地质信息约束下的河流相储层地震模式识别应用研究

发布时间：2018-12-08 13:23

【摘要】：渤海海上河流相油田油藏地质条件复杂,储层厚度薄、横向变化大、连通性较差,开发此类油田难度大,面临多方面的挑战。传统的储层地质研究方法已经不能满足当前的开发需求,尤其对小层内砂体叠置构型研究较少,传统上主要通过研究地震反射特征来判断砂体叠置状况,但是受地震分辨率影响,砂体叠置和地震响应之间不存在一一对应的关系,准确性不高,所以必须考虑使用其它方法对储层的砂体叠置特征进行研究。本文主要基于地震属性提取和优选,利用神经网络方法对砂体的叠置特征进行了一些研究,首先将渤海海上河流相油田中存在的砂体叠置特征进行总结,建立理论正演模型,提取模型地震属性,使用地震属性优选方法选取敏感属性,利用神经网络方法对所有存在的砂体叠置类型进行分类,旨在研究实际资料中所有存在的砂体叠置类型,是否能够通过基于地震属性的神经网络模式识别技术分辨出来,且能够分辨出几种类型。最终,研究得出,将工区中所有的砂体叠置样式分成了6种模式。随后,将在模型研究中取得的成果运用于渤海工区的实际资料之中,对工区区域内砂体叠置模式进行模式识别预测,取得了较好的效果,证明了本文提出并运用的对砂体叠置模式研究方案的可行性,对实际资料储层预测具有一定指导意义。
[Abstract]:The reservoir geological conditions of fluvial facies oil field in Bohai Sea are complicated, the reservoir thickness is thin, the lateral change is large, the connectivity is poor, the development of this kind of oil field is difficult, and it faces many challenges. The traditional reservoir geological research methods can not meet the current development needs, especially the research on the overlay configuration of sand bodies in the small layers. Traditionally, the study of seismic reflection characteristics is mainly used to judge the overlay situation of sand bodies. However, due to the influence of seismic resolution, there is no one-to-one correspondence between sand body overlay and seismic response, and the accuracy is not high. Therefore, other methods must be considered to study the overlay characteristics of sand body. In this paper, based on seismic attribute extraction and optimal selection, the superposition characteristics of sand bodies are studied by using neural network method. Firstly, the overlay characteristics of sand bodies in river facies oil fields in Bohai Sea are summarized, and the theoretical forward modeling is established. The model seismic attributes are extracted, the sensitive attributes are selected by the seismic attribute optimization method, and all the existing sand overlay types are classified by the neural network method. The purpose of this paper is to study all the sand overlay types in the actual data. Whether the neural network pattern recognition technology based on seismic attributes can be used to distinguish several types. Finally, it is concluded that all the sand overlay patterns in the working area are divided into six models. Then, the results obtained in the model research are applied to the actual data of the Bohai work area, and the pattern recognition and prediction of the sand body superposition pattern in the work area are carried out, and good results are obtained. It is proved that the feasibility of the research scheme of the sand body superposition model proposed and applied in this paper is of certain guiding significance to the reservoir prediction of the actual data.
【学位授予单位】：西南石油大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：P618.13

【参考文献】