Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “regression 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.

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

At least 1,765 records · Page 98

Rethinking Sensitivity Analysis of Nuclear Simulations with Topology

In nuclear engineering, understanding the safety margins of the nuclear reactor via simulations is arguably of paramount importance in predicting and preventing nuclear accidents. It is therefore crucial to perform sensitivity analysis to understand how changes in the model inputs affect the outputs. Modern nuclear simulation tools rely on numerical representations of the sensitivity information -- inherently lacking in visual encodings -- offering limited effectiveness in communicating and exploring the generated data. In this paper, we design a framework for sensitivity analysis and visualization of multidimensional nuclear simulation data using partition-based, topology-inspired regression models and report on its efficacy. We rely on the established Morse-Smale regression technique, which allows us to partition the domain into monotonic regions where easily interpretable linear models can be used to assess the influence of inputs on the output variability. The underlying computation is augmented with an intuitive and interactive visual design to effectively communicate sensitivity information to the nuclear scientists. Our framework is being deployed into the multi-purpose probabilistic risk assessment and uncertainty quantification framework RAVEN (Reactor Analysis and Virtual Control Environment). We evaluate our framework using an simulation dataset studying nuclear fuel performance.

computational topology↗

Interpreting Write Performance of Supercomputer I/O Systems with Regression Models

This work seeks to advance the state of the art in HPC I/O performance analysis and interpretation. In particular, we demonstrate effective techniques to: (1) model output performance in the presence of I/O interference from production loads; (2) build features from write patterns and key parameters of the system architecture and configurations; (3) employ suitable machine learning algorithms to improve model accuracy. We train models with five popular regression algorithms and conduct experiments on two distinct production HPC platforms. We find that the lasso and random forest models predict output performance with high accuracy on both of the target systems. We also explore use of the models to guide adaptation in I/O middleware systems, and show potential for improvements of at least 15% from model-guided adaptation on 70% of samples, and improvements up to 10× on some samples for both of the target systems.

Xie, Bing↗

HAPPA: A Modular Platform for HPC Application Resilience Analysis with LLMs Embedded

High-performance computing (HPC) systems are increasingly vulnerable to soft errors, which pose significant challenges in maintaining computational accuracy and reliability. Predicting the resilience of HPC applications to these errors is crucial for robust code protection and detailed resilience analysis. In this study, we present HAppA, a modular platform designed for HPC Application Resilience Analysis. Embedding Large Language Models (LLMs), HAppA addresses understanding the context information of long code sequences typical in HPC applications. HAppA implements a novel code representation module that chunks the code into fixed-size segments and aggregates the embeddings of these segments. Three aggregation methods have been explored: MeanPooling, MaxPooling, and LSTM-based techniques. We built a DAtaset for REsilience analysis using Fault Injection (FI), named DARE. Using our DARE dataset, HAppA is trained for regression prediction tasks. Our evaluation results demonstrate the predictive accuracy of HAppA compared to other models, particularly noting that the LSTM-based aggregation method -- HAppA-LSTM -- achieves a mean squared error (MSE) of 0.078 for SDC prediction, surpassing the existing state-of-the-art PARIS model, which recorded an MSE of 0.1172. Additionally, HAppA with the KeyBERT model extracts a list of keywords representing the source code. A comprehensive importance analysis of these keywords further elucidates the code patterns contributing to the error rate. These findings highlight the effectiveness of HAppA in analyzing the resilience of HPC applications and establish a new benchmark for predictive accuracy in resilience.

Jiang, Hailong [Kent State University]↗

Relationship Between Column-Density and Surface Mixing Ratio: Statistical Analysis of O3 and NO2 Data from the July 2011 Maryland DISCOVER-AQ Mission

To investigate the ability of column (or partial column) information to represent surface air quality, results of linear regression analyses between surface mixing ratio data and column abundances for O3 and NO2 are presented for the July 2011 Maryland deployment of the DISCOVER-AQ mission. Data collected by the P-3B aircraft, ground-based Pandora spectrometers, Aura/OMI satellite instrument, and simulations for July 2011 from the CMAQ air quality model during this deployment provide a large and varied data set, allowing this problem to be approached from multiple perspectives. O3 columns typically exhibited a statistically significant and high degree of correlation with surface data (R(sup 2) > 0.64) in the P- 3B data set, a moderate degree of correlation (0.16 < R(sup 2) < 0.64) in the CMAQ data set, and a low degree of correlation (R(sup 2) < 0.16) in the Pandora and OMI data sets. NO2 columns typically exhibited a low to moderate degree of correlation with surface data in each data set. The results of linear regression analyses for O3 exhibited smaller errors relative to the observations than NO2 regressions. These results suggest that O3 partial column observations from future satellite instruments with sufficient sensitivity to the lower troposphere can be meaningful for surface air quality analysis.

nitrogen dioxide↗

Dosimetric Predictors of Radiation-induced Acute Nausea and Vomiting in IMRT for Nasopharyngeal Cancer

Purpose: We wanted to investigate dosimetric parameters that would predict radiation-induced acute nausea and vomiting in intensity-modulated radiation therapy (IMRT) for undifferentiated carcinoma of the nasopharynx (NPC). Methods and Materials: Forty-nine consecutive patients with newly diagnosed NPC were treated with IMRT alone in this prospective study. Patients receiving any form of chemotherapy were excluded. The dorsal vagal complex (DVC) as well as the left and right vestibules (VB-L and VB-R, respectively) were contoured on planning computed tomography images. A structure combining both the VB-L and the VB-R, named VB-T, was also generated. All structures were labeled organs at risk (OAR). A 3-mm three-dimensional margin was added to these structures and labeled DVC+3 mm, VB-L+3 mm, VB-R+3 mm, and VB-T+3 mm to account for physiological body motion and setup error. No weightings were given to these structures during optimization in treatment planning. Dosimetric parameters were recorded from dose-volume histograms. Statistical analysis of parameters' association with nausea and vomiting was performed using univariate and multivariate logistic regression. Results: Six patients (12.2%) reported Grade 1 nausea, and 8 patients (16.3%) reported Grade 2 nausea. Also, 4 patients (8.2%) complained of Grade 1 vomiting, and 4 patients (8.2%) experienced Grade 2 vomiting. No patients developed protracted nausea and vomiting after completion of IMRT. For radiation-induced acute nausea, V40 (percentage volume receiving at least 40Gy) to the VB-T and V40>=80% to the VB-T were predictors, using univariate analysis. On multivariate analysis, V40>=80% to the VB-T was the only predictor. There were no predictors of radiation-induced acute vomiting, as the number of events was too small for analysis. Conclusions: This is the first study demonstrating that a V40 to the VB-T is predictive of radiation-induced acute nausea. The vestibules should be labeled as sensitive OARs, and weightings should be considered for dose sparing during optimization in the treatment planning of IMRT.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Influence of the North Atlantic SST Variability on the Atmospheric Circulation during the Twentieth Century

The ocean–atmosphere coupling in the North Atlantic is investigated during the twentieth century using maximum covariance analysis of sea surface temperature (SST) and 500-hPa geopotential height analyses and performing regressions on dynamical diagnostics such as Eady growth rate, wave activity flux, and velocity potential. The North Atlantic Oscillation (NAO) generates the so-called SST anomaly tripole. A rather similar SST anomaly tripole, with the subpolar anomaly displaced to the east and a more contracted subtropical anomaly, which is referred to as the North Atlantic horseshoe pattern, in turn influences the atmosphere. In the fall and early winter, the response is NAO like and primarily results from subpolar forcing centered over the Labrador Sea and off Newfoundland. In summer, the largest atmospheric response to SST resembles the east Atlantic pattern and results from a combination of subpolar and tropical forcing. To emphasize the interannual to multidecadal variability, the same analysis is repeated after low-pass filtering. The SST influence is dominated by the Atlantic multidecadal oscillation (AMO), which also has a horseshoe shape, but with larger amplitude in the subpolar basin. A warm AMO phase leads to an atmospheric warming limited to the lower troposphere in summer, while it leads to a negative phase of the NAO in winter. The winter influence of the AMO is suggested to be primarily forced by the Atlantic SSTs in the northern subtropics. Such influence of the AMO is found in winter instead of early winter because the winter SST anomalies have a larger persistence, presumably because of SST reemergence.

54 ENVIRONMENTAL SCIENCES↗

A parametric finite element study for determining burst strength of thin and thick-walled pressure vessels

To accurately predict the burst strength of both thin and thick-walled pressure vessels (PVs), a parametric study of PV burst strength was performed for a wide range of vessel geometries and materials using elastic-plastic finite element analysis (FEA). A valid FEA model was established through a detailed study of 2D versus 3D FEA models, the critical stress failure criterion versus the limit load criteria, and the thick-wall effect on the FEA simulations. Here, the results show that the stresses and strains at the mean diameter, rather than outside diameter, determines a more accurate burst strength for both thin and thick-walled PVs. On this basis, a parametrized FEA script using the ABAQUS Python application programming interface (API) was used to create a large database of PV burst strengths for a variety of vessel geometries and materials, demonstrating that Python scripting is a powerful technique for performing parametric studies or generating large databases. From the FEA results, using the regression method, a new burst pressure model was developed as a function of the vessel geometry (D/t ratio) and material properties (UTS and n). As validated by a large number of full-scale burst test data, the proposed burst model can very accurately predict the burst strength for both thin and thick-walled PVs.

42 ENGINEERING↗

Machine Learning-Based Investigation of the Associations between Residential Power Consumption and Weather Conditions

Machine learning feature selection analyses were conducted to understand the impact of various weather attributes on power electricity demand and its components in the northwest United States. Load profiles data were obtained from the Northwest Energy Efficiency Alliance (NEEA) and processed to yield hourly load profiles with the same temporal resolution as the weather conditions data obtained from the National Oceanic and Atmospheric Administration (NOAA) for representative weather stations. The data were divided into five seasons; for each season, for each load component, the influences of weather factors were evaluated and quantified using cross-correlation, principal component analysis, and mutual information evaluation; then predictive models were developed based on the ranked/screened factors using the regression tree (RT) approach. After multi-fold cross-validation, the optimal complexity/depth of the regress tree models is found to vary for different load components, and the prediction accuracy using RT and weather data can be higher than 80% for heating and refrigeration, or lower than 30% for vent and cooking. The study provides guidance on improvement of load profile generation, or even serves as a standalone predictive tool for approximating load profiles in balance authorities or climate zones where load profile data is absent.

load composite, load profiling, NEEA, mutual infor↗

Delegated Regressor, A Robust Approach for Automated Anomaly Detection in the Soil Radon Time Series Data

We propose a new method based on the idea of delegating regressors for predicting the soil radon gas concentration (SRGC) and anomalies in radon or any other time series data. The proposed method is compared to different traditional boosting e.g., Extreme Gradient Boosting (EGB) and simple regression methods e.g., support vector regressors with linear kernel and radial kernel in terms of accurate predictions. R language has been used for the statistical analysis of radon time series (RTS) data. The results obtained show that the proposed methodology predicts SRGC more accurately when compared to different traditional boosting and regression methods. The best correlation is found between the actual and predicted radon concentration for window size of 2 i.e., two days before and after the start of seismic activities. RTS data was collected from 05 February 2017 to 16 February 2018, including 7 seismic events recorded during the study period. Findings of study show that the proposed methodology predicts the SRGC with more precision, for all the window sizes, by overlapping predicted with the actual radon time series concentrations.

54 ENVIRONMENTAL SCIENCES↗

S AP F LOWER : an automated tool for sap flow data preprocessing, gap-filling, and analysis using deep learning

Sap flow, a critical process in plant water use and ecosystem water cycles, is often measured using thermal dissipation probes (TDP) due to their ease of installation and continuous data collection. However, sap flow data frequently include noise, outliers, and gaps, creating challenges for analysis and requiring substantial manual processing. We developed S AP F LOWER , a tool that automates data preprocessing, model training, gap-filling, sapwood area scaling and modeling, and water use analysis. It integrates autocleaning, machine learning and deep learning models (e.g. random forest, Gaussian process regression, long short-term memory (LSTM), bidirectional LSTM (BiLSTM)), and efficient workflows to process sap flow data. S AP F LOWER can remove over 90% of noisy data while preserving legitimate variations and achieve high accuracy in gap-filling based on user-determined parameters. Random forest, LSTM, and BiLSTM models reduced root mean square error to 10% or less for long-term gaps. Model training and prediction can be performed efficiently within seconds. S AP F LOWER significantly enhances the efficiency and accessibility of TDP data analysis by automating complex tasks, enabling researchers without programming expertise to employ advanced techniques. Future improvements will focus on species-specific corrections for TDP and support for additional measurement methods. S AP F LOWER is openly available on GitHub (https://github.com/JiaxinWang123/SapFlower) and Zenodo (doi: 10.5281/zenodo.13665919).

ecosystem water balance↗

Acousto-ultrasonic verification of the strength of filament wound composite material

The concept of acousto-ultrasonic (AU) waveform partitioning was applied to nondestructive evaluation of mechanical properties in filament wound composites (FWC). A series of FWC test specimens were subjected to AU analysis and the results were compared with destructively measured interlaminar shear strengths (ISS). AU stress-wave factor (SWF) measurements gave greater than 90 percent correlation coefficient upon regression against the ISS. This high correlation was achieved by employing the appropriate time and frequency domain partitioning as dictated by wave propagation path analysis. There is indication that different SWF frequency partitions are sensitive to ISS at different depths below the surface.

Kautz, H. E.↗

Experiments with Test Case Generation and Runtime Analysis

Software testing is typically an ad hoc process where human testers manually write many test inputs and expected test results, perhaps automating their execution in a regression suite. This process is cumbersome and costly. This paper reports preliminary results on an approach to further automate this process. The approach consists of combining automated test case generation based on systematically exploring the program's input domain, with runtime analysis, where execution traces are monitored and verified against temporal logic specifications, or analyzed using advanced algorithms for detecting concurrency errors such as data races and deadlocks. The approach suggests to generate specifications dynamically per input instance rather than statically once-and-for-all. The paper describes experiments with variants of this approach in the context of two examples, a planetary rover controller and a space craft fault protection system.

Artho, Cyrille↗

Surrogate model evaluation and building energy benchmarking for commercial buildings

Building energy consumption benchmarking involves challenges associated with various energy patterns for different building types; heating, ventilating, and air-conditioning (HVAC) system types; and climates. Given significant variation in energy use patterns, accurate prediction of long-term energy use using surrogate models remains challenging. Multiple linear regression (MLR) is commonly used for building energy benchmarking because of its simple structure; however, it lacks accuracy compared to other black-box models. Although many studies have compared surrogate models and offer guidance on model selection based on metrics, they do not provide detailed analysis on improving the surrogate model accuracy. In this paper, we implement a surrogate model using polynomial ridge regression (i.e., MLR with interaction terms combined with ridge regularization) for small office and retail strip mall buildings across six HVAC system types and all climate zones, for electricity and natural gas in baseline and proposed scenarios. A simulation workflow is developed using OpenStudio TM /EnergyPlus TM to generate simulation data using measures over a wide range of efficiency inputs. Enhancements based on statistical insights are used for improving the model accuracy using filters, input transformations, and change points. Surrogate models achieved average coefficient of variation of the root mean squared error (CVRMSE) values of 2.17, 1.06, 2.05, and 3.26 for proposed electricity, proposed natural gas, baseline electricity, and baseline natural gas, respectively, with enhancements reducing CVRMSE by an average of 14.9% across all combinations. We provide model interpretation via Shapley additive explanations to determine which input variables most influence energy consumption and provide supportive arguments for enhancements.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Improving air leakage prediction of buildings using the fan pressurization method with the Weighted Line of Organic Correlation

In many countries, the fan pressurization method is the most frequently chosen approach for measuring the air leakage of houses. The measurements are usually performed at pressures that far exceed pressures to which buildings are exposed to under normal conditions. A fit of these tests to the power-law formulation allows an extrapolation to data points outside the measured pressure range. With the Ordinary Least Square (OLS) fitting method, the pressure exponent and flow coefficient can be determined. However, the measurement results are highly sensitive to uncertainties induced by external factors like changing wind conditions during the tests, which is neglected by OLS. This may lead to errors in the prediction of flows at lower pressures. The Weighted Line of Organic Correlation (WLOC) is an alternative approach and takes measurement uncertainty into account. In this paper, a statistical analysis of an extensive data set of pressurization measurements has been performed. Both regression techniques have been compared for almost 7500 fan pressurization measurements of six houses in 109 different house leak configurations. The variability in predicting pressure exponent and flow coefficient for both WLOC and OLS regression was analyzed using probability density functions. It was found that the Weighted Line of Organic Correlation significantly decreases the uncertainty in predicting pressure exponent, flow coefficient, and other low-pressure air leakage metrics compared to the Ordinary Least Square fitting. Here, the authors highly recommend the implementation of WLOC in current measurement standards and test equipment.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Association of Locoregional Control With High Body Mass Index in Women Undergoing Breast Conservation Therapy for Early-Stage Breast Cancer

Purpose: Obesity, as measured by the body mass index (BMI), is a risk factor for distant recurrence and decreased survival in breast cancer. We sought to determine whether the BMI correlated with local recurrence and reduced survival in a cohort of predominantly obese women treated with breast conservation therapy. Methods and Materials: From 1998 to 2010, 154 women with early-stage invasive breast cancer and 39 patients with ductal carcinoma in situ underwent prone whole breast irradiation. Cox proportional hazards regression, Kaplan-Meier methods with the log-rank test, and multivariate analysis were used to explore the association of the outcomes with the BMI. Results: The median patient age was 60 years, and the median follow-up duration was 73 months. The median BMI was 33.2 kg/m{sup 2}; 91% of the patients were overweight (BMI ≥25 kg/m{sup 2}) and 69% of the patients were clinically obese (BMI ≥30 kg/m{sup 2}). The BMI was significantly associated with the locoregional recurrence-free interval for patients with invasive cancer and ductal carcinoma in situ (hazard ratio [HR], 1.09; P=.047). Also, a trend was seen for increased locoregional recurrence with a higher BMI (P=.09) for patients with invasive disease, which was significant when examining the outcomes with a BMI stratified by the median value of 33.2 kg/m{sup 2} (P=.008). A greater BMI was also significantly associated with decreased distant recurrence-free interval (HR, 1.09; P=.011) and overall survival (HR, 1.09; P=.004); this association remained on multivariate analysis (distant recurrence-free interval, P=.034; overall survival, P=.0007). Conclusions: These data suggest that the BMI might affect the rate of locoregional recurrence in breast cancer patients. A higher BMI predicted a worse distant recurrence-free interval and overall survival. The present investigation adds to the increasing evidence that BMI is an important prognostic factor in early-stage breast cancer treated with breast conservation therapy.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Indoor versus Outdoor Air Quality during Wildfires

The human behavioral modification recommendations during wildfire events are based on particulate matter and may be confounded by the potential risks of gas-phase pollutants such as polycyclic aromatic hydrocarbons (PAHs). Moreover, the majority of adults spend over 90 percent of their time indoors where there is an increased concern of indoor air quality during wildfire events. We address these timely concerns by evaluating paired indoor and outdoor PAH concentrations in residential locations and their relationship with satellite model-based categorization of wildfire smoke intensity. Low density polyethylene passive air samplers were deployed at six residential sites for one week in Eugene, Oregon with matched indoor and outdoor samples and twenty-four-hour time resolution. Samples were then quantitatively analyzed for 63 PAH concentrations using gas-chromatography. A probabilistic principal components analysis was used to reduce all 63 PAHs into an aggregate measure. Linear regression of the first principal component against indoor versus outdoor shows that indoor gas-phase PAH concentrations are consistently equal or greater than outdoor concentrations. Regression against a satellite-based model for wildfire smoke shows that outdoor, but not indoor gas-phase PAH concentrations are likely associated with wildfire events. Furthermore, these results point towards the need to include gas-phase pollutants such as PAHs in air pollution risk assessment.

54 ENVIRONMENTAL SCIENCES↗

DETERMINISTIC COMPONENTS IN THE LIGHT CURVE AMPLITUDE OF Y OPH

About two decades after the discovery of the amplitude decline of the light curve of the classical Cepheid Y Oph, its study is resumed using an increased amount of homogenized data and an extended time base. In our approach, the investigation of different time series concerning the light curve amplitude of Y Oph is not only the reason for the present study, but also a stimulus for developing a coherent methodology for studying long- and short-term variability phenomena in variable stars, taking into account the details of concrete observing conditions: amount of data, data sampling, time base, and individual errors of observational data. The statistical significance of this decreasing trend was estimated by assuming its linearity. We approached the decision-making process by formulating adequate null and alternative hypotheses, and testing the value of the regression line slope for different data sets via Monte Carlo simulations. A variability analysis, through various methods, of the original data and of the residuals obtained after removing the linear trend was performed. We also proposed a new statistical test, based on amplitude spectrum analysis and Monte Carlo simulations, intended to evaluate how detectible is a given (linear) trend in well-defined observing conditions: the trend detection probability. The main conclusion of our study on Y Oph is that, even if the false alarm probability is low enough to consider the decreasing trend to be statistically significant, the available data do not allow us to obtain a reasonably powerful test. We are able to confirm the light curve amplitude decline, and the order of magnitude of its slope with a better statistical substantiation. According to the obtained values of the trend detection probability, it seems that the trend we are dealing with is marked by a low detectibility. Our attempt to find signs of possible variability phenomena at shorter timescales ended by emphasizing the relative constancy of our data, within their precision limits.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗