【摘要】 模式识别储层预测技术是一种用地震资料对储层进行横向预测的技术。它利用计算机首先从地震资料中提取多种波形特征参数和有关介质吸收的参数,然后对这些参数进行综合处理,最后根据综合处理的结果对储层的岩性或含油气性进行分类。综合处理的方法有两种:一是有学习样本的判别分析法;二是无学习样本的综合参数法。1991年以来,该方法在珠江口盆地的实际应用中取得了较好的效果:为H3-NE井和H4-1井主要目的层的含油气性作出了正确的预测,对生物礁的鉴别也有其独到之处。在油气勘探中,它为用地震资料直接预测储层提供了一种有效的手段。更多还原

【Abstract】 The Pattern Recognition Reservoir Prediction Method is a method to predict the lithology or hydrocarbons in reservoirs through integral processing after extracting multiple characteristic parameters of waves and medium absorption parameters from seismic data with computers. There are two ways of integral processing: one is the discrimination analysis method in which learning samples are available; the other is the integral parameter method in which learning samples. Fairly good results have been obtained in application of this method in the Pearl River Mouth Basin since 1991, hydrocarbons in the major target zones in the wells H3-NE and H4-1 are correctly predicted. It also has a unique application in recognizing carbonate reefs. Therefore, it provides a useful means for directly predicting reservoirs with seismic data.更多还原