圈闭评价中含油气性风险依赖性的概念、方法及应用

Concept, method and application of geological risk dependency indicating petroleum discovery

盛秀杰 金之钧 肖晔 王义刚 蒋瀚

Sheng Xiujie Jin Zhijun Xiao Ye Wang Yigang Jiang Han

“至少有一个次级圈闭含有油气”的组合概率分析技术可评价圈闭的含油气性，评价结果反映了所有可能同时含有油气的圈闭组合形式。但受限于客观地质条件约束，有些圈闭是不应该同时出现的——不同圈闭组合形式对资源量计算有直接影响。为了计算符合地质模型约束的圈闭组合概率，界定了次级圈闭的含油气性定量评价模型的地质涵义，指出次级圈闭含油气性评价本质上是遵循贝叶斯分析原则，包括边际概率和条件概率两部分评价内容，分别体现全局成藏和局部成藏的可能性。首先，强调不同圈闭成藏时会存在明显的地质相关性，单个层圈闭的评价是以全局成藏可能性为前提进行评价；其次，通过对应边际概率的可能取值范围，区分了“完全独立”、“部分决定”和“完全决定”3种不同含油气性风险依赖类型，而不同依赖类型直接决定了到底哪些圈闭组合才符合当前地质认识；最后，为油气资源一体化评价软件平台（PetroV）设计了一种改进的概率树分析技术，与不确定性体积法有机结合，实现了“基于含油气性风险依赖的概率组合加和”资源量计算方法。实例证明，要想获得较为客观的圈闭定量评价结果，需要充分考虑其所属不同次级圈闭间的含油气性风险依赖类型，并依此为基础才能给出对应合理地质模型解释的不确定性油气资源量分布结果。

The combination probability, the possibility that at least one segment contains hydrocarbons, can be used to assess the discovery probability of a trap.Obviously,‘at least one segment containing oil/gas’ reflects all possibilities of trap combinations.However, some traps should not be treated as one combination due to the constraint of special geological conditions.The trap combinations have direct influences on the estimation of geological resource.In order to calculate trap combination probability under the constraint of geologic models, this paper first cleared geological meaning of current evaluation model for segment and demystified its mysterious hype.It is in fact a classical Bayes mathematical model consisting of marginal and conditional probability, which quantifies separately the overall accumulation conditions or play-level and local accumulation conditions.Geological correlations do exist among different traps when hydrocarbons accumulate in them, thus the evaluation of single trap should be based on the prerequisite of possibility of overall hydrocarbon accumulation.Secondly, corresponding to the value range of marginal probability, this paper recognizes three geological risk dependence types including full independence, partial dependence and full dependence.These dependence types can directly determine which traps are concordant with current geologic knowledge.Finally, an improved probability tree technique was designed for the petroleum integrated assessment software (PetroV).When integrated with the non-deterministic volumetric method, it can realize resource volume estimation through‘summation of probability combinations based on geological risk dependency’.Case study shows that geological risk dependence types play a key role in quantitative evaluation of traps and mapping of uncertain petroleum resource distribution corresponding to reasonable geologic model interpretations.

10.11743/ogg20150120