Fibre-optic sensor and deep learning-based structural health monitoring systems for civil structures: A review

Jayawickrema, U. M. N.; Herath, H.M.C.M.; Hettiarachchi, Nandita; Sooriyaarachchi, Harsha; Epaarachchi, Jayantha

doi:10.1016/j.measurement.2022.111543

Cited by 69 publications

(27 citation statements)

References 166 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition, deep neural networks provide a unified approach for selecting features and classifying them, removing the need to search for optimal classifiers. Although there are already applications of ML and DL approaches for event detection in DAS data in industry and structural health monitoring 59 , to our knowledge, there are not so many attempts in their use for volcano monitoring. Just recently 29,60 , DL are becoming common tools for investigating and monitoring seismic and volcanic areas exploiting the spatially and temporally large DAS data, that may contain signals from natural and anthropogenic sources.…”

Section: Events Detectionmentioning

confidence: 99%

Distributed dynamic strain sensing of very long period and long period events on telecom fiber-optic cables at Vulcano, Italy

Currenti

Allegra

Cannavò

et al. 2023

Sci Rep

View full text Add to dashboard Cite

Volcano-seismic signals can help for volcanic hazard estimation and eruption forecasting. However, the underlying mechanism for their low frequency components is still a matter of debate. Here, we show signatures of dynamic strain records from Distributed Acoustic Sensing in the low frequencies of volcanic signals at Vulcano Island, Italy. Signs of unrest have been observed since September 2021, with CO2 degassing and occurrence of long period and very long period events. We interrogated a fiber-optic telecommunication cable on-shore and off-shore linking Vulcano Island to Sicily. We explore various approaches to automatically detect seismo-volcanic events both adapting conventional algorithms and using machine learning techniques. During one month of acquisition, we found 1488 events with a great variety of waveforms composed of two main frequency bands (from 0.1 to 0.2 Hz and from 3 to 5 Hz) with various relative amplitudes. On the basis of spectral signature and family classification, we propose a model in which gas accumulates in the hydrothermal system and is released through a series of resonating fractures until the surface. Our findings demonstrate that fiber optic telecom cables in association with cutting-edge machine learning algorithms contribute to a better understanding and monitoring of volcanic hydrothermal systems.

show abstract

Section: Events Detectionmentioning

confidence: 99%

Distributed dynamic strain sensing of very long period and long period events on telecom fiber-optic cables at Vulcano, Italy

Currenti

Allegra

Cannavò

et al. 2023

Sci Rep

View full text Add to dashboard Cite

show abstract

“…For track circuit case, it was determined that LSTM network performs better than convolutional network, despite fact that convolutional network is simpler to train. Work [16] used an LSTM model to predict tool wear by encoding raw sensory data into embedding. Hybrid methods that link CNN layers as well as LSTM layers in this manner are referred to as CNN-LSTM in this article in order to extract both spatial as well as temporal information [17].…”

Section: 1mentioning

confidence: 99%

“…By referring to solution as saddle point of Lagrangian, it is possible to spot the dual issue (D) in Lagrange multipliers connected to first set of constraints by eq. (16)(17)(18):…”

Section: Kernelized Component Vector Neural Network Based Feature Sel...mentioning

confidence: 99%

Machine Learning in Molecular Communication and Applications for Health Monitoring Networks

Kumar

Sharma

et al. 2023

Preprint

View full text Add to dashboard Cite

The world has been greatly affected by increased utilization of mobile methods as well as smart devices in field of health. Health professionals are increasingly utilizing these technologies' advantages, resulting in a significant improvement in clinical health care. For this purpose, machine learning (ML)as well as Internet of Things (IoT) can be utilized effectively. This study aims to propose a novel data analysis method for a health monitoring system based on machine learning. Goal of research is to create a ML based smart health monitoring method. It lets doctors keep an eye on patients from a distance as well as take periodic actions if they need to. Utilizing wearable sensors, a set of five parameters—the electrocardiogram (ECG), pulse rate, pressure, temperature, and position detection—have been identified. Kernelized component vector neural networks are used to choose the features in the input data. Then, a sparse attention-based convolutional neural network with a structural search algorithm was used to classify the selected features. For a variety of datasets, the proposed technique attained validation accuracy 95%, training accuracy 92%, RMSE 52%, F-measure 53%, sensitivity 77%.

show abstract

“…Recent growth in the number of publications and applications related to new measurement technologies based on fiber-optic sensors (FOSs) is explained by a number of their advantages: low weight, small size, reliability, stability, resistance to external electromagnetic interference, high sensitivity, low power consumption, remote control and possibility to obtain real-time data [ 1 ]. FOSs provide the ability to measure various physical quantities: strain, acceleration, pressure, temperature and humidity [ 2 ]. The most widespread are point FOSs based on fiber Bragg gratings (FBGs), distributed fiber-optic sensors (DFOSs) based on Rayleigh, Brillouin and Raman scattering, and interferometric FOSs [ 3 , 4 ].…”

Section: Introductionmentioning

confidence: 99%

Measurement of Gradient Strain Fields with Fiber-Optic Sensors

Матвеенко

Kosheleva

Serovaev

et al. 2022

Sensors

View full text Add to dashboard Cite

The results of measuring gradient strain fields by embedded or mounted point fiber-optic sensors based on Bragg gratings and distributed fiber-optic sensors based on Rayleigh scattering are discussed. Along with the experiment, the results of numerical modeling of strain measurement errors associated with the assumption of uniaxial stress state in the area of the embedded Bragg grating and measurement errors by distributed fiber-optic sensors associated with gage length are presented. Experimental results are presented for 3D printed samples and samples made of polymer composite material. The geometry of the samples was chosen based on the results of numerical simulations, and provides different variants of non-uniform strain distribution under uniaxial tension, including the variant in which the derivative of the strain distribution function changes its sign. A good agreement of numerical results and experimental data obtained by distributed and point fiber-optic sensors in areas where the derivative of the strain distribution function keeps a sign and an increase in the error of strain measurement results by distributed fiber-optic sensors in areas where this derivative changes sign are demonstrated.

show abstract

Fibre-optic sensor and deep learning-based structural health monitoring systems for civil structures: A review

Cited by 69 publications

References 166 publications

Distributed dynamic strain sensing of very long period and long period events on telecom fiber-optic cables at Vulcano, Italy

Distributed dynamic strain sensing of very long period and long period events on telecom fiber-optic cables at Vulcano, Italy

Machine Learning in Molecular Communication and Applications for Health Monitoring Networks

Measurement of Gradient Strain Fields with Fiber-Optic Sensors

Contact Info

Product

Resources

About