基于EMD和近似熵的储层预测  

Reservoir Detection Based on EMD and Approximate Entropy

作　　者：陈胜红[1,2] 贺振华[1] 朱明 汪瑞良 陈雪芳 文晓涛[1] 姜建 曾驿 

机构地区：[1]成都理工大学,四川成都610059 [2]中海石油有限公司深圳分公司研究院,广东广州510240

出　　处：《天然气地球科学》2010年第5期828-832,共5页Natural Gas Geoscience

基　　金：国家重点基础研究发展计划项目(编号:2007CB41170501;2009CB219501)联合资助

摘　　要：通常情况下,不同地层特征对应的地震波复杂程度是有差异的。近似熵是一种反映信号复杂程度的特征量,对储层特征有一定反映。但由于地震波是有一定带宽的信号,波形受带宽内所有频率成分的影响,如果直接进行近似熵的计算,其结果往往不能很好地反映储层特征。鉴于这种情况,提出了经验模态分解法(EMD)与近似熵相结合的方法,利用EMD分解后的每一个地震波的本征模函数(IMF)分量进行近似熵计算,并与已知井对比分析后,选取对储层特征识别有效的IMF分量进行储层预测。利用该方法对HZ地区珠江组进行了试算,效果较好。In general,complexities of seismic waveforms corresponding to different strata characters are different.Approximate entropy is an index of signal complexities,which can reflect the characters of reservoir to a certain extent.However,seismic wave is a signal with definite bandwidth and its waveform is affected by all frequency components in the band.So the result would not reflect the characters of reservoir if calculating the approximate entropy directly.Here we use an advanced method integrating Empirical Mode Decomposition（EMD） and approximate entropy,which includes calculating the approximate entropy of every Intrinsic Mode Functions（IMF） of seismic wave decomposed by EMD,and choosing the effective IMF by comparing with the known well to predict the characters of reservoir.The method is effective in Zhujiang Formation of HZ area.

关 键 词：经验模态分解 近似熵 本征模函数 储层 

分 类 号：TE132.1[石油与天然气工程—油气勘探]