Abstract: Five oil and gas pool size probability distribution models are introduced and compared by using the data from west Siberian basin.They are tested on the same statistic standard and aseries of conclusion comes out.1.The probability density function of oil and gas pool size hasarather strong relationship with basin's geological features,especially the richness of hydro-carbon in basin.2.Shpilman distribution is effective in the highly rich hydrocarbon district Log-normal distribution is suitable for the intermediately rich district and Log-Log normaldistribution for the low one.The Sum-Average distribution shows a better fit than Log normal and Log-Log normal in the high to intermediately rich district.3.Jin-Shpilman distribution is the best one among these.It can describes pool size distribution which is sampled from-different exploration stage and up to the last natural population,It can give not only an accurate prediction of oil and gas pools number,but also can give a detail of the total resourceswhich graded into different size classes.It is applicable to different basin and different exploration stage.

Key words: oil and gas reservoir, resources assessment, size, probability, statis-tical analysis, west Siberian basin

摘要： 介绍五种常见的描述油气藏规模概率分布模型,应用西西伯利亚的实际资料研究比较了这些模型;并按照统一的统计检验标准对这些模型进行了检验,得到一系列重要结论.1.油气藏规模概率密度函数与沉积盆地的特征有密切关系,尤其是同烃资源密度有关.2.施比伊曼模型在资源密度高地区应用效果好;对数正态分布模型适应于资源密度中等地区;对数一对数正态分布适合于资源密度低的地区.在油气资源密度高和中等的地区均和分布好于对数正态分布及对数一对数正态分布.3.金-施分布是一个较好的数学模型,它可以描述任何勘探样本的油气藏分布,直到自然总体.它不仅能较精确地预测油气藏个数,同时能较精确地给出各个分级别的资源量,有较广的适用性.

关键词: 油气藏, 资源评价, 规模, 概率, 统计分析, 西西伯利亚盆地