改进的希尔伯特-黄变换在储层预测中的应用  

The application of improved Hilbert-Huang transform in reservoir prediction

作　　者：梁岳[1,2] 顾汉明[1,2] 姚知铭[3] 

机构地区：[1]中国地质大学(武汉)地球内部多尺度成像湖北省重点实验室,湖北武汉430074 [2]中国地质大学(武汉)地球物理与空间信息学院,湖北武汉430074 [3]中国石油化工集团公司华北分公司勘探开发研究院,河南郑州450006

出　　处：《石油物探》2016年第4期606-615,共10页Geophysical Prospecting For Petroleum

基　　金：国家科技重大专项(2011ZY05002-001)资助~~

摘　　要：希尔伯特-黄(Hilbert-Huang transform,HHT)变换是一种非线性非平稳信号处理技术,在复杂地震信号处理方面比传统的时频分析方法更为有效,但该方法存在模态混叠和端点效应等问题,导致信号处理的精度下降。为此,提出了基于自回归(AR)模型预测的完备总体经验模态分解(Complete Ensemble Empirical Mode Decomposition,CEEMD)方法对希尔伯特-黄变换加以改进:在经验模态分解(Empirical Mode Decomposition,EMD)过程中加入成对的辅助白噪声,降低了由信号中随机噪声引起模态混叠问题;并利用AR模型在信号端点预测出极值点并对其进行包络线拟合,较好地抑制了端点效应。应用改进后的方法提取实际地震记录的瞬时振幅和瞬时频率并进行储层预测,预测结果与测井资料所反映的储层信息吻合度很高,证明该方法能够更为准确有效地反映储层特征。As a new time-frequency analysis method for non-linear and non-stationary signal,Hilbert-Huang transform has the advantage in seismic data interpretation,compared with conventional time-frequency analysis method.But there are still some problems such as mode mixing and endpoint effect.These problems will reduce signal processing accuracy.We improved the Hilbert-Huang transform（HHT）based on autoregressive（AR）model and proposed complete ensemble empirical mode decomposition（CEEMD）with adding pairs of auxiliary white noises,which can reduce the mode mixing problem caused by random noise.The AR model is used to predict the extreme points in the endpoint and fit the envelope to suppress the endpoint effect.Seismic instantaneous attributes from actual seismic data were extracted by the improved HHT to conduct reservoir prediction.The predicted results was coincident with the reservoir information from logging data.It proves that the improved HHT can reflect the reservoir features more accurately and effectively.

关 键 词：时频分析 希尔伯特-黄变换 模态混叠 端点效应 AR模型 完备总体经验模态分解 储层预测 

分 类 号：P631[天文地球—地质矿产勘探]