成藏体系油气资源评价中的统计方法体系

摘要

基于成藏体系理论对传统区带油气资源评价技术进行了系统归纳与新方法的补充，深入阐述了成藏体系油气资源评价的内涵。其中，在深层次剖析上述制约因素的形成机制以及相关统计数学模型机理的基础上，通过界定不同勘探程度的成藏体系油气资源评价的关键目标（如，资源有无、多少、在哪里等），整理、归类并规范了业内主流统计方法以及在开发油气资源一体化评价软件过程中涉及的系列特色资评技术，约定了成藏体系油气资源评价统计方法体系，其内容主要包括：①面向新区的概率分析系列；②面向未成熟区（早期）的蒙氏模拟系列；③面向未成熟区（中、晚期）的数理统计系列；④面向成熟区（早、中期）的概率推理系列；⑤面向成熟区（中、晚期）的发现过程和空间分布预测系列。提出的统计方法体系，较为全面地阐述了系列主流以及最新统计方法的成因机制与适用性说明，在实际应用时，可以有效规避地质模型的差异性、增强统计方法应用的针对性以及降低计算参数的敏感性。此外，该方法体系的归类逻辑也有助于明确后续进一步值得关注的资源评价方法研究方向。%Based on the principles of petroleum accumulation system , this paper systematically summarizes and perfects the traditional techniques of petroleum resource assessment ,and expounds the connotation of resource evaluation under a petroleum accumulation system .On the basis of a deep understanding of the controlling factors of petroleum accumulation and the principles of statistical models ,we defined the key objectives of resource assessment at various exploration stages , categorized and standardized the most popular mainstream approaches used in the industry and academia ,as well as me-thods evolved during the development of PetroV ( a petroleum resource evaluation system of Sinopec ) , and proposed a methodology framework for resource assessment of a petroleum accumulation system .The categorized five methodology se-ries include 1) probabilistic inference or analogue approaches for frontier targets ,2) Monte Carlo approaches for conceptual plays,3)statistical methods for established plays in early exploration stage ,4)probabilistic reasoning models for targets in mature plays and 5) discovery process and spatial statistical models in well explored regions .We elaborate the objectives of assessment and the suitability of statistical methods under a variety of exploration scenarios through real examples .In practice ,the proposed assessment framework can effectively reduce the sensitivity on parameter selections and the impact due to differences in the geological models on output ,thus enhancing applicability of the statistical methods .In addition , the proposed framework may also provide useful information for direction on future methodology development .