Unsupervised Structural Damage Detection Technique Based on a Deep Convolutional Autoencoder

Abstract

Structural health monitoring (SHM) is a hot research topic with the main purpose of damage detection in a structure and assessing its health state. The major focus of SHM studies in recent years has been on developing vibration-based damage detection algorithms and using machine learning, especially deep learning-based approaches. Most of the deep learning-based methods proposed for damage detection in civil structures are based on supervised algorithms that require data from the healthy state and different damaged states of the structure in the training phase. As it is not usually possible to collect data from damaged states of a large civil structure, using such algorithms for these structures may be impractical. This paper proposes a new unsupervised deep learning-based method for structural damage detection based on convolutional autoencoders (CAEs). The main objective of the proposed method is to identify and quantify structural damage using a CAE network that employs raw vibration signals from the structure and is trained by the signals solely acquired from the healthy state of the structure. The CAE is chosen to take advantage of high feature extraction capability of convolution layers and at the same time use the advantages of an autoencoder as an unsupervised algorithm that does not need data from damaged states in the training phase. Applications on the two numerical models of IASC-ASCE benchmark structure and a grid structure located at the University of Central Florida, as well as the full-scale Tianjin Yonghe Bridge, prove the efficiency of the proposed algorithm in assessing the global health state of the structures and quantifying the damage.

Generative Adversarial Networks for Data Generation in Structural Health Monitoring

“…In short, with a correctly built model and proper training, DL models can show superior performance over ML models. In the civil SHM field, there are few studies, including vibration-based unsupervised DL-mostly Autoencoders (Pathirage et al, 2018;Shang et al, 2021;Rastin et al, 2021), and few supervised DL-mostly Convolutional Neural Networks (Abdeljaber et al, 2017(Abdeljaber et al, , 2018Eren, 2017;Avci et al, 2017;Yu et al, 2019).…”

Structural Health Monitoring of Dams Based on Acoustic Monitoring, Deep Neural Networks, Fuzzy Logic and a CUSUM Control Algorithm

“…More broadly, considering the health monitoring of buildings and bridges, Rastin et al [ 18 ] proposed a new unsupervised deep learning-based method for structural damage detection based on convolutional autoencoders (CAEs). Their method aimed to identify and quantify structural damage using a CAE network that takes as inputs the raw vibration signals from the structure and is trained by the signals solely acquired from the healthy state of the structure.…”

Application of Deep Learning Techniques for the Impedance-based SHM to the Oil & Gas Industry

“…In this sense, the maintenance of systems considered critical becomes even more fundamental, and, consequently, different inspection and damage assessment techniques have been researched in recent years. These monitoring and inspection techniques now make up the so-called Structural Health Monitoring (SHM) (Zhao et al [2011], Melville et al [2018], Gulgec, Takáč and Pakzad [2019], Rastin, Ghodrati Amiri and Darvishan [2021]).…”

“…SHM formulations and methods, in general, aim to provide or improve issues such as safety, operability, minimization of maintenance and repair costs, logistical efficiency, and increase in useful structural life, among others (Moura Jr and Steffen Jr [2006]). For this, SHM methods commonly employ different types of software and hardware to characterize the systems under study, acquire and manage monitoring data, and evaluate, in the long term, these systems' environmental and operational conditions (Melville et al [2018], Gulgec, Takáč and Pakzad [2019], Rastin, Ghodrati Amiri and Darvishan [2021]).…”

“…Therefore, this situation requires adopting a damage identification process with rapid prognostic capacity. For certain types of structures, it is necessary to evaluate the integrity and functioning conditions of the system in real-time (Zhao et al [2011], Rastin, Ghodrati Amiri and Darvishan [2021]).…”