Structural Health Monitoring: Latest Applications and Data Analysis

Structural health monitoring (SHM) is focused on the systematic monitoring of the state of structures to determine their safety, dependability, and durability. The process entails the utilization of diverse sensing technologies, data analysis methodologies, and computational models to consistently evaluate the structural soundness, identify any impairments, and anticipate potential malfunctions. The utilization of SHM is of essential importance in various industries such as aerospace, civil infrastructure, energy, and transportation. Its significance lies in the timely identification of damage or deterioration, which can help prevent catastrophic events, improve maintenance strategies, and reduce operational expenses.

In recent years, the field of SHM has witnessed significant advancements in multiple areas of research. The development of SHM systems is taking place according to many aspects, including advances in known or introduction of new materials, as well as the fabrication of complex structures. Additionally, it is driven by the need to elevate safety protocols. These factors affect several important domains of growth in SHM systems. Generally, a primary concern has been the advancement of novel measurement methodologies that facilitate the enhancement of data quality and the exploration of material characteristics through innovative means. The development of new technologies, on the other hand, also entails the need to develop new materials for the construction of elements constituting sensory systems. An essential aspect of modern systems pertains to the integration of complex data analysis algorithms that facilitate the identification of interdependencies and the analysis of large data sets.

Referring to the first-mentioned aspect—advancement of novel measurement methodologies, which involves the development of new sensors, measurement techniques, and miniaturization technologies—the authors featured in the presented Special Issue introduce innovative measurement techniques that improve the accuracy of material testing and evaluation. In [1], the detection of real corrosion cracks using sweep frequency eddy-current frequency–response analysis, aiming to increase the probability of crack detection compared to conventional sweep frequency methods, is investigated. A new approach incorporating innovative solutions and a multi-point detection method was proposed and tested on austenitic steel plate specimens. The results demonstrate that utilizing a multi-point detection method and additional mathematical processing of signals enhances the detection ability of sweep frequency excitation for real corrosion cracks. Paper [2] addresses the challenge of determining cable forces in bridges, emphasizing the importance of accurate assessment for evaluating the performance of structures. The proposed method presented in this study allows for the direct identification of cable tensions in a two-cable network without removing cross-ties, resulting in precise estimates with an error of less than 1.0 percent, saving time and improving accuracy. Assessment of the condition of building structure elements has become the subject of another work. In study [3], authors propose an innovative methodology using microwave thermography to determine the concrete cover thickness in a reinforced concrete wall. By employing the transmission approach with multiple angles and applying physical principles, such as the Snell–Descartes law, the proposed method accurately estimates the concrete cover thickness, spacing of steel bars, and dielectric constant of the concrete, achieving results close to the real values. Additionally, the detection of concrete cover thickness is highlighted as a notable achievement of infrared thermography methods.

Alongside the hardware aspect mentioned in the previous paragraph, there has been a notable focus on data processing and fusion. In particular, the integration of artificial intelligence techniques has emerged as a crucial element in SHM. Today, this area is one of the main branches of work on current control systems, and in the context of the rapid development of artificial intelligence algorithms, it will be an integral part of practical measurement systems. Advanced analysis methods make it possible to search for non-obvious and complex relationships and correlations between data, which allows for achieving ever-higher efficiency and accuracy rates of newly developed systems and technologies. This modern trend is also visible in the articles published in this Special Issue relating to a wide range of areas: from assessing the condition of operating building structures of various sizes to testing the properties and state of materials. In [4], the verification of the random forest algorithm for non-invasive assessment of excessively damp and salty historical brick walls is presented. The developed method, which provides quantitative assessment, was validated using archival research and experimental tests in two historical buildings. The results showed satisfactory linear correlation coefficients and low absolute and relative errors, confirming the effectiveness of the random forest algorithm for evaluating the moisture content of brick walls without destructive intervention. Advanced data processing techniques are also used in the research presented in article [5], where the authors emphasize the importance of continuous and reliable dam safety monitoring data for the evaluation and control of dam operation and security. The article proposes a cokriging spatial model based on variable importance for data repair, considering the spatial and environmental factors that influence dam safety. This model aims to improve the accuracy of data repair and ensure that monitoring data reflect the actual safety conditions of the dam. In article [6], authors investigate the use of magnetic Barkhausen noise MBN combined with classical machine learning methods to assess the grade and magnetic anisotropic properties of electrical sheets. The study compares the performance of classical machine learning and deep learning approaches using 26 algorithms. The results show that while the highest accuracy was achieved by machine learning models based on artificial neural networks and ensemble bagged trees, the deep learning model consistently outperformed classical machine learning methods by approximately 10% in terms of accuracy. A neural network approach is also utilized in [7]—this paper addresses the challenge of detecting hidden corrosion at rivet sites in aircraft structures. The proposed approach combines machine learning and electromagnetic testing to effectively identify the corrosion of various sizes and locations. The results demonstrate that the approach can detect 89.48% of hidden corrosion with a training data utilization of only 20%, which can be further improved to 99.0% with 60–80% of the training data. The results presented in the aforementioned works confirm the need to take into account artificial intelligence algorithms in the implementation of newly built SHM systems, and the wide range of presented applications indicates the large scale and importance of the problem in the context of future development.

Aside from the previously mentioned aspects, the authors of this Special Issue placed significant emphasis on subjects pertaining to the advancement of novel materials and methodologies for their fabrication. In [8], the authors discuss the ultrasonic testing of Ti6Al4V material produced using conventional and laser bed fusion methods, with a focus on additive manufacturing. Additive manufacturing, due to a different production process to standard methods, can significantly affect the properties of the materials produced, which determines the need for a different approach to the utilized monitoring techniques. The study aims to determine the acoustic properties of the material and develop an inspection scheme for adhesive joints in additive manufacturing components. The results reveal higher decibel drops in amplitude for connections made with additive technology compared to conventional methods, providing valuable insights for testing adhesive joints in additive manufacturing parts. The joints examination was also a subject of article [9], where inseparable joints in machine and vehicle construction were investigated, particularly focusing on hybrid joints combining adhesive bonding and sheet metal clinching. The results showed that sheet metal clinching significantly increased the joint strength, with shear strengths of 965 N for clinching and 476 N for adhesive bonding. The hybrid joints exhibited even higher strengths, with average forces of 1085 N for fully cured joints and 1486 N for joints made immediately after adhesive application, indicating the improved strength of the hybrid joint due to stabilization and better crosslinking conditions. Modern materials, creating the possibility of changing properties as a result of the influence of external factors and reconfiguration of their structure, became the subject of another work published in the Special Issue. Paper [10] explores the evaluation of thin dielectric layers in the THz frequency range using a tunable split-ring resonator-based metasurface. By changing the geometry of the metasurface’s unit cells, the resonant frequency can be varied, allowing for the assessment of a nearby thin dielectric layer. The study utilizes finite element method simulations and establishes relationships between resonant frequencies and dielectric parameters, providing more comprehensive information for the evaluation of the material’s permittivity compared to non-tunable metasurfaces.

In conclusion, topics presented in the Special Issue titled “Structural Health Monitoring: Latest Applications and Data Analysis” cover a wide range of areas not only related to hardware solutions, including the implementation of sensing layers and new solutions in material technology, but also, to a large extent, to the use of advanced data science algorithms. The works cover a variety of applications, including nondestructive assessment of historic brick walls, monitoring of dam safety, evaluation of electrical sheets using magnetic Barkhausen noise, corrosion detection in aircraft structures, ultrasonic testing of additively manufactured components, monitoring of hybrid joints in mechanical engineering, study of tunable split-ring resonator-based metasurfaces for thin-film dielectric structures assessment, detecting real corrosion cracks, determining the strength of tendons in bridges, and assessing the thickness of concrete cover using microwave thermography. The papers present new and novel methodologies, advanced and complex data processing techniques, and original measurement approaches to increase the accuracy and reliability of structural health monitoring and material evaluation. These works confirm the upcoming trends in the development of SHM technologies, aimed at the development of the hardware layer by introducing new materials and methods of acquisition, but above all, advanced analysis of data. Bearing in mind the existing needs of the industry and the need for multi-threaded system operation and online analysis of the examined objects state, continuous and increasing progress in the SHM area should be expected in the coming years.