Abstract

     This paper presents a rapid machine learning-based damage detection framework for identifying the damage extent of concrete shear wall buildings. For this purpose, a parametric study was carried out to determine the most efficient machine learning algorithm in classifying the damage states of the building. According to this parametric study, the K-Nearest Neighbor (KNN) learner was selected as the reference prediction model because of the higher accuracy achieved by this algorithm. Bayesian Optimization (BO) algorithm was used to tune the hyperparameters affecting the accuracy of the model. The most efficient attributes were selected from the set of damage indicators through the BO algorithm to train the model. Three different benchmark buildings, including 7-,9-, and 13-story concrete shear wall buildings, were used to evaluate the robustness of the proposed framework. A suite of 111 pair motions, originally developed for the SAC project, were employed to create a generalized dataset. These motions were uniformly scaled from 0.05 g to 1.5 g to expand the intensity range of the events. All the acceleration signals were polluted to 10% noise using white Gaussian signals to simulate the field condition. Results reveal the efficiency of the proposed framework in identifying the extent of damage in concrete shear wall elements of the building. In addition, a parametric study was conducted to illustrate the reliability of two commonly used features, called Cumulative Absolute Velocity (CAV) and the energy ratio between the acceleration response and the input excitation, in determining the damage states of the shear walls under seismic motions.

Introduction

     Assessment of structural safety is essential for post-earthquake restoration. Generally, to evaluate the post-earthquake vitality of the exposed structures, a complete visual inspection is required [1]. Coordination and implementation of the manual visual inspection needs several dedicated teams and monetary resources. In this regard, considerable efforts have been carried out to automate the visual inspection process, e.g., image-based visual inspection [2]. However, such an engineering visual inspection is only able to detect the visible defects that occurred in the structures [3]. This means that some serious invisible damages may be left latent during the visual inspection. The process of identifying and tracking the structural damage is known as the Structural Health Monitoring (SHM) [4,5]. In SHM, damage detection is related to the methods developed for identifying the probable existence, severity, and location of the structural damage. Model-based and data-driven methods are two of the most commonly used strategies proposed for damage detection. Model-based methods generally involve a system identification algorithm paired with a finite element analysis to update the structural model [6]. The performance of the model-based approaches directly depends on the accuracy of the information about the physical properties of the under-study structure. In addition, updating the finite element model based on the physical properties of the structure is computationally expensive for large-scale structures, and rapid condition monitoring could be challenging in this condition [7,8]. On the other hand, data-driven methods apply statistical learning algorithms to the vibration data captured from the structure. This method uses learning algorithms to construct a classification or regression learner for predicting structural damage [9, 10]. Tsou and Shen [11] proposed the use of Neural Networks (NNs) for predicting the severity and location of the structural damage. They used the variations in the modal properties of the structure as the damage feature to identify the damage. Worden et al.

Conclusion

     This paper presented a rapid algorithm for identifying the severity of local damage in the concrete shear wall buildings. A total number of 1884 nonlinear response history analyses were conducted for each building using the SAC motions. A suite of damage indicators was extracted from the acceleration signals to construct the prediction models. A parametric study was carried out to determine the most efficient learner for classifying the damage states of the buildings. The KNN classifier was selected to construct the predictive models because of its maximum accuracy compared to the other algorithms. The Bayesian optimization algorithm implemented to tune the hyperparameters of the classification learners. The main conclusions derived from the study are as below:

Abstract

     As a key component in composite beams, the studs are prone to fatigue failure under reciprocating load. Fatigue problem is more prominent especially under circumstances of concrete cracking, stud corrosion or with initial damage. Therefore, static and fatigue tests are carried out to study the fatigue properties of push-out specimens with initial defects, including stud damage and concrete cracking. The test results show that the damage of stud has a great influence on fatigue life, while the crack of concrete has a little influence on fatigue life. The damage of stud has a greater influence on fatigue performances compared to static properties. In addition, based on the dissipative energy theory and experimental phenomenon, the roles of stud and concrete in the process of stud fatigue failure and their interaction are summarized. The dissipative energy mainly includes the shear plastic strain energy of the stud, the plastic strain energy of concrete under compression after the stud, and the heat energy generated by the friction between the stud and concrete.

Introduction

     Steel-concrete composite beams have become popular in recent years for use in bridge engineering [1,2]. Under the complicated environment and load, there may be some damage to the composite beam. There might be some damage, such as corrosion, fractures of weld leg and longitudinal or transverse cracks in the concrete deck. The static and fatigue behavior of stud connection is sophisticated and affects the design resistance and stiffness of composite beams [3,4]. The damage of the stud and concrete may have a great influence on the shear fatigue performance of the interface.

      Welding defects significantly lower the shear capacity of studs, as is widely known. There are many types of welding defects that might appear, such as inadequate penetration, absence of fusion, and gas inclusion [5]. To determine how the damage location and degree affected the static behavior and shear capacity of stud shear connections, six specimens with identical geometrical dimensions were investigated by Qi et al [6]. The test findings showed that, in comparison to the shear capacity of normal specimen, a loss of up to 36.6 % and 62.9 % of the shank area might cause a falling shear capacity of 7.9 % and 57.2 %, respectively. The push-out test was simulated using mathematics, and the results of the test were used to validate the study. It was shown that even though the area of the stud had significantly decreased, the shear capacity was not affected by the severity of the damage when the damaged part was placed 0.5d (where d is the shank diameter) from the stud root, based on the numerical model.

Conclusions

     Static and fatigue tests were carried out on studs in steel–concrete composite study the effects of initial damage in stud and concrete on fatigue performance. The following conclusions are drawn:

     1.  Compared with nondestructive specimens, the fatigue life of transverse crack and longitudinal crack specimens were reduced by 22 % and 4.9 %, respectively, which is determined by the stress characteristics of the stud. Transverse crack makes the stud easily in a state of bending and shear composite, while longitudinal crack does not

     2.  Compared with the preset cracks of concrete, the damage of stud and weld collar had more influence on fatigue performance, in which the loss of stud section of 62.9 % leaded to only about 10 % of fatigue life. The initial damage of stud accelerates the damage accumulation in the stable fatigue development stage, while the concrete cracking leads to the increase of residual slip in the early stage.

     3. The existing calculation methods of residual slip are not completely applicable to the test results of different stresses, and most of the calculation results are too small, which requires some modification of their coefficients.

Studies on enhancing the strength and usability of steel members for high-rise buildings and long-span bridges have been conducted (Xiong and Liew, 2015; Khan et al., 2015; Mikkola et al., 2015). Furthermore, numerous experiments and numerical analyses of the damage assessment of steel columns, beams, bracings, and other structural members under repeated loading such as an earthquake load have been conducted (Nguyen et al., 2015; Park, 1993; Gao and Usami, 2000). Owing to the emergence of ultrahigh-strength steels, structural steel shapes are a very important design parameter considering the local buckling, and buckling expressions according to the structural steel shapes and load conditions have been proposed (Seif and Schafer, 2010). In addition, a trilinear kinematic hardening model was proposed to simulate the behavior of ultrahigh-strength steel columns with an H-section under cyclic loading (Wang et al., 2015; Chen et al., 2016). A steel member that has a low slenderness ratio is less affected by additional flexure owing to the axial force, and lateral buckling is caused by local buckling. Therefore, the local buckling of plate elements is the major reason for failure. According to an experimental study by Park (1993), a steel member accumulates damage during repeated loading, and local buckling leads to buckling of the entire member (Park, 1993). As a very large amount of plastic deformation repetitively occurs in the steel member, the steel member will eventually fracture. It was found that a steel member element can fail when such a large amount of plastic deformation occurs in a steel member element damaged by local buckling, even when the number of repeated loading cycles is low. In this study, a damage index is developed to quantitatively comprehend the degree of damage due to a large amount of plastic deformation under the repeated loading of steel members. By using this damage index, a method for analyzing the damage to box-type steel members is developed.