Abstract: The leakage and dispersion risk of certain chemical facility was always difficult to evaluate quantitatively because of the existing of many uncertain parameters in evaluation models. For this reason, taking weather data of Nanjing and some other cities as data sources, STATISTICA 6.0 was used to study the distribution of wind speed, humidity, air temperature and cloud cover rate, etc. And their respective distribution functions were proposed. The distributions of leakage orifice area and leakage orifice direction were also analyzed and reasonable distribution functions were gotten. As a widely used software package for leakage and dispersion, HGSYSTEM was chosen to unite Monte-Carlo method to form a quantity risk analysis method for chemical facilities. An accessory pipe (20 cm) of a propane tank (normal temperature, high pressure) was taken as a demo, and the influencing distances of two cloud blocks, with the concentration of 2.1%(lower explosion limit) and 0.015 3%(Russia MAC), were calculated for different cumulative probability, among which, when the cumulative probability is 95%, the two distances are respectively 69.7 m and 2 764 m.

Key words: Risk ,  Leakage , Dispersion , Uncertain ,  Monte-Carlo ,  HGSYSTEM

摘要： 泄漏扩散模型中, 大量不确定参数的存在严重影响了化工装置泄漏扩散定量风险分析, 为此, 以南京
等城市的气象数据为依据, 应用STATISTICA 6.0 对风速、湿度、大气温度、云层覆盖率等气象数据的分布形式进行
了统计分析, 得到了各自分布函数;对泄漏裂口面积、裂口角度等泄漏源参数分布进行了合理的假设。进一步以
HGSYST EM 工具包作为泄漏扩散后果分析的工具, 提出了一种基于蒙特卡罗方法, 用于分析特定化工装置有毒有
害物质扩散后果风险的方法。对与常温高压液态丙烷储罐相连, 直径20 cm 的管道进行试算, 得到了不同累积概率
对应的爆炸下限(2 .1 %)气云范围及前苏联M AC(0.0153%)气云范围, 发现当置信度达到95%时, 两种浓度气云的
影响范围分别达到69 .7 m 及2 764 m 。

关键词: 风险 , 泄漏 , 扩散 , 不确定性 , 蒙特卡罗 ,  HGSYST EM