基于EMD-奇异值分解的储层识别方法

Reservoir Identification Method Based on EMD-Singular Value Decomposition

阵列式声波测井是地球物理学测井技术中的一种重要的方法,其中用于声波仪记录的地质波形能够准确有效地记录地质信息,从而求出各种模式波及其他相关属性.为减少噪音背景噪声对信号储层分布的影响,得到信号储层识别的分布,笔者采用经验模态分解法(EMD)和奇异值分解法(SVD)开展基于阵列的声波信号储层识别技术研究.结果表明:将较为复杂的阵列声波检测井信号进行EMD分解后,各高频固有模态函数(IMF)的奇异值频谱区域与原始波列的奇异值频谱区域具有较好的对应关系;通过绘制散点图可清晰地显示储层奇异值的分解差距区域.经实际测井数据验证,笔者采用的方法所得结果与实际结果相一致,说明基于EMD的奇异值分解方法在阵列声波测井数据中的储层识别是有效的.

Array acoustic logging is an important method in geophysical logging technology,in which the geological waveform recorded by acoustic logging instrument can accurately and effectively record geological information,so as to calculate various modes and other related attributes.In order to reduce the influence of noise background noise on the distribution of signal reservoir,empirical mode decomposition(EMD)and singular value decomposition(SVD)were used to study the array based acoustic signal reservoir identification technology.The results show that after EMD decomposition of the complex array acoustic well signals,the singular value spectrum region of each high-frequency intrinsic mode function(IMF)has a good corresponding relationship with the singular value spectrum region of the original wave train,and the decomposition gap region of the reservoir singular value can be clearly displayed by drawing the scatter diagram.The actual logging data shows that the results obtained by the method are consistent with the actual results,which shows that the singular value decomposition method based on EMD is effective in reservoir identification in array acoustic logging data.