Long-large subway tunnel accommodates the requirements of high-speed operation of trains, but also poses challenges on smoke control in case of tunnel fire. In conjunction with full-scale fire tests and theoretical analysis, this paper studies the fire behaviors of long-large subway tunnel by focusing on maximum ceiling smoke temperature rise, and smoke temperature rise distribution. The space downstream of fire source are divided into three zones, namely, inclined plume zone (Zone I), smoke entrainment zone (Zone II) and smoke column zone (Zone III). This paper proves that ceiling smoke temperature rise (CSTR) distribution at both Zone II and Zone III are exponential functions, but with different exponents. This paper proposes a method to compute the above exponents for Zone II and Zone III respectively. Full-scale tests verify that the proposed theoretical prediction method could predict the CSTR distribution with a relative error within 20%. Particularly, the proposed method is compared with several current computation methods, and is found has better correlation with the test data. Moreover, the measured maximum CSTR is found to be between the results of Kurioka’s model and Li Ingason’s model, and is closer with the later.

大型地铁隧道既可以满足火车高速运行的要求，又在隧道着火的情况下也对烟气控制提出了挑战。结合全面的防火测试和理论分析，本文通过关注最大天花板烟气温度上升和烟气温度上升分布，研究了大型地铁隧道的火灾行为。火源下游的空间分为三个区域，即倾斜羽流区域（I区），烟雾夹带区（II区）和烟柱区（III区）。本文证明了区域II和区域III的天花板烟气温度上升（CSTR）分布是指数函数，但是具有不同的指数。本文提出了一种分别计算区域II和区域III的上述指数的方法。全面测试验证了所提出的理论预测方法可以预测CSTR分布，相对误差在20％以内。特别地，将提出的方法与几种当前的计算方法进行比较，发现与测试数据具有更好的相关性。而且，测得的最大CSTR在Kurioka模型和Li Ingason模型的结果之间，并且与后者更接近。