sr/micrometers, 313.7 square kilometers, 0.189 s, 4 optical groups per CG, and 8 optical events per CG. Invoking the multiplicity data, the median number of optical groups per stroke was 2.5, and the median number of optical events per stroke was 5.0. Median values of peak current for negative and positive CGs were -21.6 kA and 17.8 kA, respectively, and as expected, the negative CGs had a larger average multiplicity than the positive CGs. A statistical summary is provided for all CGs, for positive and negative CGs, and for CGs from different seasons. Standard two-distribution hypothesis tests were perfonned to intercompare the population means of the various lightning parameters. In particular, and to greater than the 99% confidence level, it was found that positive CGs are on average more radiant, of greater areal extent, and are longer lasting than negative CGs. Rankings from a complete set of hypothesis tests between CGs of different polarities and from different seasons are also provided. Most notably, wintertime positive CGs tend to be more radiant, of greater areal extent, and longer lasting than any other group of CGs (i.e., negative springtime CGs, positive summertime CGs, etc.).

Koshak, William J.↗

Azimuthal Sampling of Tectonic Release from Rayleigh Waves, Triggered by Shallow Explosions from the Source Physics Experiment: Theory

Underground explosions often produce Rayleigh waves with non-circular radiation patterns. The presence of significant anisotropic radiation patterns from isotropic sources is often attributed to the coincident release of tectonic strain across faults that are spatially near the explosion working point. However, poor instrument deployment geometries can lead to misinterpretations of these fields as isotropic when the are non-circular. In this work, we construct a physical representation for the first two moments of the Rayleigh wave radiation pattern triggered by a shallow, underground explosion. We then construct a hypothesis test for discriminating between the isotropic and anisotropic portion of the radiated Rayleigh wavefield, using these moments. We find that the screening power is completely quantified by a single parameter λ that depends on azimuthal deployment geometry as well as wavefield sampling. To improve radiation pattern discrimination, we present an algorithm for iteratively including receivers to sample the radiation pattern that increases the performance (power) of the hypothesis test at each iteration. Last, we discuss future applications of our method to data recorded from the Source Physics Experiment taking place at the Nevada National Security Site (NNSS).

42 ENGINEERING↗

Hierarchical top-down bottom-up calibration with consideration for uncertainty and inter-scale discrepancy of Peierls stress of bcc Fe

A hierarchical multiscale model of plasticity in single crystal bcc Fe, framed at the mesoscale pertaining to a statistically representative set of dislocations for each slip system, relies on specification of a set of parameters which are informed using both bottom-up (BU) (atomistic) and top-down (TD) (experimental) information. We term this specification process a 'connection' between BU and TD pathways to inform the mesoscale model. The connection is considered in the presence of error, uncertainty, and discrepancy between the models. Furthermore, we expand upon a previously developed reconciled TD and BU calibration method to account for anticipated discrepancy between information from different length scales. The results of a previously formulated likelihood-based connection test of the multiscale model suggest a 'missing link' may be responsible for part of the inter-scale discrepancy. In this case, this link is assumed to be a relation between the unit-process (single dislocation line) of coordinated kink-pair nucleation on a screw dislocation segment and the many-body dislocation process which manifests in the onset of experimentally observed plastic deformation in a single crystal. A physics-informed discrepancy layer is formulated to improve the connection in the presence of additional persistent uncertainties. This physics-informed hypothesis testing is demonstrated as an alternative to fully data-driven search methods, particularly applicable to datasets of limited size that are often encountered in such multiscale material modeling applications. Discrete dislocation dynamics simulations are used to investigate the effect of the Peierls stress on straight screw dislocations and in the operation of Frank–Read sources. Discussion concerns the critical role of uncertainty quantification in providing a basis for this form of incremental hypothesis testing.

42 ENGINEERING↗

Automated Identification of Characteristic Droplet Size Distributions in Stratocumulus Clouds Utilizing a Data Clustering Algorithm

Abstract Droplet-level interactions in clouds are often parameterized by a modified gamma fitted to a “global” droplet size distribution. Do “local” droplet size distributions of relevance to microphysical processes look like these average distributions? This paper describes an algorithm to search and classify characteristic size distributions within a cloud. The approach combines hypothesis testing, specifically, the Kolmogorov–Smirnov (KS) test, and a widely used class of machine learning algorithms for identifying clusters of samples with similar properties: density-based spatial clustering of applications with noise (DBSCAN) is used as the specific example for illustration. The two-sample KS test does not presume any specific distribution, is parameter free, and avoids biases from binning. Importantly, the number of clusters is not an input parameter of the DBSCAN-type algorithms but is independently determined in an unsupervised fashion. As implemented, it works on an abstract space from the KS test results, and hence spatial correlation is not required for a cluster. The method is explored using data obtained from the Holographic Detector for Clouds (HOLODEC) deployed during the Aerosol and Cloud Experiments in the Eastern North Atlantic (ACE-ENA) field campaign. The algorithm identifies evidence of the existence of clusters of nearly identical local size distributions. It is found that cloud segments have as few as one and as many as seven characteristic size distributions. To validate the algorithm’s robustness, it is tested on a synthetic dataset and successfully identifies the predefined distributions at plausible noise levels. The algorithm is general and is expected to be useful in other applications, such as remote sensing of cloud and rain properties. Significance Statement A typical cloud can have billions of drops spread over tens or hundreds of kilometers in space. Keeping track of the sizes, positions, and interactions of all of these droplets is impractical, and, as such, information about the relative abundance of large and small drops is typically quantified with a “size distribution.” Droplets in a cloud interact locally, however, so this work is motivated by the question of whether the cloud droplet size distribution is different in different parts of a cloud. A new method, based on hypothesis testing and machine learning, determines how many different size distributions are contained in a given cloud. This is important because the size distribution describes processes such as cloud droplet growth and light transmission through clouds.

54 ENVIRONMENTAL SCIENCES↗

Hybrid approach for estimating dynamic states of synchronous generators

Accurate and robust estimation of dynamic states is essential for monitoring and controlling a power grid. To achieve higher accuracy and robustness, this paper proposes a hybrid estimation approach to estimate the dynamic states of synchronous generators using multiple models, classified into a set of local models and a wide-area model. The proposed estimation approach, first, estimates dynamic states using the extended Kalman filter (EKF) based on a local model and a wide-area model. Then, at each time step, probability indexes, which quantify the likelihood of both local and wide-area models, are determined using hypothesis testing based on measurements innovations. Finally, using the weighted fix approach, the states estimated from the local and wide-area models are combined based on their probability indexes. Simulation studies using the two-area four-machine system shows that the proposed approach can improve estimation accuracy and increase robustness against model errors.

Akhlaghi, Shahrokh↗

Nuclear Corrections in Neutrino-Nucleus Deep Inelastic Scattering and their Compatibility with Global Nuclear Parton-Distribution-Function Analyses

We perform a global {chi}{sup 2} analysis of nuclear parton distribution functions using data from charged current neutrino-nucleus ({nu}A) deep-inelastic scattering (DIS), charged-lepton-nucleus (l{sup {+-}A}) DIS, and the Drell-Yan (DY) process. We show that the nuclear corrections in {nu}A DIS are not compatible with the predictions derived from l{sup {+-}A} DIS and DY data. We quantify this result using a hypothesis-testing criterion based on the {chi}{sup 2} distribution which we apply to the total {chi}{sup 2} as well as to the {chi}{sup 2} of the individual data sets. We find that it is not possible to accommodate the data from {nu}A and l{sup {+-}A} DIS by an acceptable combined fit. Our result has strong implications for the extraction of both nuclear and proton parton distribution functions using combined neutrino and charged-lepton data sets.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Spectral analysis for the M AJORNA D EMONSTRATOR experiment

The M AJORNA D EMONSTRATOR is an experiment constructed to search for neutrinoless double-beta decays in germanium-76 and to demonstrate the feasibility to deploy a ton-scale experiment in a phased and modular fashion. It consists of two modular arrays of natural and 76 Ge-enriched germanium detectors totaling 44.1kg (29.7kg enriched detectors), located at the 4850' level of the Sanford Underground Research Facility in Lead, South Dakota, USA. Data taken with this setup since summer 2015 at different construction stages of the experiment show a clear reduction of the observed background index around the ROI for 0$νββ$- decay search due to improvements in shielding. We discuss the statistical approaches to search for a 0$νββ$-signal and derive the physics sensitivity for an expected exposure of 10kg y from enriched detectors using a profile likelihood based hypothesis test in combination with toy Monte Carlo data.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Overall Performance Losses and Activated Mechanisms in Double Glass and Glass-backsheet Photovoltaic Modules with Monofacial and Bifacial PERC Cells, under Accelerated Exposures

Commercial PV modules have various packaging choices nowadays, which influence their long-term reliability. This study compared the degradation behaviors of sixteen module variants from two brands with varying encapsulant materials (EVA or POE), encapsulant types, module architectures (GB or DG), and cell types (monofacial or bifacial) using null hypothesis testing to determine statistical significant findings. The modules were exposed for 2,520 hours under two accelerated exposures: modified damp heat (mDH) and modified damp heat with full-spectrum light (mDH+FSL). For both brands, two DG module variants with UV-Cutoff rear encapsulant are found to have significantly lower average power loss than the module variants of EVA+GB with opaque rear encapsulant after each accelerated exposure. Metallization interconnect corrosion is identified as the primary degradation mechanism. Furthermore, unsupervised hierarchical clustering finds that the degradation behaviors of modules from one brand with a more strict manufacturing quality control depends on module architectures only.

14 SOLAR ENERGY↗

Particle Filter Based Inference Testing

The primary intent of PAR-FIT (Particle Filter based Inference Testing) is to provide hard inductive evidence that a machine learning model is capable and proven for an individual test input. By examining training data used to form the underlying model functional correlation, an estimate of the reliability that a model will make the correct prediction can be made. The Sequential Probability Ratio Test is used to derive a qualitative evaluation for reliability based on hypothesis testing. The PAR-FIT framework achieves this by implementing a particle filter and the sequential probability ratio test algorithms on the machine learning model training data to determine relevancy of new individual test samples to the training dataset. The kernel function evaluates the local proximity and density of training data used to derive a prediction outcome. Particles are used to probabilistically determine which training data to evaluate for proximity. For test samples that are within a close proximity to and surrounded by multiple training data points, the evaluated reliability of the prediction is high. For test samples that are anomalies not represented by the training dataset, in low density data clusters, or are far from existing data points, the evaluated reliability is low as insufficient training evidence exists to suggest the model is capable of making the correct prediction. Sequential Probability Ratio Test is further used to determine when a hypothesis on whether a signal can be rejected or accepted for use. The ratio test collects sequence information from the particle filter to test whether the signal is anomalous or normal via hypothesis testing of the underlying distributions.

Chen, Edward [Idaho National Laboratory (INL), Ida↗

Characteristics of human-climate feedbacks differ at different radiative forcing levels

The human and Earth systems are intricately linked: climate influences agricultural production, renewable energy potential, and water availability, for example, while anthropogenic emissions from industry and land use change alter temperature and precipitation. Such feedbacks have the potential to significantly alter future climate change. These feedbacks may also exert significant changes on 21st-century energy, agriculture, land use and carbon cycle projections, but little is known about their possible magnitudes, or about regional and sector dynamics under different forcing scenarios. Here we use an integrated Earth System Model (ESM) featuring bidirectional information exchange between an economically-oriented integrated assessment model and the ESM to examine how human-natural feedbacks operate under high and medium radiative forcing (RF) scenarios. Specifically, we examine the effect of changing land productivity on human systems, and the effect of changing land use/land cover and CO2 emissions on the Earth system. We find that the effect of coupling differs across radiative forcing levels and across regions, due to differences in the climate signal, human responses to those signals, and regional characteristics. In particular, we find reductions in cropland area due to feedbacks in both the medium and high RF scenarios. In the medium RF scenario, these reductions result in increased area for bioenergy and afforestation and reduced energy system CO2 emissions, as the carbon price in this scenario incentivizes low carbon energy sources and terrestrial carbon storage; these incentives are absent in the high RF scenario. These differences are key to understanding the possible future evolution pathways of the integrated Earth system in response to 21st century climate change. Additional models and hypothesis testing are needed to determine exactly when and how bidirectional feedbacks between human and Earth systems should be considered in future assessments.

Calvin, Katherine V.↗

Kernelized Generalized Likelihood Ratio Test for Spectrum Sensing in Cognitive Radio

Spectrum sensing in next-generation wireless radio networks is considered a key technology to overcome the problem of spectrum scarcity. Unfortunately, many approaches to spectrum sensing do not work well in low signal-to-noise ratio (SNR) environments. This paper proposes and analyzes a new algorithm named kernelized generalized likelihood ratio test (KGLRT) for spectrum sensing in cognitive radio systems to overcome this problem. Effectively, KGLRT uses a nonlinear kernel to map input data onto a high-dimensional feature space; then, the widely accepted (linear) generalized likelihood ratio test is used for hypothesis testing. This new algorithm gives a gain of 4 dB in SNR over its linear counterpart. A theoretical analysis for this algorithm is given for the first time and is shown analogous to algorithms used in image signal processing. The detection metrics are found to be concentrated random variables; furthermore, the probability distributions of the detection metrics are proved to follow the F -distributions, which agree with the results obtained using the concentration inequality. The analytical thresholds are derived for target false-alarm probabilities. Here, the thresholds are independent of noise power; thus, the proposed algorithm can overcome noise uncertainty issues at very low SNR levels. Simulations validate the theoretical results.

47 OTHER INSTRUMENTATION↗