论文部分内容阅读
砂岩储层孔隙中的流体识别一直是石油勘探开发过程中重要的环节,传统方法主要依赖于测井数据,但是在测井数据缺失的条件下较难得到准确的流体识别结果.本文提出一种只依靠地震数据的砂岩中流体识别的新方法,并选择地球物理方法可测或可求的地球物理参数σ、ρλ和ρμ作为流体识别因子,然后进行模型实验.首先,设置典型流体状态,用Gassmann方程进行流体替换,将得到的流体识别因子作为支持向量机的训练集数据,并定义支持向量机的分类标签;之后,设置随机流体状态,利用Gassmann方程计算流体因子,将得到的结果作为支持向量机的测试集数据.将训练集、测试集数据集输入支持向量机,进行分类,得出测试集数据的分类结果.模型实验分类结果表明,支持向量机法可以判别砂岩孔隙中流体的主要属性.
Fluid identification in sandstone reservoir pores has always been an important step in petroleum exploration and development, and the traditional methods mainly depend on well logging data, but it is difficult to obtain accurate fluid identification results in the absence of well logging data. Only rely on the new method of fluid identification in sandstone with seismic data and select the geophysical parameters σ, ρλ and ρμ that can be measured or obtained by geophysical methods as the fluid identification factor, and then carry out the model experiment.First, set the typical fluid state with Gassmann equation, the fluid identification factor is used as training set data of support vector machine, and the classification label of support vector machine is defined. Then, the random fluid state is set, the fluid factor is calculated by Gassmann equation and the result is supported The training set and test set dataset are input into the support vector machine to classify them to get the classification result of the test set data.The results of the model experiment show that the support vector machine method can judge the main fluid in the sandstone pore Attributes.