Vertical ground motions have a significant impact on the seismic response of engineering structures, making the development of reliable vertical ground motion prediction models an important topic in the field of earthquake engineering. Traditional ground motion predictions are primarily based on actual strong motion records, using least squares regression to derive seismic motion parameter prediction models. However, conventional least squares regression often assumes linear relationships or predefined functional forms between variables, which may fail to fully capture the complex nonlinear relationships inherent in seismic data. In contrast, deep learning models can learn patterns from data and provide higher prediction accuracy for complex data distributions. In this study, deep learning methods were applied, and 9 953 vertical ground motion records from the NGA-West2 database were selected for model training and prediction. The self-DNN vertical seismic response spectrum prediction model was established and its performance was compared with traditional prediction models and a DNN neural network models. The results indicate that the vertical seismic response spectrum prediction model established using deep learning algorithms achieves high accuracy and delivers excellent predictive performance. These findings and analyses provide valuable references for vertical seismic response spectrum prediction and structural seismic design.

To study the longitudinal seismic response and damage state of the high-speed railway track-isolation bridge system, a 7-span 32 m simply supported beam bridge with CRTSⅢ type ballastless track structure laid on the bridge deck was taken as the research object. A finite element model of the track-bridge system was established, and the seismic response distribution law of each key component under different seismic waves, seismic intensity and bearing types was obtained through nonlinear time-history analysis. The results show that the longitudinal displacement of the beam body presents a stepwise distribution under longitudinal seismic excitation, with the maximum value occurring at the center of the bridge span. The maximum displacement of the fastener occurs at the expansion joint of the abutment, and extreme values appear at the expansion joint at each beam end. The displacement of the fastener is significantly affected by the spectral characteristics of different seismic waves. The maximum stress of the rail occurs at the expansion joints on both sides of the side span, and the normal stress of the composite slab section is caused by the combined action of axial force and bending moment components. After the bearing and track system enter the nonlinear state, compared with the increase in seismic intensity, the increase in longitudinal deformation of the vulnerable components shows a significant amplification effect and distribution imbalance. Considering the track system, compared with the friction pendulum bearing, the same ball direction double spherical surface bearing can significantly reduce the displacement response of the fastener, beam body and bearing. The track system has a significant inhibitory effect on the displacement of the bearing.

Ground motion has significant uncertainty, and different ground motion response spectra under the same amplitude have significant differences, which have a significant impact on the estimation of seismic damage to regional buildings. This study developed a regional building seismic damage simulation program suitable for multiple-story masonry and concrete frame structures, which can conveniently and quickly simulate regional buildings seismic damage under set earthquakes, and performed regional building damage simulations and probabilistic analysis with motion uncertainties. A typical urban region in Chifeng city was selected as the research area: 30 ground motions were selected to consider their uncertainty, and the ground motions amplitudes were modulated to the set intensity (0.05 g, 0.10 g, 0.20 g, and 0.40 g). Then seismic damage simulation of regional buildings under single and multiple seismic inputs were conducted respectively, the impact of seismic uncertainty on the seismic damage results of regional buildings was analyzed. Based on the seismic damage results under the set intensities, a probability density distribution model of regional buildings damage index based on Beta distribution was established. Results indicate that multiple seismic inputs take into account the uncertainty of seismic motion, which can more scientifically and objectively reflect the seismic damage situation of regional buildings; The established Beta distribution model can be used to estimate the post-earthquake damage of buildings in similar areas. The research results can provide reference for regional buildings safety assessment and seismic fortification.

Mountain terrain of significantly alters the propagation path and energy distribution characteristics of seismic waves. Through interactions such as reflection, scattering, and diffraction, the seismic response of local sites exhibits notable spatial variability. This terrain effect has a significant impact on the seismic response of engineering structures in mountainous areas and is one of the key factors contributing to the intensification of earthquake damage. To consider the impact of terrain effects on ground motion parameters in engineering seismic design, this study uses a railway station building site as the example. A three-dimensional finite element model of the mountain area where the station building is located was established. A viscoelastic artificial boundary is set for the model, and historical seismic data recorded by observation stations in the region were used as the ground motion input. The seismic response of the mountain region was obtained, and a comparative analysis of the input seismic motion and response results was performed to analyze the impact of the mountain height difference on the terrain amplification effect. The results show that at higher elevations (such as the freight yard and station building locations), the amplification effect is significant, while at lower elevations, the amplification effect is weaker, displaying a characteristic distribution along the height difference from large to small. The highest elevation of the site is more sensitive to high-frequency (10~20 Hz) seismic motion components. The peak ground acceleration is significantly positively correlated with the height difference, indicating that the height difference of the mountain terrain is a key factor influencing the site amplification effect. The study concludes that the terrain amplification effect is closely related to the height difference and topographical variations in mountainous areas, providing important theoretical guidance for the seismic design of major engineering projects in mountainous regions.

The Great East Japan Earthquake on March 11, 2011, triggered a massive tsunami that caused devastating destruction to buildings in coastal cities. However, in areas unaffected by the tsunami, buildings experienced relatively fewer collapses or severe damage, despite the high seismic intensity. This study explores the characteristics and impacts of seismic damage from this earthquake, reveals the intrinsic relationship between ground motion features and building damage, and analyzes the seismic damage data and spatial distribution of building clusters using a vulnerability model established based on the 1995 Great Hanshin Earthquake. Through the analysis of building collapse rate, we found that the computed results closely aligned with the actual seismic damage survey outcomes. The analysis indicates that seismic damage was concentrated in coastal areas such as Miyagi, Fukushima, and Ibaraki Prefectures, particularly in narrow inland zones near the coastline. Notably, the areas most severely affected were not always the closest to the epicenter or the zones with the highest intensity. Furthermore, significant differences in collapse rates were observed across different building codes, with buildings constructed under newer regulations showing a markedly lower collapse rate compared to those built under older standards. This research contributes to a better understanding of the seismic damage characteristics associated with offshore earthquakes, providing crucial insights for earthquake defense and disaster relief efforts.

To systematically study the influence of near-fault pulse characteristics on the seismic response of skew bridges, a four-span highway skew continuous girder bridge was taken and finite element models of the skew bridges with different skew angles were established by using OpenSees. The near-fault pulse-like ground motions with high-frequency components were artificially synthesized by using the decomposition-incorporation method. The effects of the moment magnitude and fault distance to the piers, main girders, beam-abutment collisions and exterior shear keys of skew bridges are analyzed. The results show that the sensitivity of the top displacement of the middle pier to the moment magnitude is greater than that of the side piers, but the sensitivity to the fault distance is not much different. The increase of the moment magnitude increases the displacement and rotation ratio of the main girder, as well as the maximum pounding force between the girder and the abutment. The results of increasing the fault distance are opposite to the previous results and the rotation ratio of the girder varies most when the skew angles is 15°. It is suggested to strengthen the shear reinforcement of the exterior shear keys for skew bridges with skew angles greater than 15°. As the moment magnitude is greater than 6.0 or the fault distance is 2~7 km, the exterior shear keys are in a state of failure under the near-fault ground motion.

At 7:58 on April 3, 2024, an M_S7.3 earthquake occurred (23.81°N, 121.74°E) in the waters of Hualien County, Taiwan, China, which was characterized as a thrust rupture. This earthquake is the largest earthquake since the “9·21” Chi-Chi earthquake. To comprehensively understand the characteristics and disaster effects of the earthquake and learn from the experience and lessons of the earthquake disaster, the cause of the earthquake is explained in combination with the mechanism of earthquake generation. Then, 714 strong ground motions recorded by 238 stations of the Earthquake Network Center of the Meteorological Bureau of Taiwan within 32 km of the fault are selected to analyze the engineering characteristics of these ground motions. Based on disaster investigation data, the relation between the earthquake and the structural damage of civil engineering as well as the distribution pattern of earthquake damage are discussed. The results show that ground motions of this earthquake have the characteristics of significant middle-to-high frequency contents with slow attenuation of PGA and Sa over rupture distance, causing the damage to medium-to-short period structures. The distribution of earthquake damage is concentrated in Hualien County, New Taipei City and Taipei City, and along the east side of the crustal butt belt of Taitung longitudinal Valley, the distribution is linear with the development of the fault zone. Hualien County is the most serious earthquake damage due to its proximity to the focal point, while New Taipei City and Taipei City are far away from the focal point, but the earthquake damage is also more serious due to the mountain amplification effect and basin amplification effect of ground motions. The relevant study can provide reference for the research of seismic fortification and seismic regionalization of medium-to-short period structures.

Ground motion prediction models are an important foundation for seismic hazard analysis. Currently, the research on vertical ground motion prediction models in China is relatively few, and most of the existing ground motion prediction models used parametric equations, which may have limited prediction accuracy. Therefore, the development of horizontal and vertical ground motion prediction models with better prediction accuracy and reliability is necessary for further research. To address the above problems, this study, uses 1991 sets of Chinese horizontal and vertical ground motion records. The Butterworth non-causal filter method is applied to filter and reduce the noise of Chinese ground motion. The Chinese horizontal and vertical ground motion prediction model (CHV-DNN) is developed based on the deep learning method, and it is comprehensively assessed in terms of model performance, physical characteristics, and intra-and inter-event residual analyses. Finally, a correlation coefficient model for Chinese horizontal and vertical ground motion is provided. The results show that based on the residual analysis results of the CHV-DNN model, the most of the inter-event residuals are mainly distributed in the range of [-1, 1], and most of the residuals within events are mainly distributed in the range of [-1.5, 1.5], and the intra-event and inter-event residuals are both uniformly distributed on both sides of the residuals 0 baseline, which validate the reliability and accuracy of the model; The CHV-DNN model has better prediction accuracy and also has well physical characteristics; the correlation coefficient model calculated based on CHV-DNN has been more reasonable. The Chinese horizontal and vertical ground motion prediction model developed in this study will provide a research foundation for horizontal and vertical seismic hazard analysis in China.

As a research hotspot in earthquake engineering, the performance-based seismic design concept has achieved mature applications in the seismic damage assessment of bridges, but its implementation in seismic design still needs further research. This study proposes a multi-objective optimization design method for piers based on seismic reliability by integrating the probabilistic seismic risk analysis framework with response surface theory and the improved Non-dominated Sorting Genetic Algorithm (NSGA-Ⅱ). First, the method for establishing seismic reliability of bridges is elaborated by combining seismic fragility and seismic hazard theories. A mathematical optimization model is then proposed with the seismic reliability of bridges and the material cost of piers as objective functions. A systematic design workflow for seismic optimization of piers is established by embedding response surface theory and the NSGA-Ⅱ. Subsequently, a typical highway bridge is taken as a case study. In accordance with the seismic design specifications for bridges in China, the seismic hazard curve and seismic vulnerability curve are developed, and the seismic damage characteristics of the bridge are analyzed. Finally, a response surface model for seismic reliability is developed to perform seismic optimization design for the case study bridge. The results show that the response surface model based on the quadratic polynomial can accurately describe the implicit relationship between the design parameters of piers and the seismic reliability of the bridge. The proposed seismic optimization design method in this paper can improve the seismic reliability of the bridge or reduce the material cost of the piers. Incorporating seismic reliability as an objective function directly consider the influence of piers on the seismic damage risk of the bridge. In addition, the multi-objective optimization seismic design can overcome the limitations of traditional empirical design methods and achieve more refined quantitative design. Designers can flexibly obtain the optimal solution from the Pareto solution set based on different optimization strategies.

To study the impact of floor slab construction on metro-induced vibration responses and to explore the feasibility of reducing structural responses by optimizing floor construction, the conventional floor slab, the thickened floor slab and the additional sub-beam floor slab were designed and manufactured. Field model tests were carried out on the three test floor slabs respectively to study their dynamic response under on-site metro vibrations. Taking conventional floor slab as the test control group, the effects of increasing floor slab thickness or adding sub-beam on the floor slabs' characteristics and metro-induced vibration responses were compared and analyzed. The results indicate that the resonance effect is the main reason of metro-induced vibration responses of the floor slab. Increasing the floor slab thickness or adding sub-beam increases the vertical modal frequency of the floor slab, alleviate the resonance effect, and reduced the time-domain acceleration responses of the floor slab. Both increasing floor slab thickness and adding sub-beam can reduce the vibration acceleration level of the floor slab in a wide frequency band, and can reduce the weighted vibration level by 6 dB and 4 dB respectively, thus improving the vibration comfort performance of the floor slab. Adding secondary beams achieves a vibration reduction effect similar to doubling the floor slab thickness without significantly increasing the floor slab's weight and engineering cost. It is recommended that the addition of sub-beams as a floor construction measure be priority in the design of new metro adjacent structures to mitigate metro-induced vibration responses and improve vibration comfort.