Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Anomaly Detection”

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 649 records · Page 36

Anomalous change detection in imagery

A distribution-based anomaly detection platform is described that identifies a non-flat background that is specified in terms of the distribution of the data. A resampling approach is also disclosed employing scrambled resampling of the original data with one class specified by the data and the other by the explicit distribution, and solving using binary classification.

47 OTHER INSTRUMENTATION↗

Artificial intelligence techniques for ground test monitoring of rocket engines

An expert system is being developed which can detect anomalies in Space Shuttle Main Engine (SSME) sensor data significantly earlier than the redline algorithm currently in use. The training of such an expert system focuses on two approaches which are based on low frequency and high frequency analyses of sensor data. Both approaches are being tested on data from SSME tests and their results compared with the findings of NASA and Rocketdyne experts. Prototype implementations have detected the presence of anomalies earlier than the redline algorithms that are in use currently. It therefore appears that these approaches have the potential of detecting anomalies early eneough to shut down the engine or take other corrective action before severe damage to the engine occurs.

Ali, Moonis↗

A spatially regularized detector for emergent/re-emergent disease outbreaks

Early detection of outbreaks caused by emergent pathogens, using epidemiological surveillance data i.e., daily case counts, is difficult. This is because the data tend to be noisy during the early epoch of the outbreak. In contrast, the spread-rate of the disease tends to be well-behaved, as it depends only on the mixing patterns of the population and the characteristics of the pathogen, neither of which behave erratically in space-time. In this report, we explore whether the spread-rate can be used for epidemiological surveillance, conditional on case count data. Estimating the spread-rate from case count data allows us to exploit exogenous information, e.g., incubation period distributions etc., which can considerably smooth out any erratic temporal behavior. Further, epidemiological dynamics are spatially correlated, and if case counts are available for multiple areal units e.g., counties, these correlations could potentially be used to suppress noise in the early epoch data. These exogenous information and structure are not exploited by conventional syndromic surveillance detectors to extract a well-behaved latent variable for monitoring purposes. The technical challenge lies in the estimation of the spread-rate field jointly over a collection of areal units; further, the spread-rate varies over time. We develop a method based on mean-field variational inference to approximately estimate the spread-rate field, using a Gaussian Random Field Model for spatial regularization. The method is tested on the estimation of spread-rate in the thirty-three counties of New Mexico and detect the arrival of the Fall 2020 COVID-19 wave in September 2020. We find that the method is scalable, but underestimates the uncertainty in the estimated spread-rate field. We detect the arrival of the Fall 2020 wave a week ahead of conventional syndromic surveillance algorithms, but our simplistic detection algorithm, based on simple anomaly detection, suffers from a high false positive rate, similar to conventional detectors.

59 BASIC BIOLOGICAL SCIENCES↗

Neural Network Approaches for Mobile Spectroscopic Gamma-Ray Source Detection

Artificial neural networks (ANNs) for performing spectroscopic gamma-ray source identification have been previously introduced, primarily for applications in controlled laboratory settings. To understand the utility of these methods in scenarios and environments more relevant to nuclear safety and security, this work examines the use of ANNs for mobile detection, which involves highly variable gamma-ray background, low signal-to-noise ratio measurements, and low false alarm rates. Simulated data from a 2” × 4” × 16” NaI(Tl) detector are used in this work for demonstrating these concepts, and the minimum detectable activity (MDA) is used as a performance metric in assessing model performance.In addition to examining simultaneous detection and identification, binary spectral anomaly detection using autoencoders is introduced in this work, and benchmarked using detection methods based on Non-negative Matrix Factorization (NMF) and Principal Component Analysis (PCA). On average, the autoencoder provides a 12% and 23% improvement over NMF- and PCA-based detection methods, respectively. Additionally, source identification using ANNs is extended to leverage temporal dynamics by means of recurrent neural networks, and these time-dependent models outperform their time-independent counterparts by 17% for the analysis examined here. The paper concludes with a discussion on tradeoffs between the ANN-based approaches and the benchmark methods examined here.

Bilton, Kyle J. (ORCID:0000000184553689)↗

An Open, Cloud-Based Platform for Whole-Building Fault Detection and Diagnostics

Small commercial buildings in the U.S. waste an estimated 300 Trillion BTU (approximately $6 billion in energy costs) annually due to faults, but lack cost-effective automated fault detection and diagnosis (AFDD) tools. NREL and GE Global Research are partnering to develop hybrid AFDD algorithms tailored to the unique needs of small commercial buildings (</=10,000 sq. ft.). The approach combines physics-based energy modeling tools with state-of-the-art machine learning algorithms to enable diagnostics using modeled input data. NREL, in partnership with Purdue University, has adapted DOE's physics-based modeling software EnergyPlus and OpenStudio to simulate common building faults. Cloud-based implementations of these physics-based models were used to generate data for multiple climate zones. The data were then used to train machine learning algorithms, such as those available in GE's Predix platform, to detect and diagnose building faults. Key features of the system including the automated, cloud-implementation of EnergyPlus, genetic algorithms for automated model feature selection, multiclass detection and diagnosis models, and anomaly detection algorithms will be presented.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Failure diagnosis and trend‐based performance losses routines for the detection and classification of incidents in large‐scale photovoltaic systems

Abstract Fault detection and classification in photovoltaic (PV) systems through real‐time monitoring is a fundamental task that ensures quality of operation and significantly improves the performance and reliability of operating systems. Different statistical and comparative approaches have already been proposed in the literature for fault detection; however, accurate classification of fault and loss incidents based on PV performance time series remains a key challenge. Failure diagnosis and trend‐based performance loss routines were developed in this work for detecting PV underperformance and accurately identifying the different fault types and loss mechanisms. The proposed routines focus mainly on the differentiation of failures (e.g., inverter faults) from irreversible (e.g., degradation) and reversible (e.g., snow and soiling) performance loss factors based on statistical analysis. The proposed routines were benchmarked using historical inverter data obtained from a 1.8 MWp PV power plant. The results demonstrated the effectiveness of the routines for detecting failures and loss mechanisms and the capability of the pipeline for distinguishing underperformance issues using anomaly detection and change‐point (CP) models. Finally, a CP model was used to extract significant changes in time series data, to detect soiling and cleaning events and to estimate both the performance loss and degradation rates of fielded PV systems.

14 SOLAR ENERGY↗

Modeling inter-signal arrival times for accurate detection of CAN bus signal injection attacks

Modern vehicles rely on hundreds of on-board electronic control units (ECUs) communicating over in-vehicle networks. As external interfaces to the car control networks (such as the on-board diagnostic (OBD) port, auxiliary media ports, etc.) become common, and vehicle-to-vehicle / vehicle-to-infrastructure technology is in the near future, the attack surface for vehicles grows, exposing control networks to potentially life-critical attacks. This paper addresses the need for securing the CAN bus by detecting anomalous traffic patterns via unusual refresh rates of certain commands. While previous works have identified signal frequency as an important feature for CAN bus intrusion detection, this paper provides the first such algorithm with experiments on five attack scenarios. Our data-driven anomaly detection algorithm requires only five seconds of training time (on normal data) and achieves true positive / false discovery rates of 0.9998/0.00298, respectively (micro-averaged across the five experimental tests).

anomaly detection↗

Using bursty announcements for detecting BGP routing anomalies

Despite the robust structure of the Internet, it is still susceptible to disruptive routing updates that prevent network traffic from reaching its destination. Our research shows that BGP announcements that are associated with disruptive updates tend to occur in groups of relatively high frequency, followed by periods of infrequent activity. We hypothesize that we may use these bursty characteristics to detect anomalous routing incidents. In this work, we use manually verified ground truth metadata and volume of announcements as a baseline measure, and propose a burstiness measure that detects prior anomalous incidents with high recall and better precision than the volume baseline. We quantify the burstiness of inter-arrival times around the date and times of four large-scale incidents: the Indosat hijacking event in April 2014, the Telecom Malaysia leak in June 2015, the Bharti Airtel Ltd. hijack in November 2015, and the MainOne leak in November 2018; and three smaller scale incidents that led to traffic interception: the Belarusian traffic direction in February 2013, the Icelandic traffic direction in July 2013, and the Russian telecom that hijacked financial services in April 2017. Our method leverages the burstiness of disruptive update messages to detect these incidents. We describe limitations, open challenges, and how this method can be used for routing anomaly detection.

97 MATHEMATICS AND COMPUTING↗

Network Intrusion Detection and Visualization using Aggregations in a Cyber Security Data Warehouse

The challenge of achieving situational understanding is a limiting factor in effective, timely, and adaptive cyber-security analysis. Anomaly detection fills a critical role in network assessment and trend analysis, both of which underlie the establishment of comprehensive situational understanding. To that end, we propose a cyber security data warehouse implemented as a hierarchical graph of aggregations that captures anomalies at multiple scales. Each node of our pro-posed graph is a summarization table of cyber event aggregations, and the edges are aggregation operators. The cyber security data warehouse enables domain experts to quickly traverse a multi-scale aggregation space systematically. We describe the architecture of a test bed system and a summary of results on the IEEE VAST 2012 Cyber Forensics data.

anomaly detection↗

Towards deep computer vision for in-line defect detection in polymer electrolyte membrane fuel cell materials

Polymer Electrolyte Membrane (PEM) fuel cells are a promising source of alternative energy. However, their production is limited by a lack of well-established methods for quality control of their constituent materials like the membrane-electrode assembly during roll-to-roll manufacturing. One potential solution is the implementation of deep learning methods to detect unwanted defects through their detection in scanned images. Here we explore the detection of defects like scratches, pinholes, and scuffs in a sample dataset of PEM optical images using two deep learning algorithms: Patch Distribution Modeling (PaDiM) for unsupervised anomaly detection and Faster-RCNN for supervised object detection. Both methods achieve scores on performance metrics (ROC-AUC and PRO-AUC for PaDiM and AP for Faster-RCNN) that are comparable to their scores on benchmark datasets. These methods also have the potential to detect a wider range of defects compared to IR thermography and previous optical inspection methods. Overall, deep learning shows promise at detecting relevant defects of interest and has the potential to achieve real-time defect detection.

30 DIRECT ENERGY CONVERSION↗

Data Center Facility Monitoring with Physics Aware Approach

U.S. Department of Energy's National Renewable Energy Laboratory (NREL) hosts one of the world's most energy-efficient HPC data centers; this system uses component-level warm-water liquid cooling to efficiently remove heat from the data center and capture it for reuse in the building or rejection to the atmosphere. Given the complexity of this system, building data-driven tools for holistically monitoring and operating the entire data center is a priority for ensuring maximal efficiency and resiliency. In this advanced smart facility, over one million metrics are recorded per minute using state-of-the-art streaming data architecture and software to capture and process the state of the system in real time. Here we detail two efforts to effectively analyze, visualize, and interpret this large volume streaming data. We have developed a novel, flexible system for identifying and visualizing individual metric anomalies and component performance across the data center through automatic metadata extraction and physically-motivated visualization for quick interpretation. Additionally, to directly connect system maintenance to data stream processing we explore a physics informed multi-metric drift and anomaly detection application to detect scale-build up in heat exchangers.

anomaly detection↗

Spatio-temporal multivariate cluster evolution analysis for detecting and tracking climate impacts

Recent years have seen a growing concern about climate change and its impacts. While Earth System Models (ESMs) can be invaluable tools for studying the impacts of climate change, the complex coupling processes encoded in ESMs and the large amounts of data produced by these models, together with the high internal variability of the Earth system, can obscure important source-to-impact relationships. Here, this paper presents a novel and efficient unsupervised data-driven approach for detecting statistically-significant impacts and tracing spatio-temporal source-impact pathways in the climate through a unique combination of ideas from anomaly detection, clustering and Natural Language Processing (NLP). Using as an exemplar the 1991 eruption of Mount Pinatubo in the Philippines, we demonstrate that the proposed approach is capable of detecting known post-eruption impacts/events. We additionally describe a methodology for extracting meaningful sequences of post-eruption impacts/events by using NLP to efficiently mine frequent multivariate cluster evolutions, which can be used to confirm or discover the chain of physical processes between a climate source and its impact(s).

Anomaly detection↗

Chimbuko: A Workflow-Level Scalable Performance Trace Analysis Tool

Due to the sheer volume of data it is typically impractical to analyze the detailed performance of an HPC application running at-scale. While conventional small-scale benchmarking and scaling studies are often sufficient for simple applications, many modern workflow-based applications couple multiple elements with competing resource demands and complex inter-communication patterns for which performance cannot easily be studied in isolation and at small scale. This work discusses Chimbuko, a performance analysis framework that provides real-time, in situ anomaly detection. By focusing specifically on performance anomalies and their origin (aka provenance), data volumes are dramatically reduced without losing necessary details. To the best of our knowledge, Chimbuko is the first online, distributed, and scalable workflow-level performance trace analysis framework. We demonstrate the tool's usefulness on Oak Ridge National Laboratory's Summit system.

97 MATHEMATICS AND COMPUTING↗

EV SALaD 2023 Demonstration: Best Practices and Mitigations for Protecting EVSE Infrastructure

The Electric Vehicle Secure Architecture Laboratory Demonstration (EV SALaD) program is a demonstration of cybersecurity best practices for high-power electric vehicle (EV) charging infrastructure led by Idaho National Laboratory (INL), in collaboration with other DOE National Laboratories participating in the EVs at Scale Consortium.a Sandia National Laboratories (SNL) and Pacific Northwest National Laboratory (PNNL) participated in the first 2-year (FY22-23) demonstration cycle for EV SALaD. This report documents the FY23 demonstration, the second in a series of demonstrations and collaborations in deploying and operating cybersecure EV charging infrastructure. It includes a summary of improvements from the FY22 demonstration, technical analysis of the FY23 demonstration, how the research demonstrates cyber-physical and cybersecurity best practices for high-power EV charging infrastructure, and related impacts to national and energy security. For EV SALaD, the FY22 demonstration focused on the detection, ranking, and prioritization of anomalous events for high-power EV charging. The FY23 demonstration additionally included the demonstration of cybersecurity best practices, which included protection and mitigation solutions to prevent, respond, and recover from anomalous events. During the demonstrations, the multi-lab EV SALaD team conducted a Test Effect Payload (TEP)b evaluation on extreme fast charger (XFC) hardware equipped with Cerberus, a detection and response solution, to demonstrate anomaly detection and mitigation cybersecurity best practices against cyber-enabled events.

33 ADVANCED PROPULSION SYSTEMS↗

Reliable detection of fluence anomalies in EPID-based IMRT pretreatment quality assurance using pixel intensity deviations

Purpose: This work uses repeat images of intensity modulated radiation therapy (IMRT) fields to quantify fluence anomalies (i.e., delivery errors) that can be reliably detected in electronic portal images used for IMRT pretreatment quality assurance. Methods: Repeat images of 11 clinical IMRT fields are acquired on a Varian Trilogy linear accelerator at energies of 6 MV and 18 MV. Acquired images are corrected for output variations and registered to minimize the impact of linear accelerator and electronic portal imaging device (EPID) positioning deviations. Detection studies are performed in which rectangular anomalies of various sizes are inserted into the images. The performance of detection strategies based on pixel intensity deviations (PIDs) and gamma indices is evaluated using receiver operating characteristic analysis. Results: Residual differences between registered images are due to interfraction positional deviations of jaws and multileaf collimator leaves, plus imager noise. Positional deviations produce large intensity differences that degrade anomaly detection. Gradient effects are suppressed in PIDs using gradient scaling. Background noise is suppressed using median filtering. In the majority of images, PID-based detection strategies can reliably detect fluence anomalies of {>=}5% in {approx}1 mm{sup 2} areas and {>=}2% in {approx}20 mm{sup 2} areas. Conclusions: The ability to detect small dose differences ({<=}2%) depends strongly on the level of background noise. This in turn depends on the accuracy of image registration, the quality of the reference image, and field properties. The longer term aim of this work is to develop accurate and reliable methods of detecting IMRT delivery errors and variations. The ability to resolve small anomalies will allow the accuracy of advanced treatment techniques, such as image guided, adaptive, and arc therapies, to be quantified.

60 APPLIED LIFE SCIENCES↗

Application of Data Cubes for Improving Detection of Water Cycle Extreme Events

As part of an ongoing NASA-funded project to remove a longstanding barrier to accessing NASA data (i.e., accessing archived time-step array data as point-time series), for the hydrology and other point-time series-oriented communities, "data cubes" are created from which time series files (aka "data rods") are generated on-the-fly and made available as Web services from the Goddard Earth Sciences Data and Information Services Center (GES DISC). Data cubes are data as archived rearranged into spatio-temporal matrices, which allow for easy access to the data, both spatially and temporally. A data cube is a specific case of the general optimal strategy of reorganizing data to match the desired means of access. The gain from such reorganization is greater the larger the data set. As a use case of our project, we are leveraging existing software to explore the application of the data cubes concept to machine learning, for the purpose of detecting water cycle extreme events, a specific case of anomaly detection, requiring time series data. We investigate the use of support vector machines (SVM) for anomaly classification. We show an example of detection of water cycle extreme events, using data from the Tropical Rainfall Measuring Mission (TRMM).

water cycle extreme events↗