The seismic resilience of cities plays a crucial role in achieving the United Nations Sustainability Development Goal. However, despite the occurrence of elevator passenger entrapment in numerous earthquakes, there is a notable lack of studies addressing this sophisticated issue. This study aims to bridge this gap by proposing a novel urban risk assessment model designed to evaluate city-scale earthquake-induced elevator passenger entrapment. The model integrates big data and physics-based approaches. A novel mapping method was developed to estimate city-scale elevator traffic level based on population heatmap data and deep learning. A process-based parallel computing scheme was designed to accelerate the assessment. The applicability was demonstrated based on a real-world urban area comprising 619 buildings. The findings reveal that as the time of the earthquake varies, the risk exhibits significant fluctuations. Additionally, this study highlights that a simplistic correspondence between seismic intensity and passenger entrapment risk can lead to erroneous estimations.

中文翻译：

---

利用人口热图量化地震引起的电梯乘客被困的城市风险评估模型


城市的抗震能力对于实现联合国可持续发展目标发挥着至关重要的作用。然而，尽管在多次地震中都发生过电梯乘客被困的情况，但针对这一复杂问题的研究却明显缺乏。本研究旨在通过提出一种新颖的城市风险评估模型来弥补这一差距，该模型旨在评估城市范围内地震引起的电梯乘客被困情况。该模型集成了大数据和基于物理的方法。基于人口热图数据和深度学习，开发了一种新颖的地图方法来估计城市规模的电梯交通水平。设计了基于流程的并行计算方案来加速评估。该适用性基于包含 619 栋建筑物的现实城市区域进行了论证。研究结果表明，随着地震发生时间的变化，风险呈现出显着的波动。此外，这项研究强调，地震烈度和乘客被困风险之间的简单对应关系可能会导致错误的估计。