节点文献
应用模式识别方法预测油气储集层
APPLICATION OF PATTERN RECOGNITION TECHNIQUE TO LATERAL PREDICTION OF HYDROCARBON RESERVOIR
【摘要】 本文论述了模式识别在地震油气解释中的应用,并提出了一套实现的方法.在地震资料处理中提出了信息保持的思想和处理方法.从特殊处理后的地震记录中提取多种地震特征,这些特征主要来自地震道的自回归系数、最大熵频谱和自相关函数.利用聚类分析和判别分析方法对地震特征向量进行分类,实例表明,本文提出的地震模式识别方法能有效地划分油气储集层,即使在油层较薄或是反射变化不明显的地区也能奏效.
【Abstract】 In this paper an application of pattern recognition (PR) technique to seismic interpretationfor the purpose of dilimiting hydrocarbon reservoir is presented. Special attention was paid in data processing in order to preserve information necessary for PR. Features were extracted from autoregressive model, maximum entropy spectrum and auto-correlation function. Cluster testing and Fisher discrimination analysis were used for classification. Examples show that multivariate PR method is effective for reservoir interpretation even in cases of thin reservoir and subtle change of reflection characters.
【Key words】 Information preservation; Pattern recognition; Seismic interpretation.;
- 【文献出处】 地球物理学报 ,Chinese Journal of Geophysics , 编辑部邮箱 ,1992年05期
- 【被引频次】33
- 【下载频次】137