Air-supported membrane buildings may suffer from air-leakage caused by the opening of exit, failure of inflation system or environmental force majeure, which is essential for pedestrians' evacuation but receives less attention. The paper proposes a coupling evacuation model based on multi-velocity cellular automaton by incorporating dynamic air-leakage process in evacuation analysis to quantitatively assess the impact of air-leakage on evacuation. Two factors are considered to determine evacuation process in deflating building— herding behavior caused by panic among crowd and the variation of moving velocity. An experiment was conducted to obtain individual velocity under different elevations of membrane surface to provide practical velocity for the model. Coupling simulation results show that evacuation efficiency can be raised in the deflating building with small population density, which is reduced as population density grows. The maximum capacity of the double-exit building is improved by 24% of the single-exit building. This work offers a possible method to predict crowd movement process and assess the level of security of air-supported membrane buildings suffered air-leakage under emergency from the aspect of evacuation dynamics.

空气支撑的膜建筑物可能会因出口的打开，充气系统的故障或环境不可抗力而导致漏气，这对于行人疏散是必不可少的，但很少受到关注。本文提出了一种基于多速度元胞自动机的耦合疏散模型，将动态漏气过程纳入疏散分析中，定量评估漏气对疏散的影响。考虑两个因素来确定建筑物排空过程中的疏散过程-人群中的恐慌和移动速度的变化引起的羊群行为。进行了实验以获得膜表面不同高度处的个体速度，从而为模型提供实用速度。耦合模拟结果表明，人口密度较小的放气建筑物可以提高疏散效率，人口密度越高，疏散效率越高。双出口建筑物的最大容量比单出口建筑物提高了24％。这项工作提供了一种可能的方法，可以从疏散动态的角度预测人群移动过程并评估在紧急情况下遭受空气泄漏的空气支撑膜建筑物的安全性。