Tight sandstone reservoirs in Sulige gasfield of Ordos Basin have the characteristics of complicated pore structure,low porosity,low permeability and strong heterogeneity,and the fluid identification of reservoir is difficult.According to the dipole acoustic logging data,based on rock elastic parameters of tight sandstone,fluid properties of tight sandstone reservoir identified by index method,crossplot method and support vector machine were discussed.The results show that ratio of time differences of longitudinal and shear waves,ratio of compressibility and poisson ratio and ratio of differences of elasticity moduli of longitudinal wave can be used to identify the gas reservoir;the gas-sensitive elastic parameters are selected optimally, which are fluid sensibility evaluation parameters including product of Lame coefficient and density, Lame coefficient,bulk modulus,ratio of Lame coefficient and shear modulus and elasticity modulus, and the accuracy of fluid identified by support vector machine is improved;compared with the gas production test results of reservoir sample,the prediction matching rate reaches 92% by support vector machine.Finally,well X350 in Sulige gasfield was used to verify the gas production test results.It is verified that the gas-sensitive elastic parameters selected optimally are used to identify the reservoir fluid by the means of support vector machine.%鄂尔多斯盆地苏里格气田致密砂岩储层的孔隙结构复杂，储层低孔、低渗、非均质性特征明显，储层流体识别困难。基于偶极声波测井资料，以致密砂岩岩石弹性参数为基础，利用指标法、交会图分析法、支持向量机进行储层流体性质识别的方法研究。结果表明：纵横波时差比、地层压缩系数与泊松比比值、纵波弹性模量差比可以作为气层识别的指标；引入流体敏感度评价参数，优选出能敏感反映储层含气性的弹性参数，如拉梅系数与密度之积、拉梅系数、体积模量、拉梅系数与剪切模量之比、弹性模量，结合支持向量机对流体进行识别，能够提高流体识别精度；对有试气结论的储层样本进行实际分类处理，利用支持向量机判别样本的成功率为92％。通过对苏里格气田的X350井进行实际分类处理，并经过试气结论验证，证明优选能敏感反映储层含气性的弹性参数，结合支持向量机对储层流体进行识别的方法是切实可行的。
展开▼
