---

Abstract In this paper, using Artificial Neural Networks (ANNs) and Finite Element Method (FEM), health monitoring of damaged cantilever beams having longitudinal cracks is discussed. The main focus is devoted to the nonlinear behavior (breathing) of crack, which, to our knowledge, is taken into account in the crack detection of structures using ANNs, for the first time. Thus nonlinear behavior of crack is modeled using FEM.The changes in the natural frequencies (due to crack) of various vibration modes were implemented as input for training and testing of ANNs. By producing various scenarios for sound and damaged beams (with different damage location and severity), two specific classes of ANNs were trained to predict the location and length of longitudinal cracks. The Results showed a promising prediction for the length of cracks by the proposed methodology. Also a considerable approximation observed in the prediction of cracks location.

---

In this study, the structural health of a thick steel beam, made of ST-52, is inspected by ultrasonic guided wave propagation method using piezoelectric wafer active sensors that is one of the most important techniques of on-line structural health monitoring. The key parameters of the diagnostic waveform such as excitation frequency and cycle number are determined in relation to beam dimensions as well as pulse-echo configuration of PZT active sensors attached to the beam. Finite element simulations were conducted to characterize wave propagation in the beam, and the signals of wave propagation were experimentally measured. For signal processing and feature extraction, continuous wavelet transform and scaled average wavelet power technique are used. Using the extracted features, probable existing damage in the structure is detected, localized, and intensified. The acquired results are representing a higher precision of the implemented method for damage identification and characterization with respect to a previous study.

---

One of the most important challenges in the field of structural health monitoring and non-destructive testing is to assess some features of damages in structures, like the shape of damaged region. To reconstruct the shape of damage there are different methods in tomography. Already, researchers have used two general types of shape reconstruction techniques: transform based methods and algebraic reconstruction methods. Both methods suffer from some disadvantages like high sensitivity to incomplete data sets, bulky and expensive scanning hardware or low image resolution. In this work, a novel method to find the shape of damage via polygon reconstruction technique in tomography using the Radon transform is introduced. In this technique, damaged region is approximated by a polygon which the number of its sides is chosen arbitrarily, and the aim is to find this polygon’s vertices. To achieve this goal, an aluminum plate with a triangular hole as the damage was modeled in software. Then beams of guided Lamb wave were propagated toward the damaged region using arrays of piezoelectric transducers in just a few numbers of angles. Finally the polygon’s vertices were determined by processing the reflected signals from the damaged region. The results confirmed the efficiency of the proposed method.

---

Structural damage identification can be considered as the main step in Structural Health Monitoring (SHM). We can find many different methods which use structural dynamic responses for damage prognosis. Although some of them are concentrated on solving an inverse problem for damage identification, others suggest a direct procedure for defect detection. Despite the good performance of these methods in damage identification, researchers are attempting to find efficient and simple methods for damage identification with high level of accuracy. This paper presents a reference-free method for structural damage identification under earthquake excitation. Damages are defined as some changes in the special instants during an earthquake occurrence and structural time history responses are used as an input signal for discrete wavelet analysis. Finally the “detail coefficients” are inspected for determining the damage characteristics, such as the appearance, the time sequence, and the location of damage(s). Although the peak values in the detail coefficients can show the existence and time sequence of damage, for determining damage location we should inspect these peaks for finding the maximum value. As a result, the associated element with a signal which has maximum peak, can be considered as the damaged element. The applicability of the presented method is demonstrated by studying three numerical examples. First example is devoted for damage identification in a four-story shear frame. It is assumed that we have equipped all of the stories by sensors for recording structural responses. Three different damage scenarios with single and multiple damage cases under two samples of earthquake records, namely El-Centro (1940), and Northridge (1994) earthquakes, are studied. In addition, we study the effect of using different wavelet mother functions and different input signals, such as displacement and velocity responses. All of the obtained results emphasize the applicability of the presented method in damage identification. In second example, we consider a concrete simple beam with ten elements by simulating two different damage scenarios. In this case, we inspect the applicability of the method by considering only the transitional degrees of freedom (DOF) as the equipped DOFs by sensors. This can be interpreted as using limited number of sensors. In addition we use the displacement time histories for damage identification. For having a clear strategy in damage localization, we propose two rules for judging about elements’ health which are based on seeking maximum values of the wavelet coefficients in the damaged instants. Obtained results show the good performance of the presented method in finding time sequence of damage occurrence and damage location. In the third example, we investigate the applicability of the presented method in the presence of complex models of damages by defining bilinear stiffness reduction. In this case, although damage can cause some reduction in the effective stiffness of damaged structure, this reduction is different in positive and negative displacements. Two different damage scenarios are simulated on a single DOF structure under different excitations, namely earthquake excitations and generated White Noise excitation. Obtained results reveal the robustness of the presented method in damage prognosis in the presence of complex damage models.

---

In the current study, seismic cracking identification of concrete dams is conducted based on extended finite element method (XFEM) and Wavelet (WT) transform. First, the dam is numerically modeled and analyzed using the finite element method (FEM). Then cracking capability to the dam structure is added by applying the XFEM without introducing the initial crack, and the dam is analyzed under the seismic excitation. In fact, the whole dam structure is potentially under damage risk, and any zone reaching the fracture limit, begins to crack, which grows in the structure. This crack is usually unpredictable and is not easy to detect, therefore the structural modal parameters and their variation should be investigated based on structure response by using time-frequency transform. Results show that, investigating time-frequency window of the structure response and model parameters obtained from the numerical model, the history of physical changes occurred in the structure, cracking initiation time and damage localization is performed from comparing the intact and damaged vibration modes. Moreover, investigating the first natural modal indices of the intact and damaged structure, damage initiation and its location on Koyna dam height is easily detected, while for the second indices it is not performable.

---

Operational modal analysis (OMA), as a branch of the system identification, plays a very important and practical role in determining the dynamic characteristics of the structures. In the operational approach that is implemented based on the ambient vibration test, the ambient and operation loads are considered as the excitation source of the structure. In the present research, an integrative method composed of frequency domain decomposition (FDD) and wavelet transform (WT) called FDD-WT is proposed in order to identify the natural frequencies and damping ratios of an arch concrete dam. Furthermore, the wavelet transform of the seismic responses is also calculated in order to validate and compare the results. For this purpose, the time and frequency position of the system modes during the different earthquakes is evaluated using the time-frequency representation obtained from wavelet transform. In this paper, Pacoima arch concrete dam located in California, US is selected as the case study and the seismic records related to 1994 Northridge, 2001 San Fernando and 2008 Chino Hills earthquakes are also used to evaluate the dynamic characteristics and structural health monitoring during the period between 1994 to 2008. Investigation of changes in the natural frequencies of the structure indicates that the dam had taken serious damage during 1994 Northridge earthquake (about the fourth second), while the vibrations of the concrete structure has been almost linear during the first 4 seconds of the earthquake and also in 2001 and 2008 earthquakes.

---

According to the vital role that bridges play in transportation system and also communications of a society, monitoring their structural safety and keeping theme in service is crucial. Numerous methods have been proposed for detecting probable damages in bridges. Unfortunately most of them are based on comparison between the response of bridge in an intact and damaged state. Therefore intact state response must be known. However, not always it’s true in practice. So proposing a method which can determine and localize damages without prior knowledge of intact state is necessary. Such a method which was proposed by Sun et al. is studied carefully. Through the aforementioned method, the dynamic displacement response of a simply supported beam was decomposed into a dynamic component and a quasi-static component. Using Maxwell-Betti law of reciprocal deflection, the quasi-static component was attributed to the static deflection of the beam. Later damage which is defined by loss of stiffness, could be localized based on the abrupt changes in the static deflection curvature as it is related to bending moment and flexural stiffness of a beam. It is found out that the decomposition approach proposed by Sun et al. is restricted to fact that only one mode of oscillation must be dominant and also the natural frequency of motion must be determined through experimental measuring. Another limitation is that the abrupt changes in the curvature diagram cannot be related to damage essentially as curvature is also affected by the bending moment. In this study two modifications were proposed to get more accurate results in localizing the imposed damages. The first modification is the use of EMD method in order to decompose the displacement response into its intrinsic mode functions. Hence the aforementioned method could be used in real bridge displacement responses as higher modes corporations can also be determined and extracted through EMD process and finally the quasi-static component is determined as the residue of EMD algorithm. Also the ambient noise may be decomposed from the original signal, improving the method to work in real situations. The second modification is creating an imaginary constant moment length in the beam by the use of super position principle. So sudden increase in the curvature diagram is essentially a damage. Different scenarios of damage were studied and both methods have been used to detect damage in each scenario. Results show a great improvement in detection and localization of damage using the improved algorithm rather than the original proposed method. Eventually a five span real bridge model was taken into study. The improved damaged detection method could clearly determine the longitudinal position of the damage.

---

---

---

In recent decades, the science of structural health monitoring has played a key role in preventing damage and extending the life of structures. To conduct behavioral assessment, it is desirable to use tools that achieve sufficient accuracy with low cost. The processing of behavioral data requires methods that are able to identify and correctly troubleshoot different levels of damage from existing information.
Nowadays, sensors are used to measure the behavior of structures including deformations and displacements and even deflections, but these sensors have some weak points. For example, Risk of damage to the sensor, pointwise and one-dimensional measuring, their data is difficult to analyze and using multiple or high-tech sensors becomes expensive.  
Optical behavior measurement and close-range photogrammetric operations have recently received attention due to their low cost and good accuracy. This method has some advantages like Indirect contact with objects, high-speed image capture, easy access to convenient digital cameras, low viewing costs, and the ability to process composite and instant data with easy operation. In addition, the high flexibility of this method in measuring accuracy and design capability to achieve predetermined accuracy is an important feature of this tool.
Analytical methods are based on rules or equations that provide a clear definition of the problem. These methods work well in the cases which the rules are accurately clear and defined but there are many practical cases for which the rules are not known or it is very difficult to discover that calculations cannot be performed using analytical methods.
Neural network is a generalizable model, which is based on the experience of a set of training data and therefore free of explicit law. Neural networks have the ability to collect, store, analyze, and process large amounts of data from numerical analyzes or experiments. Therefore, they have the ability to predict and build diagnostic models to solve various engineering problems and tasks
In this paper, an attempt has been made to use this method to measure and troubleshoot laboratory model of a scaled suspension bridge that has a relatively complex behavior. For this purpose, the structure was subjected to uniform static loading in three step levels with three states: healthy and damaged in the deck and cables. Damages were created quite intentionally in the laboratory model, and from the information obtained, a database of bridge behavior in various situations was created. In order to assess the feasibility of using different methods in data processing and troubleshooting, first the data in the database were used in a simple linear method (direct comparison) and training in algorithms of machine learning methods. After that, deliberate damage was done again in the laboratory structure to allow testing the efficiency and accuracy of different methods. Finally, the accuracy, precision, and stability of the data processing methods of the support vector machine and artificial neural network were compared.
The results showed that by object bundle justification of two-dimensional optical behaver measurement with close-range photogrammetry, a guaranteed accuracy of 0.0021 mm could be achieved. Using intensity image processing seems helpful to ease the calculation. Using high number of nodes in hidden layer makes it more difficult and time-consuming to train the neural network. In the first level of processing, the detection of the presence or absence of damage was associated with the complete superiority of neural networks with 100% accuracy and in the second level, the detection of the affected area, depending on the type of processing, the neural network with hyperbolic tangent transfer function archived 93% accuracy and the support vector machine archived 68% of the accuracy.

---

   Civil structures may experience unexpected loads and consequently damages during their life cycle. Damage identification has been a challenging inverse problem in structural health monitoring. The main difficulty is characterizing the unknown relation between the measurements and damage patterns. Such damage indicators would ideally be able to identify the existence, location, and severity of damages. In order to solve such problems, biologically inspired soft-computing techniques have gained traction. The most widely used soft-computing method, called neural networks is designed such that it can learn from data without a need of feature design process. Damage pattern can be detected using neural network. A deep unsupervised neural network can recognize patterns and extract features from data. In this paper a methodology is described for global and local health condition assessment of structural systems using vibration response of the structure. The model incorporates Fast Fourier Transform and unsupervised deep Boltzmann machine to extract features from the frequency domain of the recorded signals. Restricted boltzmann machine is a shallow neural network with two layer. First layer of restricted boltzmann machine called input layer and second layer of restricted boltzmann machine called hidden layer.Deep Boltzmann machine created by setting some restricted Boltzmann machine sequentional. Hidden layer of each restricted boltzmann machine is input layer of next restricted boltzmann machine. Each layer of restricted Boltzmann machine extract features form input data Recorded data divided to smaller vectors. Fast fourier transformation used to transform divided vectors into frequency domain.  A benefit of the proposed model is that it does not require costly experimental results to be obtained from a scaled version of the structure to simulate different damage states of the structure and only vibration response of the healthy structure is needed to training deep neural network. The input consists of a set of records obtained from the healthy state of the structure and another set of records with unknown health states. The model extracts information from both healthy and unknown sets to determine the health states of the unknown set. The healthy records are low intensity vibrations of the structure at least in one planar direction in the healthy state in the form of time series signals and The unknown records are low intensity vibrations of the structure on unknown state of health. Ambient vibrations can be due to wind, traffic, or human/pedestrian activities. An appropiate health index is defined and calculated for each part of the structure. The value of this index is between 0 and 1. The closer the value is to 1 the healthier the structure. To evaluate the efficiency of the proposed method a building structures with 35 story has been simulated in OPENSEES. Data collection should be selected appropriately to prevent errors. Obtained result demonstrate that proposed method has about 95 percent efficiency to predict damages and their severity. Different damage state put on due to three earthquakes with different severity. Structural health index calculated after each earthquake. Calculated structural health index demonstrate efficieency of proposed method for detecting damages and severity of damages.

---

The health of structures, provision of safety, and the sense of security are among constant requirements and perpetual challenges of engineering and managers in the field of crisis management. Erosion and occurrence of minor local damage to structures and structural members in the early stages of construction or during operation, especially in critical structures such as power plants, tall buildings, stairs, dams, airports, and hospitals, have always been among major problems. As time passes, Structures are affected by a variety of natural and non-natural destructive factors such as earthquakes, non-systematic excavations, dynamic vibrations resulting from explosions and heavy vehicle traffic. In addition, factors such as serviceability expectation beyond the design capacity of structural elements and failure to meet the latest expectations imposed by regulations, use of poor-quality materials and execution problems will reduce efficiency and, consequently the service life of structures. Also, the spread of local damages in structures can impair the overall health of the structure. Undoubtedly, knowledge of structural health and safety is of vital importance and structural health monitoring is recognized as one of the most important subjects that has received a lot of attention from researchers. Plates are one of the most important structural elements that can, when damaged, progressively transfer damages to other elements and lead to overall structural damage incurring irreparable social and economic costs. Due to the increasing applications of steel plates, especially in building structures (as steel plate shear walls) in the present study attempts were made to focus on damage detection and localization as one of the most important steps of health monitoring using modal dynamic data (natural frequencies and mode shapes) and a proposed diagnostic method based on two-dimensional discrete wavelet analysis. To this end, the modeled steel plate was subjected to frequency analysis in ABAQUS finite element analysis software and the modal data associated with damaged and non-damaged states were extracted. The results showed differences between the frequencies and lack of correlation between primary and secondary vibration mode shapes based on the modal assurance criterion (MAC) and the angle between the primary and secondary mode shape vectors. Using a propoed damage localization index (DLI) based on the wavelet coefficients obtained from the diameter details of the two-dimensional wavelet analysis of the primary and secondary vibration mode shapes, the damage zones were detected by creating a maximum relative jumps in the DLI diagram. Studies showed that DLI values are sensitive to the damage severity of the damage zone and with increasing the damage severity, these values increase in fixed spatial coordinates in the damaged zone. Also, the DLI of one damaged zone is independent of the damage severity of the other damaged zones, and this is a positive advantage in the damage determination process. Otherwise, failure to detect one damaged zone may affect the detection of other damaged zones, and consequently pose problems in the process of damage detection and localization in cases where we are dealing with multiple damage zones. According to the results of the present study, DLI can be proposed as an efficient and effective index in detection and localization of damages in steel plate elements.

---

The health of structures, provision of safety, and the sense of security are among constant requirements and perpetual challenges of engineering and managers in the field of crisis management. Erosion and occurrence of minor local damage to structures and structural members in the early stages of construction or during operation, especially in critical structures such as power plants, tall buildings, stairs, dams, airports, and hospitals, among others, have always been among major problems. In case the damage sites are not identified timely and decisions are not made appropriately, substantial irreparable damage is expectable. Structures are always affected by various natural or unnatural factors such as earthquakes, explosions, and unprincipled excavations, which can aggravate the local damage in them and lead to their destruction, hence substantial human and financial losses. Therefore, it is highly crucial to monitor the health of structures and structural members. Therefore, health monitoring in structures and structural members is highly important. The column is one of the most significant members of engineering structures, especially in building structures and bridges, so that the instability of one of these members can lead to instability and destruction of the structure. Hence, design engineers expect columns to be the last members of structures to be damaged. In this paper, the health monitoring of the column as a structural member was performed by considering the effect of axial load on modal dynamic responses (i.e., natural frequencies and mode shapes). The results showed that the natural frequencies of all modes in both healthy and damaged states decreased with increasing axial load in proportions of the base critical load (the worst-case limit load). Also, at the same loads, the frequency of the healthy sample was always higher than that of the damaged sample so that the frequency difference between healthy and damaged states increased with greater severity of the damage. By introducing a Damage Detection Index (DDI) based on the wavelet coefficients obtained from the details of wavelet analyses of damaged and undamaged modes, the damage sites could be identified with a simple check and high accuracy by observing vibrations in DDI. Also, studies have shown that the DDIs of different damaged sites are independent of each other and are only affected by the severity of the damage and that the effects of axial load on DDI are very small and negligible. The independence of the DDIs of different damaged sites indicates the effectiveness of the proposed method in identifying damaged sites. Otherwise, failure to identify one damaged site may affect the identification of other damaged sites. The damage detection capability using the proposed DDI was investigated in columns with different support sections and conditions, and successful troubleshooting results were obtained. Moreover, investigations were performed with other wavelet functions, and the damage site was successfully identified. The proposed damage detection indicator is an efficient index in the column structures under the effect of axial load with axial buckling-prone support conditions and is proposed as a reliable method in identifying column damage sites in practical health monitoring of structures.

---

One of the active areas of research in concrete structure health monitoring is the detection of cracking in structural elements. Image classification and diagnosis have attracted the attention of many researchers nowadays. Due to the advancement of artificial neural networks and their fast processing, a convolution neural network has been established to detect these cracks. In this study, crack detection in concrete structures has been studied using a convolutional neural network, which can be generalized to all concrete structures for example dams, canals, bridges, shells, road infrastructure, foundations and concrete frames. Convolution neural network training was performed by the SGDM method with the ReLU activator function. Also, 250 iterations were employed for convolution neural network training, which gradually reduced the error rate and increased the accuracy of detecting cracked and uneaten concrete. The convolutional neural network is trained and validated with these 250 iterations. First, images with 32-pixel window dimensions are converted and separated. Then, the 32-pixel window, the 16-pixel, and the 8-pixel windows filter the images. A total of 3 stages of 32, 16, and 8-pixel filter images are analyzed and interpreted. During the training process, validation is performed every 20 iterations, and a diagram related to the accuracy of convolution network estimation and data classification error is drawn and completed. In convolutional neural networks, where the output is in pairs, the cracked and uncracked images of the network architecture are almost identical, differing only in minor specifications. The database of this research includes 20,000 images of cracked concrete and 20,000 uncracked concrete with dimensions of 3×227×227 pixels, 80% of it is used for training and the remaining 20% is used for validation of the convolution neural network. The accuracy of distinguishing cracked concrete from uncracked ones is about 98.16%, which is acceptable for operation and is considered practical. To evaluate the accuracy and performance of the proposed algorithm, each classification was performed against the overall accuracy, the confusion matrix was used for the validation data. According to the clutter matrix, 3861 images, in other words, 48.3% have been predicted to be correctly cracked, and 3992 images, equivalent to 49.9%, have been predicted to be correctly uncracked, and a total of 147 incorrect images have been predicted, which is equivalent to 1.8 percent. Images that are cracked and not accidentally cracked are predicted. They had crack lines in the corner of the image or cracks with a very small width, which the proposed convolutional neural network was mistaken for due to a very small crack width or crack position. Also, the results of the present study showed that the accuracy of this research has the best accuracy in less analysis time compared to previous studies. It should be noted that this method and its associated database can be used to produce a crack detection application on a smartphone, to be able to make a good initial estimate of the structure in question, such as a bridge or building after an unusual loading event, such as an earthquake or explosion.

---

Infrastructures such as bridges, buildings, pipelines, marine structures, etc., play an important role in human life. Since major disasters in these structures, such as the collapse of bridges or buildings, often result in many casualties, damages, and social and economic problems, most industrialized countries allocate significant funds to monitor their health. Failure detection strategies and continuous monitoring of the structure's condition, especially after natural and manufactured disasters, make necessary measures to be taken in the early stages of failure and can reduce the cost of maintenance and the possibility of collapse. Structural health monitoring methods often provide an opportunity to reduce maintenance, repair, and retrofit costs during the structure's life cycle. Most of the structural health monitoring methods proposed and implemented to identify possible damages depend on the structure's dynamic characteristics. One of the most practical methods, which uses the results of time domain system identification to detect failure, is the damage locating vector (DLV) method. The DLV method aims to identify load combinations that result in zero strain fields for damaged members in both healthy and damaged structures. To accomplish this, we find a vector in the null space of the difference between the plasticity matrices of the two structures. The singular value analysis method is used on the plasticity difference matrix to calculate this space. The method involves applying the space vectors to the healthy structure and recording the internal stresses of the members, which are then converted into weighted normal stress (WSI) using statistical tools. The member with a lower WSI is more likely to be damaged. Since truss structures are usually used in bridges, long-span structures, as well, as a wide range of steel buildings with simple and braced frames, this research uses the covariance-based random subspace optimal method in identifying the modal characteristics, which is very efficient in low excitations, has been taken into consideration to check and monitor health during operation. To investigate the capability of the DLV method in the damage detection of these structures, a 5-story residential building with a simple steel frame was subjected to the Centro earthquake. According to the desired damage scenario, the second and fifth floors were introduced as the damaged floors in this earthquake by applying a 30 and 50% reduction in the cross-section. To account for uncertainty in the data collection, we included the mean root square of the second sensor's data in the results for sensors 3 and 5. As a result of this uncertainty, the damping error between 5 and 10% has been shown in the damaged and healthy structure. Using the method (SSI_ORT), it was observed that two DLV vectors were extracted. Further, with the increasing uncertainty of the random vibration test results, it was observed that the extraction DLVs could extract the possible damaged elements with high accuracy. Next, the effect of input and output noises on the results obtained from the DLV method was investigated. This study found that by increasing the SNR of the outputs by 15% while increasing the error of the extracted modal characteristics, the extracted DlVs also lose sufficient accuracy in diagnosing structural damage.

---

Civil structures inevitably undergo damage over time due to various reasons such as environmental changes, material aging, load variations, and insufficient maintenance. Monitoring these structures, especially aging ones, is crucial to detect damage early on and implement suitable retrofitting measures, ensuring their continued safe and reliable operation without unexpected failures. Consequently, there has been significant research in this field, focusing on damage detection in both simple and complex structures. Health monitoring of highway bridges is essential for achieving a reliable transportation system. The vibration-based damage detection method uses changes in the vibrational properties of structures to detect damages and ensure a healthy state. In this study, the absolute value of the modal flexibility damage index and the modal strain energy damage index simultaneously are utilized to prevent unsafe decisions.
These absolute values of modal strain energy and flexibility damage indexes are utilized as the bases for training deep neural networks (DNNs). These indexes are applied to provide safe decisions and reliable damage evaluation in steel girder of the highway bridges. The convolution neural network (CNN) is utilized for damage quantification estimation. The CNN is one of the deep learning models that can currently be applied in 2D dominant approaches, such as pattern recognition and speech recognition. In addition, these networks can utilize the 1D time domain and vibrational signal data via the convolutional layer. The initial stage of CNN model comprises combined convolutional and pooling layers that apply different filters to extract features. Following this, fully connected layers, similar to a hidden layer of a multilayer perceptron are incorporated. Ultimately, these layers are classified together with a softmax layer. The convolution layer acts as a filter that convolutes the input layer with a set of weights, adding bias and applying an activation function to the outcome. Gradient descent momentum methods (SGDM) can be employed to optimize the parameters in CNN network architecture. SGDM estimates the gradient with high velocity in any dimension. This method mitigates issues such as jittering and saddle points by utilizing high-velocity inconsistent gradient dimensions and the SGD gradients, respectively. Additionally, when the Current gradient approaches zero, the SGDM provides some momentum.
The convolution neural network is trained to utilize damage indexes obtained from numerical simulation of the validated finite element model of the bridge. The damage indexes as the inputs for the neural network, which are achieved from different damage scenarios. Once network training and validation are completed, a well-trained neural network is used to detect, localize, and quantify the intensity of unknown damages. The proposed method overcomes previous damage detection problems such as false positive indications, the unreliability of a single damage index, and insufficient precision in determining the intensity.  The results revealed that the presented method, based on the dual updated damage indexes and CNN, practically and accurately identified unspecified single damages' location and severity in multi-span beams. The new training method of deep neural network systems overcomes some shortcomings in ANN. Moreever, this deep neural network training scheme can reduce the need for huge amounts of input data and enhance the accuracy of network training. The method is capable in predicting single damage scenarios in steel beam.   

 

---

Health monitoring allows small failures and damages to be identified and fixed before they turn into major and irreparable damages, to prevent loss of life and to make it possible to reinforce or improve it at the lowest cost. Currently, in the field of civil engineering, the health monitoring of structures is done in sensitive structures. One of the parts of the structure that may suffer initial damage before loading and during implementation due to difficulty in implementation is Concrete Filled Tube (CFT)columns.  One of the most likely damages in CFT columns is interface debonding damage. This damage causes the column to become weak and not benefit from the characteristics of steel and concrete together. Accordingly, in the present study, this damage and its severity in seismic (dynamic) parameters have been investigated. The results of the study show that damage causes changes in the mode shape of the structure, and it has caused a 6.38% reduction in the frequency in the first (main) mode of the structure. Also, the damping of the damaged specimen is reduced by approximately 12% compared to the healthy specimen. On the other hand, the results show that the severity of damage is very effective in changing seismic parameters. So that by doubling the damage area, the frequency decreased by approximately 0.35% and reached from 873.27 Hz to 20.870.20 Hz, but the mode shape of damage did not affect the frequency