论文部分内容阅读
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。
Pattern recognition Reservoir prediction is a technique for horizontal prediction of reservoirs using seismic data. It uses computer to extract various waveform characteristic parameters and parameters related to medium absorption firstly from seismic data. Then, these parameters are comprehensively processed. Finally, the lithology or hydrocarbon-bearing property of reservoirs is classified according to the result of comprehensive processing. There are two methods of comprehensive treatment: one is discriminant analysis with learning samples; the other is comprehensive parameter method without learning samples. Since 1991, this method has achieved good results in the practical application of the Pearl River Mouth Basin. It has correctly predicted the hydrocarbon-bearing properties of the main strata of the wells H3-NE and H4-1, and also has identified the reefs Its unique. In oil and gas exploration, it provides an effective means of predicting reservoirs directly using seismic data.