Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “probability and statistical methods”

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 235 records · Page 13
All sourcesNASA NTRSDOE OSTI
NASA NTRS1984

A study of environmental characterization of conventional and advanced aluminum alloys for selection and design. Phase 2: The breaking load test method

A technique is demonstrated for accelerated stress corrosion testing of high strength aluminum alloys. The method offers better precision and shorter exposure times than traditional pass fail procedures. The approach uses data from tension tests performed on replicate groups of smooth specimens after various lengths of exposure to static stress. The breaking strength measures degradation in the test specimen load carrying ability due to the environmental attack. Analysis of breaking load data by extreme value statistics enables the calculation of survival probabilities and a statistically defined threshold stress applicable to the specific test conditions. A fracture mechanics model is given which quantifies depth of attack in the stress corroded specimen by an effective flaw size calculated from the breaking stress and the material strength and fracture toughness properties. Comparisons are made with experimental results from three tempers of 7075 alloy plate tested by the breaking load method and by traditional tests of statistically loaded smooth tension bars and conventional precracked specimens.

Sprowls, D. O.↗
NASA NTRS1991

The spectral relationships between NEA and the meteorites: An overview using principal components analysis

One of the primary reservoirs for meteorites is probably the planet-crossing Aten, Apollo, and Amor asteroids. Comparing the spectral characteristics of these two populations with each other and with the spectra of the main belt asteroids would provide insight into the dynamical processes that deliver meteorites to Earth. One method for obtaining an overview of general relationships in a large spectral data set is the statistical technique of principal components analysis. This technique quantifies general spectral similarities and reprojects the data in a plot of major axes of variation where distance is a measure of relative similarity. A major caveat should be kept in mind, however, spectra are sensitive to particle size and viewing geometry effects, and near Earth asteroids (NEA's) are probably significantly different from main belt asteroids in both these factors. The analysis was restricted to the spectral range of ECAS filters and included 116 meteorite spectra from the Gaffey (1976) survey and 417 asteroids from the Zellner et. al. (1985) survey of which 13 are planet-crossers. Although thirteen asteroids are not much of a sample on which to base conclusions, a few inferences can be drawn from this exercise. First, the NEA spectral characteristics are, on average, more consistent with the spectra of meteorites than are the main belt asteroids. Second, the S-type NEA's tend to be spectrally more similar to the ordinary chondrite meteorites than the main belt S-types. This suggests that the planet-crossing S-types do not represent the spectral range of the main belt S-type population and that the planet-crossing S-types are on average more like the ordinary chondrites than the main belt S-types. Third, the only direct asteroidal ordinary chondrite analog, the Q-type asteroid, 1862 Apollo, plots well within the field of the ordinary chondrite meteorites and represents the most common meteorite fall type. Finally, it is interesting that the planet-crossing asteroids occupy similar PCA space with the general trend of the meteorites while most of the main belt objects are offset from the meteorites in statistical space.

Britt, D. T.↗
NASA NTRS2000

A simulator for evaluating methods for the detection of lesion-deficit associations

Although much has been learned about the functional organization of the human brain through lesion-deficit analysis, the variety of statistical and image-processing methods developed for this purpose precludes a closed-form analysis of the statistical power of these systems. Therefore, we developed a lesion-deficit simulator (LDS), which generates artificial subjects, each of which consists of a set of functional deficits, and a brain image with lesions; the deficits and lesions conform to predefined distributions. We used probability distributions to model the number, sizes, and spatial distribution of lesions, to model the structure-function associations, and to model registration error. We used the LDS to evaluate, as examples, the effects of the complexities and strengths of lesion-deficit associations, and of registration error, on the power of lesion-deficit analysis. We measured the numbers of recovered associations from these simulated data, as a function of the number of subjects analyzed, the strengths and number of associations in the statistical model, the number of structures associated with a particular function, and the prior probabilities of structures being abnormal. The number of subjects required to recover the simulated lesion-deficit associations was found to have an inverse relationship to the strength of associations, and to the smallest probability in the structure-function model. The number of structures associated with a particular function (i.e., the complexity of associations) had a much greater effect on the performance of the analysis method than did the total number of associations. We also found that registration error of 5 mm or less reduces the number of associations discovered by approximately 13% compared to perfect registration. The LDS provides a flexible framework for evaluating many aspects of lesion-deficit analysis.

NASA Program Biomedical Research and Countermeasur↗
DOE OSTI2022

Rapid attribution analysis of the extraordinary heat wave on the Pacific coast of the US and Canada in June 2021

Towards the end of June 2021, temperature records were broken by several degrees Celsius in several cities in the Pacific Northwest areas of the US and Canada, leading to spikes in sudden deaths and sharp increases in emergency calls and hospital visits for heat-related illnesses. Here we present a multi-model, multi-method attribution analysis to investigate the extent to which human-induced climate change has influenced the probability and intensity of extreme heat waves in this region. Based on observations, modelling and a classical statistical approach, the occurrence of a heat wave defined as the maximum daily temperature (TXx) observed in the area 45–52 °N, 119–123 °W, was found to be virtually impossible without human-caused climate change. The observed temperatures were so extreme that they lay far outside the range of historical temperature observations. This makes it hard to state with confidence how rare the event was. Using a statistical analysis that assumes that the heat wave is part of the same distribution as previous heat waves in this region led to a first-order estimation of the event frequency of the order of once in 1000 years under current climate conditions. Using this assumption and combining the results from the analysis of climate models and weather observations, we found that such a heat wave event would be at least 150 times less common without human-induced climate change. Also, this heat wave was about 2 °C hotter than a 1-in-1000-year heat wave would have been in 1850–1900, when global mean temperatures were 1.2 °C cooler than today. Looking into the future, in a world with 2 °C of global warming (0.8 °C warmer than today), a 1000-year event would be another degree hotter. Our results provide a strong warning: our rapidly warming climate is bringing us into uncharted territory with significant consequences for health, well-being and livelihoods. Adaptation and mitigation are urgently needed to prepare societies for a very different future.

54 ENVIRONMENTAL SCIENCES↗
NASA NTRS1993

Certain and possible rules for decision making using rough set theory extended to fuzzy sets

Uncertainty may be caused by the ambiguity in the terms used to describe a specific situation. It may also be caused by skepticism of rules used to describe a course of action or by missing and/or erroneous data. To deal with uncertainty, techniques other than classical logic need to be developed. Although, statistics may be the best tool available for handling likelihood, it is not always adequate for dealing with knowledge acquisition under uncertainty. Inadequacies caused by estimating probabilities in statistical processes can be alleviated through use of the Dempster-Shafer theory of evidence. Fuzzy set theory is another tool used to deal with uncertainty where ambiguous terms are present. Other methods include rough sets, the theory of endorsements and nonmonotonic logic. J. Grzymala-Busse has defined the concept of lower and upper approximation of a (crisp) set and has used that concept to extract rules from a set of examples. We will define the fuzzy analogs of lower and upper approximations and use these to obtain certain and possible rules from a set of examples where the data is fuzzy. Central to these concepts will be the idea of the degree to which a fuzzy set A is contained in another fuzzy set B, and the degree of intersection of a set A with set B. These concepts will also give meaning to the statement; A implies B. The two meanings will be: (1) if x is certainly in A then it is certainly in B, and (2) if x is possibly in A then it is possibly in B. Next, classification will be looked at and it will be shown that if a classification will be looked at and it will be shown that if a classification is well externally definable then it is well internally definable, and if it is poorly externally definable then it is poorly internally definable, thus generalizing a result of Grzymala-Busse. Finally, some ideas of how to define consensus and group options to form clusters of rules will be given.

Dekorvin, Andre↗
NASA NTRS1997

Modeling the effects of exercise during 100% oxygen prebreathe on the risk of hypobaric decompression sickness

BACKGROUND: Several previous studies indicated that exercise during prebreathe with 100% O2 decreased the incidence of hypobaric decompression sickness (DCS). We report a meta-analysis of these investigations combined with a new study in our laboratory to develop a statistical model as a predictive tool for DCS. HYPOTHESIS: Exercise during prebreathe increases N2 elimination in a theoretical 360-min half-time compartment decreasing the incidence of DCS. METHODS: A dose-response probability tissue ratio (TR) model with 95% confidence limits was created for two groups, prebreathe with exercise (n = 113) and resting prebreathe (n = 113), using nonlinear regression analysis with maximum likelihood optimization. RESULTS: The model predicted that prebreathe exercise would reduce the residual N2 in a 360-min half-time compartment to a level analogous to that in a 180-min compartment. This finding supported the hypothesis. The incidence of DCS for the exercise prebreathe group was significantly decreased (Chi-Square = 17.1, p < 0.0001) from the resting prebreathe group. CONCLUSIONS: The results suggested that exercise during prebreathe increases tissue perfusion and N2 elimination approximately 2-fold and markedly lowers the risk of DCS. Based on the model, the prebreathe duration may be reduced from 240 min to a predicted 91 min for the protocol in our study, but this remains to be verified. The model provides a useful planning tool to develop and test appropriate prebreathe exercise protocols and to predict DCS risks for astronauts.

NASA Program Environmental Health↗
NASA NTRS2012

Blocking Losses With a Photon Counter

It was not known how to assess accurately losses in a communications link due to photodetector blocking, a phenomenon wherein a detector is rendered inactive for a short time after the detection of a photon. When used to detect a communications signal, blocking leads to losses relative to an ideal detector, which may be measured as a reduction in the communications rate for a given received signal power, or an increase in the signal power required to support the same communications rate. This work involved characterizing blocking losses for single detectors and arrays of detectors. Blocking may be mitigated by spreading the signal intensity over an array of detectors, reducing the count rate on any one detector. A simple approximation was made to the blocking loss as a function of the probability that a detector is unblocked at a given time, essentially treating the blocking probability as a scaling of the detection efficiency. An exact statistical characterization was derived for a single detector, and an approximation for multiple detectors. This allowed derivation of several accurate approximations to the loss. Methods were also derived to account for a rise time in recovery, and non-uniform illumination due to diffraction and atmospheric distortion of the phase front. It was assumed that the communications signal is intensity modulated and received by an array of photon-counting photodetectors. For the purpose of this analysis, it was assumed that the detectors are ideal, in that they produce a signal that allows one to reproduce the arrival times of electrons, produced either as photoelectrons or from dark noise, exactly. For single detectors, the performance of the maximum-likelihood (ML) receiver in blocking is illustrated, as well as a maximum-count (MC) receiver, that, when receiving a pulse-position-modulated (PPM) signal, selects the symbol corresponding to the slot with the largest electron count. Whereas the MC receiver saturates at high count rates, the ML receiver may not. The loss in capacity, symbol-error-rate (SER), and count-rate were numerically computed. It was shown that the capacity and symbol-error-rate losses track, whereas the count-rate loss does not generally reflect the SER or capacity loss, as the slot-statistics at the detector output are no longer Poisson. It is also shown that the MC receiver loss may be accurately predicted for dead times on the order of a slot.

Moision, Burce E.↗
DOE OSTI2022

High bias machine learning for antineutrino-based safeguards for small reactors

The statistical methods used for antineutrino detection will need to be improved to effectively monitor the inventory of next-generation nuclear reactors. In this sensitivity study, we evaluate machine learning models compared to previously used statistical approaches to identify diversion scenarios in a simulated Advanced Fast Reactor (AFR)-100. A chi-square goodness-of-fit technique, which individually compares the simulated antineutrino yields to the expected antineutrino yield, resulted in precise but low diversion detection probability. Various support vector machine (SVM) models were applied with diverse training datasets to evaluate the robustness of the method towards unexpected or “unseen” diversion scenarios. Furthermore, our results indicate that while the SVM models significantly improved the detection probability of near-field antineutrino-based safeguards, up to a probability of ~0.04, for the simulated small reactor, the detection system still needs improvements to reach the 0.2 detection limit established by the International Atomic Energy Agency.

98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL P↗
DOE OSTI2026

Unsupervised anomaly detection in MeV ultrafast electron diffraction

MeV ultrafast electron diffraction (MUED) is a pump-probe technique used to study the dynamic structural evolution of materials. An ultrashort laser pulse triggers structural changes, which are then probed by an ultrashort relativistic electron beam. To overcome low signal-to-noise ratios, diffraction patterns are averaged over thousands of shots. However, shot-to-shot instabilities in the electron beam can distort individual patterns, introducing uncertainty. Improving MUED accuracy requires detecting and removing these anomalous patterns from large datasets. In this work, we developed a fully unsupervised methodology for the detection of anomalous diffraction patterns. Using a convolutional autoencoder, we calculate the reconstruction mean squared error of the diffraction patterns. Based on the statistical analysis of this error, we provide the user an estimation of the probability that the pattern is normal, which also allows a posterior visual inspection of the images that are difficult to classify. This method has been trained with only 100 diffraction patterns and tested on 1521 patterns, resulting in a false positive rate between 0.2% and 0.4%, with a training time of 10 s per image and a test time of about 1 s per image. Here, the proposed methodology can also be applied to other diffraction techniques in which large datasets are collected that include faulty images due to instrumental instabilities.

43 PARTICLE ACCELERATORS↗
NASA NTRS1983

Statistical mechanics of neocortical interactions - Dynamics of synaptic modification

A recent study has demonstrated that several scales of neocortical interactions can be consistently analyzed with the use of methods of modern nonlinear nonequilibrium statistical mechanics. The formation, stability, and interaction of spatial-temporal patterns of columnar firings are explicitly calculated, to test hypothesized mechanisms relating to information processing. In this context, most probable patterns of columnar firings are associated with chemical and electrical synaptic modifications. It is stressed that synaptic modifications and shifts in most-probable firing patterns are highly nonlinear and interactive sets of phenomena. A detailed scenario of information processing is calculated of columnar coding of external stimuli, short-term storage via hysteresis, and long-term storage via synaptic modification.

Ingber, L.↗
NASA NTRS2012

Refinement of a Method for Identifying Probable Archaeological Sites from Remotely Sensed Data

To facilitate locating archaeological sites before they are compromised or destroyed, we are developing approaches for generating maps of probable archaeological sites, through detecting subtle anomalies in vegetative cover, soil chemistry, and soil moisture by analyzing remotely sensed data from multiple sources. We previously reported some success in this effort with a statistical analysis of slope, radar, and Ikonos data (including tasseled cap and NDVI transforms) with Student's t-test. We report here on new developments in our work, performing an analysis of 8-band multispectral Worldview-2 data. The Worldview-2 analysis begins by computing medians and median absolute deviations for the pixels in various annuli around each site of interest on the 28 band difference ratios. We then use principle components analysis followed by linear discriminant analysis to train a classifier which assigns a posterior probability that a location is an archaeological site. We tested the procedure using leave-one-out cross validation with a second leave-one-out step to choose parameters on a 9,859x23,000 subset of the WorldView-2 data over the western portion of Ft. Irwin, CA, USA. We used 100 known non-sites and trained one classifier for lithic sites (n=33) and one classifier for habitation sites (n=16). We then analyzed convex combinations of scores from the Archaeological Predictive Model (APM) and our scores. We found that that the combined scores had a higher area under the ROC curve than either individual method, indicating that including WorldView-2 data in analysis improved the predictive power of the provided APM.

Tilton, James C.↗
DOE OSTI2024

Topological Interpretability for Deep Learning

With the growing adoption of AI-based systems across everyday life, the need to understand their decision-making mechanisms is correspondingly increasing. The level at which we can trust the statistical inferences made from AI-based decision systems is an increasing concern, especially in high-risk systems such as criminal justice or medical diagnosis, where incorrect inferences may have tragic consequences. Despite their successes in providing solutions to problems involving real-world data, deep learning (DL) models cannot quantify the certainty of their predictions. These models are frequently quite confident, even when their solutions are incorrect. This work presents a method to infer prominent features in two DL classification models trained on clinical and non-clinical text by employing techniques from topological and geometric data analysis. We create a graph of a model's feature space and cluster the inputs into the graph's vertices by the similarity of features and prediction statistics. We then extract subgraphs demonstrating high-predictive accuracy for a given label. These subgraphs contain a wealth of information about features that the DL model has recognized as relevant to its decisions. We infer these features for a given label using a distance metric between probability measures, and demonstrate the stability of our method compared to the LIME and SHAP interpretability methods. This work establishes that we may gain insights into the decision mechanism of a DL model. This method allows us to ascertain if the model is making its decisions based on information germane to the problem or identifies extraneous patterns within the data.

Spannaus, Adam↗
NASA NTRS1989

Review of probabilistic analysis of dynamic response of systems with random parameters

The various methods that have been studied in the past to allow probabilistic analysis of dynamic response for systems with random parameters are reviewed. Dynamic response may have been obtained deterministically if the variations about the nominal values were small; however, for space structures which require precise pointing, the variations about the nominal values of the structural details and of the environmental conditions are too large to be considered as negligible. These uncertainties are accounted for in terms of probability distributions about their nominal values. The quantities of concern for describing the response of the structure includes displacements, velocities, and the distributions of natural frequencies. The exact statistical characterization of the response would yield joint probability distributions for the response variables. Since the random quantities will appear as coefficients, determining the exact distributions will be difficult at best. Thus, certain approximations will have to be made. A number of techniques that are available are discussed, even in the nonlinear case. The methods that are described were: (1) Liouville's equation; (2) perturbation methods; (3) mean square approximate systems; and (4) nonlinear systems with approximation by linear systems.

Kozin, F.↗
NASA NTRS1983

Transition probability of the Si III 189.2-nm intersystem line

Measurement of the lifetime of the metastable 3s3p(3)P(0)1 level of Si(2+) (Si III), which decays by photon emission at 189.2 nm to the 3s2(1)S0 state, is reported. The data were taken from spontaneous emission from metastable Si III stored in an RF ion trap. The Si III ions were produced through electron bombardment of SiH4 and SiF4 at pressures of 1/100,000,000-1/10,000,000 Torr. A photomultiplier was employed to count the photon emissions from the transitions. A total of 11 decay curves were generated for analysis, with Poisson statistics used to set the uncertainties at within 8 pct. Significant systematic effects were controlled, and the lifetime was found to be within 3.6 microsec of 59.9 microsec. The method used is concluded valid for determining the lifetimes of metastable levels of low-Z ions with low charge, and thereby the transition probabilities.

Kwong, H. S.↗
NASA NTRS1989

Probabilistic structural analysis methods for improving Space Shuttle engine reliability

Probabilistic structural analysis methods are particularly useful in the design and analysis of critical structural components and systems that operate in very severe and uncertain environments. These methods have recently found application in space propulsion systems to improve the structural reliability of Space Shuttle Main Engine (SSME) components. A computer program, NESSUS, based on a deterministic finite-element program and a method of probabilistic analysis (fast probability integration) provides probabilistic structural analysis for selected SSME components. While computationally efficient, it considers both correlated and nonnormal random variables as well as an implicit functional relationship between independent and dependent variables. The program is used to determine the response of a nickel-based superalloy SSME turbopump blade. Results include blade tip displacement statistics due to the variability in blade thickness, modulus of elasticity, Poisson's ratio or density. Modulus of elasticity significantly contributed to blade tip variability while Poisson's ratio did not. Thus, a rational method for choosing parameters to be modeled as random is provided.

Boyce, L.↗
DOE OSTI2024

Optimal Zeno Dragging for Quantum Control: A Shortcut to Zeno with Action-Based Scheduling Optimization

The quantum Zeno effect asserts that quantum measurements inhibit simultaneous unitary dynamics when the “collapse” events are sufficiently strong and frequent. This applies in the limit of strong continuous measurement or dissipation. It is possible to implement a dissipative control that is known as “Zeno dragging” by dynamically varying the monitored observable, and hence also the eigenstates, which are attractors under Zeno dynamics. This is similar to adiabatic processes, in that the Zeno-dragging fidelity is highest when the rate of eigenstate change is slow compared to the measurement rate. We demonstrate here two theoretical methods for using such dynamics to achieve control of quantum systems. The first, which we shall refer to as “shortcut to Zeno,” is analogous to the shortcuts to adiabaticity (counterdiabatic driving) that are frequently used to accelerate unitary adiabatic evolution. In the second approach, we apply the Chantasri-Dressel-Jordan stochastic action [PRA 88, 042110 (2013)], and demonstrate that the extremal-probability readout paths derived from this are well suited to setting up a Pontryagin-style optimization of the Zeno-dragging schedule. A fundamental contribution of the latter approach is to show that an action suitable for measurement-driven control optimization can be derived quite generally from statistical arguments. Implementing these methods on the Zeno dragging of a qubit, we find that both approaches yield the same solution, namely, that the optimal control is a unitary that matches the motion of the Zeno-monitored eigenstate. We then show that such a solution can be more robust than a unitary-only operation and we comment on solvable generalizations of our qubit example embedded in larger systems. These methods open up new pathways toward systematically developing dynamic control of Zeno subspaces to realize dissipatively stabilized quantum operations. Published by the American Physical Society 2024

Physics↗