Next Article in Journal
Interleukin-33 and Obesity-Related Inflammation and Cancer
Previous Article in Journal
Empowering Healthcare Leaders: Coaching for Leadership Development
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Entry

AI-Driven Non-Destructive Testing Insights

by
Amine el Mahdi Safhi
*,
Gilberto Cidreira Keserle
and
Stéphanie C. Blanchard
Englobe Corp., Soil and Materials Engineering—Expertise, Laval, QC H7S 2E4, Canada
*
Author to whom correspondence should be addressed.
Encyclopedia 2024, 4(4), 1760-1769; https://doi.org/10.3390/encyclopedia4040116
Submission received: 7 October 2024 / Revised: 4 November 2024 / Accepted: 13 November 2024 / Published: 21 November 2024
(This article belongs to the Section Engineering)

Definition

:
Non-destructive testing (NDT) is essential for evaluating the integrity and safety of structures without causing damage. The integration of artificial intelligence (AI) into traditional NDT methods can revolutionize the field by automating data analysis, enhancing defect detection accuracy, enabling predictive maintenance, and facilitating data-driven decision-making. This paper provides a comprehensive overview of AI-enhanced NDT, detailing AI models and their applications in techniques like ultrasonic testing and ground-penetrating radar. Case studies demonstrate that AI can improve defect detection accuracy and reduce inspection times. Challenges related to data quality, ethical considerations, and regulatory standards were discussed as well. By summarizing established knowledge and highlighting advancements, this paper serves as a valuable reference for engineers and researchers, contributing to the development of safer and more efficient infrastructure management practices.

1. Introduction

NDT plays a critical role in civil engineering, allowing for the assessment of structural integrity and safety without causing damage to the tested materials. This is particularly important for large-scale infrastructure such as bridges, buildings, dams, tunnels, pipelines, and power plants. Traditional NDT methods have proven to be effective in identifying defects, assessing material properties, and guiding maintenance decisions [1].
However, these traditional NDT techniques are often labor-intensive, dependent on the expertise of the operator, and can be prone to human error and subjectivity. Additionally, the increasing complexity and scale of modern infrastructure projects demand faster, more accurate, and scalable inspection processes. The integration of AI into NDT provides transformative potential by automating defect detection, improving accuracy and consistency, enabling predictive maintenance, and allowing for data-driven decision-making [2,3]. This paper explores how AI enhances traditional NDT techniques by automating data analysis, improving defect detection accuracy, enabling predictive maintenance, and facilitating data-driven decision-making. It offers insights into specific AI algorithms, their applications, and the challenges faced, serving as a valuable reference for engineers and researchers interested in the intersection of AI and NDT.
Ensuring the safety, durability, and serviceability of critical infrastructure has never been more important. Failures in large structures can lead to catastrophic consequences, making accurate and timely inspections essential. While traditional NDT methods have served the industry well, the adoption of computational intelligence opens new horizons for more intelligent, autonomous, and precise inspection systems that can deliver insights far beyond what is possible through manual assessments alone [4,5,6].
A search was conducted in Scopus on 19 August 2024, using the query “non AND destructive AND testing AND concrete” within the title, abstract, and keywords fields. The search returned 4737 documents with a total of 22,137 keywords. To focus on the relevant terms, the analysis set the minimum keyword occurrence threshold at 20, resulting in 542 keywords. These keywords were visualized using Vosviewer 1.6.20 (Figure 1). This tool is used for constructing and visualizing bibliometric networks—widely used [7,8,9] to analyze and map relationships within the scientific literature. The visualization generated four clusters, with 55,120 total links and an overall link strength of 245,644. The central node, labeled “nondestructive examination”, indicates that this concept is at the core of the visualization, connecting to numerous related topics across the NDT domain.
The visualization is color-coded with four different clusters representing various subfields or areas within NDT. Here are some key clusters:
  • Red cluster (159 items): Material properties. This group focuses on fracture testing, self-healing materials, digital image correlations, and interferometry. These represent more specialized and emerging techniques in the NDT domain.
  • Green cluster (150 items): Mechanical properties. This section focuses on concrete-related testing methods, aggregate analysis, compressive strength, and mechanical properties. The presence of terms like geopolymer concrete, recycled aggregate, and curing mixtures suggests a focus on sustainable and advanced concrete technologies.
  • Blue cluster (133 items): Acoustic NDT methods. This cluster focuses on methods like acoustic emission, ultrasonic waves, defect detection, radar imaging, and damage detection. These are common techniques used to identify internal flaws in materials.
  • Yellow cluster (100 items): Structure quality. This section appears to emphasize infrastructure assessment and condition monitoring, including ground-penetrating radar (GPR), retrofitting, quality control, and maintenance.
The dense interconnections between clusters suggest that NDT is an interdisciplinary field that combines various methods, materials, and application areas. Nodes with thicker connections indicate key areas where multiple methodologies or topics converge, reflecting the integrated approach often needed in NDT.
Several nodes stand out due to their larger size or central position, indicating their importance, including acoustic emission testing, bridges and decks, radar measurements, and ultrasonic waves, among others. Based on visualization, topics like AI integration, digital twin technology, and smart infrastructure management are not as prominent, indicating potential areas for innovative research and application within the NDT field.

2. The Role of AI in NDT

AI techniques are transforming NDT by automating data analysis, enhancing defect detection, predicting failures, and optimizing maintenance strategies [4]. Traditional NDT methods have long relied on manual interpretation and often need a correlation analysis, which can lead to inconsistent results and human error. By incorporating AI, these processes are becoming more efficient, reliable, and scalable. The integration of AI into specific NDT techniques is driving significant improvements across civil engineering inspections. Figure 2 summarizes the role of AI in NDT.
AI algorithms can automate defect detection by analyzing image data and signals from various NDT methods [10,11]. They consistently identify subtle defects—such as micro-cracks [12,13,14], internal voids [15,16], and corrosion—that manual inspections might miss. In ultrasonic testing, AI systems analyze waveforms to detect debonding in concrete structures [17]. AI can enhance image and signal analysis by efficiently processing large datasets acquired from NDT methods. These datasets can be collected using advanced sensors and imaging equipment during inspections. Data pre-processing steps, such as noise reduction through filtering techniques and normalization, are crucial to handle the inherent noise and variability in NDT data. Additionally, data augmentation methods, like rotation and scaling of images, help in increasing the diversity of the training data, improving the robustness of AI models against noisy or incomplete data. Recognizing patterns in signals, AI models detect flaws not easily found through traditional analysis, speeding up the process and allowing for larger inspections within shorter timeframes—a critical advantage for large-scale infrastructure projects.
For instance, convolutional neural networks have been effectively utilized to analyze image data from ultrasonic testing and radiographic testing, identifying defects such as cracks and voids with higher accuracy than manual inspections. Support vector machines and random forest algorithms are also employed to classify signal patterns in acoustic emission testing, enhancing the detection of micro-cracks and corrosion. These AI models are trained on labeled datasets of NDT results, learning to recognize patterns associated with specific types of defects.
AI can enable predictive maintenance by analyzing historical inspection data to forecast potential failures, allowing proactive strategies that extend infrastructure lifespan, reduce downtime, and lower costs. For example, predictive models using half-cell potential testing data can evaluate a moment of high corrosion probability, optimizing maintenance efforts where they are most needed. ML plays a pivotal role in Prognostics and Health Management systems, as highlighted by recent studies [18,19,20]. These works discuss how ML algorithms improve predictive maintenance strategies, which aligns with the advancements in AI-powered NDT methods discussed in this paper.
These predictive models can be trained using historical measurements. The data are divided into training and testing sets to develop and validate the models. Techniques such as cross-validation ensure that the models generalize well to new data. Evaluation metrics like mean squared error for regression models or accuracy and F1-score for classification models are used to assess performance. This systematic approach ensures that the predictive models are reliable for forecasting potential failures.
AI’s ability to process vast amounts of data and identify meaningful correlations can transform decision-making in NDT. By integrating data from multiple methods, AI systems can enable comprehensive, data-driven strategies that optimize inspection schedules, resource allocation, and maintenance priorities. The use of digital twins—virtual models replicating physical assets in real time—is gaining popularity [21]. This AI-driven technique can provide monitoring of health infrastructure, predict performance, and enable proactive maintenance, enhancing the safety and longevity of critical structures.
For example, all these roles of AI in NDT can be exemplified in the use of GPR testing (Figure 3). AI can enhance defect detection by automatically interpreting radargrams to identify subsurface anomalies such as voids, cracks, and rebar positions with greater accuracy than manual analysis. Predictive maintenance can enable analyzing historical GPR data to forecast potential structural issues before they become critical, allowing for proactive interventions. AI-driven image and signal analysis can process the large datasets generated by GPR efficiently, recognizing patterns and anomalies that might be overlooked by human inspectors. Furthermore, integrating GPR data into AI systems can facilitate data-driven decision-making and the development of digital twins, providing real-time monitoring and predictive insights into the health of infrastructure assets. Thus, AI’s roles in NDT notably enhance the efficiency of GPR in civil engineering inspections.

3. Overview of AI Applications in NDT

The practical integration of AI into NDT has led to significant advancements across various industries, enhancing inspection efficiency, accuracy, and predictive maintenance. This section explores several real-world applications demonstrating how AI-enhanced NDT exploration techniques are transforming traditional practices.
In the USA, bridge inspections have benefited greatly from AI integration. Traditional inspections are labor-intensive, time-consuming, and often pose safety risks due to the need for inspectors to access hard-to-reach areas. By employing unmanned aerial vehicles equipped with high-resolution cameras and AI algorithms, inspections have become more efficient and safer. AI models analyze captured images to detect the defects. This AI-powered approach accelerates the inspection process, enhances accuracy by identifying defects that might be missed during manual inspections, and facilitates predictive maintenance through detailed structural health records over time [23].
Another sector benefiting from AI integration is the oil and gas industry, ensuring the integrity of pipelines is critical to prevent leaks and environmental hazards. Ultrasonic Testing (UT) is commonly used to detect internal flaws, but interpreting UT data requires expert knowledge and can be time-consuming. An energy company implemented an AI system to automate the analysis of UT data collected from pipeline inspections. ML models trained on historical datasets recognize patterns associated with defects like corrosion, cracks, and weld anomalies. This results in faster inspections, reduced costs, and improved detection rates. Integrating AI analysis with predictive maintenance programs allows for proactive scheduling of repairs, reducing the risk of pipeline failures and associated environmental impacts [24].
In the renewable energy sector, wind turbine blade inspections have been revolutionized by AI-driven NDT solutions. Traditional methods involve manual visual inspections, which are risky and time-consuming. Renewable energy companies have adopted AI-based solutions using drones equipped with high-resolution cameras and thermal imaging sensors. Collected data are processed using AI algorithms capable of detecting surface cracks, delamination, etc. ML models analyze historical inspection data to develop predictive maintenance schedules. This proactive approach reduces unexpected downtime and maintenance costs while increasing the reliability of energy production [25].
Digital twin technology is transforming structural health monitoring. Smart building projects are incorporating AI-enhanced NDT data into digital twins to monitor the health of high-rise structures. Sensors embedded in buildings collect data on vibrations, strains, and other parameters. AI algorithms analyze these data in real time, detecting anomalies and predicting potential structural issues. The digital twin provides a virtual representation of the building’s condition, allowing engineers to visualize and assess the impact of various stressors. This facilitates data-driven decision-making, optimizes maintenance strategies, and enhances the safety and resilience of structures [26].
These case studies collectively demonstrate the significant impact of integrating AI into NDT practices across diverse sectors. Those solutions offer improved defect detection accuracy, increased efficiency, predictive maintenance capabilities, and enhanced safety. By automating data analysis and leveraging ML models, companies can optimize maintenance strategies. This leads to cost reduction and extends the lifespan of critical infrastructure and components. The successful application of AI in these cases highlights its transformative potential in redefining traditional NDT and advancing industry standards.
After implementation, considerations for computational resources are crucial. For example, the AI algorithms used for pipeline inspections are deployed on high-performance computing platforms that can handle large volumes of ultrasonic testing data. Integration with existing NDT tools is achieved through customized software interfaces and data formats, ensuring seamless operation within established workflows. This practical approach demonstrates how AI methods can be effectively applied in real-world scenarios.

4. Benefits of Implementing AI in NDT

4.1. Enhanced Accuracy and Reliability

AI algorithms reduce the subjectivity and inconsistency associated with human inspection by providing consistent and repeatable results. By learning from large datasets, these systems improve their accuracy over time, leading to better detection. Automated systems also mitigate the risk of human error, which is particularly critical in high-stakes environments such as nuclear plants, bridges, and aerospace components, where even minor flaws can have serious consequences.

4.2. Increased Efficiency and Cost Savings

AI-driven analysis can process data faster than traditional methods, significantly reducing inspection times. This leads to quicker decision-making and minimizes the disruption caused by lengthy inspection procedures. For instance, AI-based systems can analyze thousands of images or signals in a fraction of the time it would take for a human inspector, making them ideal for large-scale projects where timely assessments are crucial. AI-powered solutions reduce costs by predicting zones with high damage risk, helping to prevent failures and optimize maintenance schedules. This efficiency minimizes operational costs and emergency repairs. By targeting areas that require attention rather than adhering to fixed schedules, organizations, particularly in critical sectors like transportation and energy, can achieve significant long-term savings while avoiding delayed repairs.

4.3. Scalability

AI can easily scale to handle large and complex datasets, making them suitable for use in extensive infrastructure projects where traditional inspection methods would be too resource-intensive. For example, in large-scale bridge inspection projects, AI systems can process data from thousands of sensors and multiple NDT methods simultaneously, providing comprehensive and detailed assessments of structural health. Scalability also extends to automated inspection systems, such as drones or robotic devices equipped with AI-powered sensors. To address computational requirements for processing large-scale data, AI systems leverage parallel computing and cloud-based platforms. Technologies like distributed computing frameworks (e.g., Apache Hadoop or Spark) enable the handling of vast amounts of NDT data from multiple sources.

4.4. Integration with Digital Twins

Digital twin technology—virtual replicas of physical structures—represents a major advancement in infrastructure management. By integrating NDT data with digital twins, AI systems can monitor structures in real-time and provide insights based on historical data. This enables a shift from periodic inspections to continuous monitoring, ensuring that potential issues are detected early.

4.5. Enhanced Safety

The use of AI in NDT notably improves safety by minimizing the need for human direct intervention. Autonomous drones and robots can inspect areas that are difficult or dangerous for human inspectors to access, such as high-rise structures, underwater pipelines, or nuclear facilities. Also, AI models can analyze data from sensors to detect early signs of deterioration or failure, enabling timely intervention before conditions become unsafe. By finding risks early, AI-enhanced NDT helps prevent accidents and reduces the likelihood of catastrophic failures.

5. Challenges in Integrating AI in NDT

While the benefits of integrating AI into NDT are clear, several challenges must be addressed to fully realize their potential.

5.1. Data Quality and Availability

The performance of AI models relies on the availability of high-quality labeled data. In many cases, NDT data are inconsistent, making it difficult to train accurate models. For example, data from GPR testing may contain artifacts or be influenced by environmental factors, leading to inaccurate interpretations. Building comprehensive datasets requires investment in data collection and preprocessing. Also, the need for labeled data can be a bottleneck, as creating high-quality labels often requires expert knowledge. Addressing these challenges requires collaboration between AI specialists and NDT experts to develop standardized datasets and methodologies.
To mitigate the impact of poor-quality or insufficient data, data augmentation and synthetic data generation techniques are employed. For example, generative adversarial networks can create realistic synthetic NDT data to supplement limited datasets. Furthermore, transfer learning allows AI models to leverage knowledge from related domains, enhancing performance even when data are scarce. Robust data preprocessing pipelines are essential to clean and standardize data before they are used for training AI models.

5.2. Skill Gaps and Workforce Training

The integration of AI into NDT necessitates a workforce with skills in both data science and traditional NDT practices. This requires substantial investment in training and upskilling current employees to effectively use AI tools and interpret their results. To bridge the skill gap, organizations must invest in training programs that cover both the technical aspects of AI and the practical application of NDT. Collaborative efforts between academia, industry, and professional organizations can also help develop specialized training programs that address the unique needs of AI-enhanced NDT.

5.3. Ethical and Operational Challenges

AI models can sometimes act as “black boxes”, making it difficult to interpret their decision-making processes. This raises concerns regarding transparency especially when critical safety decisions are at stake. For instance, in cases where AI models predict potential structural failures, engineers must understand the reasoning behind these predictions to make informed decisions. Moreover, biases in the training data can lead to inaccurate outcomes, requiring careful monitoring and validation of AI systems. Operational challenges, such as integrating AI systems with existing NDT tools and workflows, also need to be carefully managed to ensure seamless adoption.
The ‘black box’ nature of complex AI models, particularly deep learning networks, poses ethical concerns in safety-critical applications. Lack of transparency can hinder the acceptance of AI decisions in industries where accountability is paramount. To address this, explainable AI (XAI) techniques are being developed to provide insights into the model’s decision-making process. For instance, methods like Layer-wise Relevance Propagation (LRP) or SHAP values can highlight which inputs most influenced the AI’s conclusions, aiding in validation and trust-building among stakeholders.

5.4. Regulatory and Standardization Issues

The adoption of AI in NDT is still in its early stages, and regulatory frameworks have not yet fully adapted to these advancements. Industry standards and guidelines for AI-driven NDT methods are needed to ensure consistency, reliability, and safety. Regulatory bodies must work closely with industry stakeholders to develop standards that accommodate the use of AI. Standardization is particularly important in critical industries such as nuclear energy and aerospace, where the consequences of failure can be catastrophic.
Developing universally accepted protocols and guidelines for AI-driven NDT will be key to fostering widespread adoption and ensuring the safe and effective use of these technologies. The lack of established regulatory frameworks and industry standards for AI applications in NDT presents a significant hurdle. Organizations such as ISO and ASTM are beginning to explore guidelines for AI integration, but comprehensive standards are still in development. Compliance with existing safety and quality regulations must be maintained, requiring AI systems to undergo rigorous validation and certification processes. Collaboration between regulatory bodies, industry experts, and researchers is essential to develop standards that ensure the effective use of AI in NDT.

6. Outlook for Integrating AI into NDT

The future of AI in NDT is promising, with several advancements on the horizon.

6.1. Integration with Digital Twin Technology

The concept of digital twins—virtual models that replicate physical assets in real time—is gaining traction in the NDT industry. AI can play a crucial role in analyzing data from digital twins, providing real-time insights into asset health and enabling predictive maintenance. Digital twins, coupled with AI-driven NDT systems, can simulate the impact of different scenarios on infrastructure, allowing for better planning and risk management. For example, in smart cities, digital twins of critical infrastructure such as bridges, tunnels, and water distribution systems can be continuously updated with real-time NDT data. AI models can analyze this data to predict maintenance needs, optimize resource allocation, and enhance overall safety. The combination of AI, digital twins, and advanced NDT techniques will redefine how infrastructure is monitored and maintained.

6.2. Edge Computing and IoT-Enabled NDT

Edge computing brings computation and data storage closer to the location where it is needed, improving response times and saving bandwidth. In the context of NDT, integrating AI with edge computing and Internet of Things (IoT) devices means that data collected by sensors can be analyzed immediately on-site. For example, a smart sensor installed on a bridge can process vibration data in real time to detect anomalies, without needing to send all the data to a central server. This enables faster decision-making and reduces the amount of data that needs to be transmitted.
This reduces latency and enhances the responsiveness of NDT processes, particularly in remote or challenging environments. Edge computing enables AI models to process data locally, without the need for cloud connectivity, making it possible to deploy AI-enhanced NDT systems in areas with limited internet access. IoT-enabled NDT devices equipped with AI algorithms can continuously monitor the health of infrastructure assets and send alerts when anomalies are detected. For example, sensors embedded within a concrete structure could continuously monitor parameters such as moisture content, temperature, and electrical resistivity. AI models can analyze these data in real time to detect early signs of deterioration and recommend corrective actions.

6.3. AI-Powered Autonomous Inspection Systems

The development of autonomous drones and robotic systems equipped with AI capabilities is set to transform the way inspections are conducted. These systems can perform inspections in hazardous or hard-to-reach areas, reducing the need for human intervention while increasing safety and efficiency. For example, autonomous drones can be used to inspect the exterior of high-rise buildings, bridges, or offshore wind turbines, providing detailed imagery and data for analysis. AI-powered robots are also being developed for the inspection of underground pipelines, nuclear facilities, and other critical infrastructure. These robots can navigate complex environments and perform NDT using techniques such as ultrasonic testing, GPR, and magnetic flux leakage. By integrating AI into autonomous systems, inspections can be conducted more frequently, providing a higher level of monitoring and reducing the risk of unexpected failures.

6.4. Advanced AI Algorithms for Multi-Modal NDT Data Fusion

As NDT techniques generate diverse types of data, ranging from images and signals to 3D models, there is a growing need for advanced AI algorithms capable of fusing multi-modal data. Data fusion techniques allow for the integration of data from multiple NDT methods, providing a more comprehensive and accurate assessment of structural integrity. For example, AI models can combine ultrasonic, radiographic, and electrical resistivity data to create a unified representation of a structure’s health. The development of deep learning models specifically designed for multi-modal data fusion will be a key area of research in the coming years. These models will enable engineers to gain deeper insights into the condition of infrastructure assets and make more informed decisions. As AI algorithms become more sophisticated, they will be able to extract even more valuable information from complex NDT datasets.

7. Conclusions and Prospects

The integration of AI into non-destructive testing is transforming the industry by enhancing the accuracy, efficiency, and predictive capabilities of inspections. From automated defect detection and predictive maintenance to real-time data analysis and autonomous inspection systems, AI is revolutionizing how infrastructure is monitored and maintained. As AI continues to evolve, its role in NDT will expand, enabling more sophisticated and scalable inspection solutions.
By addressing challenges related to data quality, ethical considerations, and regulatory compliance, the engineers can fully leverage AI’s potential in NDT. Collaboration between human expertise and AI-driven analysis is vital for advancing infrastructure safety and reliability, leading to smarter, more sustainable management practices.
In conclusion, integrating AI into NDT offers significant benefits, including enhanced accuracy, efficiency, and predictive capabilities. By addressing challenges related to data quality, ethical considerations, and regulatory compliance, the engineering community can fully leverage AI’s potential in NDT. The collaboration between human expertise and AI-driven analysis is vital for advancing infrastructure safety and reliability, paving the way for smarter, more sustainable management practices. By embracing the possibilities of AI, the field of NDT can achieve new levels of performance, paving the way for smarter, more sustainable infrastructure management.

Author Contributions

Conceptualization, A.e.M.S.; validation, S.C.B.; investigation, A.e.M.S.; writing—original draft preparation, A.e.M.S.; writing—review and editing, G.C.K. and S.C.B.; visualization, A.e.M.S. All authors have read and agreed to the published version of the manuscript.

Funding

This research project did not receive any kind of grant from any source.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The data in this article were derived from the published literature.

Acknowledgments

The authors are thankful to the reviewers for helping to improve this paper.

Conflicts of Interest

All authors are employees of Englobe Corp., and declare no conflicts of interest. The opinions and conclusions expressed in this paper are solely those of the authors and do not necessarily represent the official policies or positions of the affiliated institutions.

References

  1. Ottosen, L.M.; Kunther, W.; Ingeman-Nielsen, T.; Karatosun, S. Non-Destructive Testing for Documenting Properties of Structural Concrete for Reuse in New Buildings: A Review. Materials 2024, 17, 3814. [Google Scholar] [CrossRef] [PubMed]
  2. Azimi, M.; Eslamlou, A.D.; Pekcan, G. Data-Driven Structural Health Monitoring and Damage Detection through Deep Learning: State-of-the-Art Review. Sensors 2020, 20, 2778. [Google Scholar] [CrossRef] [PubMed]
  3. Taheri, H.; Gonzalez Bocanegra, M.; Taheri, M. Artificial Intelligence, Machine Learning and Smart Technologies for Nondestructive Evaluation. Sensors 2022, 22, 4055. [Google Scholar] [CrossRef] [PubMed]
  4. Sarkar, K.; Shiuly, A.; Dhal, K.G. Revolutionizing Concrete Analysis: An in-Depth Survey of AI-Powered Insights with Image-Centric Approaches on Comprehensive Quality Control, Advanced Crack Detection and Concrete Property Exploration. Constr. Build. Mater. 2024, 411, 134212. [Google Scholar] [CrossRef]
  5. Zhang, C.; Yan, Q.; Zhang, Y.; Liao, X.; Xu, G.; He, Z. Nondestructive Detection of Fiber Content in Steel Fiber Reinforced Concrete through Percussion Method Coordinated with a Hybrid Deep Learning Network. J. Build. Eng. 2024, 86, 108857. [Google Scholar] [CrossRef]
  6. Alavi, S.A.; Noel, M.; Moradi, F.; Layssi, H. Development of a Machine Learning Model for On-Site Evaluation of Concrete Compressive Strength by SonReb. J. Build. Eng. 2024, 82, 108328. [Google Scholar] [CrossRef]
  7. Safhi, A.e.M.; Mejjad, N.; El FadilI, H.; Bortali, M. Dredged Materials in Morocco: Current Practices, Policies, and Roadmap for Sustainable Management. Case Stud. Constr. Mater. 2024, 20, e03045. [Google Scholar] [CrossRef]
  8. Safhi, A.e.M.; Dabiri, H.; Soliman, A.; Khayat, K.H. Prediction of Self-Consolidating Concrete Properties Using XGBoost Machine Learning Algorithm: Part 1–Workability. Constr. Build. Mater. 2023, 408, 133560. [Google Scholar] [CrossRef]
  9. Benaicha, M. AI-Driven Prediction of Compressive Strength in Self-Compacting Concrete: Enhancing Sustainability through Ultrasonic Measurements. Sustainability 2024, 16, 6644. [Google Scholar] [CrossRef]
  10. Rahman, M.R.; Tabassum, S.; Haque, E.; Nishat, M.M.; Faisal, F.; Hossain, E. CNN-Based Deep Learning Approach for Micro-Crack Detection of Solar Panels. In Proceedings of the 2021 3rd International Conference on Sustainable Technologies for Industry 4.0 (STI), Dhaka, Bangladesh, 18–19 December 2021; pp. 1–6. [Google Scholar]
  11. Yang, F.; Huo, J.; Cheng, Z.; Chen, H.; Shi, Y. An Improved Mask R-CNN Micro-Crack Detection Model for the Surface of Metal Structural Parts. Sensors 2024, 24, 62. [Google Scholar] [CrossRef]
  12. Xu, G.; Yue, Q.; Liu, X. Deep Learning Algorithm for Real-Time Automatic Crack Detection, Segmentation, Qualification. Eng. Appl. Artif. Intell. 2023, 126, 107085. [Google Scholar] [CrossRef]
  13. Ding, Z.; Luo, L.; Wang, X.; Liu, Y.; Zhang, W.; Wu, H. An Artificial Intelligence-Based Method for Crack Detection in Engineering Facilities around Subways. Appl. Sci. 2023, 13, 11002. [Google Scholar] [CrossRef]
  14. Laxman, K.C.; Tabassum, N.; Ai, L.; Cole, C.; Ziehl, P. Automated Crack Detection and Crack Depth Prediction for Reinforced Concrete Structures Using Deep Learning. Constr. Build. Mater. 2023, 370, 130709. [Google Scholar] [CrossRef]
  15. Yan, Q.; Zhang, Y.; Liao, X.; Xu, Y.; Zhang, C.; Liu, X.; Zhang, Z. Detection and Diagnosis of Concrete Void Defect Using Percussion-Based Method Combined with Convolutional Neural Network. Measurement 2024, 231, 114585. [Google Scholar] [CrossRef]
  16. Dong, W.; Wu, Z.; Zhou, X.; Tan, Y. Experimental Studies on Void Detection in Concrete-Filled Steel Tubes Using Ultrasound. Constr. Build. Mater. 2016, 128, 154–162. [Google Scholar] [CrossRef]
  17. Soleimanpour, R.; Yassin, M.H.; Mohammad, N.K.; Bo Arki, M.K.; Sweid, M.N. Detecting Debonding in FRP Retrofitted Concrete Beams Using Nonlinear Ultrasonic Waves. In Proceedings of the 8th International Conference on Civil Engineering, Singapore, 22–24 March 2024; Strauss, E., Ed.; Springer Nature: Singapore, 2025; pp. 113–125. [Google Scholar]
  18. Polverino, L.; Abbate, R.; Manco, P.; Perfetto, D.; Caputo, F.; Macchiaroli, R.; Caterino, M. Machine Learning for Prognostics and Health Management of Industrial Mechanical Systems and Equipment: A Systematic Literature Review. Int. J. Eng. Bus. Manag. 2023, 15, 18479790231186848. [Google Scholar] [CrossRef]
  19. Cuesta, J.; Leturiondo, U.; Vidal, Y.; Pozo, F. Challenges on Prognostics and Health Management for Wind Turbine Components. J. Phys. Conf. Ser. 2024, 2745, 012003. [Google Scholar] [CrossRef]
  20. Song, H.; Sun, Y.; Gao, H.; Guo, L.; Wu, T. A Novel HI Construction Method Based on Healthy-State Data Training for Rotating Machinery Components. Struct. Health Monit. 2024, 14759217241279784. [Google Scholar] [CrossRef]
  21. Saback, V.; Popescu, C.; Blanksvärd, T.; Täljsten, B. Asset Management of Existing Concrete Bridges Using Digital Twins and BIM: A State-of-the-Art Literature Review. Nord. Concr. Res. 2022, 66, 91–111. [Google Scholar] [CrossRef]
  22. Varevac, D.; Guljaš, I.; Ištoka Otković, I.; Radočaj, D. Challenges in Ground-Penetrating Radar Application in Structural Elements: Determination of the Dielectric Constant of Glued Laminated Timber Case Study. Electronics 2024, 13, 3718. [Google Scholar] [CrossRef]
  23. Kulhandjian, H. AI-Based Bridge and Road Inspection Framework Using Drones; San Jose State University, College of Business, Mineta Transportation Institute: San Jose, CA, USA, 2023. [Google Scholar]
  24. Yao, J.; Liang, W.; Xiong, J. Novel Intelligent Diagnosis Method of Oil and Gas Pipeline Defects with Transfer Deep Learning and Feature Fusion. Int. J. Press. Vessels Pip. 2022, 200, 104781. [Google Scholar] [CrossRef]
  25. Wang, L.; Zhang, Z. Automatic Detection of Wind Turbine Blade Surface Cracks Based on UAV-Taken Images. IEEE Trans. Ind. Electron. 2017, 64, 7293–7303. [Google Scholar] [CrossRef]
  26. Koutamanis, A. From Building Information Modelling to Digital Twins: Digital Representation for a Circular Economy. In A Circular Built Environment in the Digital Age; De Wolf, C., Çetin, S., Bocken, N.M.P., Eds.; Springer International Publishing: Cham, Switzerland, 2024; pp. 3–20. ISBN 978-3-031-39675-5. [Google Scholar]
Figure 1. Keyword co-occurrence network of ‘non-destructive testing in concrete’ research generated using VOSviewer 1.6.20, illustrating relationships between key topics in NDT and highlighting areas AI can significantly impact.
Figure 1. Keyword co-occurrence network of ‘non-destructive testing in concrete’ research generated using VOSviewer 1.6.20, illustrating relationships between key topics in NDT and highlighting areas AI can significantly impact.
Encyclopedia 04 00116 g001
Figure 2. The role of AI in NDT.
Figure 2. The role of AI in NDT.
Encyclopedia 04 00116 g002
Figure 3. Data visualization recorded from a GPR scan (radargram) after Time-to-Depth conversion, bandpass filtering and Hilbert transformation—600 MHz antenna [22].
Figure 3. Data visualization recorded from a GPR scan (radargram) after Time-to-Depth conversion, bandpass filtering and Hilbert transformation—600 MHz antenna [22].
Encyclopedia 04 00116 g003
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Safhi, A.e.M.; Keserle, G.C.; Blanchard, S.C. AI-Driven Non-Destructive Testing Insights. Encyclopedia 2024, 4, 1760-1769. https://doi.org/10.3390/encyclopedia4040116

AMA Style

Safhi AeM, Keserle GC, Blanchard SC. AI-Driven Non-Destructive Testing Insights. Encyclopedia. 2024; 4(4):1760-1769. https://doi.org/10.3390/encyclopedia4040116

Chicago/Turabian Style

Safhi, Amine el Mahdi, Gilberto Cidreira Keserle, and Stéphanie C. Blanchard. 2024. "AI-Driven Non-Destructive Testing Insights" Encyclopedia 4, no. 4: 1760-1769. https://doi.org/10.3390/encyclopedia4040116

APA Style

Safhi, A. e. M., Keserle, G. C., & Blanchard, S. C. (2024). AI-Driven Non-Destructive Testing Insights. Encyclopedia, 4(4), 1760-1769. https://doi.org/10.3390/encyclopedia4040116

Article Metrics

Back to TopTop