【摘要】 随着测井技术及大数据分析技术的快速发展,自动测井解释技术可以有效辅助人工快速开展储层划分、油水层解释等工作。为了提升储层划分及油水层识别准确度,本文提出了一种基于有监督学习的多粒度聚类识别方法,该方法通过对标准测井曲线及分层结果的学习提取不同分层测井曲线特征,在划分出储层的基础上再进行油水层识别。与已有方法相比,本文方法通过对真实测井曲线进行多种处理,从而融合曲线多层次特征,有利于取得更加准确的分层结果。实验结果表明,该方法可以对测井曲线进行自动分层,提高了曲线自动分层的效率,在真实测井曲线上能够取得较好的分层识别结果。更多还原

【Abstract】 With the rapid development of logging technology and big data analysis technology, automatic logging interpretation technology can effectively auxiliary artificial rapid reservoir classification and oil-water reservoir interpretation. In order to improve the accuracy of reservoir classification and oil-water layer identification, this paper gives a method based on multi-granularity of supervised learning. This method extracts the characteristics of different layers of logging curves by learning the standard logging curves and the layered results, and then identifies the oil and water layers based on the reservoir division. Compared with the existing methods, the proposed method can fully explore the characteristics of the real logging curve by processing the real logging curve through a variety of treatments, which is conducive to obtain more accurate stratification results. The experimental results show that the proposed method can automatically separate the logging curves, improve the efficiency of the automatic slicing of the curves, and can obtain good results of the layer identification on the real logging curves.更多还原