Damage detection in smart structures using neural networks and finite-element analyses

This thesis is devoted to present some novel techniques in Structural Health Monitoring (SHM). SHM is a developing field that tries to monitor structures to make sure that they remain in their desired condition to avoid any catastrophe. SHM includes different levels from damage detection area to prognosis field. This work is dedicated to the first level, which might be considered the main and most important level. New techniques presented in this work are based on different statistical and signal processing methods such as Principal Component Analysis and its robust counterpart, Wavelet Transform, Fuzzy similarity, Andrew plots, etc. These techniques are applied on the propagated waves that are activated and captured in the structure using appropriate transducers. Piezoceramic (PZT) devices are chosen in this work to capture the signals due to their special characteristics such as high performance, low energy consumption and reasonable price. To guarantee the efficiency of the suggested techniques, they are tested on different laboratory and real scale test benchmarks, such as aluminum and composite plates, fuselage, wing skeleton, tube, etc. Because of the variety of tested benchmarks, this thesis is called damage detection in smart structures. This variety may promise the ability and capability of the proposed methods on different fields such as aerospace and gas/oil industry. In addition to the normal laboratory conditions, it is shown in this work that environmental changes can affect the performance of the damage detection and wave propagation significantly. As such, there is a vital need to consider their effect. In this work, temperature change is chosen as it is one of the main environmental fluctuation factors. To scrutinize its effect on damage detection, first, the effect of temperature is considered on wave propagation and then all the proposed methods are tested to check whether they are sensitive to temperature change or not. Finally, a temperature compensation method is applied to ensure that the proposed methods are stable and robust even when structures are subjected to variant environmental conditions. La presente tesis doctoral se dedica a la exploración y presentación de técnicas novedosas para la Monitorización y detección de defectos en estructuras (Structural Health Monitoring -SHM-) SHM es un campo actualmente en desarrollo que pretende asegurarse que las estructuras permanecen en su condición deseada para evitar cualquier catástrofe. En SHM se presentan diferentes niveles de diagnóstico, Este trabajo se concentra en el primer nivel, que se considera el más importante, la detección de los defectos. Las nuevas técnicas presentadas en esta tesis se basan en diferentes métodos estadísticos y de procesamiento de señales tales como el Análisis de Componentes Princpales (PCA) y sus variaciones robustas, Transformada wavelets, lógica difusa, gráficas de Andrew, etc. Estas técnicas de aplican sobre las ondas de vibración que se generan y se miden en la estructura utilizando trasductores apropiados. Dispositivos piezocerámicos (PZT's) se han escogido para este trabajo ya que presentan características especiales tales como: alto rendimiento, bajo consumo de energia y bajo costo. Para garantizar la eficacia de la metodología propuesta,se ha validado en diferentes laboratorios y estructuras a escala real: placas de aluminio y de material compuesto, fuselage de un avión, revestimiento del ala de un avóin, tubería, etc. Debido a la gran variedad de estructuras utilizadas, su aplicación en la industria aeroespacial y/o petrolera es prometedora. Por otra parte, los cambios ambientales pueden afectar al rendimiento de la detección de daños y propagación de la onda significativamente . En este trabajo , se estudia el efecto de las variaciones de temperatura ya que es uno de los principales factores de fluctuación del medio ambiente . Para examinar su efecto en la detección de daños, en primer lugar, todos los métodos propuestos se prueban para comprobar si son sensibles a los cambios de temperatura o no. Finalmente , se aplica un método de compensación de temperatura para garantizar que los métodos propuestos son estables y robustos incluso cuando las estructuras se someten a condiciones ambientales variantes

Structural Health Monitoring (SHM) of aircraft structures, especially composite structures, has assumed increased significance on considerations of safety and costs. With the advent of co-cured structures, wherein bonded joints are replacing bolted joints there is a concern regarding skin-stiffener separation, which might not be detected unless a rigorous non-destructive testing (NDT) is done. It would hence be necessary to be able to detect and assess skin-stiffener separation in composite structures before it reaches the critical size. One of the health monitoring strategies is through strain monitoring using fibre optic strain sensors such as Fibre Bragg Grating (FBG) sensors. The first aspect that needs to be addressed is the characterization of the FBG sensors. Issues of embedment in composites have also to be addressed. Before evolving a damage detection strategy, the sensitivity of the structural strain to skin-stiffener separations must be clearly understood and quantified. This paper presents the analysis and experiments done with a composite test box to study the effect of skin-stiffener separation on the strain behaviour. The box consists of two skins stiffened with spars made of Bi-Directional (BD) glass-epoxy prepreg material. The spars are bolted to the skins and removing suitable number of bolts simulates 'de-bonds'. The strains of the healthy box are compared with the unhealthy box. The strains in the experiments are monitored using both strain gauges and Fibre Bragg Grating (FBG) sensors. The experimental results show that there is significant change in the measured strain near and away from the debond location. The finite element analysis of the box is done using ABAQUS and the analysis is validated with the experimental results. A neural network based methodology is developed here to detect skin-stiffener debonds in structures. A multi-layer perceptron (MLP) neural network with a feed forward back propagation algorithm is used to determine the size/severity of damage. The FE model is used to generate the neural network training data for various sizes of debonds. The results show that the network is able to predict the damage size well. The network is implemented for a specified load. However, it is seen that the damage size predicted is independent of the applied load and the network performance is dependent on the fidelity of the finite element model used to train the network.

In the past 10 years, full scale fatigue tests were undertaken for two types of aircraft groups, and correspondently, the acoustic emission based integrated non-destructive testing means was used for the early detection of fatigue damages throughout the test. What accompanied the full scale fatigue test was eventually a health monitoring process for the aircraft. During the test, researchers needed to know when the aircraft “got ill” and what remedy measure was required to make the aircraft “recover.” A variety of techniques have found important applications in the structural monitoring. Studies were made of various means and the role of health monitoring of aircraft structures and of some special problems to be solved for each method. Comparisons were made for various monitoring methods and the most suitable one was chosen for the full scale aircraft under test. A carefully designed NDT scheme plus AE monitoring was eventually selected as the first choice to complete the task and was shown rather effective in early damage detection for the aircraft under fatigue test although very high risk was confronted throughout the test.

The advancement of machine learning algorithms has opened a wide scope for vibration-based SHM (Structural Health Monitoring). Vibration-based SHM is based on the fact that damage will alter the dynamic properties viz., structural response, frequencies, mode shapes, etc of the structure. The responses measured using sensors, which are high dimensional in nature, can be intelligently analyzed using machine learning techniques for damage assessment. Neural networks employing multilayer architectures are expressive models capable of capturing complex relationships between input-output pairs but do not account for uncertainty in network outputs. A BNN (Bayesian Neural Network) refers to extending standard networks with posterior inference. It is a neural network with a prior distribution on its weights. Deep learning architectures like CNN (Convolutional neural network) and LSTM(Long Short Term Memory) are good candidates for representation learning from high dimensional data. The advantage of using CNN over multi-layer neural networks is that they are good feature extractors as well as classifiers, which eliminates the need for generating hand-engineered features. LSTM networks are mainly used for sequence modeling. This paper presents both a Bayesian multi-layer perceptron and deep learning-based approach for damage detection and location identification in beam-like structures. Raw frequency response data simulated using finite element analysis is fed as the input of the network. As part of this, frequency response was generated for a series of simulations in the cantilever beam involving different damage scenarios. This case study shows the effectiveness of the above approaches to predict bending rigidity with an acceptable error rate.

Wireless sensor networks are increasingly adopted both in aircraft and civil structural health monitoring. However, due to the limited resources on a wireless sensor node, structural evaluation ability of the wireless sensor network is also limited, especially for large-scale structures. This paper presents a multi-agent cooperation system based on mobile agent technology for wireless sensor network-based structural health monitoring which deals with the evaluation of large-scale structure. Mobile agent with the signal processing or evaluation algorithm is designed to migrate among different sensor nodes and perform network self-organization, data processing, information fusion, and damage estimation. This method can reduce the request to the sensor nodes and achieve a better monitoring and evaluation performance through multiagent cooperation. The proposed mobile multiagent system is evaluated by a large dimension aerospace aluminum structure monitoring experiment in which strain distribution and joint failure are monitored by different sensors. Experimental results show that the system can significantly improve the monitoring performance for large-scale structures.

Artificial Neural Networks (ANNs) have been applied in structural damage detection as a classifier, but generally a capable ANNs has to be trained with a certain amount of samples. When both damage locations and damage extents are to be identified, the amount of training samples is tremendous because of the combinations of damage locations and extents. By wavelet transform of the structure free motion equations, the Residual Wavelet Coefficient Vector (RWCV) is deduced. A damage feature parameter is defined as the ratio between RWCVs in two different frequency bands. This parameter has a unique property that it's sensitive only to damage locations, and is independent of damage extents. The damage feature parameters are then fed to the neural network for damage localization. After the damage sites are detected, the damage extent is further identified by another neural network with RWCVs as inputs. This two-phase approach for damage localization and extent identification can simply the neural network and reduce the training samples tremendously. Finally a numerical example is given for damage detection of a 10 DOFs system using the proposed approach.

In the present work, the design of thermoplastic induced welded joints was carried out considering different aspects related to structural performance, manufacturability and structural health monitoring capability. Three lap-joints configurations were considered: two of them having welded plates with two different carbon fibers plies stacked layups and one obtained by welding carbon fibers fabric laminates. A numerical analysis was carried out by performing (i) structural simulations to assess the mechanical properties of the joints, (ii) induction welding process simulations to evaluate temperature distribution in the welded interface area and (iii) damage detection assessment by reproducing the response of a structural health monitoring sensor bonded in correspondence of the joint overlap area. Electromagnetic-based structural health monitoring techniques were employed to evaluate local variation of electromagnetic characteristics due to defect presence. The results obtained show that the design choices have a significant influence on the physical properties of the welded material such as the electrical conductivity anisotropy. The different induced currents distribution obtained during both the fabrication and damage inspection phase, affects respectively the manufacturing and detection capability. An experimental validation was carried out by mechanical tests under quasi-static loading for the case with the best tradeoff between structural performance, manufacturing feasibility and damage detection capability. The proposed approach show the importance of taking into account the requirement of damage detectability from an early stage of the design.

Vibration-based damage detection methods, leveraging structural response data to identify structural damage, play an important role in structural health monitoring, particularly for critical infrastructure like bridges and high-rise buildings. In recent years, deep learning, which can autonomously learn damage-sensitive features and classify damage from complex datasets, has been widely used in structural damage detection. This paper reviews deep learning architectures, including Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and autoencoders, applied in vibration-based structural damage detection for civil structures over the past ten years, highlighting their advantages and disadvantages. The challenges and future trends for vibration-based damage detection by using deep learning are summarized.

In civil engineering, health monitoring of structure have been treated as an important subject. Health monitoring by using experimental measurements such as strain, acceleration, velocity, displacement, rotation and other parameters has been used for detecting changes, which may indicate damage or degradation. In recent, the applications of health monitoring of large structure have been facilitated by the developments in measurement, communications and computational technology gradually. Especially, large civil structures as bridges are damaged by fatigue loads or unexpected load such as earthquakes and finally reach to failure. Thus, a structural health monitoring and damage detection methods have been developed and attracted a great deal of public attention in recent decades. Actually many research efforts have been conducted to develop a health monitoring and damage detection techniques of existing civil structures. However, these methods are needed to improve to apply to a real civil structure. The purpose of this study is to develop a practical instrument to identify dynamic parameters of real civil structural systems.

Advances in optic fiber sensing technology provide easy and reliable way for the vibration-based strain measurement of engineering structures. As a typical optic fiber sensing techniques with high accuracy and resolution, long-gauge Fiber Bragg Grating (FBG) sensors have been widely employed in health monitoring of civil engineering structures. Therefore, the development of macro strain-based identification methods is crucial for damage detection and structural condition evaluation. In the previous study by the authors, a damage detection algorithm for a beam structure with the direct use of vibration-based macro-strain measurement time history with neural networks had been proposed and validated with experimental measurements. In this paper, a damage locating and quantifying method was proposed using modal macrostrain vectors (MMSVs) which can be extracted from vibration induced macro-strain response measurement time series from long-gage FBG sensors. The performance of the proposed methodology for damage detection of a beam with different damage scenario was studied with numerical simulation firstly. Then, dynamic tests on a simply-supported steel beam with different damage scenarios were carried out and macro-strain measurements were employed to detect the damage severity. Results show that the proposed MMSV based structural identification and damage detection methodology can locate and identify the structural damage severity with acceptable accuracy.

In this paper, an improvement of the elliptical triangulation method for damage detection using Lamb waves is presented. The damage is the main cause of structural failure and often occurs on structures. In order to avoid sudden failure, a special attention was given in the past decades to the damage detection in structures. In order to obtain efficient damage detection techniques, the structural health monitoring has been the main research topic of many scientists worldwide. The elliptical triangulation method, proposed in this paper, is a non-destructive method based on measurement of Lamb waves. This method, through the calculation of the time of flight of the signals and the actuator-sensor positioning, allows to identify position and dimension of the damage. The application of the method to the metallic structures and to the composite material structures is presented in this paper. The complexity connected with the uncertainty of the waves’ propagation speed due to the anisotropy of the composite materials has been explored through an iterative approach. The initialization of the wave propagation speed at first tentative iteration is the key issue for the convergence of the method. Seven different conditions were used to validate the method on both metallic and composite structures combining two damage shapes, two damage dimensions (effective damaged area), and three different positions. Upon evaluating the effectiveness, the method has been applied at two composite panels in order to detect by test the post-impact damages. Tests results have been compared with the numerical ones. The feasibility of the elliptical triangulation method to detect the damage (evaluating the damage position and area) has been proved using the ultrasonic C-Scan.

Experimental system identification of the dynamics of a vibro-impact beam with a view towards structural health monitoring and damage detection

Efficient wind energy harvesting becomes more and more important as a consequence of the increasing interest in renewable energy in the European Union [1]. This leads to growing sizes of wind turbines (WTs), and with it, larger WT blades (WTBs). The structural designs of these WTBs are created to optimize the potential energy output, where low mass is a key requirement. However, high flexibilities and lower buckling capacities are further results of these developments [2], thus certain damage scenarios become significant. Intelligently designed structural health monitoring (SHM) systems can help to reduce the associated operations and maintenance costs. Even though, several techniques are already developed for structural damage detection (SDD) in WTBs, the majority of these methods is not suitable for inservice measurements. This paper presents a SDD and structural damage localization (SDL) method based on the partial autocorrelation function (PACF) of vibration responses. The approach is applied to a numerical model of a large WTB, where the acceleration responses are obtained from transient dynamic simulations with a simplified aerodynamic loading approach. The novel damage sensitive feature (DSF) is developed as the Mahalanobis distance between a baseline and current vector of PACF coefficients. First, numerical modal analysis of the finite element (FE) WTB model is performed in order to estimate the effect of a disbonding damage scenario on the vibration characteristics. Second, the behaviour of the PACF for time series of the healthy system is discussed. Third, the SDD results on the basis of statistical hypothesis testing are assessed for two selected sensor locations and increasing damage extents. Finally, the performance of the proposed DSF with respect to SDL is illustrated for multiple locations on the WTB’s surface. This study demonstrated the efficiency of a DSF based on the PACF for SDD and SDL, which is promising for future developments of vibration-based SHM techniques in WTBs.

As newer and more reliable ways of construction were developed, civilization began to spread out further and retain functional infrastructure for longer periods of time.[...]

In the present paper, Structural health monitoring has become an evolving area of research in last few decades with increasing need of online monitoring the health of large structures. The damage detection by visual inspection of the structure can prove impractical, expensive and ineffective in case of large structures like offshore platforms, multistoried buildings and bridges. Structural health monitoring is defined as the process of detecting damage in a structural system. Damage in the system causes a change in dynamic properties of a system. The structural damage is typically a local phenomenon, which tends to be captured by higher frequency signals. Most of vibration-based damage detection methods require the modal properties that are obtained from measured signals through the system identification techniques. However, the modal properties such as natural frequencies and mode shapes are not such a good sensitive indication of structural damage. Structural damage detection and damage localization of jacket platforms, based on wavelet packet transforms is presented in this paper. Dynamic signals measured from the structure by the finite element software package ANSYS are first decomposed into wavelet packet components. Component energies are then calculated and used for damage assessment. The results show that the WPT-based component energies are good candidate indices that are sensitive to structural damage. These component energies can be used for damage assessment including identifying damage occurrence and location.