Search NASA⌕ Search

SEARCH · Search NASA

Results for “automatic loading”

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 163 records · Page 9

Testbed-based Performance Evaluation of Attack Resilient Control for AGC

The modern electric power grid is a complex cyber-physical system whose reliable operation is enabled by a wide-area monitoring and control infrastructure. This infrastructure, supported by an extensive communication backbone, enables several control applications functioning at multiple time scales to ensure the grid is maintained within stable operating limits. Recent events have shown that vulnerabilities in this infrastructure may be exploited to manipulate the data being exchanged. Such a scenario could cause the associated control application to mis-operate, potentially causing system-wide instabilities. There is a growing emphasis on looking beyond traditional cybersecurity solutions to mitigate such threats. In this paper we perform a testbed-based validation of one such solution - Attack Resilient Control (ARC) - on Iowa State University's \textit{PowerCyber} testbed. ARC is a cyber-physical security solution that combines domain-specific anomaly detection and model-based mitigation to detect stealthy attacks on Automatic Generation Control (AGC). In this paper, we first describe the implementation architecture of the experiment on the testbed. Next, we demonstrate the capability of stealthy attack templates to cause forced under-frequency load shedding in a 3-area test system. We then validate the performance of ARC by measuring its ability to detect and mitigate these attacks. Our results reveal that ARC is efficient in detecting stealthy attacks and enables AGC to maintain system operating frequency close to its nominal value during an attack. Our studies also highlight the importance of testbed-based experimentation for evaluating the performance of cyber-physical security and control applications.

attack resilience↗

Real-Time Cyber-Physical False Data Attack Detection in Smart Grids Using Neural Networks

Attacks on cyber-physical systems have recently increased in frequency, impact, and publicity. In this paper, a cyber-physical false data attack detection mechanism is proposed to protect the operation of power transmission and distribution systems by automatically inferring underlying physical relationships using cross-sensor analytics in order to detect sensor failures, replay attacks, and other data integrity issues in real-time. We investigate a neural network based mechanism acting on voltage and current readings resulting from a wide variety of load conditions on the IEEE 30-bus power system standard and compare its performance with a support vector machine based mechanism. Experiments showed that 99% detection accuracy of replay attacks was achieved using the proposed neural network mechanism. More importantly, we showed that the best approach was to not create physics-based features using what we knew about the system, but rather to use a neural network to automatically learn the laws, and then use the outputs of that to build a classifier to identify whether and where data spoofing occurs. Thus, it is preferable to infer and exploit the physics using a single machine learning solution

Ferragut, Erik M.↗

Web-based Preprocessing and Visualization of 3D FIB Tomography Data for Nuclear Fuel Characterization

Three-dimensional (3D) focused ion beam (FIB) tomography enables reconstruction of internal nuclear fuel features that can't be fully evaluated through surface imaging alone. This capability supports characterization of fuel constituents and defects under thermal and irradiation conditions relevant to microreactor development. However, large tomography datasets can create data-handling, loading, and visualization challenges, especially when image-stack preparation and file conversion must be completed with separate tools. The Computational Ultraspatial Tomography Toolkit for High-Resolution Object Analysis Tools (CUTTRHOAT) is an open-source web application being developed to display FIB tomography datasets available through the Nuclear Research Data System (NRDS). The current alpha version requires prepared HDF5 datasets and has limited integrated data-preparation capabilities. This project improves CUTTHROAT by adding dataset-folder selection, automatic input detection, dataset scanning, missing-slice identification, blank-slice insertion, and image-stack-to-HDF5 conversion. Two applications will be compared: the baseline CUTTHROAT alpha workflow and the updated application containing the integrated data-handling and preprocessing functions. Evaluation will consider dataset detection accuracy, conversion success, loading time, rendering responsiveness, application stability, and user interaction. Preliminary results demonstrate successful loading of existing HDF5 files and converted image stacks, while testing also identified performance reductions caused by excessive blank-slice generation. The updated workflow reduces reliance on external preparation tools and supports more direct movement from image stacks to color-code 3D visualization. Future work includes refining missing-slice handling, integrating additional preprocessing functions, like a denoising feature, parsing TIFF metadata for automatic voxel scaling, and adding manual X, Y, and Z voxel-spacing inputs for PNG and JPEG.

36 - MATERIALS SCIENCE↗

Deactivation and decommissioning web log analysis using big data technology - 15710

The D and D Knowledge Management Information Tool (D and D KM- IT) is a web-based knowledge management information tool built for the D and D user community. Big data is a massive volume of structured and unstructured datasets which is so large that it's difficult to process using traditional database techniques. Web logs are the repository of files which are generated automatically for any operation on the web site. Web log files generated from the D and D KM-IT will be processed using the Apache Hadoop Framework to extract meaningful data. Hadoop is an open-source software framework for storing and processing big data in a distributed environment on large clusters of commodity hardware. Hadoop framework consists of two main layers: Hadoop distributed file system (HDFS) and the Execution engine (MapReduce). Automatic parallelization and distribution; clean and simple programming abstraction. Implementation: Web log files from the D and D KM-IT server are fetched using Apache Flume and loaded into the Hadoop distributed file system. Using the Apache Hcatalog tool, the web log data is parsed and converted into a structured table format. A Pig program is developed to query the tables to retrieve and store the significant information onto the HDFS. The extracted data is fed to business visualization tools such as Microsoft Excel for analysis and reporting. Results The reports generated through the web log data processing will be analyzed to improve the usability and performance of the D and D KM-IT. Conclusion/Future Work: User keyword information can be explored to add new trending topics to the D and D KM-IT application. Statistics generated can be compared to analysis reports from existing analytical tools.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

TH-CD-207A-02: Implementation of Live EPID-Based Inspiration Level Assessment (LEILA) for Deepinspiration Breath-Hold (DIBH) Monitoring Using MV Fluoroscopy

Purpose: As prior work has shown that current DIBH monitoring approaches using surrogate measures (marker block on chest) do not always correspond with the clinical quantity of interest (lung depth, LD), a software tool and workflow are introduced to use MV fluoroscopy during treatment for real-time / Live EPID-based Inspiration Level Assessment (LEILA). Methods: A prototype software tool calculates and displays the LD during the treatment of left sided breast cancer. Calculations are based on MV cine images which are acquired with the treatment beam thereby not incurring any additional imaging dose. Image capture and processing are implemented using a dedicated frame grabber computer. The calculation engine automatically detects image orientation and includes provisions for large treatment fields that exceed the size of the EPID panel. LD is measured along a line profile in the middle of the field. LEILA’s interface displays the current MV image, a reference image (DRR), the current LD, as well as a trace of LD over treatment time. The display includes patient specific LD tolerances. Tolerances are specified for each field and loaded before the treatment. A visual warning is generated when the tolerance is exceeded. LEILA is initially run in parallel with current DIBH techniques. When later run by itself DIBH setup will be done using skin marks and room laser. Results: Offline tests of LEILA confirmed accurate automatic LD measurement for a variety of patient geometries. Deployment of the EPID during all left sided breast treatments was well tolerated by patients and staff during a multi-month pilot. The frame grabber provides 11 frames-per-second; the MATLAB based LEILA prototype software can analyze five frames-per-second standalone on standard desktop hardware. Conclusion: LEILA provides an automated approach to quantitatively monitor LD on MV images during DIBH treatment. Future improvements include a database and further speed optimization.

60 APPLIED LIFE SCIENCES↗

American Recovery and Reinvestment Act (ARRA) FEMP Technical Assistance - United States Pacific Command (PACOM) Guam, Task 3.3: Building Retuning Training

Document describes an onsite workshop and building retuning training conducted in Guam in August 2010. Document reports on issues identified during an audit of several buildings and recommendations to save energy throughout the site. During the workshop, it became apparent that as site personnel maintain the facilities at Guam, the following retuning efforts and strategies should be prioritized: (1) Controlling the mechanical systems operational hours and zone temperature set points appeared to present the best opportunities for savings; (2) Zone temperature set points in some buildings are excessively low, especially at night, when the zone temperatures are so cold that they approached the dewpoint; and (3) Manually-set outside air dampers are providing excessive outside air, especially for spaces that are unoccupied. Two of the larger schools, one on the Naval Base and one on Anderson AFB, are in need of a significant recommissioning effort. These facilities are relatively new, with direct digital controls (DDC) but are significantly out of balance. The pressure in one school is extremely negative, which is pulling humid air through the facility each time a door is opened. The draft can be felt several feet down the halls. The pressure in the other school is extremely positive relative to the outside, and you can stand 20-feet outside and still feel cool drafts of air exiting the building. It is recommended that humidity sensors be installed in all new projects and retrofitted into exist facilities. In this humid climate, control of humidity is very important. There are significant periods of time when the mechanical systems in many buildings can be unloaded and dehumidification is not required. The use of CO{sub 2} sensors should also be considered in representative areas. CO{sub 2} sensors determine whether spaces are occupied so that fresh air is only brought into the space when needed. By reducing the amount of outside air brought into the space, the humidity load is also substantially reduced. CO{sub 2} and humidity sensors, combined with outside air sensors, can be used to predict whether conditions are amenable to mold growth and to automatically adjust systems to help prevent mold without using extra energy. The goal of this training is to give the building operators the knowledge needed to make positive changes in the operation of building systems. As class participants apply this knowledge, building systems will run more efficiently, occupant comfort should improve, while saving energy and reducing greenhouse gas emissions.

54 ENVIRONMENTAL SCIENCES↗

Heave-Plate Hydrodynamic Coefficients for Floating Offshore Wind Turbines - A Compilation of Data: Preprint

The National Renewable Energy Laboratory's OpenFAST software is utilized by academics and industry professionals alike to simulate offshore wind turbines. The software's modeling of hydrodynamic loads on heave plates attached to these structures relies on user specified hydrodynamic coefficients. To guide the proper selection of these coefficients and potentially develop a new functionality within OpenFAST that automatically prescribes and/or adjust the heave-plate hydrodynamic coefficients, this paper reviews past literature to examine the dependence of the added mass, damping, and drag coefficients on various relevant non-dimensional parameters, including the Keulegan-Carpenter number, the frequency parameter, and the plate thickness ratio. Existing data in the literature show strong dependence of the hydrodynamic coefficients on the Keulegan-Carpenter number. Consistent trends are observed across a range of different plate geometry, plate porosity, and flow conditions. Secondary dependence of the coefficients on the frequency parameter and plate thickness ratio is also present.

added mass↗

Benefits of digitalizing and employing simulation to increase plant system performance and ensure compliance with technical specifications - 109

The essential services water system in a BWR nuclear power plant was affected by corrosion, soiling and blockages due to ageing. Those episodes affected safety margins hindering the Technical Specifications compliance during periodic system testing. Tecnatom designed, implemented and tested a design modification for the on-line monitoring of this system by integrating it in the existing Digital Control System (DCS). Key system parameters were acquired in real time to be displayed in the Human System Interface, used in performing calculations and stored their historical evolution. The objective was to optimize monitoring and surveillance of the essential service water system. Furthermore, Tecnatom developed an engineering simulator ('What if' simulator) consisting of a hydraulic model of the system. This simulator takes as inputs each heat exchanger performance parameters from the plant DCS, allows the user to change the essential services water system configuration (valves position, UHS level...) and calculates theoretical process values predicting the system real behavior. In a second stage, more system instruments were wired to the DCS for Technical Specifications compliance automatic surveillance regarding opening and closing times of system valves. Operating experience shows important benefits after the digitalization. It enhances on-line information of system operation and automates calculation of load loss factor of each heat exchanger. It eases the evaluation of blockages and soiling in the system allowing optimization of cleaning proceedings. System Surveillance Requirements and Test Requirements reports have been automated optimizing the operation workload. The simulator use enables to identify optimal system configuration increasing safety margins of the system and ensuring Technical Specifications fulfillment. (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

PySIDT: Subgraph Isomorphic Decision Trees for Molecular Property Prediction

Accurate molecular property prediction is important across all fields of chemistry. Deep neural networks (DNNs) have become increasingly popular due to their ability to train automatically, avoiding the incredibly tedious process of constructing and extending traditional property estimation schemes. However, DNNs require large amounts of training data, are challenging to interpret, require large amounts of memory to load even during inference, and have severe difficulties incorporating qualitative chemical knowledge, which are often desired for molecular property prediction tasks. Here, in this study, we present PySIDT (https://github.com/zadorlab/PySIDT), a software for training and running inference on Subgraph Isomorphic Decision Trees (SIDTs). SIDTs are graph-based decision trees made of nodes associated with molecular substructures. Inference is done by descending target molecular structures down the decision tree to nodes with matching subgraph isomorphic substructures and making predictions based on the final (most specific) nodes matched. SIDTs scale down well to dataset sizes much smaller than is feasible for DNNs. As trees of molecular substructures, SIDTs are inherently readable and easy to visualize, making them easy to analyze. They are also straightforward to extend and retrain, facilitate uncertainty estimation, and enable easy integration of expert knowledge. We demonstrate the SIDT approach discussing its application to a diverse range of molecular prediction tasks: rate coefficient estimation, diffusion coefficient estimation, thermochemistry estimation, transition state bond stretch prediction, p K a prediction, stability of molecular structures, stability of surface structures, and prediction of surface lateral interaction energetics. Additionally, we demonstrate the power of the SIDT algorithms in two direct learning curve vanilla comparisons with the popular DNN-based software Chemprop and the popular gradient boosted trees-based software XGBoost on enthalpy of formation and rate coefficient prediction tasks. In particular, in the enthalpy of formation case, vanilla PySIDT is able to outperform vanilla Chemprop and XGBoost across the full range of training/validation set sizes out to 11,560 data points.

Johnson, Matthew Sean [Sandia National Laboratorie↗

Multi-channel programmable power supply with temperature compensation for silicon sensors

Silicon Photo-Multipliers (SiPMs) are increasingly becoming popular for discrete photon counting applications due to the wealth of advantages they offer over conventional photo-detectors such as photo-multiplier tubes and hybrid photo-diodes. SiPMs are used in variety of applications ranging from high energy physics and nuclear physics experiments to medical diagnostics. The gain of a SiPM is directly proportional to the difference between applied and breakdown voltage of the device. However, the breakdown voltage depends critically on the ambient temperature and has a large temperature co-efficient in the range of 40-60 mV/°C resulting in a typical gain variation of 3%-5%/°C [Dinu et al., in IEEE Nuclear Science Symposium, Medical Imaging Conference and 17th Room Temperature Semiconductor Detector Workshop (IEEE, 2010), p. 215]. We plan to use the SiPM as a replacement for PMT in the cosmic ray experiment (GRAPES-3) at Ooty [Gupta et al., Nucl. Instrum. Methods Phys. Res., Sect. A 540, 311 (2005)]. There the SiPMs will be operated in an outdoor environment subjected to temperature variation of about 15 °C over a day. A gain variation of more than 50% was observed for such large variations in the temperature. To stabilize the gain of the SiPM under such operating conditions, a low-cost, multi-channel programmable power supply (0-90 V) was designed that simultaneously provides the bias voltage to 16 SiPMs. The programmable power supply (PPS) was designed to automatically adjust the operating voltage for each channel with a built-in closed loop temperature feedback mechanism. The PPS provides bias voltage with a precision of 6 mV and measures the load current with a precision of 1 nA. Using this PPS, a gain stability of 0.5% for SiPM (Hamamatsu, S10931-050P) has been demonstrated over a wide temperature range of 15 °C. The design methodology of the PPS system, its validation, and the results of the tests carried out on the SiPM is presented in this article. The proposed design also has the capability of gain stabilization of devices with non-linear thermal response.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Heave Plate Hydrodynamic Coefficients for Floating Offshore Wind Turbines - A Compilation of Data

The National Renewable Energy Laboratory's OpenFAST software is utilized by academics and industry professionals alike to simulate offshore wind turbines. The software's modeling of hydrodynamic loads on heave plates attached to these structures relies on user-specified hydrodynamic coefficients. To guide the proper selection of these coefficients and potentially develop a new functionality within OpenFAST that automatically prescribes and/or adjusts the heave-plate hydrodynamic coefficients, we review past literature to examine the dependence of the added mass, damping, and drag coefficients on various relevant nondimensional parameters, including the Keulegan-Carpenter number, the frequency parameter, and the plate thickness ratio. Existing data in the literature show strong dependence of the hydrodynamic coefficients on the Keulegan-Carpenter number. We observe consistent trends across a range of different plate geometry, plate porosity, and flow conditions. Secondary dependence of the coefficients on the frequency parameter and plate thickness ratio is also present.

added mass↗

Field Validation of Air-Source Heat Pumps for Cold Climates

Heating energy is the largest end-use for U.S. residential buildings accounting for approximately one-third of residential building energy consumption (EIA 2021). Historically, air-source heat pumps have been limited to temperate climates because of subpar performance at extremely cold outdoor air temperatures. However, recent advances to cold-climate air-source heat pump technology, which typically rely on inverter-driven, variable-speed compressors and variable-speed fans, have significantly improved low-temperature heat pump performance enabling the technology to save energy for many homes in cold climates. The primary objective of this project was to measure in-field performance of centrally ducted, variable-capacity air-source heat pumps in cold climates to validate performance and develop field-based performance maps. The project focused on quantifying heat pump performance at cold temperatures. The sites identified for the study were primarily located in the Northwest United States since homes in the region tend to have all-electric space heating systems and high-efficiency heat pumps have been incentivized in the region for several years. NREL partnered with Ecotope, Inc., a small energy consulting firm located in Seattle, WA, for site recruitment, monitoring equipment installation, data quality management. All the sites included in the study had previously installed a high-efficiency, central heat pump system. One site was in a Denver, CO suburb, which was the only dual fuel heat pump in the study. We used airside and power measurements, collected at 5-second intervals, to quantify heat pump capacity, coefficient of performance (COP), and auxiliary heater energy consumption. We developed algorithms to automatically determine the heat pump operating mode including defrost and auxiliary heating operation. A whole-house thermal and duct audit was completed during the initial site visit to estimate winter heating loads and assess heat pump sizing. Whole-home heating design loads were calculated at ASHRAE 99% design temperatures and compared to manufacturer-reported maximum capacities to assess the heat pump sizing at each site.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

NV Energy Large-Scale Photovoltaic Integration Study: Intra-Hour Dispatch and AGC Simulation

The uncertainty and variability with photovoltaic (PV) generation make it very challenging to balance power system generation and load, especially under high penetration cases. Higher reserve requirements and more cycling of conventional generators are generally anticipated for large-scale PV integration. However, whether the existing generation fleet is flexible enough to handle the variations and how well the system can maintain its control performance are difficult to predict. The goal of this project is to develop a software program that can perform intra-hour dispatch and automatic generation control (AGC) simulation, by which the balancing operations of a system can be simulated to answer the questions posed above. The simulator, named Electric System Intra-Hour Operation Simulator (ESIOS), uses the NV Energy southern system as a study case, and models the system’s generator configurations, AGC functions, and operator actions to balance system generation and load. Actual dispatch of AGC generators and control performance under various PV penetration levels can be predicted by running ESIOS. With data about the load, generation, and generator characteristics, ESIOS can perform similar simulations and assess variable generation integration impacts for other systems as well. This report describes the design of the simulator and presents the study results showing the PV impacts on NV Energy real-time operations.

Energy↗

Proposal for a Vehicle Level Test Procedure to Measure Air Conditioning Fuel Use

The air-conditioning (A/C) compressor load significantly impacts the fuel economy of conventional vehicles and the fuel use/range of plug-in hybrid electric vehicles (PHEV). A National Renewable Energy Laboratory (NREL) vehicle performance analysis shows the operation of the air conditioner reduces the charge depletion range of a 40-mile range PHEV from 18% to 30% in a worst case hot environment. Designing for air conditioning electrical loads impacts PHEV and electric vehicle (EV) energy storage system size and cost. While automobile manufacturers have climate control procedures to assess A/C performance, and the U.S. EPA has the SCO3 drive cycle to measure indirect A/C emissions, there is no automotive industry consensus on a vehicle level A/C fuel use test procedure. With increasing attention on A/C fuel use due to increased regulatory activities and the development of PHEVs and EVs, a test procedure is needed to accurately assess the impact of climate control loads. A vehicle thermal soak period is recommended, with solar lamps that meet the SCO3 requirements or an alternative heating method such as portable electric heaters. After soaking, the vehicle is operated over repeated drive cycles or at a constant speed until steady-state cabin air temperature is attained. With this method, the cooldown and steady-state A/C fuel use are measured. This method can be run at either different ambient temperatures to provide data for the GREEN-MAC-LCCP model temperature bins or at a single representative ambient temperature. Vehicles with automatic climate systems are allowed to control as designed, while vehicles with manual climate systems are adjusted to approximate expected climate control settings. An A/C off test is also run for all drive profiles. This procedure measures approximate real-world A/C fuel use and assess the impact of thermal load reduction strategies.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Automatic Modeling of VVER-1000 Power Maneuvers

A special version of the IR code is described. The purpose of this code is to generate recommendations for the reactor operator when operating in a maneuvering regime, automate boron regulation, and study the maneuvering characteristics of VVER-1000 fuel loads. The results of modeling of operation in a daily load schedule are presented for the example of a stationary run of VVER-TOI. The impact of variation of the parameters of power maneuvers (run, operating time a reduced power, loading rate) on the boron regulation parameters is analyzed.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

High capacity of lithium-sulfur batteries at low electrolyte/sulfur ratio enabled by an organosulfide containing electrolyte

We present that lithium-sulfur (Li-S) battery is a promising energy storage technology to replace lithium ion batteries for higher energy density and lower cost. Dissolution of lithium polysulfide intermediates in conventional Li-S electrolytes is known as one of the key technical barriers to the development of Li-S, because it promotes redistribution and irreversible deposition of Li 2 S, and also forces large amounts of electrolyte to be used, shortening cycling life and driving down cell energy density. Recently, dimethyl disulfide as a functional co-solvent has been demonstrated to show an alternate electrochemical reaction pathway for sulfur cathodes by the formation of dimethyl polysulfides and lithium organosulfides as intermediates and reduction products. In this work, comprehensive studies show that this new pathway not only provides high capacity but also enables excellent capacity retention through a built-in automatic discharge shutoff mechanism by tuning carbon/sulfur ratio in sulfur cathodes to reduce unfavorable Li 2 S formation. Furthermore, this new electrolyte system is also found to enable high capacity of high-sulfur-loading cathodes with low electrolyte/sulfur (E/S) ratios, such as a stable specific capacity of around 1000 mAh g -1 using a low electrolyte amount (i.e, E/S ratio of 5 mL g -1 ) and highsulfur-loading (4 mg cm -2 ) cathodes. This electrolyte system almost doubles the capacity obtained with conventional electrolytes under the same harsh conditions. In conclusion, these results highlight the practical potential of this electrolyte system to enable high-energy-density Li-S batteries.

25 ENERGY STORAGE↗

Ground heat exchanger design tool with rowwise placement of boreholes

Simulation-based design tools have been used since the late 1980s for designing ground heat exchangers (GHE) used with ground source heat pump (GSHP) systems. The ground heat exchanger simulations used in these tools rely on thermal response functions known as g-functions. Because of the significant computational burden in computing g-functions for even a single configuration, the design tools have relied on libraries of pre-computed g-functions. These g-functions were available for standard configuration shapes, such as lines, rectangles, open rectangles, L-shapes, and U-shapes. Standard shapes are often sub-optimal. For any building on a site, the available land may preclude use of a standard shape. For large GSHP systems with significantly imbalanced annual heat rejection and extraction loads, large rectangular fields may experience significant heat build-up (or heat draw-down) in the interior of the field. This paper describes a new ground heat exchanger design tool capable of automatically selecting and sizing both standard and irregular configurations. The focus of this paper is a method for creating, selecting and sizing irregular configurations where the available land area and "no-go" zones are described as irregular polygons.

Spitler, Jeffery↗

Automatic identification and quantification of dense microcracks in high-performance fiber-reinforced cementitious composites through deep learning-based computer vision

Highlights: • A method is presented to detect, locate, quantify, and visualize dense microcracks in HPFRCC. • Quantification of dense microcracks is realized using deep learning method for the first time. • The presented method uses deep learning models that are trained using a realistic dataset size. • The presented method has a high computation efficiency for identifying and quantifying cracks. • The presented method provides crack width with errors up to 50 μm and a R{sup 2} value of 0.984. High-performance fiber-reinforced cementitious composites (HPFRCCs) feature high mechanical strengths, crack resistance, and durability. Under excessive loading, HPFRCCs demonstrate dense microcracks that are difficult to identify using existing methods. This study presents a computer vision method for identification, quantification, and visualization of microcracks in HPFRCCs based on deep learning. The presented method integrates multiple deep learning models and computer vision techniques in a hierarchical architecture. The crack pattern (e.g., number, width, and spacing of cracks) are automatically determined from pictures without human intervention. This study shows that the presented method achieves an accuracy of 0.992 for crack detection and an accuracy finer than 50 μm (R{sup 2} > 0.984) for quantification of crack width when deep learning models are trained using only 200 pictures of HPFRCCs and 200 pictures of conventional concrete with incorporation of data augmentation. The presented method is expected to be also applicable to other materials featuring complex cracks.

36 MATERIALS SCIENCE↗