Diagnosis of internal cracks in concrete using vibro-acoustic modulation and machine learning

This paper investigates the utility of physics-informed machine learning models for vibro-acoustic modulation (VAM)–based damage localization in concrete structures. Vibro-acoustic modulation is a nonlinear dynamics-based non-destructive testing method, which was initially developed to perform damage detection and later extended to accomplish damage localization. The VAM-based damage (hidden crack) diagnosis is performed by analyzing the damage index pattern on the surface of the component to arrive at the size and location of the hidden damage. Past investigations have employed heuristically selected damage index thresholds as well as computationally expensive Bayesian estimation methods for VAM-based damage localization in two (surface) dimensions. Compared to these studies, the proposed methodology automates the threshold selection (algorithmic instead of heuristic), increases the speed of the probabilistic damage diagnosis process, and enables the estimation of damage depth. We generate training data (damage index) for the machine learning models using the pertinent nonlinear dynamics (finite element) models using different combinations of test parameters. The (supervised) machine learning models are thus informed by computational physics models. These include two types of artificial neural network (ANN) models: classification models that identify whether a sensor location is damaged or not and regression models that enable Bayesian estimation to obtain the posterior probability distribution of damage location and size. The accuracy of machine learning-based diagnosis is evaluated using both numerical and laboratory experiments. The proposed physics-informed machine learning models for VAM-based damage diagnosis are able to achieve an accuracy of about 60–64% in the validation experiments, indicating the potential of these methods for internal crack detection. The results show that for complex (nonlinear dynamics-driven) diagnostic methods, damage index patterns learned from physics models could be successfully used for damage detection as well as localization.

Internal Damage Analysis of Braided Composites Embedded in Carbon Nanotube Yarn

Carbon nanotube (CNT) yarn sensors were embedded in 3D braided composites in the form of arrays to detect the internal damage of specimens and study the internal damage monitoring of the 3D braided composites. The signals collected by the sensor array of CNT yarn were preprocessed using the dynamic wavelet threshold algorithm. The exact position of the damage was calculated based on the main features of the resistance signal matrix, which was calculated using the quadratic matrix singular value. The results show that the internal damage localization of the specimens was consistent with the actual damage. The localizations in this study can provide a basis for enhancing the structural health monitoring of smart 3D braided composites.

Potential of Neural Networks for Structural Damage Localization

<p>Fabrication technology and structural engineering states-of-art have led to a growing use of slender structures, making them more susceptible to static and dynamic actions that may lead to some sort of damage. In this context, regular inspections and evaluations are necessary to detect and predict structural damage and establish maintenance actions able to guarantee structural safety and durability with minimal cost. However, these procedures are traditionally quite time-consuming and costly, and techniques allowing a more effective damage detection are necessary. This paper assesses the potential of Artificial Neural Network (ANN) models in the prediction of damage localization in structural members, as function of their dynamic properties – the three first natural frequencies are used. Based on 64 numerical examples from damaged (mostly) and undamaged steel channel beams, an ANN-based analytical model is proposed as a highly accurate and efficient damage localization estimator. The proposed model yielded maximum errors of 0.2 and 0.7 % concerning 64 numerical and 3 experimental data points, respectively. Due to the high-quality of results, authors’ next step is the application of similar approaches to entire structures, based on much larger datasets.</p>

Surface-Mounted Smart PZT Sensors for Monitoring Damage Using EMI-Based Multi-Sensing Technique

In this article, a multi-sensing technique on surface-mounted PZT sensors is proposed. The investigation was performed on concrete structures to detect and localize structural damage. Multiple smart sensing units (SSU) were adhesively bonded on the top surface of a concrete beam. As each PZT sensor features a small zone of influence, the use of multiple smart sensors is recommended for effective damage detection. The conductance signatures were obtained at different stages in the frequency range of 0–450 kHz. This article also presents an effective methodology for damage localization, which assumes the parallel connection of SSUs in MISO mode. The methodology adopted for structural damage detection is effective, as it is verified by the experimental results performed on concrete structures with multiple surface-mounted PZT sensors.