Search

In the field of structural health monitoring, vibration-based damage identification remains a formidable challenge. Key to this challenge is the establishment of a reliable association between observed vibration characteristics and the actual state of structural damage (e.g. stiffness reduction). This association not only accurately indicates the presence of damage, but also the location and severity of the damage. To solve this complex pattern identification problem, a large number of approaches, including deep learning, have emerged in recent years. In this paper, we propose a new structural damage identification method that utilizes the vibration information of the structure and a convolutional neural network based on Alex NET improvement. The method consists of calculating the acceleration response power spectral density of damaged and undamaged structures under impact loading separately, and then making a difference between the two power spectral data, and subsequently introducing these power spectral difference data into the convolutional neural network for training. The use of power spectral density analysis as a preprocessing step converts the time-domain signals into frequency-domain signals, and this conversion allows the convolutional neural network to capture and learn from the specific frequency characteristics of the data, thus facilitating the learning process of the neural network model. In this paper, the effectiveness of the method is critically evaluated through numerical simulation and experimental validation, and 3% and 5% noise are added to the numerical study to test the robustness of the method. During the convolution neural network training process, the optimal training mean squared error (MSE) is 5 × 10−6 in the case of no noise addition; the optimal training MSE is 1.3 × 10−5 in the case of noise addition. Both the results of simulations and experiments confirm the high accuracy and good robustness of the method in localizing structural damage. [ABSTRACT FROM AUTHOR]

The corrosion of longitudinal rebar and stirrups will cause serious degradation to the bond performance between steel bars and concrete. To investigate the impact of both longitudinal rebar and stirrup corrosion on the bond behavior, 35 reinforced concrete (RC) pull-out specimens were subjected to accelerated corrosion tests of electroosmosis-constant current-dry wet cycles and pull-out tests, and then the influence laws of the degree of corrosion of the longitudinal reinforcement, degree of corrosion of the stirrups, cover thickness, stirrup spacing and longitudinal reinforcement diameter on the bond performance were studied. The bond behavior between corroded steel bars and concrete was analyzed from the aspects of corrosion form, failure mode and bond strength. Subsequently, based on the experimental data, a modified bond-slip constitutive model that considers the design parameters, stirrup corrosion and longitudinal rebar corrosion was built and verified. Finally, based on the established modified bond-slip constitutive model, the corroded RC column was simulated, and the accuracy of the established modified bond-slip constitutive model was verified at the component level. [ABSTRACT FROM AUTHOR]

• This paper proposed a substructure damage identification approach based on sensitivity analysis of power spectral density. • The IRS condensation of substructures has reduced the degrees of freedom of the structure and simplified the structure model which raised computational efficiency. • In the process of solving the damage parameters, the number of frequency points are increased to reduce the selection of measuring points. • The damage identification method proposed in this paper is verified by numerical simulation and experimental model respectively. In this paper, a new method for substructure damage identification under random excitation is proposed, which is based on the sensitivity of the power spectral density (PSD) analysis. The finite element model is divided into different substructures and combined with the improved reduced system (IRS) method to achieve model condensation. When the system is subjected to excitations, the sensitivity of dynamic response of power spectral density function to structural damage parameters is obtained. The substructure is regarded as an independent structure, and the structural damage identification is realized by using the method of model updating. By increasing the number of frequency points used for damage identification to reduce the selection of measuring points. The identified results are obtained by updating the damaged parameters using sensitivity of PSD with few measurement points. The assembled finite element model is obtained by condensation and the number of elements is greatly reduced. The substructure damage identification method based on the PSD sensitivity is proposed in this paper, it not only reduces the size of analytical model involved in model updating but also decreases the uncertain parameters that need updating, thus resulting in a rapid convergence of the iterative process. The numerical and experimental model of a six-story steel frame structure with different damage conditions demonstrates the satisfactory identification results from the proposed method. [ABSTRACT FROM AUTHOR]