scholarly journals Damage Detection in Structural Health Monitoring using Hybrid Convolution Neural Network and Recurrent Neural Network

2021 ◽  
Vol 16 (59) ◽  
pp. 461-470
Author(s):  
Thanh Bui-Tien ◽  
Dung Bui-Ngoc ◽  
Hieu Nguyen-Tran ◽  
Lan Nguyen-Ngoc ◽  
Hoa Tran-Ngoc ◽  
...  
Keyword(s):  
Neural Network ◽  
Time Series ◽  
Structural Health Monitoring ◽  
Damage Detection ◽  
Recurrent Neural Network ◽  
Health Monitoring ◽  
Damage Identification ◽  
Convolution Neural Network ◽  
Series Data ◽  
Structural Health

The process of damage identification in Structural Health Monitoring (SHM) gives us a lot of practical information about the current status of the inspected structure. The target of the process is to detect damage status by processing data collected from sensors, followed by identifying the difference between the damaged and the undamaged states. Different machine learning techniques have been applied to attempt to extract features or knowledge from vibration data, however, they need to learn prior knowledge about the factors affecting the structure. In this paper, a novel method of structural damage detection is proposed using convolution neural network and recurrent neural network. A convolution neural network is used to extract deep features while recurrent neural network is trained to learn the long-term historical dependency in time series data. This method with combining two types of features increases discrimination ability when compares with it to deep features only. Finally, the neural network is applied to categorize the time series into two states - undamaged and damaged. The accuracy of the proposed method was tested on a benchmark dataset of Z24-bridge (Switzerland). The result shows that the hybrid method provides a high level of accuracy in damage identification of the tested structure.

scholarly journals Statistical guided-waves-based structural health monitoring via stochastic non-parametric time series models

2021 ◽  
pp. 147592172110245
Author(s):  
Ahmad Amer ◽  
Fotis P Kopsaftopoulos
Keyword(s):  
Decision Making ◽  
Time Series ◽  
Structural Health Monitoring ◽  
Damage Detection ◽  
Health Monitoring ◽  
Guided Waves ◽  
State Of The Art ◽  
Time Series Models ◽  
Active Sensing ◽  
Structural Health

Damage detection in active-sensing, guided-waves-based structural health monitoring (SHM) has evolved through multiple eras of development during the past decades. Nevertheless, there still exist a number of challenges facing the current state-of-the-art approaches, both in the industry as well as in research and development, including low damage sensitivity, lack of robustness to uncertainties, need for user-defined thresholds, and non-uniform response across a sensor network. In this work, a novel statistical framework is proposed for active-sensing SHM based on the use of ultrasonic guided waves. This framework is based on stochastic non-parametric time series models and their corresponding statistical properties in order to readily provide healthy confidence bounds and enable accurate and robust damage detection via the use of appropriate statistical decision-making tests. Three such methods and corresponding statistical quantities (test statistics) along with decision-making schemes are formulated and experimentally assessed via the use of three coupons with different levels of complexity: an Al plate with a growing notch, a carbon fiber-reinforced plastic (CFRP) plate with added weights to simulate local damage, and the CFRP panel used in the Open Guided Waves project, all fitted with piezoelectric transducers under a pitch-catch configuration. The performance of the proposed methods is compared to that of state-of-the-art time-domain damage indices (DIs). The results demonstrate the increased detection sensitivity and robustness of the proposed methods, with better tracking capability of damage evolution compared to conventional approaches, even for damage-non-intersecting actuator–sensor paths. In particular, the Z statistic emerges as the best damage detection metric compared to conventional DIs, as well as the other proposed statistics. Overall, the proposed statistics in this study promise greater damage sensitivity across different components, with enhanced robustness to uncertainties, as well as user-friendly application.

scholarly journals Deep residual network framework for structural health monitoring

2020 ◽  
pp. 147592172091837 ◽  
Author(s):  
Ruhua Wang ◽  
Chencho ◽  
Senjian An ◽  
Jun Li ◽  
Ling Li ◽  
...  
Keyword(s):  
Neural Network ◽  
Structural Health Monitoring ◽  
Convolutional Neural Network ◽  
Health Monitoring ◽  
Damage Identification ◽  
State Of The Art ◽  
Experimental Studies ◽  
The State ◽  
Residual Network ◽  
Structural Health

Convolutional neural networks have been widely employed for structural health monitoring and damage identification. The convolutional neural network is currently considered as the state-of-the-art method for structural damage identification due to its capabilities of efficient and robust feature learning in a hierarchical manner. It is a tendency to develop a convolutional neural network with a deeper architecture to gain a better performance. However, when the depth of the network increases to a certain level, the performance will degrade due to the gradient vanishing issue. Residual neural networks can avoid the problem of vanishing gradients by utilizing skip connections, which allows the information flowing to the next layer through identity mappings. In this article, a deep residual network framework is proposed for structural health monitoring of civil engineering structures. This framework is composed of purely residual blocks which operate as feature extractors and a fully connected layer as a regressor. It learns the damage-related features from the vibration characteristics such as mode shapes and maps them into the damage index labels, for example, stiffness reductions of structures. To evaluate the efficacy and robustness of the proposed framework, an intensive evaluation is conducted with both numerical and experimental studies. The comparison between the proposed approach and the state-of-the-art models, including a sparse autoencoder neural network, a shallow convolutional neural network and a convolutional neural network with the same structure but without skip connections, is conducted. In the numerical studies, a 7-storey steel frame is investigated. Four scenarios with considering measurement noise and finite element modelling errors in the data sets are studied. The proposed framework consistently outperforms the state-of-the-art models in all the scenarios, especially for the most challenging scenario, which includes both measurement noise and uncertainties. Experimental studies on a prestressed concrete bridge in the laboratory are conducted. The proposed framework demonstrates consistent damage prediction results on this beam with the state-of-the-art models.

scholarly journals A new approach for structural damage detection exploring the singular spectrum analysis

2016 ◽  
Vol 28 (9) ◽  
pp. 1160-1174 ◽  
Author(s):  
Mario A de Oliveira ◽  
Jozue Vieira Filho ◽  
Vicente Lopes ◽  
Daniel J Inman
Keyword(s):  
Time Series ◽  
Structural Health Monitoring ◽  
Damage Detection ◽  
Health Monitoring ◽  
Spectrum Analysis ◽  
Structural Damage ◽  
Singular Spectrum Analysis ◽  
Reconstruction Process ◽  
Structural Health ◽  
Singular Spectrum

This article presents a novel approach for damage detection applied to structural health monitoring systems exploring the residues obtained from singular spectrum analysis. In this technique, a lead zirconate titanate patch acting as actuator excites the structure, and three other patches are used as sensors to receive the structural responses. This method is based on a high-frequency excitation range in order to overcome the problem caused when the low-vibration modes are excited. In this method, a wideband chirp signal, with low amplitude and variable frequency, is used to excite the structure. The response signals are acquired in the time domain, and the singular spectrum analysis procedure is performed. The residues obtained between the reconstructed and original time series are used to compute statistical metrics. The residues calculated from singular spectrum analysis are used to compute the root mean square deviation and correlation coefficient deviation metric indices, rendering the damage detection approach more reliable. Tests were carried out on an aluminum plate, and the results have demonstrated the effectiveness of the proposed method making it an excellent approach for structural health monitoring applications. The results exploring different numbers of components used during the reconstruction process of time series are obtained, and the highlights are presented.

Toward data anomaly detection for automated structural health monitoring: Exploiting generative adversarial nets and autoencoders

2020 ◽  
pp. 147592172092460 ◽  
Author(s):  
Jianxiao Mao ◽  
Hao Wang ◽  
Billie F Spencer
Keyword(s):  
Structural Health Monitoring ◽  
Anomaly Detection ◽  
Damage Detection ◽  
Health Monitoring ◽  
Sensor System ◽  
Series Data ◽  
Generative Adversarial Networks ◽  
Learning Methods ◽  
Structural Health ◽  
Anomalous Data

Damage detection is one of the most important tasks for structural health monitoring of civil infrastructure. Before a damage detection algorithm can be applied, the integrity of the data must be ensured; otherwise results may be misleading or incorrect. Indeed, sensor system malfunction, which results in anomalous data (often called faulty data), is a serious problem, as the sensors usually must operate in extremely harsh environments. Identifying and eliminating anomalies in the data is crucial to ensuring that reliable monitoring results can be achieved. Because of the vast amounts of data typically collected by a structural health monitoring system, manual removal of the anomalous data is prohibitive. Machine learning methods have the potential to automate the process of data anomaly detection. Although supervised methods have been proven to be effective for detecting data anomalies, two unresolved challenges reduce the accuracy of anomaly detection: (1) the class imbalance and (2) incompleteness of anomalous patterns of training dataset. Unsupervised methods have the potential to address these challenges, but improvements are required to deal with vast amounts of monitoring data. In this article, the generative adversarial networks are combined with a widely applied unsupervised method, that is, autoencoders, to improve the performance of existing unsupervised learning methods. In addition, the time-series data are transformed to Gramian Angular Field images so that advanced computer vision methods can be included in the network. Two structural health monitoring datasets from a full-scale bridge, including examples of anomalous data caused by sensor system malfunctions, are utilized to validate the proposed methodology. Results show that the proposed methodology can successfully identify data anomalies with good accuracy and robustness, hence can overcome one of the key difficulties in achieving automated structural health monitoring.

Everyday Life Quantification Using mDFA: Heart Health Monitoring and Structural Health Monitoring

2015 ◽  
Author(s):  
Toru Yazawa
Keyword(s):  
Time Series ◽  
Structural Health Monitoring ◽  
Health Monitoring ◽  
Natural World ◽  
Series Data ◽  
Fluctuation Analysis ◽  
Scaling Exponent ◽  
Heart Health ◽  
Structural Health ◽  
Wheel Vibration

The aim of this study was to make a method usable in an early detection of malfunction, e.g., abnormal vibration/fluctuation in recorded signals. We conducted experimentations of heart health and structural health monitoring. We collected natural world signals, e.g., heartbeat fluctuation and mechanical vibration. For the analysis, we used modified detrended fluctuation analysis (mDFA) method that we have made recently. mDFA calculated the scaling exponent (SI, the acronym SI is derived from the scaling indices) from the time series data, e.g., R-R interval time series obtained from electrocardiograms. In the present study, peaks were identified by our own method. In every single mDFA computation, we identified ∼2000 consecutive peaks from a data: “2000” was necessary number to conduct mDFA. mDFA was able to distinguish between normal and abnormal behaviors: Normal healthy hearts exhibited an SI around 1.0, which is a phenomena comparable to 1/f fluctuation. Job-related stressful hearts and extrasystolic hearts both exhibited a low SI such as 0.7. Normally running car’s vibration — recorded steering wheel vibration — exhibited an SI around 0.5, which is white noise like fluctuation. Normally spinning ball-bearings (BB) exhibited an SI around 0.1, which belongs to the anti-correlation phenomena. A malfunctioning BB showed an increased SI. At an SI value over 0.2, an inspector must check BB’s correct functioning. Here we propose that healthiness in various cyclic vibration behaviors can be quantitatively analyzed by mDFA.

Sign in / Sign up

Export Citation Format

Share Document