基于地震属性的储层参数估算方法研究

发布时间：2018-04-29 20:16

本文选题：角度域共成像点道集 + 地震属性优选　；参考：《中国石油大学(华东)》2015年硕士论文

【摘要】：叠前地震数据中含有十分丰富的物性、岩性、流体等信息,需要综合利用多种方法把反映这些储层信息的参数提取出来,地震属性是获得这些参数的一个重要途径,通过对提取的地震属性进行分析可以获得较为准确的储层信息。地下地层的厚度、孔隙度、渗透率、饱和度等特征参数只能通过钻井信息来确定,而从井中所得到的各种特征参数仅是井点上的信息,只能确定一些离散位置点上的储层信息。因此,要了解整个三维空间上储层特征参数分布的规律,必须依靠地震数据在空间上进行展开,这就要建立地震数据与由测井数据获得的各种储层特征参数之间的关系。首先,通过相关性分析发现,相比基于叠后地震数据的常规属性,在叠前角度域共成像点道集(ADCIGs)上提取的地震属性包含更加丰富的原始地震信息,叠加会损失叠前地震资料里微弱的有效信息,因此在ADCIGs上提取常规地震属性,同时提取叠前属性P、G,与测井数据一起组成样本集,采用标准差标准化方法对样本集数据进行标准化处理。进而,提出了将遗传算法、支持向量回归机算法结合的属性优选方法(GA-SVR),同时将禁忌搜索算法(TS)引入到属性优选中,将GA-SVR与TS进行对比、分析,选择了GA-SVR方法对样本集进行地震属性优选。最后,利用支持向量回归机(SVR)建立属性优选后的样本集与Gamma、Porosity测井数据等储层特征参数之间的非线性估算模型,进而对三维工区进行储层特征参数估算。针对Opendtect中的F3演示数据和胜利油田某工区实际数据进行了方法试算,估算结果在井位置处与实际测井数据有较高的一致性,在远离井的位置,与地震数据、地质背景相吻合,验证了本研究内容的有效性,为油田的石油勘探提供了参考。
[Abstract]:The prestack seismic data contain abundant physical property, lithology, fluid and so on, so it is necessary to extract the parameters reflecting these reservoir information by comprehensive use of many methods. Seismic attribute is an important way to obtain these parameters. More accurate reservoir information can be obtained by analyzing the extracted seismic attributes. The characteristic parameters of underground formation such as thickness porosity permeability and saturation can only be determined by drilling information. Therefore, in order to understand the distribution law of reservoir characteristic parameters in the whole three-dimensional space, we must rely on the spatial expansion of seismic data, which is necessary to establish the relationship between seismic data and various reservoir characteristic parameters obtained from logging data. First of all, through correlation analysis, it is found that compared with the conventional attributes based on post-stack seismic data, the seismic attributes extracted from the common imaging point gathers in prestack angle domain (ADCIGs) contain more abundant original seismic information. Stacking will lose the weak and effective information in the prestack seismic data, so the conventional seismic attributes are extracted from the ADCIGs, and the prestack attributes are extracted to form the sample set together with the logging data. The standard deviation standardization method is used to standardize the sample set data. Furthermore, a new attribute selection method, which combines genetic algorithm and support vector regression algorithm, is proposed. At the same time, Tabu search algorithm is introduced into attribute selection, and GA-SVR and TS are compared and analyzed. The GA-SVR method is selected to select the seismic attributes of the sample set. Finally, the nonlinear estimation model between the sample set after attribute selection and the reservoir characteristic parameters such as Gamma-Porosity logging data is established by using support vector regression machine (SVR), and then the reservoir characteristic parameters are estimated in 3D working area. Based on the F3 demonstration data in Opendtect and the actual data in a certain working area of Shengli Oilfield, the estimated results are in good agreement with the actual logging data at the well location, the location far away from the well, the seismic data and the geological background. The validity of this study is verified and a reference for oil field exploration is provided.
【学位授予单位】：中国石油大学(华东)
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：P631.4;P618.13

【参考文献】