提高低信噪比核磁共振测井弛豫信息真实性回波处理方法  

Echoes Processing Method to Improve Relaxation Information Authenticity for Low SNR NMR Log Data

作　　者：王鹏[1] 谭茂金[1,2] 

机构地区：[1]中国地质大学地球物理与信息技术学院,北京100083 [2]地下信息探测技术与仪器教育部重点实验室,北京100083

出　　处：《测井技术》2014年第4期400-406,共7页Well Logging Technology

基　　金：油气资源与探测国家重点实验室开放课题(2012009);中央高校基本科研业务费专项页岩气储层响应机理研究(2012ZY28);国家重大专项大型油气田及煤层气开发(2012ZX05014-001-001)

摘　　要：核磁共振(NMR)测井采集的回波信号(CPMG Echo Trains)不可避免地要受到噪声的影响,表现为在大中孔隙可动流体成分造成基线偏移、在微小孔隙束缚流体成分产生畸变,给反演解释工作带来极大的不确定性。针对低信噪比的核磁共振测井数据处理提出了基于奇异值分解法(SVD)滤波的核磁共振数据变换反演方法 BRD(Butler Reeds Dawson)。该方法利用Hankel矩阵与回波信号的相互构建关系,把回波信号的Hankel矩阵作为SVD滤波处理的信号矩阵,实现了对低信噪比回波信号的分解和滤波。结合BRD反演算法,利用SVD处理后的高信噪比全波列数据反演出流体的T2分布。在随后的模拟数据试验和实际数据处理中分解和滤波在确保可动流体成分和提高束缚流体成分在反演信息准确恢复上是有效的。Nuclear magnetic resonance （NMR） logging has become an important geophysical exploration method.Practically,the echo signal （CPMG Echo Trains） collected is inevitably affected by noise,with the characteristics of baseline offset in movable fluid component of large/medium pores and distortion in bound fluid component of micro/small pore,it brings a huge uncertainty to inversion and explanation work.For low SNR NMR logging data inversion processing,a new approach based on SVD filtering-BRD （Butler Reeds Dawson） inversion is proposed in this paper.The method uses transforming relationship of Hankel matrix and echo signal with the signal＇s Hankel matrix as the object,thereby,realizing decomposition and filtering of low SNR signal.Combining with BRD inversion algorithm,T2 spectrum of fluids is constructed using high SNR full echo trains treated by SVD filtering.In the following simulation experiment and practical data processing,the movable fluid component is definitely ensured and bound fluid component is effectively retrieved in inversed spectrum,that validates the method is effective.

关 键 词：核磁共振测井 回波信号 信噪比 奇异值分解法 变换反演方法 横向弛豫时间 

分 类 号：P631.83[天文地球—地质矿产勘探]