基于支持向量机分类的有效储层识别

摘要

The physical properties of tight sandstone reservoirs are lower and the pore structure is more complex than conventional ones, which lead to the accuracy of effective reservoir identification hard to meet the requirement along with the difficulties in solving permeability of tight sandstone reservoir. The pore structure index is used in this paper for scientific classification of the tight sandstone reservoir based on support vector machine (SVM) in flow unit and the porosity-permeability interpretation models are established respectively. The results of interpretation improves the accuracy and creates the favorable conditions for recognition of effective reservoir.%相比常规储层致密砂岩储层不仅物性更低，而且孔隙结构更加复杂，致密砂岩储层渗透率求解较常规储层远为困难的现状导致有效储层识别精度难以达到要求。本文基于支持向量机理论在流动单元分类识别中的应用，采用孔隙结构指数对致密砂岩储层进行科学分类后，分别建立适用于该种条件下储层孔隙结构的孔-渗解释模型，所得渗透率解释精度大为提高，对有效储层识别创造了有利条件。