基于测井曲线斜率的KNN分类算法常规测井裂缝识别——以普光气田礁滩相储层为例

Fracture Identification in Conventional Log through KNN Classification Algorithm Based on Slope of Logging Curve——A Case Study of Reef Flat Facies Reservoir in Puguang Gasfield

常规测井相对岩心和成像测井来说,具有普遍性和经济性,在裂缝识别研究过程中具有岩心和成像测井不可替代的作用。如何利用常规测井资料对裂缝进行有效识别,特别是对于储层物性好、裂缝发育的边水油(气)藏,是防止边水顺裂缝突进、保障油(气)藏高效开发的关键。以普光气田长兴组和飞仙关组岩心裂缝观察为基础,从裂缝测井响应原理出发,提出对测井曲线求斜率,放大裂缝测井响应特征的方法,并利用对类域的交叉或重叠较多的待分样本集具有较好分类作用的KNN分类算法,进行裂缝识别。通过对普光气田礁滩相储层裂缝常规测井识别的运用,证实该思路对储层物性好、裂缝和孔隙同时发育的双重孔隙介质储层的有效裂缝和无效裂缝具有很好的识别效果,常规测井识别结果和岩心观察裂缝识别结果之间具有很高的吻合性。

Compared with core and imaging log,conventional log is nonreplaceable during the fracture identi- fication research for its universal and economic nature.How to improve fracture identification accuracy conventional log information is the key to prevent the edge water breakthrough along the fractures and through ensure efficient development of the reservoirs,especially the oil-gas reservoir with active edge water,naturally frac- tured and good reservoir property.Based on the core fracture observation in Changxing formation and Feixian- guan formation of Puguang gasfield ,a novel method is presented that calculating the slope of logging curve to magnify the response feature of the fractures according to the principle of KNN classification algorithm is applied to fracture identification for its good with more cross or overlap in the class field.After its application to iden fracture in logging response.And classification role on the samples tifying reef beach facies reservoir with conventional log in Puguang gasfield ,its outstanding effect is proved for identifying effective and invalid fractures of the reservoir with good property,fracture-pore double porosity media.Identification results of the conventional log and that of the core fracture observation are almost the same.