Paleo-Channel Recognition Based on Multi-Attribute PCA Fusion

Abstract

Summary The distribution of channel sandstone has great relationship with the well deployment in clastic rock reservoirs. The transversal migration and cutoff of the paleo-channels led to the complex geological deposition of sand-shale thin interbeds with frequent lateral variation. Because of the insufficient samples of drilling and logging data and the low resolution of seismic data, it is difficult to recognize the paleo-channels with high accuracy. PCA fusion based multi-attribute analysis can solve the problem of multiply solutions when using individual seismic attributes. The variance of the principal component of each attribute is obtained as the weight coefficient of fusion. The principal components of multiple attributes are stacked under the fusion rule constrained by their weight coefficients to get the fused seismic attribute. Drilling data, logging curves are analyzed with the fused seismic attribute to predict the distribution of channel sandstone. Practical application indicates that the predicted distribution of channel sandstone fits well with the drilling and logging data, the periods of paleo-channel series and overlay deposition can also be identified. The characterization of paleo-channels by PCA fusion provides a significant reference to well deployment in sandstone reservoirs.