Search

Smart structural health monitoring (SHM) for large-scale infrastructure is an intriguing subject for engineering communities thanks to its significant advantages such as timely damage detection, optimal maintenance strategy, and reduced resource requirement. Yet, it is a challenging topic as it requires handling a large amount of collected sensors data continuously, which is inevitably contaminated by random noises. Therefore, this study developed a practical end-to-end framework that makes use of physical features embedded in raw data and an elaborated hybrid deep learning model, namely 1-DCNN-LSTM, featuring two algorithms—convolutional neural network (CNN) and long-short term memory (LSTM). In order to extract relevant features from sensory data, the method combines various signal processing techniques such as the autoregressive model, discrete wavelet transform, and empirical mode decomposition. The hybrid deep learning 1-DCNN-LSTM is designed based on the CNN’s capacity of capturing local information and the LSTM network’s prominent ability to learn long-term dependencies. Through three case studies involving both experimental and synthetic data sets, it is demonstrated that the proposed approach achieves highly accurate damage detection, as accurate as the powerful 2-D CNN, but with a lower time and memory complexity, making it suitable for real-time SHM. Note to Practitioners —This article aims to develop a practical data-driven method for automatically monitoring the operational state of structures. In order to achieve consistently and highly accurate results in performing different tasks for diverse structures, we combine underlying features in both time and frequency domains extracted from measured signal vibration data. Three popular data featuring methods are combined to achieve the diversity gain which would not be possible with each individual method. As the vibration is usually measured by long time-series signals, the most efficient deep learning architecture for time-series signal, namely long-short term memory (LSTM), is considered for this work. Besides, each structure has its own dynamic properties, i.e., eigenfrequencies, around which the most relevant information is in the frequency domain, thus convolutional neural network specifically designed for capturing local information is used in combination with LSTM, forming a hybrid deep learning architecture. The applicability and effectiveness of the proposed approach are supported by three case studies with different types of structures, showing highly accurate damage detection with reduced resource requirements. These advantages can be valuable for developing a model for live monitoring of structural health in the future life-line infrastructures.

Artificial neural network (ANN) is the study of computer algorithms that can learn from experience to improve performance. ANN employs backpropagation (BP) algorithms using gradient descent (GD)-based learning methods to reduce the discrepancies between predicted and real targets. Even though these differences are considerably decreased after each iteration, the network may still face major risks of being entrapped in local minima if complex error surfaces contain too numerous the best local solutions. To overcome those drawbacks of ANN, numerous researchers have come up with solutions to local minimum prevention by choosing a beneficial starting position that relies on the global search capability of other algorithms. This strategy possibly assists the network in avoiding the first local minima. However, a network often has many local bests widely distributed. Hence, the solution of choosing good starting points may no further be beneficial because the particles are probably entrapped in other local optimal solutions throughout the process of looking for the global best. Therefore, in this work, a novel ANN working parallel with the stochastic search capacity of evolutionary algorithms, is proposed. Additionally, to increase the efficiency of the global search capacity, a hybrid of particle swarm optimization and genetic algorithm (PSOGA) is applied during the process of seeking the best solution, which effectively guarantees to assist the network of ANN in escaping from local minima. This strategy gains both benefits of GD techniques as well as the global search capacity of PSOGA that possibly solves the local minima issues thoroughly. The effectiveness of ANNPSOGA is assessed using both numerical models consisting of various damage cases (single and multiple damages) and a free-free steel beam with different damage levels calibrated in the laboratory. The results demonstrate that ANNPSOGA provides higher accuracy than traditional ANN, PSO, and other hybrid ANNs (even a higher level of noise is employed) and also considerably decreases calculational cost compared with PSO.

Artificial neural network (ANN) has been commonly used to deal with many problems. However, since this algorithm applies backpropagation algorithms based on gradient descent (GD) technique to look for the best solution, the network may face major risks of being entrapped in local minima. To overcome those drawbacks of ANN, in this work, we propose a novel ANN working parallel with metaheuristic algorithms (MAs) to train the network. The core idea is that first, (1) GD is applied to increase the convergence speed. (2) If the network is stuck in local minima, the capacity of the global search technique of MAs is employed. (3) After escaping from local minima, the GD technique is applied again. This process is applied until the target is achieved. Additionally, to increase the efficiency of the global search capacity, a hybrid of particle swarm optimization and genetic algorithm (PSOGA) is employed. The effectiveness of ANNPSOGA is assessed using both numerical models and measurement. The results demonstrate that ANNPSOGA provides higher accuracy than traditional ANN, PSO, and other hybrid ANNs (even a higher level of noise is employed) and also considerably decreases calculational cost compared with PSO.

Railway bridges are susceptible to the problems of fracture and fatigue because they endure millions of stress cycles under moving train load during their service life. This may lead to stiffness reduction in truss joints of the truss bridges and reduce the operational effectiveness of the bridge. In this paper, a hybrid algorithm based on Artificial Neural Network (ANN) coupled with an evolutionary algorithm, i.e. Genetic Algorithm (GA) is proposed to analyze the stiffness reduction of truss joints in a railway bridge. GA is employed to determine training parameters and overcome the local minimum problems of ANN. Natural frequencies are selected as input data, whereas output data is damage characteristics (locations and levels) of truss joints. The results demonstrate that ANN-GA determines vibration behavior and damages of the considered structure accurately, comprising single stiffness reduction of truss joints as well as multiple stiffness reduction scenarios.

Recently, Structural Health Monitoring (SHM) has emerged to be one of the most effective tools for the diagnosis of damages in structures. Early identification and localization of damage not only help to reduce the maintenance cost but also extend the life cycle of the structures. In this paper, a novel approach using Artificial Neural Networks (ANNs) combined with Genetic Algorithms (GA) is proposed to increase the capacity of damage detection in SHM system. ANNs can make use of different algorithms such as recognition algorithms and regression algorithms to classify, detect, localize and evaluate the severity of the damage. Meanwhile, GA can be applied to identify training parameters as well as to solve the local minima problems from ANNs. To demonstrate the method, an analysis of a bridge is performed. Finite Element (FE) model of the bridge is created using measured vibration data and it is employed as training data for the combined ANN-GA method in the model updating process. The updated model will then be used as a baseline model for damage identification. The result shows that the proposed ANN-GA algorithm provides a high level of accuracy and efficiency in detecting damage in the considered structure.

This paper proposes a hybrid optimization algorithm combining particle swarm optimization (PSO) with genetic algorithm (GA) to update a railway bridge. PSO is an evolutionary optimization algorithm based on global search techniques to look for the best solution. Nevertheless, since PSO relies crucially on the quality of initial particles, it may reduce its effectiveness and robustness in tacking optimization issues. If the positions of initial populations are too far from the global best, it is challenging to determine the best solution. To overcome these shortcomings, we propose a hybrid optimization algorithm applying the advantages of both PSO and GA. GA is first used to generate the most elite populations based on its crossover and mutation characteristics. Those populations are then employed to seek the best solution based on the global search capacity of PSO. The experimental measurements of the railway bridge are carried out under ambient vibrations used to validate the proposed algorithm (PSO-GA). The result demonstrates that PSO-GA, GA, and PSO possibly determine uncertain parameters of the bridge exactly and PSO-GA surpasses GA alone and PSO alone in terms of convergence level and accuracy.

In this paper, an enhanced damage indicator using a flexibility index is presented and applied to different complex structures to predict the exact location of damage. Finite Element Method (FEM) is used to model three complex structures, namely a 37-bar planar truss, a 52-bar planar truss, and a 52-bar space truss (Dome structure) to study the effectiveness of the proposed indicator. Single and multiple damage scenarios with different damage levels are considered. The results show that the proposed indicator provides an accurate location of damage. Next, to quantify the damage and assess its severity, two optimization techniques, namely Atom Search Optimization (ASO) and Salp Swarm Optimizer (SSA), which are recently invented, are used to solve an inverse problem. The objective function is based on the measured and calculated enhanced damage indicators. Both optimization techniques provide good results, however the convergence performance and CPU time are better for ASO than SSA. Finally, the proposed approach is tested using a benchmark structure, namely a high-rise tower (Guangzhou TV Tower) to predict the damage location at different floors. The results indicate that the proposed methodology is accurate and fast to predict single and multiple damages.

This paper presents an approach for stiffness identification of node joints of a large-scale truss bridge (Nam O Bridge in Vietnam) based on vibration measurements and model updating. Vibrations are recorded under ambient conditions using piezoelectric sensors. Excitation is due to wind, micro-tremors, or train passage. From these vibrations, modal parameters are extracted. Modal analysis is also performed using a finite element model created in MATLAB. Afterwards, model updating is applied to minimize the discrepancy between numerical and experimental modal analysis results. Evolutionary algorithm such as particle swarm optimization (PSO) based on a global search technique is employed. Three scenarios of boundary conditions of truss joints (pin, rigid, and semi-rigid) are considered. The result of model updating shows that semi-rigid (using rotational springs) joint conditions represents correctly the dynamic behavior of the considered bridge.

This paper presents an approach for damage identification in a steel structure using Cuckoo Search (CS) algorithm. CS is an evolutionary algorithm based on global search techniques, which provides a higher opportunity for seeking the best solution and avoid local minima. A steel beam calibrated on experimental modal analysis is applied to assess the efficiency of the proposed algorithm. While a finite element (FE) model is created using MATLAB to estimate structural dynamic behavior, measurement is carried out using excitation sources of a hammer. Dynamic characteristics are selected as an objective function to minimize the discrepancy between the results of numerical model and measurements. The results show that the proposed algorithm can accurately identify damage location and extents in the considered structure.

In this paper, we propose an efficient Artificial Neural Network (ANN) based on the global search capacity of evolutionary algorithms (EAs) to identify damages in laminated composite structures . With remarkable advances, ANN has taken off over the last decades. However, ANN also has major drawbacks relating to local minima issues because it applies backpropagation algorithms based on gradient descent (GD) techniques. This leads to a substantial reduction in the effectiveness and accuracy of ANN. Some researchers have been come up with some solutions to tackle the local minimal problems of ANN by looking for starting beneficial points to eliminate initial local minima based on the global search capacity of stochastic algorithms. Nevertheless, it is commonly acknowledged that those solutions are no longer useful or even counterproductive in some cases if the network contains too many local minima distributed deeply in the search space. Hence, we propose a novel approach applying the fast convergence speed of GD techniques of ANN and the global search capacity of EAs to train the network. The core idea is that EAs are employed to work parallel with ANN during the process of training the network. This guarantees that the network possibly determines the best solution fast and avoids getting stuck in local minima. To enhance the efficiency of the global search capacity, in this work, a hybrid metaheuristic optimization algorithm (HGACS) of EAs is also proposed, which possibly gains the advantages of both Genetic Algorithm (GA) and Cuckoo Search (CS). GA is applied to generate initial populations with the best quality derived from the ability of crossover and mutation operators , whereas CS with global search capacity is used to seek the best solution. Moreover, to deal with the large amount of data utilized to train the network, a vectorization technique is applied for the data of the objective function, which considerably decreases the computational cost. The obtained results prove that the proposed method is superior to traditional ANN, other hybrid-ANNs, and HGACS in terms of accuracy, and significantly reduces computational time compared with HGACS.

With recent ground-breaking advances, machine learning (ML) has been applied widely in numerous fields in this day and age. However, because of the application of backpropagation algorithms based on gradient descent (GD) techniques, the network of ML may be trapped in local minima, especially if its starting point is not on the same side of the global best or the network contains too many local minima. This drawback may reduce the accuracy and effectiveness of ML. To transcend these limitations of ML, numerous researchers have employed algorithms based on global search techniques to eliminate initial local minima of the network by looking for a beneficial starting point. Nevertheless, those solutions are only valid under certain circumstances when the network only contains a few local minima and they are distributed on the same side. With complex problems such as structural health monitoring (SHM), the network always exists of different error surfaces with numerous widely distributed local minima. The approach of the selection of a good starting position for the network may no longer be useful. Therefore, this paper proposes a novel machine-learning based on an evolutionary algorithm, namely Cuckoo search (CS) to solve the local minimum problem of ML in the most radical way. CS algorithm based on the global search technique is employed to work parallel with ML during the process of training the network. This win-win approach has both advantages of GD techniques (fast convergence) and stochastic search techniques (avoiding being trapped in local minima). The core idea of the proposed method is recapped as follows: (1) ML using the GD technique is first applied to speed up convergence; (2) if the network gets stuck in local minima, CS with global search capability is applied to assist particles in escaping from local minima; (3) the GD technique is applied again to increase the convergence speed. Steps 2 and 3 are repeated until the target is achieved. Additionally, to handle the large amount of data used to train the network, we also apply a vectorization technique for the data of the objective function, which significantly reduces the computational cost. This is another contribution of this work. To assess the performance of the proposed approach, both numerical and experimental models with different damage scenarios are considered. The results showed that the proposed approach completely outperforms CS, ML, and other hybrid ML in terms of accuracy and considerably reduces calculational costs compared to CS.

In this paper, a novel approach to model updating for a large-scale railway bridge using orthogonal diagonalization (OD) coupled with an improved particle swarm optimization (IPSO) is proposed. Particle swarm optimization (PSO) is a well-known and widely applied evolutionary algorithm. However, as other evolutionary algorithms (EAs), PSO has two main drawbacks that may reduce its capability to tackle optimization issues. A fundamental shortcoming of PSO is premature convergence. On the other hand, since PSO employs all populations to seek the best solution through iterations, it is very time-consuming. This makes PSO as well as EAs difficult to apply for optimization problems of large-scale structural models. In order to overcome those drawbacks, we propose coupling OD with IPSO (ODIPSO). OD is applied to arrange the position of particles and to select only particles with the best solution for next iterations, which helps to reduce the computational cost dramatically. There are several significant features of ODIPSO: (1) IPSO is employed to tackle the problem of premature convergence of PSO; (2) only one guide is used to update the velocity of particles instead of utilizing both guides, consisting of the local best and the global best; and (3) in each iteration, only the velocity and the position of the best particles are updated. In order to assess the effectiveness of the proposed approach, a large-scale railway bridge calibrated on the field is employed. This paper also introduces the use of wireless triaxial sensors (replacing classical wired systems) to obtain structural dynamic characteristics. The appearance of wireless triaxial transducers increases significantly the freedom in designing an ambient vibration test. The results show that ODIPSO not only outperforms PSO, IPSO and OD combined with PSO (ODPSO) in terms of accuracy, but also dramatically reduces the computational time compared to PSO and IPSO.

This paper presents an effective method for crack identification to improve the training of Artificial Neural Networks (ANN) parameters using Jaya algorithm. Dynamic and static datasets are introduced using eXtended IsoGeometric Analysis (XIGA) to improve the accuracy of the proposed application based on the frequency and strain measurements. Based on the concept used in our previous works, XIGA provided more accurate results for fracture mechanics applications than other modelling techniques. Therefore, XIGA datasets of cracked plate are used to improve ANN technique for static and dynamic analyses. Model updating of the cracked plate is considered by introducing the mass of accelerometers and identifying Young’s modulus of the plate and stiffness of springs using Jaya algorithm. The difference between measured and calculated frequencies is used as an objective function to calibrate the XIGA model. The crack length is predicted using an adaptive approach without any previous knowledge based on the data provided from a numerical model. Jaya algorithm is used to optimize the most important parameters of ANN. Several numerical examples with different crack scenarios and different boundary conditions are studied in order to evaluate the proposed approach. The results show that the proposed application is able to predict all considered scenarios and accurately identify the crack length. Experimental data of cracked plates are used to validate the numerical predictions. Hence, this application is found to be robust and accurate for crack identification in plates.

This paper presents a new approach for damage detection in structures by applying a flexible combination based on an artificial neural network (ANN) and cuckoo search (CS) algorithm. ANN has become one of the most powerful tools employing computational intelligence techniques to tackle complex problems in numerous fields. However, due to the application of backpropagation algorithms based on gradient descent, a major drawback of ANN is the common problem of local minima that acts as a great hindrance to the search for the best solution. To overcome this disadvantage, we propose to combine ANN with evolutionary algorithms based on global search techniques. This paper employs CS to improve ANN training parameters (weight and bias) by minimizing the difference between real and desired outputs and then using these parameters to generate the network. Two numerical models, comprising a steel beam calibrated using experimental measurements and a large-scale truss bridge, are used to assess the robustness of the proposed approach. The results demonstrate that ANN combined with CS (ANN-CS) is accurate and requires a lower computational time than ANN, and evolutionary algorithm (EA) alone in terms of structural damage localization and quantification.