Machine learning and structural health monitoring overview with emerging technology and high-dimensional data source highlights

Abstract

Conventional damage detection techniques are gradually being replaced by state-of-the-art smart monitoring and decision-making solutions. Near real-time and online damage assessment in structural health monitoring (SHM) systems is a promising transition toward bridging the gaps between the past’s applicative inefficiencies and the emerging technologies of the future. In the age of the smart city, Internet of Things (IoT), and big data analytics, the complex nature of data-driven civil infrastructures monitoring frameworks has not been fully matured. Machine learning (ML) algorithms are thus providing the necessary tools to augment the capabilities of SHM systems and provide intelligent solutions for the challenges of the past. This article aims to clarify and review the ML frontiers involved in modern SHM systems. A detailed analysis of the ML pipelines is provided, and the in-demand methods and algorithms are summarized in augmentative tables and figures. Connecting the ubiquitous sensing and big data processing of critical information in infrastructures through the IoT paradigm is the future of SHM systems. In line with these digital advancements, considering the next-generation SHM and ML combinations, recent breakthroughs in (1) mobile device-assisted, (2) unmanned aerial vehicles, (3) virtual/augmented reality, and (4) digital twins are discussed at length. Finally, the current and future challenges and open research issues in SHM-ML conjunction are examined. The roadmap of utilizing emerging technologies within ML-engaged SHM is still in its infancy; thus, the article offers an outlook on the future of monitoring systems in assessing civil infrastructure integrity.

Concise Historic Overview of Strain Sensors Used in the Monitoring of Civil Structures: The First One Hundred Years

“…An important research direction that is also currently being explored is full understanding of data collected by currently existing strain sensing techniques and the assessment of capabilities of strain measurements to be used for the prediction of future structural behaviors using data-driven approaches (statistics and machine learning, e.g., [ 126 , 127 ]). While this direction does not directly address the development of sensors, it is extremely important as it emphasizes the importance of sensors beyond simple strain measurement and has the potential of enabling smart structures and transforming them into cyber-physical systems (e.g., [ 128 ]).…”

A Review of Machine Learning Methods Applied to Structural Dynamics and Vibroacoustic

“…The book by Farrar and Worden [28] presents an in-depth analysis of all aspects of SHM related to ML, including main applications, data collection and processing, and ML algorithms. More recent surveys on SHM driven by ML were carried out by Khan and Yairi [29], Azimi et al [30], Toh and Park [31], Lin et al [32], Bao and Li [33], Malekloo et al [34], Lei et al [247] and Zhao et al…”

“…Varanis and Pederiva [252] compares some feature selection methods and concludes that Linear Discriminant Analysis (LDA) is suitable for non-stationary cases, while Principal component analysis (PCA) is convenient for stationary signals and independent component analysis (ICA) for problems with combined faults. Malekloo et al [34] also reviews several supervised and unsupervised feature selection methods. The classical ML approach in SHM is addressed by Worden and Manson [260] for damage detection, localization, and assessment problems.…”

“…Malekloo et al [34] provides a complete review of all processes that should be considered in an ML-based SHM analysis, from the excitation source, going to data acquisition, data normalization, data cleaning, data compression, feature extraction and selection, data fusion, and the prediction. Hereafter, some ML-based SHM works using vibroacoustic signals are presented to exemplify applications in damage detection and diagnosis -including damage location, extent, and typeand in the prognosis of RUL and mission planning.…”

A Novel Baseline-Free Damage Detection Method Based on Path Scanning of Lamb Waves Using Mobile Transducers

“…Meanwhile, the corresponding signal processing methods such as wavelet transform and empirical mode decomposition has also been substantially studied and applied [ 4 ]. Recently, the application of smartphones, high-resolution cameras, unmanned aerial vehicles, and other non-contact sensing technologies gained prominent growth in SHM due to their advantages of low labor cost, high applicative efficiency, and so forth [ 5 ]; in line with these advancements, a myriad of machine learning and deep learning-based algorithms, typified by support vector machine, Gaussian mixture, convolutional neural network, and long short-term memory network, have been studied and proposed to process sheer amount and various dimensions of data collected, as well as to provide intelligent solutions for conventional damage detection methods [ 6 , 7 ].…”