基于支持向量机法识别砂岩中流体类型

Fluid type identification of sandstone based on support vector machine

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.

Pore fluid identification in sandstone reservoir has always been a important step in the process of petroleum exploration and development, conventional methods depend on the logging data. Generally speaking, it is hard to get accurate fluid identification result under the condition of lack of logging data. In this paper, we present a new fluid identification method based on Support Vector Machine. This new method can achieve the fluid identification with only seismic data. We set geophysical parameters a, p）, and p/~ which can be ohserved or rectifiable by geophysical methods as fluid identification factor, then do model experiments. First, set up typical fluid states and do the fluid substitution with Gassmann equation, then set the fluid identification factors we get as training data for support vector machine. After that, we set up random fluid states, using Gassmann equation to calculate fluid identification factors, the results will be set as test data for support vector machine. Model experiments indicate that support vector machine method can determine the major fluid properties in the sandstone porosity.