Abstract:

To design good policies for managing accident risk in petrochemical industry requires a sound understanding on the accident risks in the petrochemical industry. The authors estimated the parameters of the distributions of the frequencies and consequences of accident risks in the petrochemical industry by using Hierarchical Bayesian Analysis, based on the information of 1479 accidents in petrochemical industry from 2006 to 2015 collected from multiple sources. It was concluded that the frequency of the accidents in China’s petrochemical industry follows Poisson distribution with its mean being 156.00, the fatalities of the accidents follows generalized Pareto distribution with its scale parameter being 1.40 and its shape parameter being 0.36. The parameters of the distributions of the frequencies and consequences of accident risks vary sharply across regions.

Key words: petrochemical accident, accident risk, risk assessment

摘要：

为了针对石油化工行业事故风险制定管理政策, 需要了解事故的发生状况和地区差异。通过多个来源, 收集我国石油化工行业2006—2015年发生的1479起事故数据, 用分层贝叶斯模型估算事故的发生频率和事故后果的统计分布函数的参数。结果表明, 我国石油化工行业事故频率服从均值参数为156.00的泊松分布, 事故的死亡人数服从尺度参数为1.40、形状参数为0.36的广义帕累托分布, 我国不同地区间石油化工行业的事故频率和事故后果分布有较大差异。

关键词: 石油化工事故, 事故风险, 风险分析