Due to massive flammable chemical materials and dense layout, domino accident chain always occurs in major accidents of atmospheric storage tank farm. In this paper, 136 domino effect accidents of atmospheric tanks are analyzed to achieve relative probability by ETA method. Based on accident probability, an analysis model is proposed to define the most likely accident chain and the key tank in the atmospheric tank farm. To get the real accident scenario, the method adopts the heat radiation model of pool fire under the fire extinguishing conditions. Combined with the dynamic Bayesian network, the escalation model of accident chain is established to figure out the dynamic probability of the tank under the condition of fire extinguishing. Finally, an example is used to verify the feasibility of the method. It is of great theoretical significance and application value for the treatment of domino accident in atmospheric tank farm.

常压储罐场重大事故中，由于大量易燃化工材料、密集布局，多米诺事故链不断发生。本文通过ETA方法对136起大气罐多米诺骨牌效应事故进行分析，得到相对概率。提出了基于事故概率的分析模型，定义了大气罐区最可能发生的事故链和关键罐。为了得到真实的事故场景，该方法采用热辐射灭火条件下的池火模型。结合动态贝叶斯网络，建立事故链升级模型，计算罐体在灭火条件下的动态概率。最后通过实例验证了该方法的可行性。对常压罐区多米诺骨牌事故的处理具有重要的理论意义和应用价值。