Abstract: Rapid seismic-damage assessment of buildings on a regional scale is of considerable significance for emergency response and recovery endeavors following earthquakes. Deep-learning methods provide a critical means for rapid seismic-damage assessment. However, this method faces the challenge of insufficient ground motion (GM) samples with high destructive power owing to the limited availability of strong GM data. To address this issue, a rapid seismic-damage assessment method for buildings on a regional scale based on spectrum-compatible data augmentation and deep learning is proposed. First, the continuous wavelet transform (CWT) method is used to construct a response-spectrum-compatible GM for augmenting strong-motion data. The augmented strongmotion database is then employed in conjunction with deep-learning algorithms to predict building damage in urban areas. The proposed method is illustrated through an analysis of individual building cases and regional scenarios. A comparative study is also conducted by comparing the proposed method with the widely used acceleration-amplitude-scaling-based data-augmentation method. The main conclusions are as follows. (1) The GM generated by the proposed data-augmentation method exhibits a smaller degree of dispersion in terms of duration and seismic responses for individual buildings and building clusters compared to those of the acceleration-amplitude-scaling-based data-augmentation method. (2) The proposed spectrum-compatible GM data-augmentation method demonstrates improved accuracy in deep-learning predictions, surpassing the traditional amplitude-scaling-based data-augmentation method. These results highlight the significance of employing the proposed GM data-augmentation method for deep-learning-based seismic-damage assessment. (3) The proposed method achieves a seismic-damage assessment accuracy of 87.4 % in a computation time of less than 1 s, resulting in a remarkable 1000-fold improvement in efficiency compared with time-history analysis method, which effectively fulfills the requirements for rapid post-earthquake emergency assessments of building clusters in a region.
Link to this article: https://doi.org/10.1016/j.soildyn.2024.108504
How to cite this article: CHENG Qingle, LI Aiqun, REN Haotian, CHEA Cheav Por, LIAO Wenjie, XIE Linlin. (2024). Rapid seismic-damage assessment method for buildings on a regional scale based on spectrum-compatible data augmentation and deep learning. Soil Dynamics and Earthquake Engineering. 178, 108504,
Abstract: Under the combined effects of in-plane deformation and out-of-plane acceleration during an earthquake, cracking occurs in the exterior infill walls of buildings causing debris to fall. The falling debris ineluctably blocks roads and hinders fast pedestrian evacuations. To identify the failure criterion for exterior infill walls, specialized test apparatus was designed, and pseudo-static experiments of infill walls were conducted in this work. The influences of different height-to-thickness ratios, drift ratios and out-of-plane accelerations on falling debris were taken into consideration in the proposed apparatus design. The experiment indicated that the infill blocks fell sooner and the falling area was larger for a thinner infill wall with a larger tilting angle. Based on the experimental results, a prediction model was proposed to evaluate the percentage of falling area of infill walls. Using the proposed failure criterion, a pedestrian evacuation simulation was conducted on the Tsinghua University campus as a case study. Moreover, the debris distribution and evacuation results under different earthquake intensities were also analyzed. The results of the case study revealed that a considerable area of roads located in densely built-up areas was covered by debris, resulting in a more than 30% increase in time for people in the nearby buildings to evacuate. Compared to the scenario without falling debris, the evacuation time with debris was 5% longer when subjected to MCE (Maximal Considered Earthquake). The proposed failure criterion for masonry infill walls can be employed in debris distribution calculation, which offers technical support for urban planning and evacuation drills.
Link to this article: https://doi.org/10.1016/j.ijdrr.2019.101372
How to cite this article: LU Xinzheng, YANG Zhebiao, CHEA Cheav Por, GUAN Hong. (2019). Experimental study on earthquake-induced falling debris of exterior infill walls and its impact to pedestrian evacuation. International Journal of Disaster Risk Reduction, p.101372.
Abstract: Earthquake-induced hazard of falling exterior infill walls is disastrous. Structural inter-story deformation and out-of-plane acceleration are the main factors induce damage and falling of infill walls. In order to study the mechanism of damage and falling of the infill walls under these two factors, a pseudo-static test of infill walls was designed in this paper. By placing the wall at different tilt angles, different out-of-plane accelerations of the infill walls during earthquake were simulated by using the acceleration component of gravity. Percentage of damage and falling area of the infill wall under different out-of-plane accelerations and in-plane drift is studied, and the fitting formula is provided. Then, aiming at RC-frame structure with exterior infill wall in different numbers of stories, the models of nonlinear structural analysis are established respectively to calculate the history response of acceleration, velocity and displacement of each floor. According to the maximum response of these structures and the experimental formula obtained from the test, the number of falling debris in different width-thickness ratio of infill wall and seismic acceleration is calculated, and the distribution of the debris after falling in projectile motion is calculated. Finally, according to the number and distribution of falling debris, the range of injury risk caused by damage and falling of the exterior infill walls under the earthquake is estimated.
Link to this article: https://doi.org/10.13577/j.jnd.2018.0103
How to cite this article: CHEA Cheav Por, LU Xinzheng, YANG Zhebiao, XU Zhen, ZHANG Xin. (2018). Experimental research and range of injury risk analysis of earthquake-induced hazard of falling exterior infill walls. Journal of Natural Disasters, 27(1): 17-26. (in Chinese)
Abstract: It is of profound significance to assess the national-scale seismic risk of urban buildings for seismic disaster prevention and mitigation. Based on the dynamic elasto plastic analysis of urban buildings, building seismic risk assessment of mainland cities was conducted in this work. First, the urban buildings database was established according to the macroscopic indicators provided by the Nation wide Census and China City Statistical Yearbook. Then, groundmotion intensities of the specific sites was obtained from the Global Earthquake Activity Rate Model (GEAR) method and the seismic groundmotion parameters zonation map of China. Then, site classifications of the cities were obtained using the corresponding relationship betweenthe ground slope and the shear wave velocity, through which the ground motion input could be further determined. Finally, based on the dynamic elastoplastic analysis of urban buildings, the seismic economic loss risk and extensive damage and collapse risk were assessed. Using the methods cited above, a map of building seismic risk of mainland citiesis given. The results showed that: (1) the method proposed in this work can assess the building seismic risk of mainland cities based on the publicly available data; (2) high seismic economic loss risk regions are mainly located in areas where the intensity of designbased earthquakes given by the 5th seismic groundmotion parameters zonation map of China is more than 0.3g; (3) after consideration of urban population and GDP, the building seismic risk of the eastern and central cities will increase significantly due to higher density of population and wealth; (4) differently selected groundmotion shave little impact on the seismic economic loss risk but has a greater impact on the extensive damage and collapse risk. The outcomes of this work can provide a useful method for future works on building seismic risk assessment of mainland cities.
Link to this article: https://doi.org/10.3969/j.issn.1000-0844.2019.02.299
How to cite this article: CHEA Cheav Por, XIE Linlin, LIN Yuanqing, LU Xinzheng. (2019) Study on seismic performance and collapse-resistant capacity of typical frame-core tube structures with single and dual lateral-force resisting system. Engineering Mechanics, 36(10): 40-49. (in Chinese)