Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “standard work specifications”

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 253 records · Page 14

Evaluating the sustainability of the 2017 US biofuel industry with an integrated techno-economic analysis and life cycle assessment

The United States produced more than 66.2 billion Liters of biofuel, primarily in the form of corn ethanol (61 billion Liters) and soybean biodiesel (6 billion Liters), for the transportation industry in 2017. No work currently exists that evaluates both the economic and environmental outcomes of this industry with adequate geospatial resolution and national scope. In this study, a model framework is constructed to perform Techno-Economic Analysis and Life Cycle Assessment using high-resolution input datasets to provide a granular estimation of corn ethanol and soybean biodiesel sustainability performance in most counties in the United States in 2017 (the most recent year for which data is available). Results show a total life cycle, greenhouse gas emission volume of 99.5 million Mg CO 2-eq produced in 2017 by these two biofuels (94.8 million Mg from corn ethanol, and 4.7 million Mg from soybean biodiesel), excluding indirect land use change emissions. Further, this volume corresponds to a combined carbon intensity of 65.3 gCO 2-eq /MJ, which is a greenhouse gas improvement compared to petroleum-based gasoline (the carbon intensity of conventional gasoline is 93.0 gCO 2-eq /MJ). However, these biofuels are produced at an additional economic cost ($1.8 billion over the displaced gas/diesel). Additionally, these biofuels generate unfavorable impacts in other environmental categories when compared to petroleum-based fuels, particularly in areas of water, land, and nitrogen intensity. This work enables stakeholders to assess the specific value of biofuels as a means of emission reduction and determine if the attained improvements are worth the associated opportunity costs.

09 BIOMASS FUELS↗

Toward the Detection of Polyglot Files

Standardized file types play a key role in the development and use of computer software. However, it is possible to confound standardized file type processing by creating a file that is valid in multiple file types. The resulting polyglot (many languages) file can confuse file type identification, allowing elements of the file to evade analysis. This is especially problematic for malware detection systems that rely on file type identification for feature extraction. Although work has been done to identify file types using more comprehensive methods than file signatures, accurate identification of polyglot files remains an open problem. Since malware detection systems routinely perform file type-specific feature extraction, polyglot files need to be filtered out prior to ingestion by these systems. Otherwise, malicious content could pass through undetected. To address the problem of polyglot detection we assembled a data set using the mitra tool. We then evaluated the performance of the most commonly used file identification tools, including file, polydet, binwalk, and TrID. Our analysis demonstrates that existing file type detection tools fail to provide reliable polyglot detection. We then evaluated the ability of a range of machine and deep learning models to detect polyglot files. The most performant models were MalConv2 and Catboost, which demonstrated the highest recall on our data set with 95.16% and 95.45%, respectively. These models outperformed existing methods and could be incorporated into a malware detector’s file processing pipeline to filter out potentially malicious polyglots before file type-dependent feature extraction takes place.

Koch, Luke↗

Toward the Detection of Polyglot Files

Standardized file types play a key role in the development and use of computer software. However, it is possible to confound standardized file type processing by creating a file that is valid in multiple file types. The resulting polyglot (many languages) file can confuse file type identification, allowing elements of the file to evade analysis. This is especially problematic for malware detection systems that rely on file type identification for feature extraction. Although work has been done to identify file types using more comprehensive methods than file signatures, accurate identification of polyglot files remains an open problem. Since malware detection systems routinely perform file type-specific feature extraction, polyglot files need to be filtered out prior to ingestion by these systems. Otherwise, malicious content could pass through undetected. To address the problem of polyglot detection we assembled a data set using the mitra tool. We then evaluated the performance of the most commonly used file identification tools, including file, polydet, binwalk, and TrID. Our analysis demonstrates that existing file type detection tools fail to provide reliable polyglot detection. We then evaluated the ability of a range of machine and deep learning models to detect polyglot files. The most performant models were MalConv2 and Catboost, which demonstrated the highest recall on our data set with 95.16% and 95.45%, respectively. These models outperformed existing methods and could be incorporated into a malware detector’s file processing pipeline to filter out potentially malicious polyglots before file type-dependent feature extraction takes place.

Koch, Lucas↗

A Step Toward Working with Untrusted Ground Stations

We are witnessing a shift toward outsourcing satellite and ground station services to third-party commercial entities. As with any enterprise, these third parties can be vulnerable to cyber compromise, including image tampering and deepfake injection. The multimedia community is beginning to establish standards and technology to enable authenticity verification of multimedia created and edited by others. While appealing to the remote sensing domain, the nature of raw satellite imagery is incompatible with the proposed change verification tools, resulting in the need for a means to validate updates made to image products. We present a simple method for verifying a specific class of algorithms. Our inverse processing approach eliminates the need to see the original image as the reversed data can be checked against an original digital signature. We demonstrate our approach on basic image restoration routines and conclude with a discussion on open challenges and next steps.

97 MATHEMATICS AND COMPUTING↗

Microparticle Impact Testing at High Precision, Higher Temperatures, and with Lithographically Patterned Projectiles

Abstract In the first decade of high‐velocity microparticle impact research, hardly any modification of the original experimental setup has been necessary. However, future avenues for the field require advancements of the experimental method to expand both the impact variables that can be quantitatively assessed and the materials and phenomena that can be studied. This work explores new design concepts for the launch pad (the assembly that launches microparticles upon laser ablation) that can address the root causes of many experimental challenges that may limit the technique in the future. Among the design changes contemplated, the substitution of a stiff glass launch layer for the standard elastomeric polymer layer offers a number of improvements. First, it facilitates a reduction of the gap between launch pad and target from hundreds to tens of micrometers and thus unlocks a reproducibility in targeting a specific impact location better than the diameter of the test particle itself (±1.75 µm for SiO 2 particles 7.38 µm in diameter). Second, the inert glass surface enables experiments at higher temperatures than previously possible. Finally—as demonstrated by the launch of thin‐film Au disks—a launch pad made of materials standard in microfabrication paves the way to facile microfabrication of advanced impactors.

36 MATERIALS SCIENCE↗

SuperCam Calibration Targets: Design and Development

Abstract SuperCam is a highly integrated remote-sensing instrumental suite for NASA’s Mars 2020 mission. It consists of a co-aligned combination of Laser-Induced Breakdown Spectroscopy (LIBS), Time-Resolved Raman and Luminescence (TRR/L), Visible and Infrared Spectroscopy (VISIR), together with sound recording (MIC) and high-magnification imaging techniques (RMI). They provide information on the mineralogy, geochemistry and mineral context around the Perseverance Rover. The calibration of this complex suite is a major challenge. Not only does each technique require its own standards or references, their combination also introduces new requirements to obtain optimal scientific output. Elemental composition, molecular vibrational features, fluorescence, morphology and texture provide a full picture of the sample with spectral information that needs to be co-aligned, correlated, and individually calibrated. The resulting hardware includes different kinds of targets, each one covering different needs of the instrument. Standards for imaging calibration, geological samples for mineral identification and chemometric calculations or spectral references to calibrate and evaluate the health of the instrument, are all included in the SuperCam Calibration Target (SCCT). The system also includes a specifically designed assembly in which the samples are mounted. This hardware allows the targets to survive the harsh environmental conditions of the launch, cruise, landing and operation on Mars during the whole mission. Here we summarize the design, development, integration, verification and functional testing of the SCCT. This work includes some key results obtained to verify the scientific outcome of the SuperCam system.

Manrique, J. A. (ORCID:0000000220532819)↗

Probing New physics with high-redshift quasars: axions and non-standard cosmology

The Hubble diagram of quasars, as candidates to “standardizable” candles, has been used to measure the expansion history of the Universe at late times, up to very high redshifts ( z ~ 7). It has been shown that this history, as inferred from the quasar dataset, deviates at ≳ 3 σ level from the concordance (ΛCDM) cosmology model preferred by the cosmic microwave background (CMB) and other datasets. In this article, we investigate whether new physics beyond ΛCDM (BΛCDM) or beyond the Standard Model (BSM) could make the quasar data consistent with the concordance model. We first show that an effective redshift-dependent relation between the quasar UV and X-ray luminosities, complementing previous phenomenological work in the literature, can potentially remedy the discrepancy. Such a redshift dependence can be realized in a BSM model with axion-photon conversion in the intergalactic medium (IGM), although the preferred parameter space is in tension with various other astrophysical constraints on axions, at a level depending on the specific assumptions made regarding the IGM magnetic field. We briefly discuss a variation of the axion model that could evade these astrophysical constraints. On the other hand, we show that models beyond ΛCDM such as one with a varying dark energy equation of state ( w CDM) or the phenomenological cosmographic model with a polynomial expansion of the luminosity distance, cannot alleviate the tension. The code for our analysis, based on emcee [1] and corner.py [2], is publicly available at github.com/ChenSun-Phys/high_z_candles.

79 ASTRONOMY AND ASTROPHYSICS↗

Reactive Power: sorted, perhaps

The measurement of reactive power has been challenging for over a century. By the 1920s, the IEC and the AIEE were seeking new definitions for reactive power, specifically to handle distorted waveforms and unbalanced systems. Solutions had not and have not been found. This paper shows there are two aspects to the problem. First, the conceptual underpinning of reactive power provides a very shaky foundation for the design of an instrument. Second, the kind of instrument needed is one that has been largely unrecognized in the power engineering field. It is an operational measurement. These aspects of the measurement have been taken into account in IEEE Standard 1459, presently undergoing revision by a working group of the IEEE PES Power System Instrumentation and Measure¬ments Committee.

reactive power, measurement theory, operational, D↗

Development of a Lower Cost, High Volume, Commercially Available, Precursor for Lower Cost Carbon Fiber for Automotive and Wind Blade Applications

Essentially all carbon fibers used in structural applications are presently manufactured from solution spun polyacrylonitrile (PAN) precursor. Solution spinning is accomplished at relatively low temperatures using very large amounts of environmentally unfriendly solvents that impart major impact to line speed and costs involved with handling and reclaiming the solvents and associated “waste streams” required in spinning and extracting those solvents. Dralon and others have demonstrated the dry spinning of PAN in the production of some acrylic fibers for textile applications, but to this point it has not been demonstrated in the production of carbon fibers. In dry spinning, the PAN polymer is also dissolved into similar solvents, although typically less solvent is utilized to maintain a somewhat higher solids content of 20-50%. The use of lower overall solvent quantities and elimination of the need for solvent/water separation simplifies and reduces the cost of solvent recovery operations. Other portions of the precursor production and carbon fiber manufacturing processes are quite similar. The work described in this report demonstrated that dry spun acrylic fiber can be an attractive candidate for carbon fiber precursor of potential interest to applications looking to capture the specific stiffness (stiffness per unit mass) and specific strength (strength per unit mass) of carbon fiber when produced at a discount relative to commercial carbon fibers in the market today. Dralon has determined that the dry spinning process utilized for producing fibers used as precursors in this project requires 27% less energy versus producing comparable fibers via the industry standard wet spinning process. Identification of carbon fiber manufacturers, as well as applications development and targeted marketing, will be required to achieve significant in-roads into mission areas of interest to DOE, such as automotive applications and wind turbine blades. In support of that activity, initial data produced in this project indicate the innovative approach as described in this report can meet at least the baseline property requirements identified for these applications.

36 MATERIALS SCIENCE↗

Transmission Operator Workflows for Real-Time Reliability Studies: A Review of Control Room Practices and Naturalistic Decision Making

This report provides an overview of real-time reliability study tools and their use by power system operators in the control room environment. After introducing some of the nuances of the control room environment and the differences in perspectives between power system engineers and operators, the roles and responsibilities of key entities involved in RTCA workflows are introduced. These are specifically the transmission system operator (TOP) and reliability coordinator (RC), which are required to run tools such as real-time contingency analysis (RTCA) as part of a real-time reliability assessment every 30 minutes, as dictated by a series of standards issued by the North American Electric Reliability Corporation (NERC). The process by which power systems operators operate the grid is discussed in terms of naturalistic decision making (NDM) and the recognition-primed decision-making (RPD) model. This cognitive model describe how experts working in high-risk, high-stress environments make safety-critical decisions under uncertainty and time pressure. For power system operators, the mental simulations involved in the traditional RPD model are supplemented by physics-based simulations using numerical tools, such as RTCA, to improve situational awareness and effectiveness of control actions. Next, a generic workflow is introduced to describe operator decision making for running RTCA tools and responding to system violations on a pre-contingent basis. The types of analysis performed and control actions chosen by power system operators are described in detail. The overall high-level workflow is then expanded in subsequent sections, with special attention given to high-voltage violations, low-voltage violations, and thermal overloads. Each type of violation is described in detail, with explanations of common causes, impacts on equipment and customers, and mitigation strategies. An additional workflow diagram is provided for each type of violation.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Feature Engineering and Process Mining to Enable Hazard Detection in Health Information Technology

Motivation for this work originates from a recognized need for developing data-driven approaches through analysis of complex electronic health records. Specifically, we aim to enhance the reliability of health information technology (HIT) systems by detection of plausible HIT hazards in clinical order transactions. In the absence of well-defined event logs in corporate data warehouses, our proposed approach identifies relevant timestamped data fields that could indicate transactions in the clinical order life cycle generating raw event sequences. Subsequently, we adopt state transitions of the OASIS Human Task standard to map the raw event sequences and simplify the complex process that clinical radiology orders go through. We describe how the current approach provides the potential to investigate areas of improvement and potential hazards in HIT systems. The discussion concludes with a use case and opportunities for future applications.

Ozmen, Ozgur↗

Modeling Approach for the Aluminum-clad Dry Storage Pilot using HFIR Fuel

To confirm that the dry storage of aluminum-clad research reactor spent nuclear fuel (ASNF) will remain within the safety envelope after applied drying schemes and that the resulting evolution of the gas space composition, temperature, and pressure conditions are understood, a dry storage pilot project is being established. The pilot will incorporate an instrumented lid for discrete interval or for on-demand gas composition and temperature monitoring of two DOE Standard Canisters (DSCs) loaded with three High Flux Isotope Reactor (HFIR) inner cores per DSC. Each DSC would be subjected to a separate alternative candidate drying scheme. Canisters will undergo 1 to 5 years of monitoring, including internal temperature and gas sampling to track pressure and composition changes. This report outlines the approach for modeling the ASNF-in-canister behavior in terms of evolving gas space conditions for the ASNF dry storage pilot using HFIR fuel. The ASNF has an adherent surface oxyhydroxide layer comprised of boehmite/bayerite that generates hydrogen when subjected to irradiation. Three-dimensional multi-physics computational fluid dynamics simulations will be executed to compute the thermal field within the DSC and provide inputs to a chemical model employed to compute pressure buildup as hydrogen is generated in the system. Implemented in Cantera, the chemical model solves gas phase and aluminum oxyhydroxide surface-mediated radiolysis reactions. Gas phase reactions are sourced from Wittman and Hanson (2015), whereas surface-mediated reactions are incorporated by fitting experimental data using an optimization algorithm (Abboud, 2023). Water radiolysis reactions from Wren and Ball (2001) are adopted with modifications as described in Abboud (2023c). Understanding the effect of the hydrogen buildup over time is important for long-term storage safety considerations. Modeling results will include the canister pressure, temperature, and composition evolution from the initial helium backfill with the addition of radiolytically-evolved chemical species (e.g., hydrogen and oxygen). The specific HFIR cores for the pilot program have not yet been selected, and the overall design is still in development. The CFD-chemical model used for this work will be based on prior models with necessary updates to allow for improved accuracy and efficiency. The experimental data obtained from the HFIR demonstration will be used to improve and validate the computational models to predict the ASNF-in-canister behavior.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Comparative clumped isotope temperature relationships in freshwater carbonates

Lacustrine, riverine and spring carbonates represent archives of terrestrial climates and their geochemistry has been used to study palaeoenvironments. Clumped isotope thermometry is an emerging tool that has been applied to freshwater carbonates. Limited work has been done to evaluate comparative relationships between clumped isotopes and temperature in different types of modern freshwater carbonates. This study assembles an extensive calibration data set with 135 samples of modern freshwater carbonates from 96 sites and constrains the relationship between independent observations of water temperature and the clumped isotopic composition of carbonates (denoted by Δ 47 ), including new measurements, and recalculates published data in accordance with current community-defined standard values. For temperature reconstruction, the study reports a composite freshwater calibration and material-specific calibrations for biogenic carbonates (freshwater gastropods and bivalves), fine-grained carbonate (e.g. micrites), biologically mediated carbonates (microbialites and tufas) and travertines. Material-specific calibration trends show a convergence of slopes that are in agreement with recently published syntheses, but statistically significant differences in intercepts occur between some materials (e.g. some biogenics, fine-grained carbonates). These differences may arise due to unresolved seasonal biases, kinetic isotope effects and/or varying degrees of biological influence. The impact of different calibrations is shown through application to new data for glacial and deglacial age travertines from Austria and published data sets. While material-specific calibrations may yield more accurate results for biogenic and fine-grained carbonate samples, the use of material-specific and the composite freshwater calibrations generally produces values within 1.0–1.5°C of each other, and typically fall within calibration uncertainty given limitations of precision.

54 ENVIRONMENTAL SCIENCES↗

A novel, steerable, low-energy proton source for detector characterization

We report on the conversion of the Manitoba II mass spectrometer into a versatile low-energy proton beam facility. This infrastructure is adaptable to any detector-under-test (DUT), and has proven itself effective with the characterization of silicon detectors used in subatomic beyond-the-Standard-Model (BSM) searches, namely the Nab experiment. A pencil beam of mono-energetic protons can be produced in a range from 25 keV to 35 keV, achieving a beam current of ~1 x 10 -18 A. Electrostatic steering plates were constructed to direct the Gaussian-profile proton beam over a 117 mm diameter area-of-interest with full-width at half-maxima (FWHM) ranging from 0.6 mm to 1.26 mm. This work discusses the modifications and subsequent tests to confirm the beam specifications met the demands of the aforementioned detectors.

Fundamental neutron physics↗

Fast-airflow tumble clothes dryer with small thermoelectric heat pump: Experimental evaluation

Residential clothes drying accounts for about 5% of the total residential-sector energy consumption in the United States. Most dryers use electric resistance heaters to dry clothes and have low efficiencies. Higher-efficiency dryers that use vapor compression heat pumps are expensive and complex and have not gained a large market share in the United States. A novel tumble clothes dryer using a small thermoelectric heat pump with faster airflow than typical dryers is presented in this work. The benchtop performance of the thermoelectric heat pump and high-speed blower are presented, and the development of the prototype dryer is described. The dryer was tested for efficiency and dry time for a range of airflow rates and applied currents to the thermoelectric heat pump. The combined efficiency factor was 5.09–6.29 lb BDW /kWh (specific moisture extraction rate of 1.23–1.53 kg w /kWh) with 100–138 min dry times for these tests. The measured efficiency was 36 %–68 % greater than the minimum efficiency standard in the United States, and compared with vapor compression heat pump–based clothes dryers, the prototype dryer had less expensive, less complex components and did not use refrigerants. Finally, the performance of this small thermoelectric heat pump clothes dryer is also compared with previous iterations of the thermoelectric tumble clothes dryer described in the literature.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Initial Efforts Organizing WPNCS SG-8: Preservation of Expert Knowledge and Judgement Applied to Criticality Benchmarks

The Working Party on Nuclear Criticality Safety (WPNCS) under the guidance of the Organization for Economic Co-operation and Development (OECD) Nuclear Energy Agency (NEA) has over 20 years of experience addressing concerns related to static and transient configurations encountered within the nuclear fuel cycle: fuel fabrication, transportation, reprocessing, storage, and geological disposal. One of the cornerstone activities of the WPNCS is the International Criticality Safety Benchmark Evaluation Project (ICSBEP), which was established to identify a comprehensive set of criticality benchmark data, evaluate the data, including quantification of overall uncertainties; compile the data into a standardized format, perform sample calculations utilizing modern nuclear data sets and codes utilized in nuclear criticality safety, and formally document the work into a single source of verified benchmark data. Annually, members of the ICSBEP Technical Review Group (TRG) contribute evaluated benchmark data that undergoes comprehensive technical review prior to publication in the ICSBEP Handbook. In the years since the ICSBEP was established, there has been much work to prepare benchmark data to support validation activities in nuclear criticality safety. The 2020 edition of the ICSBEP Handbook contains acceptable benchmark specifications for 5,053 critical, subcritical, or near-critical configurations in 582 benchmark evaluations. Modern benchmark development benefits from decades of experienced international participants, a well-established handbook format, supplementary guides to deal with uncertainty quantification, and a comprehensive review process based upon independent reviews from international experts. The ICSBEP Handbook also contains 838 configurations deemed unacceptable to support criticality safety efforts. They are recorded, with the reasoning for their rejection, to preserve the experimental data, prevent reevaluation of data that are incomplete or contain known errors, and/or to potentially allow future reevaluation of the experiment pending the identification of sufficient data to resolve identified inconsistencies and errors. Users of the ICSBEP Handbook today might notice that the rigor and quality of modern criticality safety benchmarks is much greater than those prepared within the initial decade of the project. Benchmarks with 1s uncertainties in k eff greater than 1% were traditionally rejected unless they were identified as unique experiment types that encompassed materials, fuels, or designs not available from other benchmark experiments. However, benchmarks developed using modern experimental techniques and practices typically have uncertainties on the order of a few tenths of a percent. There have been ongoing efforts to improve the overall quality of previously published benchmark evaluations. Seventy-eight evaluations, containing approximately 600 configurations, have been revised just within the past decade. An additional eleven benchmarks are under revision for updated release in the 2020 edition of the ICSBEP Handbook. If some of the historic benchmarks were resubmitted in their current form to the TRG today, they would be rejected due to lack of data, missing components in the uncertainty analysis, or incomplete benchmark model development. The use of historic criticality safety benchmarks that underestimate the total uncertainty, lack properly quantified biases, or provide inadequate benchmark specifications do not sufficiently support modern criticality safety and nuclear data efforts. Although the ICSBEP Handbook is recognized by regulating bodies to support criticality safety, users are required to justify their reasons to ignore historic benchmark data and include additional safety margins within their designs. Discussions were held at the WPNCS 23rd Annual Meeting in September 2019 regarding the aforementioned issues. The resultant decision was to establish Subgroup 8 (SG-8): Preservation of Expert Knowledge and Judgement Applied to Criticality Benchmarks. The current activities of SG-8 are discussed herein.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Reducing Uncertainty of Fielded Photovoltaic Performance (Final Technical Report)

Improved analysis and reporting of photovoltaic (PV) field performance increases the certainty of owners and financiers that systems will perform as expected. Advanced module technologies (e.g., PERC, HJT, and bifacial) introduce new degradation mechanisms and performance characteristics. The FY19-21 Reducing Uncertainty project leveraged data from the ever-increasing PV fleet to develop models and understanding of the field performance of existing and new technologies. Specifically, we accomplished: report on field performance and degradation rates for high-efficiency silicon (HJT, PERC, IBC) and more conventional technologies; developed automated analysis techniques to quantify system performance (performance ratio, energy yield) and production shortfalls (soiling, degradation, availability); refined the RdTools software toolkit to bring standard, validated analysis techniques to bear on third-party data; analyzed and reported on large datasets including Treasury data and Lawrence Berkeley National Laboratory's Utility-Scale dataset to expand the high-quality degradation-rate histogram published previously; worked with industry partners and the DuraMAT data hub to enable private parties to share and aggregate PV production data anonymously, leveraging cloud-based data analysis infrastructure and publishing on US fleet-scale performance comprising over 7GW of operating systems. (https://www.nrel.gov/pv/fleet-performance-data-initiative.html). Through our industry collaborations we have engaged in NDA-covered data transfer with twelve PV fleet owners as of January 2022, with more agreements in negotiation. Our scalable cloud-based time series database contains over 30 billion rows (20TB) of PV time series data, representing over 1700 commercial and utility-scale systems, and over 7.2 GW of DC capacity (Fig 1). Initial field performance results have been distributed in several public reports. Because our fleet composition and data quality methods are continually improving, annual updates to these results are published to our PV Fleet webpage [ https://www.nrel.gov/pv/fleet-performance-data-initiative.html ] and DuraMAT data hub [DOI: 10.21948/1842958]. Another existing dissemination channel used for observed soiling losses is a map we maintain for soiling losses. Additional products developed include a report detailing fleet-wide performance index, availability, startup loss and snow loss factors, a detailed report on the 1603 grant dataset comprising over 100,000 PV systems with failure and performance details and a utility-scale report coauthored with LBNL on 31 GW of system performance.

14 SOLAR ENERGY↗

Capsule network-based semantic segmentation model for thermal anomaly identification on building envelopes

Thermography technology is widely used to inspect thermal anomalies in building façade systems. Computer vision-based techniques provide opportunities to autonomously detect such heat anomalies to significantly improve the efficiency of decision-making for building envelope retrofitting and maintenance. Here, in this work, we propose a novel Capsule Network-based deep learning model – CapsLab – that detects and identifies thermal anomalies by semantic segmentation. CapsLab is built based on our proposed prediction-tuning capsule (PT-Capsule) layer. Different from a traditional capsule layer, which consists of part-whole transformation and capsule-routing process, the proposed layer is composed of a prediction and tuning process, which helps decreasing the number of model parameters significantly. While the applicability of traditional Capsule Networks (CapsNets) has been limited to simpler tasks and smaller datasets due to their scalability issue, we can leverage the lightweight of the proposed PT-Capsule layer, and apply it to the semantic segmentation task. In this work, we also employ our previously presented performance metric, referred to as the Anomaly Identification Metric (AIM) (Kakillioglua et al. 2021), to evaluate the segmentation outputs. Traditional performance metrics do not accurately reflect the true performance of the segmentation models in thermal anomaly identification due to the high subjectivity in the annotation process and higher overlap ratio sensitivity of the standard metrics. AIM, on the other hand, is robust to these drawbacks. Experimental results show, both qualitatively and quantitatively, that our proposed segmentation method can effectively segment the thermal anomalies. Specifically, our model provides 9.38% and 13.53% improvements over the baseline model – DeepLabV3+ – based on traditional mIoU score and the AIM score, respectively, while requiring less model parameters and less computation at the same time. In addition, the scores that the AIM metric generates better align with the scores provided by building performance experts.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗