基于主成分分析的多属性聚类方法在裂缝预测中的应用

Multi-attribute clustering method based on principal component analysis and its aapplication in crack prediction

准确预测储层裂缝分布是勘探开发中的重点。地震几何属性对裂缝预测的效果最好,但不同的裂缝属性往往是地下不同裂缝特征的响应。综合考虑多种属性预测结果,提出基于主成分分析的多属性聚类方法,提取多种裂缝属性相互独立的主成分分量,再利用这些分量进行聚类分析,最终获得能够代表多种属性的裂缝预测成果。实际资料表明,该方法具有一定的实用价值。

The accurate prediction of reservoir fracture distribution is the focus of exploration and development. The seismic geometric attributes have the best effect on fracture prediction, but different fracture attributes are often the response of different fracture characteristics in the subsurface. Comprehensively considering the prediction results of multiple attributes, a multi-attribute clustering method based on principal component analysis is proposed. The mutual independence principal component components of the different crack attributes are extracted, and then these components are used for cluster analysis. Finally, the prediction results of the crack which is capable of representing multiple attributes are obtained. The actual data shows that the method has certain practical value.