Search NASA⌕ Search

SEARCH · Search NASA

Results for “data investigations”

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
DOE OSTI2025

Data Quality Assessment Process for Real-Time Data-Driven Traffic Microsimulation of Smart Corridor

Smart corridor digital twins are often created for the development and evaluation of emerging intelligent transportation systems and Connected and Autonomous Vehicle (CAV) technologies. However, limited guidance exists for data quality assessment for digital twin development. To address this, this paper discusses the data quality assessment utilized to develop data-driven real-time microscopic simulation models, i.e., digital twins, for two separate smart corridors: the North Avenue Smart Corridor in Atlanta, GA, and the Martin Luther King Smart Corridor in Chattanooga, Tennessee. This paper provides a summary of the author’s investigations of data requirements and data characteristics for the given smart corridor digital twin development efforts. With a focus on data, this summary includes a description of the data investigation process, key data issues observed, and strategies to address observed issues. Discussion is provided to help expand the lessons from these studies to other digital twin development efforts.

Saroj, Abhilasha [ORNL] (ORCID:0000000191178063)↗
NASA NTRS2012

An Intelligent Strain Gauge with Debond Detection and Temperature Compensation

The harsh rocket propulsion test environment will expose any inadequacies associated with preexisting instrumentation technologies, and the criticality for collecting reliable test data justifies investigating any encountered data anomalies. Novel concepts for improved systems are often conceived during the high scrutiny investigations by individuals with an in-depth knowledge from maintaining critical test operations. The Intelligent Strain Gauge concept was conceived while performing these kinds of activities. However, the novel concepts are often unexplored even if it has the potential for advancing the current state of the art. Maturing these kinds of concepts is often considered to be a tangential development or a research project which are both normally abandoned within the propulsion-oriented environment. It is also difficult to justify these kinds of projects as a facility enhancement because facility developments are only accepted for mature and proven technologies. Fortunately, the CIF program has provided an avenue for bringing the Intelligent Strain Gauge to fruition. Two types of fully functional smart strain gauges capable of performing reliable and sensitive debond detection have been successfully produced. Ordinary gauges are designed to provide test article data and they lack the ability to supply information concerning the gauge itself. A gauge is considered to be a smart gauge when it provides supplementary data relating other relevant attributes for performing diagnostic function or producing enhanced data. The developed strain gauges provide supplementary signals by measuring strain and temperature through embedded Karma and nickel chromium (NiCr) alloy elements. Intelligently interpreting the supplementary data into valuable information can be performed manually, however, integrating this functionality into an automatic system is considered to be an intelligent gauge. This was achieved while maintaining a very low mass. The low mass enables debond detection and temperature compensation to be performed when the gauge is utilized on small test articles. It was also found that the element's mass must be relatively small to avoid overbearing the desired thermal dissipation characteristics. Detecting the degradation of a gauge s bond was reliably achieved by correlating thermal dissipation with the bond s integrity. This was accomplished by precisely coupling a NiCr element with a Karma element for accurately interjecting and quantifying thermal energy. A finite amount of thermal energy is consistently placed in the gauge by electrically powering the NiCr element. The energy will only be temporarily stored before it begins to dissipate into the surrounding structure through the gauge bond. The ability to transmit the energy into the structure becomes greatly inhibited by any discontinuity in the bond s substrate. Therefore, the way the thermal dissipation occurs will reveal even the slightest change in the integrity of the bond.

Jensen, Scott L.↗
DOE OSTI2024

Deep learning with mixup augmentation for improved pore detection during additive manufacturing

In additive manufacturing (AM), process defects such as keyhole pores are difficult to anticipate, affecting the quality and integrity of the AM-produced materials. Hence, considerable efforts have aimed to predict these process defects by training machine learning (ML) models using passive measurements such as acoustic emissions. This work considered a dataset in which keyhole pores of a laser powder bed fusion (LPBF) experiment were identified using X-ray radiography and then registered both in space and time to acoustic measurements recorded during the LPBF experiment. Due to AM’s intrinsic process controls, where a pore-forming event is relatively rare, the acoustic datasets collected during monitoring include more non-pores than pores. In other words, the dataset for ML model development is imbalanced. Moreover, this imbalanced and sparse data phenomenon remains ubiquitous across many AM monitoring schemes since training data is nontrivial to collect. Hence, we propose a machine learning approach to improve this dataset imbalance and enhance the prediction accuracy of pore-labeled data. Specifically, we investigate how data augmentation helps predict pores and non-pores better. This imbalance is improved using recent advances in data augmentation called Mixup, a weak-supervised learning method. Convolutional neural networks (CNNs) are trained on original and augmented datasets, and an appreciable increase in performance is reported when testing on five different experimental trials. When ML models are trained on original and augmented datasets, they achieve an accuracy of 95% and 99% on test datasets, respectively. We also provide information on how dataset size affects model performance. Lastly, we investigate the optimal Mixup parameters for augmentation in the context of CNN performance.

42 ENGINEERING↗
DOE OSTI2024

Deep learning with mixup augmentation for improved pore detection during additive manufacturing

In additive manufacturing (AM), process defects such as keyhole pores are difficult to anticipate, affecting the quality and integrity of the AM-produced materials. Hence, considerable efforts have aimed to predict these process defects by training machine learning (ML) models using passive measurements such as acoustic emissions. This work considered a dataset in which keyhole pores of a laser powder bed fusion (LPBF) experiment were identified using X-ray radiography and then registered both in space and time to acoustic measurements recorded during the LPBF experiment. Due to AM’s intrinsic process controls, where a pore-forming event is relatively rare, the acoustic datasets collected during monitoring include more non-pores than pores. In other words, the dataset for ML model development is imbalanced. Moreover, this imbalanced and sparse data phenomenon remains ubiquitous across many AM monitoring schemes since training data is nontrivial to collect. Hence, we propose a machine learning approach to improve this dataset imbalance and enhance the prediction accuracy of pore-labeled data. Specifically, we investigate how data augmentation helps predict pores and non-pores better. This imbalance is improved using recent advances in data augmentation called Mixup, a weak-supervised learning method. Convolutional neural networks (CNNs) are trained on original and augmented datasets, and an appreciable increase in performance is reported when testing on five different experimental trials. When ML models are trained on original and augmented datasets, they achieve an accuracy of 95% and 99% on test datasets, respectively. We also provide information on how dataset size affects model performance. Lastly, we investigate the optimal Mixup parameters for augmentation in the context of CNN performance.

36 MATERIALS SCIENCE↗
NASA NTRS1974

A Monte Carlo investigation of experimental data requirements for fitting polynomial functions

This report examines the extent to which sample size affects the accuracy of a low order polynomial approximation of an experimentally observed quantity and establishes a trend toward improvement in the accuracy of the approximation as a function of sample size. The task is made possible through a simulated analysis carried out by the Monte Carlo method, in which data are generated by using several transcendental or algebraic functions as models. Contaminated data of varying amounts are fitted to linear quadratic or cubic polynomials, and the behavior of the mean-squared error of the residual variance is determined as a function of sample size. Results indicate that the effect of the size of the sample is significant only for relatively small sample sizes and diminishes drastically for moderate and large amounts of experimental data.

Canavos, G. C.↗