SEARCH · Engineering Papers

Results for “variance component analysis”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

SourceResearch connectionFrom yearThrough year

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

At least 73 records · Page 4

All sources NASA NTRS DOE OSTI

DOE OSTI2018

Identifying structural changes with unsupervised machine learning methods

In this work, unsupervised machine learning methods are used to identify structural changes using the melting point transition in classical molecular dynamics simulations as an example application of the approach. Dimensionality reduction and clustering methods are applied to instantaneous radial distributions of atomic configurations from classical molecular dynamics simulations of metallic systems over a large temperature range. Principal component analysis is used to dramatically reduce the dimensionality of the feature space across the samples using an orthogonal linear transformation that preserves the statistical variance of the data under the condition that the new feature space is linearly independent. From there, k-means clustering is used to partition the samples into solid and liquid phases through a criterion motivated by the geometry of the reduced feature space of the samples, allowing for an estimation of the melting point transition. This pattern criterion is conceptually similar to how humans interpret the data but with far greater throughput, as the shapes of the radial distributions are different for each phase and easily distinguishable by humans. The transition temperature estimates derived from this machine learning approach produce comparable results to other methods on similarly small system sizes. These results show that machine learning approaches can be applied to structural changes in physical systems.

42 ENGINEERING↗

DOE OSTI2015

A variance-based decomposition and global sensitivity index method for uncertainty quantification: Application to retrieved ice cloud properties

This study develops a novel uncertainty quantification (UQ) method for cloud microphysical property retrievals using variance-based decomposition and global sensitivity index. In this UQ framework, empirical orthogonal function (EOF) analysis is applied to the U.S. Department of Energy Atmospheric Radiation Measurement (ARM) ground-based observations, which are the inputs for the cloud retrieval studied here. The principal components (PCs) in the EOF expansion are parameterized as random input variables, and hence, the input dimension is greatly reduced (up to a factor of 50), allowing large ensemble of random samplings. The EOF expansion improves the accuracy of the uncertainty estimation by taking into account the cross correlations in the input data profiles. Here, this method enables a probabilistic representation of a retrieval process by adding normally distributed perturbations into PCs of sample means of input data profiles within a time window. Therefore, it effectively facilitates objective validation of climate models against cloud retrievals under a probabilistic framework for rigorous statistical inferences. Moreover, the variance-based global sensitivity index analysis, part of this method, attributes the output uncertainties to each individual source, thus providing directions for improving retrieval algorithms and observation strategies. For demonstration, we apply this method to quantify the uncertainties of the ARM program's baseline cloud retrieval algorithm for an ice cloud case observed at the Southern Great Plains site on 9 March 2000.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2018

An adaptive least-squares global sensitivity method and application to a plasma-coupled combustion prediction with parametric correlation

We introduce an efficient non-intrusive surrogate-based methodology for global sensitivity analysis and uncertainty quantification. Modified covariance-based sensitivity indices (mCov-SI) are defined for outputs that reflect correlated effects. The overall approach is applied to simulations of a complex plasma-coupled combustion system with disparate uncertain parameters in sub-models for chemical kinetics and a laser-induced breakdown ignition seed. The surrogate is based on an Analysis of Variance (ANOVA) expansion, such as widely used in statistics, with orthogonal polynomials representing the ANOVA subspaces and a polynomial dimensional decomposition (PDD) representing its multi-dimensional components. The coefficients of the PDD expansion are obtained using a least-squares regression, which both avoids the direct computation of high-dimensional integrals and affords an attractive flexibility in choosing sampling points. This facilitates importance sampling using a Bayesian calibrated posterior distribution, which is fast and thus particularly advantageous in common practical cases, such as our large-scale demonstration, for which the asymptotic convergence properties of polynomial expansions cannot be realized due to computation expense. Effort, instead, is focused on efficient finite-resolution sampling. Standard covariance-based sensitivity indices (Cov-SI) are employed to account for correlation of the uncertain parameters. Magnitude of Cov-SI is unfortunately unbounded, which can produce extremely large indices that limit their utility. Alternatively, mCov-SI are then proposed in order to bound this magnitude $\in$ [0, 1]. The polynomial expansion is coupled with an adaptive ANOVA strategy to provide an accurate surrogate as the union of several low-dimensional spaces, avoiding the typical computational cost of a high-dimensional expansion. It is also adaptively simplified according to the relative contribution of the different polynomials to the total variance. Finally, the approach is demonstrated for a laser-induced turbulent combustion simulation model, which includes parameters with correlated effects.

42 ENGINEERING↗

DOE OSTI2019

Experiments and computations of microfluidic liquid–liquid flow patterns

We study two-phase liquid–liquid flow patterns in a 500 μm capillary microchannel for four biphasic systems: ethyl acetate/water, 2-pentanol/water, methyl isobutyl ketone/water, and heptane/water. Flow visualization experiments using laser induced fluorescence (LIF) reveal a total of 7 different flow patterns for all solvent pairs, namely slug flow, droplet flow, slug-droplet flow, parallel, annular, dispersed, and irregular flow. A map of different flow patterns was built to delineate the origin of their formation. We find conventional dimensionless groups are insufficient to uniquely identify the flow patterns. Computational fluid dynamics (CFD) modeling in OpenFOAM shows agreement with the experimental flow patterns for most of the two-phase flows. Principal component analysis reduces the dimensionality of potential descriptors of flow patterns and, unlike prior work using two dimensionless numbers, determines six important features that describe >95% of the variance of the experimental flow patterns. These include the total flow rate, the flow rate ratio between the two phases, the capillary and Ohnesorge numbers of the aqueous phase, and the Weber number and velocity of the organic phase. Lastly, we build a decision-tree model to further regress the data and identify the critical features and demonstrate an accuracy in predicting the flow patterns of up to 93%.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2023

A bi-level data-driven framework for fault-detection and diagnosis of HVAC systems

Long-term operation of heating, ventilation, and air conditioning (HVAC) systems will eventually lead to a range of HVAC system failures, resulting in excessive energy consumption and maintenance costs. Here, to avoid HVAC malfunctioning, fault detection diagnostic (FDD) is utilized as a common practice. Machine learning methods have lately received considerable interest for FDD analysis of HVAC systems due to their high detection accuracy. Meanwhile, HVAC malfunctions are regarded as rare occurrences, hence normal operating data samples are much more accessible than data samples in faulty and malfunctioning conditions. The dominating frequency of normal operation in HVAC datasets has also led to heavily biased classification algorithms within the literature. Moreover, the focus of previous literature has been on increasing the accuracy of the models which leads to a high number of false positives (misleading alarms) in the system. In order to enhance the performance of diagnostic procedures and fill the mentioned gaps, this study proposes a novel data-driven framework. A bi-level machine learning framework is developed for diagnosing faults in air handling units (AHUs) and rooftop units (RTUs) based on principal component analysis (PCA), time series anomaly detection, and random forest (RF). It is shown that PCA can reduce the dataset dimension with one principal component accounting for 95% of data variance. Also, the random forest could classify the faults with 89% precision for single-zone AHU, 85% precision for RTU, and 79% for multi-zone AHU. By proposing this framework, three persistent challenges are addressed: (I) minimizing false positives; (II) accounting for data imbalance; and (III) normal condition monitoring of equipment.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

DOE OSTI2024

Selecting Appropriate Model Complexity: An Example of Tracer Inversion for Thermal Prediction in Enhanced Geothermal Systems

Abstract A major challenge in the inversion of subsurface parameters is the ill‐posedness issue caused by the inherent subsurface complexities and the generally spatially sparse data. Appropriate simplifications of inversion models are thus necessary to make the inversion process tractable and meanwhile preserve the predictive ability of the inversion results. In this study, we investigate the effect of model complexity on fracture aperture inversion and thermal performance prediction in a field‐scale EGS model. Principal component analysis was used to map the aperture field to a low‐dimensional latent space. The complexity of the inversion model was quantitatively represented by the percentage of total variance in the original aperture fields preserved by the latent space. Tracer, pressure and flow rate data were used to invert for fracture aperture through an ensemble‐based inversion method, and the inferred aperture field was used to predict thermal performance. With an over‐simplified aperture model, ensemble collapse occurred. The inverted aperture models failed to resolve necessary flow and transport features, leading to a biased thermal performance prediction. A complex aperture model involved excessive features and was prone to overinterpreting the inversion data. Both the tracer/pressure/flow rate data reproduction and thermal prediction showed significant uncertainties, making it difficult to properly estimate long‐term thermal performance. Fortunately, our results indicate that there exists an appropriate model complexity which can simultaneously match inversion data and predict thermal performance with an acceptable uncertainty. The quality of the fit of tracer data appears to be a useful indicator of such an appropriate model complexity.

15 GEOTHERMAL ENERGY↗

DOE OSTI2024

Insights into Tetravalent Np Speciation in HNO 3 through Spectroelectrochemistry and Multivariate Analysis

In situ optical spectroscopy, spectropotentiometry, and multivariate analysis were applied to the Np(IV) nitrate system to better understand speciation and quantify HNO 3 concentration. Thin-layer spectropotentiometry, or spectroelectrochemistry, was leveraged to isolate and stabilize Np(IV) without compromising the solution conditions and generate representative Vis-NIR absorption spectra from 0.5 to 10 M HNO 3 and benchmark the corresponding Np(IV) molar absorptivity coefficients. Spectra were described with principal component analysis (PCA) to identify the purest Np(IV) absorbance spectra among other oxidation states [e.g., Np(V/VI)] at each acid concentration and then to identify the primary sources of variance within each Np(IV) spectrum with respect to Np(IV) nitrate complexes. Then, partial least-squares regression (PLSR) and support vector regression (SVR) models were built to predict HNO 3 concentration from the Np(IV) spectral data. The nonlinear SVR model outperformed the linear PLSR model for the HNO 3 concentration predictions. Finally, the inclusion of spectra collected in edge and center point HNO 3 concentrations in the calibration set was determined to be crucial for producing models with strong predictive capabilities. The multivariate approach used in this study makes it possible to quantify HNO 3 concentration solely based on Np(IV) absorption spectra, which is essential to quantifying processing streams in various online monitoring applications.

38 RADIATION CHEMISTRY, RADIOCHEMISTRY, AND NUCLEA↗

DOE OSTI2022

Machine-learning from Pseudomonas putida KT2440 transcriptomes reveals its transcriptional regulatory network

Bacterial gene expression is orchestrated by numerous transcription factors (TFs). Elucidating how gene expression is regulated is fundamental to understanding bacterial physiology and engineering it for practical use. In this study, a machine-learning approach was applied to uncover the genome-scale transcriptional regulatory network (TRN) in Pseudomonas putida KT2440, an important organism for bioproduction. We performed independent component analysis of a compendium of 321 high-quality gene expression profiles, which were previously published or newly generated in this study. We identified 84 groups of independently modulated genes (iModulons) that explain 75.7% of the total variance in the compendium. With these iModulons, we (i) expand our understanding of the regulatory functions of 39 iModulon associated TFs (e.g., HexR, Zur) by systematic comparison with 1993 previously reported TF-gene interactions; (ii) outline transcriptional changes after the transition from the exponential growth to stationary phases; (iii) capture group of genes required for utilizing diverse carbon sources and increased stationary response with slower growth rates; (iv) unveil multiple evolutionary strategies of transcriptome reallocation to achieve fast growth rates; and (v) define an osmotic stimulon, which includes the Type VI secretion system, as coordination of multiple iModulon activity changes. Taken together, this study provides the first quantitative genome-scale TRN for P. putida KT2440 and a basis for a comprehensive understanding of its complex transcriptome changes in a variety of physiological states.

09 BIOMASS FUELS↗

DOE OSTI2023

Structuring Nutrient Yields throughout Mississippi/Atchafalaya River Basin Using Machine Learning Approaches

To minimize the eutrophication pressure along the Gulf of Mexico or reduce the size of the hypoxic zone in the Gulf of Mexico, it is important to understand the underlying temporal and spatial variations and correlations in excess nutrient loads, which are strongly associated with the formation of hypoxia. This study’s objective was to reveal and visualize structures in high-dimensional datasets of nutrient yield distributions throughout the Mississippi/Atchafalaya River Basin (MARB). For this purpose, the annual mean nutrient concentrations were collected from thirty-three US Geological Survey (USGS) water stations scattered in the upper and lower MARB from 1996 to 2020. Eight surface water quality indicators were selected to make comparisons among water stations along the MARB over the past two decades. Principal component analysis (PCA) was used to comprehensively evaluate the nutrient yields across thirty-three USGS monitoring stations and identify the major contributing nutrient loads. The results showed that all samples could be analyzed using two main components, which accounted for 81.6% of the total variance. The PCA results showed that yields of orthophosphate (OP), silica (SI), nitrate–nitrites (NO 3 -NO 2 ), and total suspended sediment (TSS) are major contributors to nutrient yields. It also showed that land-planted crops, density of population, domestic and industrial discharges, and precipitation are fundamental causes of excess nutrient loads in MARB. These factors are of great significance for the excess nutrient load management and pollution control of the Mississippi River. It was found that the average nutrient yields were stable within the sub-MARB area, but the large nitrogen yields in the upper MARB and the large phosphorus yields in the lower MARB were of great concern. t-distributed stochastic neighbor embedding (t-SNE) revealed interesting nonlinear and local structures in nutrient yield distributions. Clustering analysis (CA) showed the detailed development of similarities in the nutrient yield distribution. Moreover, PCA, t-SNE, and CA showed consistent clustering results. This study demonstrated that the integration of dimension reduction techniques, PCA, and t-SNE with CA techniques in machine learning are effective tools for the visualization of the structures of the correlations in high-dimensional datasets of nutrient yields and provide a comprehensive understanding of the correlations in the distributions of nutrient loads across the MARB.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2019

Applications of principal component analysis for energy reconstruction in position-sensitive semiconductor detectors

We report semiconductor detectors with pixelated anodes offer a desirable combination of position sensitivity and energy resolution that are suitable for numerous applications in gamma-ray detection and imaging. Pixelization of electrodes also entails a non-uniform amplitude response as a function of gamma-ray interaction position. Energy calibrations as a function of interaction position improve energy resolution, but systematic errors in the energy reconstruction process persist and limit energy resolution at fixed electronic noise. Digitization of signals induced on collecting and adjacent electrodes offers rich information about gamma-ray interactions and their charge drift and collection as a function of time. Despite the abundance of signals for each interaction, human intuition and rule-based approaches fail to utilize the entire set of observed signals to mitigate the sources of systematic error. In order to maximize the utility of digitized signals and improve energy resolution, principal component analysis is applied towards digitized signals observed in semiconductor detectors with pixelated anodes. Principal components identified in the process form the basis for position-specific energy corrections. By leveraging this additional information encoded within digitized signals, energy resolution improves by factors between 10 and 15% with respect to full-width at half and tenth maximum values. The feature that accounts for the maximum amount of explained variance between interactions in any given pixel correlates strongly with the depth of interaction.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

DOE OSTI2018

Plasmonic Conducting Metal Oxide-Based Optical Fiber Sensors for Chemical and Intermediate Temperature-Sensing Applications

The demand for real-time sensors in harsh environments at elevated temperature is significant and increasing. In this manuscript, the chemical and temperature sensing using the optical response through the practical fiber platform is demonstrated, and principle component analysis is coupled with targeted experimental film characterization to understand the fundamental sensing layer properties, which dominate measured gas sensing responses in complex gas mixtures. More specifically, tin-doped indium oxide-decorated sensors fabricated with the sol–gel method show stable and stepwise transmission responses varying over a wide range of H2 concentration (5–100%) at 250–350 °C as well as responses to CH4 and CO to a lesser extent. Measured responses are attributed to modifications to the surface plasmon resonance absorption in the near-infrared range and are dominated by the highest concentrations of the most-reducing analyte based upon systematic mixed gas stream experiments. Principal component analysis is utilized for this type of sensor to improve the quantitative and qualitative understanding of responses, clearly identifying that the dominant principle component (PC #1) accounts for ~78% of total data variance. Correlations between PC #1 and the experimentally derived free carrier concentration confirm that this material property plays the strongest role on the ITO gas sensing mechanism, while correlations between the free carrier mobility and the second most important principle component (PC #2) suggest that this quantity may play a significant but secondary role. As such, the results presented here clarify the relationship between generalized principle components and fundamental sensing materials properties thereby suggesting the pathway toward improved multicomponent gas speciation through sensor layer engineering. Furthermore, the work presented represents a significant step toward the ultimate objective of optical waveguide sensors integrated with multivariate data analytics for multiparameter monitoring with a single sensor element.

47 OTHER INSTRUMENTATION↗

DOE OSTI2014

A fully Bayesian method for jointly fitting instrumental calibration and X-ray spectral models

Owing to a lack of robust principled methods, systematic instrumental uncertainties have generally been ignored in astrophysical data analysis despite wide recognition of the importance of including them. Ignoring calibration uncertainty can cause bias in the estimation of source model parameters and can lead to underestimation of the variance of these estimates. We previously introduced a pragmatic Bayesian method to address this problem. The method is 'pragmatic' in that it introduced an ad hoc technique that simplified computation by neglecting the potential information in the data for narrowing the uncertainty for the calibration product. Following that work, we use a principal component analysis to efficiently represent the uncertainty of the effective area of an X-ray (or γ-ray) telescope. Here, however, we leverage this representation to enable a principled, fully Bayesian method that coherently accounts for the calibration uncertainty in high-energy spectral analysis. In this setting, the method is compared with standard analysis techniques and the pragmatic Bayesian method. The advantage of the fully Bayesian method is that it allows the data to provide information not only for estimation of the source parameters but also for the calibration product—here the effective area, conditional on the adopted spectral model. In this way, it can yield more accurate and efficient estimates of the source parameters along with valid estimates of their uncertainty. Provided that the source spectrum can be accurately described by a parameterized model, this method allows rigorous inference about the effective area by quantifying which possible curves are most consistent with the data.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2024

Evaporation and Transpiration From Multiple Proximal Forests and Wetlands

Abstract Climate change is intensifying the hydrologic cycle and altering ecosystem function, including water flux to the atmosphere through evapotranspiration (ET). ET is made up of evaporation (E) via non‐stomatal surfaces, and transpiration (T) through plant stomata which are impacted by global changes in different ways. E and T are difficult to measure independently at the ecosystem scale, especially across multiple sites that represent different land use and land management strategies. To address this gap in understanding, we applied flux variance similarity (FVS) to quantify how E and T differ across 13 different ecosystems measured using eddy covariance in a 10 × 10 km area from the CHEESEHEAD19 experiment in northern Wisconsin, USA. The study sites included eight forests with a large deciduous broadleaf component, three evergreen needleleaf forests, and two wetlands. Average T/ET for the study period averaged nearly 52% in forested sites and 45% in wetlands, with larger values after excluding periods following rain events when evaporation from canopy interception may be expected. A dominance analysis revealed that environmental variables explained on average 69% of the variance of half‐hourly T, which decreased from summer to autumn. Deciduous and evergreen forests showed similar E trajectories over time despite differences in vegetation phenology, and vapor pressure deficit explained some 13% of the variance E in wetlands but only 5% or less in forests. Retrieval of E and T within a dense network of flux towers lends confidence that FVS is a promising approach for comparing ecosystem hydrology across multiple sites to improve our process‐based understanding of ecosystem water fluxes.

Environmental Sciences & Ecology↗

DOE OSTI2017

A Piecewise Local Partial Least Squares (PLS) Method for the Quantitative Analysis of Plutonium Nitrate Solutions

We have developed a piecewise local (PL) partial least squares (PLS) analysis method for total plutonium measurements by absorption spectroscopy in nitric acid-based nuclear material processing streams. Instead of using a single PLS model that covers all expected solution conditions, the method selects one of several local models based on an assessment of solution absorbance, acidity, and Pu oxidation state distribution. The local models match the global model for accuracy against the calibration set, but were observed in several instances to be more robust to variations associated with measurements in the process. The improvements are attributed to the relative parsimony of the local models. Not all of the sources of spectral variation are uniformly present at each part of the calibration range. Thus, the global model is locally overfitting and susceptible to increased variance when presented with new samples. A second set of models quantifies the relative concentrations of Pu(III), (IV), and (VI). Standards containing a mixture of these species were not at equilibrium due to a disproportionation reaction. Therefore, a separate principal component analysis is used to estimate of the concentrations of the individual oxidation states in these standards in the absence of independent confirmatory analysis. The PL analysis approach is generalizable to other systems where the analysis of chemically complicated systems can be aided by rational division of the overall range of solution conditions into simpler sub-regions.

47 OTHER INSTRUMENTATION↗

DOE OSTI2016

Democratizing an electroluminescence imaging apparatus and analytics project for widespread data acquisition in photovoltaic materials

In this work, we present a description of an electroluminescence (EL) apparatus, easily sourced from commercially available components, with a quantitative image processing platform that demonstrates feasibility for the widespread utility of EL imaging as a characterization tool. We validated our system using a Gage R&R analysis to fnd a variance contribution by the measurement system of 80.56%, which is typically unacceptable, but through quantitative image processing and development of correction factors a variance contribution by the measurement system of 2.41% was obtained. We further validated the system by quantifying the signal-to-noise ratio (SNR) and found values consistent with other systems published in the literature, at SNR values of 10-100, albeit at exposure times of greater than 1 s compared to 10 ms for other systems. Finally, this SNR value range is acceptable for image feature recognition, providing the opportunity for widespread data acquisition and large scale data analytics of photovoltaics.

14 SOLAR ENERGY↗

DOE OSTI2024

Time series methods for the analysis of soundscapes and other cyclical ecological data

Biodiversity monitoring has entered an era of ‘big data’, exemplified by a near-continuous collection of sounds, images, chemical and other signals from organisms in diverse ecosystems. Such data streams have the potential to help identify new threats, assess the effectiveness of conservation interventions, as well as generate new ecological insights. However, appropriate analytical methods are often still missing, particularly with respect to characterizing cyclical temporal patterns. Here, we present a framework for characterizing and analysing ecological responses that represent nonstationary, complex temporal patterns and demonstrate the value of using Fourier transforms to decorrelate continuous data points. In our example, we use a framework based on three approaches (spectral analysis, magnitude squared coherence, and principal component analysis) to characterize differences in tropical forest soundscapes within and across sites and seasons in Gabon. By reconstructing the underlying, cyclic behaviour of the soundscape for each site, we show how one can identify circadian patterns in acoustic activity. Soundscapes in the dry season had a complex diel cycle, requiring multiple harmonics to represent daily variation, while in the wet season there was less variance attributable to the daily cyclic patterns. Our framework can be applied to most continuous, or near-continuous ecological data collected at a fine temporal resolution, allowing ecologists to explore patterns of temporal autocorrelation at multiple levels for biologically meaningful trends. Such methods will become indispensable as biological big data are used to understand the impact of anthropogenic pressures on biodiversity and to inform efforts to mitigate them.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2023

BigNeuron: a resource to benchmark and predict performance of algorithms for automated tracing of neurons in light microscopy datasets

BigNeuron is an open community bench-testing platform with the goal of setting open standards for accurate and fast automatic neuron tracing. We gathered a diverse set of image volumes across several species that is representative of the data obtained in many neuroscience laboratories interested in neuron tracing. Here, we report generated gold standard manual annotations for a subset of the available imaging datasets and quantified tracing quality for 35 automatic tracing algorithms. The goal of generating such a hand-curated diverse dataset is to advance the development of tracing algorithms and enable generalizable benchmarking. Together with image quality features, we pooled the data in an interactive web application that enables users and developers to perform principal component analysis, t-distributed stochastic neighbor embedding, correlation and clustering, visualization of imaging and tracing data, and benchmarking of automatic tracing algorithms in user-defined data subsets. The image quality metrics explain most of the variance in the data, followed by neuromorphological features related to neuron size. Furthermore, we observed that diverse algorithms can provide complementary information to obtain accurate results and developed a method to iteratively combine methods and generate consensus reconstructions. The consensus trees obtained provide estimates of the neuron structure ground truth that typically outperform single algorithms in noisy datasets. However, specific algorithms may outperform the consensus tree strategy in specific imaging conditions. Finally, to aid users in predicting the most accurate automatic tracing results without manual annotations for comparison, we used support vector machine regression to predict reconstruction quality given an image volume and a set of automatic tracings.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2020

Leveraging Machine Learning Capabilities for the Characterization of Irradiated Uranium: A Case Study of Analysis Methods for Nuclear Safeguards and Nuclear Forensics

Nondestructively determining the initial enrichment of irradiated uranium is a complex and laborious multivariable problem due to the presence of fission products. This work demonstrates the capabilities of machine learning to analyze gamma-ray spectral data to determine initial enrichment without knowledge of the decay time of the sample. The approach developed is agnostic to the particular scenario and is applicable to a wide variety of applications in nuclear forensics and nuclear safeguards. We irradiated 5 mg uranium standard reference materials at discrete enrichment values ranging from 0.02% to 97% 235 U (weight percent) in UT Austin’s Nuclear Engineering Teaching Laboratory TRIGA Mark II 1.1 MW research reactor, allowed each to decay for 8 hours, and then measured each sample via gamma-ray spectrometry for 50 hours post-irradiation yielding 1,400 individual gamma-ray spectra discretized into 8,192 energy bins. We then trained decision trees models to analyze individual gamma-ray spectra and estimate the associated initial enrichment without knowledge of the time since end of irradiation. We evaluated the performance of the models with a reserved test set not used for training or calibrating the model. A decision tree model constructed with this procedure achieved a mean absolute error in initial enrichment determination of 2.3% (weight percent 235 U). Next, we implemented a principal component analysis pre-processing routine of the gamma-ray spectrometry data to reduce the dimensionality of the dataset from 8,192 channels in the spectrum to 10 principal components while retaining over 99% of the inherent variance in the data. Decision tree models constructed with these data demonstrated decreased mean absolute error in enrichment determination, reduced computational time, and decreased complexity. A single decision tree model constructed with this procedure achieved a mean absolute error in initial enrichment determination of 0.05% (weight percent 235 U). Furthermore, we analyzed these models with learning curves to ensure that overfitting did not occur. The capabilities provided by these models can be naturally extended to other application-focused measurements in the fields of nuclear safeguards, nuclear forensics, and nuclear non-proliferation.

Drescher, Adam↗