基于KL-E的地质力学模型参数反演及应用  

Inversion of Geomechanical Model Parameters Based on KL Expansion and Its Application

作　　者：覃建华[1] 杨琨[2] 丁艺[1] 张博宁 唐慧莹 QIN Jianhua;YANG Kun;DING Yi;ZHANG Boning;TANG Huiying(Exploration and Development Institute,Xinjiang Oilfield Company,PetroChina,Karamay,Xinjiang 834000,China;Development Company,Xinjiang Oilfield Company,PetroChina,Karamay,Xinjiang 834000,China;Chengdu North Petroleum Exploration and Development Technology Company Limited,Chengdu,Sichuan 610051,China;State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation,Southwest Petroleum University,Chengdu,Sichuan 610500,China)

机构地区：[1]中国石油新疆油田分公司勘探开发研究院,新疆克拉玛依834000 [2]中国石油新疆油田分公司开发公司,新疆克拉玛依834000 [3]成都北方石油勘探开发技术有限公司,四川成都610051 [4]油气藏地质及开发工程国家重点实验室·西南石油大学,四川成都610500

出　　处：《西南石油大学学报（自然科学版）》2022年第2期65-78,共14页Journal of Southwest Petroleum University(Science & Technology Edition)

基　　金：国家自然科学基金(51874251,51904257);四川省科技厅苗子工程重点项目(2019JDRC0089);油气藏地质及开发工程国家重点实验室项目(PLM201819)。

摘　　要：地应力及地质力学参数分布是石油工业井壁稳定性分析、水力压裂以及防砂措施制定等所需的关键参数。提出一种反演地质力学模型参数的方法,该方法通过优化算法依次调整地质力学模型的边界条件与力学参数场,以匹配模型计算地应力与硬数据(一维室内实验测量地应力或一维测井综合解释地应力)为优化目标。利用条件Karhunen-Loève展开(KL-E)生成满足特定位置地质力学属性硬数据的地质力学参数随机场,可以将基于网格单元的地质力学参数量减少为一组一维随机变量,加快算法收敛速度。通过测试算例和四川盆地某页岩气井区的现场实例,验证了该方法的适用性。该方法不仅提高了模型计算结果与硬数据的匹配程度,同时更准确地描述了未知位置的地质力学参数分布情况。与地质力学参数场的反演相比,边界条件反演过程收敛速度更快。The distribution of in-situ stresses,as well as geomechanical properties,is critical for wellbore stability analysis,hydraulic fracturing,and sand control in the petroleum industry.In this paper,we propose a data assimilation procedure for matching the in-situ stresses with the hard data(1 D measured or interpreted in-situ stresses)by sequentially adjusting the fields of geomechanical properties and the boundary conditions with optimization algorithms.The distribution of geomechanical properties are generated with the conditional Karhunen-Loeve expansion(KL-E),which can reduce the number of unknowns from element-based geomechanical properties to a small set of random variables that significantly save the computing time and accelerate the convergence of the optimization algorithms.The applicability of the proposed procedure has been systematically tested by several synthetic cases and a field case study of a shale gas reservoir in Sichuan Basin.Not only the improvement of stress matching but also a more accurate description of the distribution of geomechanical properties at unmeasured locations is observed.In addition,compared to the inversion of the geomechanical properties,the inversion of boundary conditions is much faster.

关 键 词：地质力学建模 Karhunen-Loève展开 地质力学性质分布 数据同化 页岩气 

分 类 号：TE142[石油与天然气工程—油气勘探]