Structural damage identification based on fast S-transform and convolutional neural networks

Ghahremani, Behzad; Bitaraf, Maryam; Ghorbani‐Tanha, Amir K.; Fallahi, Reza

doi:10.1016/j.istruc.2020.11.068

Cited by 18 publications

(4 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Examples of data-driven approaches are the studies presented in refs. [ 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 ]. Liu et al [ 7 ] applied neural network and multi-sensor feature fusion theory to classify the recorded sensor’s data.…”

Section: Introductionmentioning

confidence: 99%

“…Azimi and Pekcan [ 10 ] studied the application of convolutional neural networks in damage detection of realistic large-scale systems. Ghahremani et al [ 11 ] investigated damage detection in a three-story model using convolutional neural network and recorded acceleration signals. Bao et al [ 12 ] proposed a two-step procedure to identify the anomalies caused by structural damage using a deep neural network.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Bridge Health Monitoring Using Strain Data and High-Fidelity Finite Element Analysis

Ghahremani

Enshaeian

Rizzo

2022

Sensors

Self Cite

View full text Add to dashboard Cite

This article presented a physics-based structural health monitoring (SHM) approach applied to a pretensioned adjacent concrete box beams bridge in order to predict the deformations associated with the presence of transient loads. A detailed finite element model was generated using ANSYS software to create an accurate model of the bridge. The presence of concentrated loads on the deck at different locations was simulated, and a static analysis was performed to quantify the deformations induced by the loads. Such deformations were then compared to the strains recorded by an array of wireless strain gauges during a controlled truckload test performed by an independent third party. The test consisted of twenty low-speed crossings at controlled distances from the bridge parapets using a truck with a certified load. The array was part of a SHM system that consisted of 30 wireless strain gauges. The results of the comparative analysis showed that the proposed physics-based monitoring is capable of identifying sensor-related faults and of determining the load distributions across the box beams. In addition, the data relative to near two-years monitoring were presented and showed the reliability of the SHM system as well as the challenges associated with environmental effects on the strain reading. An ongoing study is determining the ability of the proposed physics-based monitoring at estimating the variation of strain under simulated damage scenarios.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Bridge Health Monitoring Using Strain Data and High-Fidelity Finite Element Analysis

Ghahremani

Enshaeian

Rizzo

2022

Sensors

Self Cite

View full text Add to dashboard Cite

show abstract

“…Vision-based methods mainly rely on damaged images for detection. In recent studies, the analysis of damaged images is highly reliant on DL algorithms, such as the YOLO algorithm [122], CNN algorithm [123,124], and GAN algorithm [101]. Figure 4 shows the general process for damage detection using DL models.…”

Section: Damage Detectionmentioning

confidence: 99%

Deep Learning for Earthquake Disaster Assessment: Objects, Data, Models, Stages, Challenges, and Opportunities

Jia,

2023

Remote Sensing

View full text Add to dashboard Cite

Earthquake Disaster Assessment (EDA) plays a critical role in earthquake disaster prevention, evacuation, and rescue efforts. Deep learning (DL), which boasts advantages in image processing, signal recognition, and object detection, has facilitated scientific research in EDA. This paper analyses 204 articles through a systematic literature review to investigate the status quo, development, and challenges of DL for EDA. The paper first examines the distribution characteristics and trends of the two categories of EDA assessment objects, including earthquakes and secondary disasters as disaster objects, buildings, infrastructure, and areas as physical objects. Next, this study analyses the application distribution, advantages, and disadvantages of the three types of data (remote sensing data, seismic data, and social media data) mainly involved in these studies. Furthermore, the review identifies the characteristics and application of six commonly used DL models in EDA, including convolutional neural network (CNN), multi-layer perceptron (MLP), recurrent neural network (RNN), generative adversarial network (GAN), transfer learning (TL), and hybrid models. The paper also systematically details the application of DL for EDA at different times (i.e., pre-earthquake stage, during-earthquake stage, post-earthquake stage, and multi-stage). We find that the most extensive research in this field involves using CNNs for image classification to detect and assess building damage resulting from earthquakes. Finally, the paper discusses challenges related to training data and DL models, and identifies opportunities in new data sources, multimodal DL, and new concepts. This review provides valuable references for scholars and practitioners in related fields.

show abstract

“…The presence of damage is detected by variations in these parameters in the damaged configuration compared to the reference state (undamaged) evaluated before the event (sometimes seconds before an event for continuously monitored structures). There are several approaches that allow such comparisons [30][31][32][33][34][35][36][37][38][39][40][41][42] to be made. It has been observed that methods based on the analysis of variations in the modal shapes and/or their derivatives, such as the mode curvature, are very effective [1] and can also be used as a diagnostic tool for structural and/or non-structural damage localization and quantification.…”

Section: Introductionmentioning

confidence: 99%

Damage Detection and Localization on Real Structures Subjected to Strong Motion Earthquakes Using the Curvature Evolution Method: The Navelli (Italy) Case Study

et al. 2021

View full text Add to dashboard Cite

In recent years, structural health monitoring (SHM) has received increasing interest from both research and professional engineering communities. This is due to the limitations related to the use of traditional methods based on visual inspection for a rapid and effective assessment of structures and infrastructures when compared with the great potential offered by newly developed automatic systems. Most of these kinds of systems allow the continuous estimation of structural modal properties that are strictly correlated to the mechanical characteristics of the monitored structure. These can change as a result of material deterioration and structural damage related to earthquake shaking. Furthermore, a suitable configuration of a dense sensor network in a real-time monitoring system can allow to detect and localize structural and non-structural damage by comparing the initial and a final state of the structure after a critical event, such as a relevant earthquake. In this paper, the modal curvature evaluation method, used for damage detection and localization on framed structures, considering the mode curvature variation due to strong earthquake shaking, is further developed. The modified approach is validated by numerical and experimental case studies. The extended procedure, named “Curvature Evolution Method” (CEM), reduces the required computing time and the uncertainties in the results. Furthermore, in this work, an empirical relationship between curvature variation and damage index has been defined for both bare and infilled frames.

show abstract

Structural damage identification based on fast S-transform and convolutional neural networks

Cited by 18 publications

References 30 publications

Bridge Health Monitoring Using Strain Data and High-Fidelity Finite Element Analysis

Bridge Health Monitoring Using Strain Data and High-Fidelity Finite Element Analysis

Deep Learning for Earthquake Disaster Assessment: Objects, Data, Models, Stages, Challenges, and Opportunities

Damage Detection and Localization on Real Structures Subjected to Strong Motion Earthquakes Using the Curvature Evolution Method: The Navelli (Italy) Case Study

Contact Info

Product

Resources

About