基于支持向量回归机与井导向的三角洲岩性油气藏储层参数预测

三角洲岩性油气藏中储层河道砂体小、散、薄，连通性差，且与泥岩互层，导致反射信号弱，难以进行综合/精细解释。据此，本文提出基于支持向量回归机与井导向的三角洲岩性油气藏储层参数预测方法，用于刻画此类砂体分布特征。从测井资料中提取揭示储层特征的参数作为储层预测的导向，利用支持向量回归机建立多种属性与储层参数之间的映射关系，进而开展储层预测。针对CF区实例，通过估算储层的伽马参数和R4参数预测河道砂体的分布特征，所得结果与钻井揭示的实际岩性有较高吻合度。因此，本文方法适用于三角洲岩性油气藏预测。

Delta stratigraphic reservoir parameter estimation based on support vector regression machine and well logging data

Delta stratigraphic reservoirs such as stream course sand body are thin and interbedded with mudstone, and they have poor connectivity. Feeble reflected signals from this kind of reservoirs on seismic section are not helpful for sand body distribution interpretation. Therefore we propose a method for reservoir parameter estimation based on support vector regression machine and well logging data. First we extract parameters from well logging data, which can reveal reservoir features as a guide. Then we establish relations between reservoir feature parameters and some seismic attributes by support vector regression machine for reservoir prediction. We apply this method to a project in the Block Changfeng, Huabei Oilfield. Estimated gamma parameter and R4 parameter are used to predict sand body distribution. The results reveal that the sand body distribution feature is well consistent with that of real well drilling, which proves that the proposed method can predict delta stratigraphic reservoirs and sand body distribution features.