频谱分解技术在古河道识别中的应用

The Application of Spectrum Decomposition Technology to Ancient Channel Prediction

河流相储集层一般具有单层厚度薄、横向变化快、储层预测难的特点。传统地震预测方法是根据全频段信息进行研究,模糊了不同频率信号所携带的不同地质信息,降低了其预测精度。基于短时窗傅里叶变换的频谱分解技术,可将地震数据由时间域变换到频率域,生成的振幅频率体和相位体可用来检测薄层及地质异常体横向不连续性,提高了储层预测的精度。基于此,首先根据已钻井的测井、录井等资料,建立了河道砂体正演模型,得到正演记录,并对正演记录进行时频分解,分析河道处响应特征,用于河道砂体的追踪识别。然后,对目的层段进行时频转换,提取振幅体频率切片,刻画出古河道的形态特征,并提取相位体切片,识别出断裂分布特征,指导了有利储层的预测。该技术在古河道砂体预测中的成功运用,值得进一步推广。

The reservoir of fluvial facies generally has the characteristics of thin single layer, rapid lateral variation, difficulties in predicting reservoir. The traditional method of reservoir prediction is studied according to the full band information, so that different geological information carried by different frequency is complicated, which will limit prediction accuracy. The principle of spectrum decomposition is Fourier transform in short window, seismic data are converted from time domain to frequency domain, the amplitude data are used to detect thin layer, and the phase data are used to detect lateral discontinuity of geologic body. This technology can improve the precision of reservoir prediction. In this paper, according to well logging data in the area, forward model of channel sand is established. Then the results of forward record are analyzed by spectrum decomposition technology, which can be used to analyzing the responsed characteristic of river channel for channel sand body recognition in working area. Furthermore, time frequency is transformed into target layer. Amplitude frequency slices are extracted to depict morphological characteristics of the ancient river. The fracture was also successfully identified by phase frequency slice. By use of low frequency slice, the plane distribution of main faults can be observed. Using the high frequency slice, the spatial distribution of small level faults can be observed. So this technology is favourable for reservoir prediction, which should be used in the areas with similar sedimentary characteristics.