SEARCH · Engineering Papers

Results for “error 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.

SourceResearch connectionFrom yearThrough year

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

At least 1,585 records · Page 88

All sources NASA NTRS DOE OSTI

NASA NTRS2005

Adaptive Numerical Dissipation Control in High Order Schemes for Multi-D Non-Ideal MHD

The required type and amount of numerical dissipation/filter to accurately resolve all relevant multiscales of complex MHD unsteady high-speed shock/shear/turbulence/combustion problems are not only physical problem dependent, but also vary from one flow region to another. In addition, proper and efficient control of the divergence of the magnetic field (Div(B)) numerical error for high order shock-capturing methods poses extra requirements for the considered type of CPU intensive computations. The goal is to extend our adaptive numerical dissipation control in high order filter schemes and our new divergence-free methods for ideal MHD to non-ideal MHD that include viscosity and resistivity. The key idea consists of automatic detection of different flow features as distinct sensors to signal the appropriate type and amount of numerical dissipation/filter where needed and leave the rest of the region free from numerical dissipation contamination. These scheme-independent detectors are capable of distinguishing shocks/shears, flame sheets, turbulent fluctuations and spurious high-frequency oscillations. The detection algorithm is based on an artificial compression method (ACM) (for shocks/shears), and redundant multiresolution wavelets (WAV) (for the above types of flow feature). These filters also provide a natural and efficient way for the minimization of Div(B) numerical error.

Yee, H. C.↗

NASA NTRS2010

The Second Swift BAT Gamma-Ray Burst Catalog

We present the second Swift Burst Alert Telescope (BAT) catalog of gamma-ray bursts (GRBs), which contains 476 bursts detected by the BAT between 2004 December 19 and 2009 December 21. This catalog (hereafter the BAT2 catalog) presents burst trigger time, location, 90% error radius, duration, fluence, peak flux, time-averaged spectral parameters and time-resolved spectral parametert:; measured by the BAT. In the correlation study of various observed parameters extracted from the BAT prompt emission data, we distinguish among long-duration GRBs (L-GRBs), short-duration GRBs (S-GRBs), and short-duration GRBs with extended emission (S-GRBs with E.E.) to investigate differences in the prompt emission properties. The fraction of L-GRBs, S-GRBs and S-GRBs with E.E. in the catalog are 89%, 8% and 2% respectively. We compare the BAT prompt emission properties with the BATSE, BeppoSAX and HETE-2 GRB samples. We also correlate the observed prompt emission properties with the redshifts for the GRBs with known redshift. The BAT T90 and T50 durations peak at 70 s and 30 s, respectively. We confirm that the spectra of the BAT S-GRBs are generally harder than those of the L-GRBs. The time-averaged spectra of the BAT S GRBs with E.E. are similar to those of the L-GRBs. Whereas, the spectra of the initial short spikes of the S-GRBs with E.E. are similar to those of the S-GRBs. We show that the BAT GRB samples are significantly softer than the BATSE bright GRBs, and that the time-averaged E obs/peak of the BAT GRBs peaks at 80 keV which is significantly lower energy than those of the BATSE sample which peak at 320 keV. The time-averaged spectral properties of the BAT GRB sample are similar to those of the HETE-2 GRB samples. By time-resolved spectral analysis, we find that 10% of the BAT observed photon indices are outside the allowed region of the synchrotron shock model. The observed durations of the BAT high redshift GRBs are not systematically longer than those of the moderate red shift GRBs. Furthermore, the observed spectra of the BAT high red shift GRBs are similar to or harder than the moderate red shift GRBs. The T90 and T50 distributions measured at the 140-220 keY band in the GRB rest frame form the BAT known redshift GRBs peak at 19 sand 8 s, respectively. We also provide an update on the status of the on-orbit BAT calibrations.

Barthelmy, S. D.↗

NASA NTRS2011

Enhanced Fuel-Optimal Trajectory-Generation Algorithm for Planetary Pinpoint Landing

An enhanced algorithm is developed that builds on a previous innovation of fuel-optimal powered-descent guidance (PDG) for planetary pinpoint landing. The PDG problem is to compute constrained, fuel-optimal trajectories to land a craft at a prescribed target on a planetary surface, starting from a parachute cut-off point and using a throttleable descent engine. The previous innovation showed the minimal-fuel PDG problem can be posed as a convex optimization problem, in particular, as a Second-Order Cone Program, which can be solved to global optimality with deterministic convergence properties, and hence is a candidate for onboard implementation. To increase the speed and robustness of this convex PDG algorithm for possible onboard implementation, the following enhancements are incorporated: 1) Fast detection of infeasibility (i.e., control authority is not sufficient for soft-landing) for subsequent fault response. 2) The use of a piecewise-linear control parameterization, providing smooth solution trajectories and increasing computational efficiency. 3) An enhanced line-search algorithm for optimal time-of-flight, providing quicker convergence and bounding the number of path-planning iterations needed. 4) An additional constraint that analytically guarantees inter-sample satisfaction of glide-slope and non-sub-surface flight constraints, allowing larger discretizations and, hence, faster optimization. 5) Explicit incorporation of Mars rotation rate into the trajectory computation for improved targeting accuracy. These enhancements allow faster convergence to the fuel-optimal solution and, more importantly, remove the need for a "human-in-the-loop," as constraints will be satisfied over the entire path-planning interval independent of step-size (as opposed to just at the discrete time points) and infeasible initial conditions are immediately detected. Finally, while the PDG stage is typically only a few minutes, ignoring the rotation rate of Mars can introduce 10s of meters of error. By incorporating it, the enhanced PDG algorithm becomes capable of pinpoint targeting.

Acikmese, Behcet↗

NASA NTRS2012

The Second SWIFT Burst Alert Telescope (BAT) Gamma-Ray Burst Catalog

We present the second Swift Burst Alert Telescope (BAT) catalog of gamma-ray bursts. (GRBs), which contains 476 bursts detected by the BAT between 2004 December 19 and 2009 December 21. This catalog (hereafter the BAT2 catalog) presents burst trigger time, location, 90% error radius, duration, fluence, peak flux, time-averaged spectral parameters and time-resolved spectral parameters measured by the BAT. In the correlation study of various observed parameters extracted from the BAT prompt emission data, we distinguish among long-duration GRBs (L-GRBs), short-duration GRBs (S-GRBs), and short-duration GRBs with extended emission (S-GRBs with E.E.) to investigate differences in the prompt emission properties. The fraction of L-GRBs, S-GRBs and S-GRBs with E.E. in the catalog are 89%, 8% and 2% respectively. We compare the BAT prompt emission properties with the BATSE, BeppoSAX and HETE-2 GRB samples.. We also correlate the observed prompt emission properties with the redshifts for the GRBs with known redshift. The BAT T(sub 90) and T(sub 50) durations peak at 70 s and 30 s, respectively. We confirm that the spectra of the BAT S-GRBs are generally harder than those of the L-GRBs.

Sakamoto, T.↗

NASA NTRS2010

Mapping Successional Stages in a Wet Tropical Forest Using Landsat ETM+ and Forest Inventory Data

In this study, we test whether an existing classification technique based on the integration of Landsat ETM+ and forest inventory data enables detailed characterization of successional stages in a wet tropical forest site. The specific objectives were: (1) to map forest age classes across the La Selva Biological Station in Costa Rica; and (2) to quantify uncertainties in the proposed approach in relation to field data and existing vegetation maps. Although significant relationships between vegetation height entropy (a surrogate for forest age) and ETM+ data were detected, the classification scheme tested in this study was not suitable for characterizing spatial variation in age at La Selva, as evidenced by the error matrix and the low Kappa coefficient (12.9%). Factors affecting the performance of the classification at this particular study site include the smooth transition in vegetation structure between intermediate and advanced successional stages, and the low sensitivity of NDVI to variations in vertical structure at high biomass levels.

forest structure↗

NASA NTRS2001

Cloud Overlapping Detection Algorithm Using Solar and IR Wavelengths with GOES Data Over ARM/SGP Site

One of the most perplexing problems in satellite cloud remote sensing is the overlapping of cloud layers. Although most techniques assume a one layer cloud system in a given retrieval of cloud properties, many observations are affected by radiation from more than one cloud layer. As such, cloud overlap can cause errors in the retrieval of many properties including cloud height, optical depth, phase, and particle size. A variety of methods have been developed to identify overlapped clouds in a given satellite imager pixel. Baum et al used CO2 slicing and a spatial coherence method to demonstrate a possible analysis method for nighttime detection of multilayered clouds. Jin and Rossow also used a multispectral CO2 slicing technique for a global analysis of overlapped cloud amount. Lin et al. used a combination infrared (IR), visible (VIS), and microwave data to detect overlapped clouds over water. Recently, Baum and Spinhirne proposed a 1.6 and 11 micron bispectral threshold method. While all of these methods have made progress in solving this stubborn problem none have yet proven satisfactory for continuous and consistent monitoring of multilayer cloud systems. It is clear that detection of overlapping clouds from passive instruments such as satellite radiometers is in an immature stage of development and requires additional research. Overlapped cloud systems also affect the retrievals of cloud properties over the Atmospheric Radiation Measurement (ARM) domains and hence should be identified as accurately as possible. To reach this goal, it is necessary to determine which information can be exploited for detecting multilayered clouds from operational meteorological satellite data used by ARM. This paper examines the potential information available in spectral data available on the Geostationary Operational Environmental Satellite (GOES) imager and the National Oceanic Atmospheric Administration (NOAA) Advanced Very High Resolution Radiometer (AVHRR) used over the ARM Program's Southern Great Plains (SGP), and North Slope of Alaska (NSA) sites to study the capability of detecting overlapping clouds.

Kawamoto, K.↗

DOE OSTI2021

Geostatistical interpolation of streambed hydrologic attributes with addition of left censored data and anisotropy

Spatial geostatistical interpolation of point measurements of streambed attributes in the hyporheic zone may be constrained by the streambed anisotropy, and data density and spatial distribution may significantly impact the results. Spatial clustering and low spatial data density can be caused by bedrock outcropping at the streambed limiting installation of in-stream piezometers. This study examines parameter error variability of the geostatistical interpolation using anisotropic interpolation methods and increasing the data density by adding left censored values (i.e., data below measurement limit) to locations where measurements were limited by exposed bedrock lining the streambed. The reduction in relative standard error of the interpolation was determined for the spatial distributions of streambed attributes including hydraulic conductivity, seepage flux, and mercury solute flux measured in two different years along a study reach in East Fork Poplar Creek, Tennessee, USA. Here, two methods to impute the left censored values were compared including the conventional half the detection limit substitution method, and the Stochastic Approximation of Expectation-Maximization (SAEM) algorithm, which both had comparable results. Imputing left censored data increased the data density to recommended ranges, reduced data clustering, increased the spatial dependence for some attributes, and reduced the standard error for each of the three attributes. For the reach considered herein, addition of the left censored values resulted in a larger error reduction than the consideration of anisotropy within the interpolation, which confirms the benefit of data addition to increase data density within data-limited river corridors.

58 GEOSCIENCES↗

NASA NTRS2001

Cloud Overlapping Detection Algorithm Using Solar and IR Wavelengths With GOSE Data Over ARM/SGP Site

One of the most perplexing problems in satellite cloud remote sensing is the overlapping of cloud layers. Although most techniques assume a 1-layer cloud system in a given retrieval of cloud properties, many observations are affected by radiation from more than one cloud layer. As such, cloud overlap can cause errors in the retrieval of many properties including cloud height, optical depth, phase, and particle size. A variety of methods have been developed to identify overlapped clouds in a given satellite imager pixel. Baum el al. (1995) used CO2 slicing and a spatial coherence method to demonstrate a possible analysis method for nighttime detection of multilayered clouds. Jin and Rossow (1997) also used a multispectral CO2 slicing technique for a global analysis of overlapped cloud amount. Lin et al. (1999) used a combination infrared, visible, and microwave data to detect overlapped clouds over water. Recently, Baum and Spinhirne (2000) proposed 1.6 and 11 microns. bispectral threshold method. While all of these methods have made progress in solving this stubborn problem, none have yet proven satisfactory for continuous and consistent monitoring of multilayer cloud systems. It is clear that detection of overlapping clouds from passive instruments such as satellite radiometers is in an immature stage of development and requires additional research. Overlapped cloud systems also affect the retrievals of cloud properties over the ARM domains (e.g., Minnis et al 1998) and hence should identified as accurately as possible. To reach this goal, it is necessary to determine which information can be exploited for detecting multilayered clouds from operational meteorological satellite data used by ARM. This paper examines the potential information available in spectral data available on the Geostationary Operational Environmental Satellite (GOES) imager and the NOAA Advanced Very High Resolution Radiometer (AVHRR) used over the ARM SGP and NSA sites to study the capability of detecting overlapping clouds

Kawamoto, Kazuaki↗

NASA NTRS2022

Building Lunar Maps for Terrain Relative Navigation and Hazard Detection Applications

Terrain Relative Navigation (TRN) systems that localize a spacecraft with respect to a map of the surface by comparing descent imagery to that reference map can only be as accurate as the reference map itself. Accurate map products that are based on orbital reconnaissance data must be validated for navigation applications to ensure that all relevant error sources are minimized. Currently available map products have been generated for scientific applications, so the need for accurate TRN maps remains a gap to be filled for upcoming lunar lander missions, in particular missions to the South Pole region. Additionally, representative high-resolution maps that contain lander-scale features are needed for successful development and testing of Hazard Detection (HD) systems. This paper describes one of NASA’s current efforts to develop benchmark data sets that can be used for developing and testing TRN and HD algorithms as well as suggested processes and metrics for generating and validating lunar maps that can be used for navigation and hazard detection.

Beyer, Ross A.↗

NASA NTRS2014

Erratum: Wise Detections of Known Qsos at Redshifts Greater than Six

In the published version of this paper, Roberto Assef was mistakenly affiliated with the Division of Astronomy and Astrophysics at the University of California, Los Angeles. This is incorrect. Dr. Assef's affiliation correctly appears in this erratum as the Nucleo de Astronomıa de la Facultad de Ingenierıa, Universidad Diego Portales, Av. Ejercito 441, Santiago, Chile. IOP Publishing sincerely regrets this error.

Wise↗

NASA NTRS1997

Using Laser Altimetry to Detect Topographic Change in Long Valley Caldera, California

Long Valley caldera California, is a site of extensive volcanism, persistent seismicity, and uplift of a resurgent dome, currently at a rate of about 3 cm/year. Airborne laser altimetry was used to determine the surface topography of the region in 1993. A repeat mission occurred in 1995. Three different laser altimeters were flown, dubbed ATLAS, SLICER and RASCAL. Data processing consists of the combination of the aircraft trajectory and attitude data with the laser range, the determination of an atmospheric delay, laser pulse timing errors, laser system biases, and data geolocation to obtain the position of the laser spot on the ground. Results showed that using the ATLAS and SLICER instruments, the elevation of an overflown lake is determined to precisions of 3.3 cm and 2.9 cm from altitudes of 500 m and 3 km above the ground, and about 10 cm using the RASCAL instrument from 500 m above ground. Comparison with tide gauge data showed the laser measurements are able to resolve centimeter-level changes in the lake elevation over time. Repeat pass analysis of tracks over flat surfaces indicate no systematic biases affect the measurement procedure of the ATLAS and SLICER instruments. Comparison of GPS and laser-derived elevations of easily-identifiable features in the caldera confirm the horizontal accuracy of the measurement is within the diameter of the laser footprint, and vertical accuracy is within the error inherent in the measurement. Crossover analysis shows that the standard error of the means at track intersection points within the caldera, and dome (i.e., where zero and close to the maximum amount of uplift is expected) are about I cm, indicating elevation change at the 3 cm/year level should be detectable. We demonstrate one of the powerful advantages of scanning laser altimetry over other remote sensing techniques; the straightforward creation of precise digital elevation maps of overflown terrain. Initial comparison of the 1993-1995 data indicates uplift occurred, but filtering is required to remove vegetation effects. Although research continues to utilize the full potential of laser altimetry data, the results constitute a successful demonstration that the technique may be used to perform geodetic monitoring of surface topographic change.

Hofton, M. A.↗

NASA NTRS1997

Using Laser Altimetry to Detect Topographic Change at Long Valley Caldera, California

Long Valley caldera, California, is a site of extensive volcanism, persistent seismicity, and uplift of a resurgent dome, currently at a rate of approximately 3 cm/year. Airborne laser altimetry was used to determine the surface topography of the region in 1993. A repeat mission occurred in 1995. Three different laser altimeters were flown, dubbed ATLAS, SLICER and RASCAL. Data processing consists of the combination of the aircraft trajectory and attitude data with the laser range, the determination of an atmospheric delay, laser pulse timing errors, laser system biases, and data geolocation to obtain the position of the laser spot on the ground. Results showed that using the ATLAS and SLICER instruments, the elevation of an overflown lake is determined to precisions of 3.3 cm and 2.9 cm from altitudes of 500 m and 3 km above the ground, and approximately 10 cm using the RASCAL instrument from 500 m above ground. Comparison with tide gauge data showed the laser measurements are able to resolve centimeter-level changes in the lake elevation over time. Repeat pass analysis of tracks over flat surfaces indicate no systematic biases affect the measurement procedure of the ATLAS and SLICER instruments. Comparison of GPS and laser-derived elevations of easily-identifiable features in the caldera confirm the horizontal accuracy of the measurement is within the diameter of the laser footprint, and vertical accuracy is within the error inherent in the measurement. Crossover analysis shows that the standard error of the means at track intersection points within the caldera and dome (i.e., where zero and close to the maximum amount of uplift is expected) are about 1 cm, indicating elevation change at the 3 cm/year level should be detectable. We demonstrate one of the powerful advantages of scanning laser altimetry over other remote sensing techniques; the straightforward creation of precise digital elevation maps of overflown terrain. Initial comparison of the 1993-1995 data indicates uplift occurred, but filtering is required to remove vegetation effects. Although research continues to utilize the full potential of laser altimetry data, the results constitute a successful demonstration that the technique may be used to perform geodetic monitoring of surface topographic change.

Hofton, M. A.↗

DOE OSTI2025

Sensitive detection of structural dynamics using a statistical framework for comparative crystallography

Chemical and conformational changes are crucial to protein function and its pharmacological control. X-ray crystallography can reveal these changes in atomic detail, but standard analysis methods, which refine separate datasets, often overlook differences that are subtle or arise in only a subset of molecules. Direct comparison of crystallographic datasets is, in principle, more powerful, but systematic errors (“scales”) often mask changes in the crystallographic observables (“structure factors”). Machine learning algorithms that jointly estimate scales and structure factors can address this limitation. Here, we augment this approach with multivariate, structured priors derived from crystallographic theory, implemented in the variational deep learning framework Careless. Doing so strongly improves the detection of protein dynamics, element-specific anomalous signals, and the binding of drug candidates, offering a robust approach to comparative crystallography and, potentially, to detection of protein dynamics by other structure determination methods.

Hekstra, Doeke R. [Harvard Univ., Cambridge, MA (U↗

NASA NTRSDate not supplied

Intelligent Change Detection System: Autonomous Intelligent Machine Agent Model Development

NASA’s significant role in facilitating the harmonious integration of unmanned aircraft systems (UAS), with other aerial vehicles operating in the National Airspace System (NAS), has revealed a need for more advanced technological tools than are being utilized currently. This technology would lend itself to significantly assisting Direct-Action Aviation Personnel (DAAP) with the ingress and egress of UAS operations within the NAS. Providing research findings that would reduce technical barriers, associated with UAS-NAS integration, has been a persistent effort by both NASA and the FAA. One such research effort, pursued by NASA’s Transformative Tools and Technologies – Revolutionary Aviation Mobility (T3-RAM) project, is the development of an autonomous intelligent machine (AIM) agent that would aid DAAP functioning as implemented in remote ground control stations (RGCS). This evolution of the “human-machine” symbiosis, within the aviation environment, is necessary for many reasons. For example, there are projections of large increases to the 864,000 registered UAS and 45,000 aviation operations taking place in the NAS each day. With a data output range from 1 to 20 terabytes per flight or each aerial vehicle, which is projected to have proportional rate increase to that of registered UASs. It is evident, that due to the projected increase of UASs and their generated data, the human-agent’s data managing capabilities will be quickly overwhelmed by the enormous amounts of data emanating in the NAS. The research efforts presented in this paper puts forward results from the development, assessment, and verification of a previously conceptualized AIM-Agent that combats actionable-data (information) errors resulting from the visual perception phenomenon known as “change blindness” (CB). CB has been identified as one of the main culprits of information erroring encountered within the ground control station operator (GCSO) community.

Change Blindness↗

DOE OSTI2023

High-dimensional isotomics, part 2: Observations of over 100 constraints on methionine's isotome

The abundances of different isotopic forms of a compound, or isotopologues, will vary based on its physical and chemical history. The number of isotopologues increases combinatorically with the size of a molecule, and even small molecules such as amino acids have thousands of potentially observable isotopic variants. However, due to the analytical challenges of separating and observing isotopologues, only a few dimensions of isotopic diversity are routinely measured. Overcoming these challenges requires both an experimental method to observe many isotopic properties and a theoretical framework for interpreting these experiments. In Part 1, we presented such a theoretical framework; here, we demonstrate an experimental method, which we apply to methionine. Our approach uses a Q Exactive HF Orbitrap to perform several “M + N experiments”, where a sample is ionized, a subset of its isotopologues with cardinal mass N daltons greater than the unsubstituted isotopologue is selected and fragmented, and the proportions of all detectable isotopic forms of those fragment ions are quantified. We perform M + 1, M + 2, M + 3, and M + 4 experiments of a methionine sample and standard where the sample has a 100 ‰ enrichment of 13C at the methyl carbon relative to the natural 13C abundance at that position in the standard, and is otherwise identical to the standard. We observe isotopic forms of 8 fragment ion species for each version of the M + N experiment. With the assistance of a forward model of expected mass spectra, we identify isotopic peaks for each fragment ion based on observed mass and abundance, screen these for data quality, and quantify abundances for 146 unique isotopic peaks at precisions of ≈ 0.3–3 ‰. We present our direct observations and use them to reconstruct the concentrations of 19 individual singly, doubly, and triply-substituted isotopologues; doing so gives fewer constraints and broader error bars than working with the direct observations, but may be more interpretable for some applications. We also examine possibilities for measuring additional peaks, which are primarily limited by the detection limit of the Orbitrap-IRMS method. We then suggest some possible uses of our direct measurements for chemical forensics and hypothesis testing. Furthermore, our results demonstrate the diversity of isotopic constraints currently observable and interpretable for organic molecules.

58 GEOSCIENCES↗

DOE OSTI2024

EHR-BERT: A BERT-based model for effective anomaly detection in electronic health records

Objective: Physicians and clinicians rely on data contained in electronic health records (EHRs), as recorded by health information technology (HIT), to make informed decisions about their patients. The reliability of HIT systems in this regard is critical to patient safety. Consequently, better tools are needed to monitor the performance of HIT systems for potential hazards that could compromise the collected EHRs, which in turn could affect patient safety. In this paper, we propose a new framework for detecting anomalies in EHRs using sequence of clinical events. This new framework, EHR-Bidirectional Encoder Representations from Transformers (BERT), is motivated by the gaps in the existing deep-learning related methods, including high false negatives, sub-optimal accuracy, higher computational cost, and the risk of information loss. EHR-BERT is an innovative framework rooted in the BERT architecture, meticulously tailored to navigate the hurdles in the contemporary BERT method; thus, enhancing anomaly detection in EHRs for healthcare applications.Methods: The EHR-BERT framework was designed using the Sequential Masked Token Prediction (SMTP) method. This approach treats EHRs as natural language sentences and iteratively masks input tokens during both training and prediction stages. This method facilitates the learning of EHR sequence patterns in both directions for each event and identifies anomalies based on deviations from the normal execution models trained on EHR sequences.Results: Extensive experiments on large EHR datasets across various medical domains demonstrate that EHR-BERT markedly improves upon existing models. It significantly reduces the number of false positives and enhances the detection rate, thus bolstering the reliability of anomaly detection in electronic health records. This improvement is attributed to the model’s ability to minimize information loss and maximize data utilization effectively.Conclusion: EHR-BERT showcases immense potential in decreasing medical errors related to anomalous clinical events, positioning itself as an indispensable asset for enhancing patient safety and the overall standard of healthcare services. The framework effectively overcomes the drawbacks of earlier models, making it a promising solution for healthcare professionals to ensure the reliability and quality of health data.

96 KNOWLEDGE MANAGEMENT AND PRESERVATION↗

DOE OSTI2020

Perturbation-Based Diagnosis of False Data Injection Attack Using Distributed Energy Resources

Modern smart grid relies on various sensor measurements for its operational control. In a successful false data injection attack, the attacker manipulates the measurements from the grid sensors such that undetected errors are introduced into the estimates of the system parameters leading to catastrophic situations. This paper proposes a novel perturbation based false data injection attack detection mechanism that utilizes inverter based distributed energy resources (DERs) to create low magnitude perturbation signal in the distribution system voltage that is inconsequential to the normal grid operation. Two voltage sensitivity analysis based algorithms are designed to identify the optimal set of DERs that can create the voltage perturbation signal of desired magnitude. An analytical method of voltage sensitivity analysis is used to compute the magnitude of voltage perturbation signal at each node in a computationally efficient manner. Then, a detection mechanism is developed that checks for the presence of the perturbation sequence in each sensor measurement. A sensor measurement is deemed authentic if the voltage perturbation signal is present in the data. In case of sensor malfunction or cyber-attack, the perturbation signal will not be present in the measurement data. Performance of the proposed attack detection mechanism is validated via simulation of the IEEE 69 bus test system.

24 POWER TRANSMISSION AND DISTRIBUTION↗

DOE OSTI2021

CrossCheck: Rapid, Reproducible, and Interpretable Model Evaluation

Evaluation beyond aggregate performance metrics, e.g. F1-score, is crucial to both establish an appropriate level of trust in machine learning models and identify future model improvements. In this paper we demonstrate CrossCheck, an interactive visualization tool for rapid crossmodel comparison and reproducible error analysis. We describe the tool and discuss design and implementation details. We then present three use cases (named entity recognition, reading comprehension, and clickbait detection) that show the benefits of using the tool for model evaluation. CrossCheck allows data scientists to make informed decisions to choose between multiple models, identify when the models are correct and for which examples, investigate whether the models are making the same mistakes as humans, evaluate models’ generalizability and highlight models’ limitations, strengths and weaknesses. Furthermore, CrossCheck is implemented as a Jupyter widget, which allows rapid and convenient integration into data scientists’ model development workflows.

Arendt, Dustin L.↗