基于核Fisher判别的碎屑岩储层流体识别

The method on the fluid identification of rock elastic properties by Fisher Discriminant Analysis in fragmental rock

流体识别是致密碎屑岩储层预测中的难点之一.本文把核Fisher判别分析用于致密碎屑岩储层流体识别,首先较全面的介绍了核Fisher判别分析理论基础,再利用测井属性参数计算出弹性属性参数,并将这两类参数作为特征变量,最后进行核Fisher判别分析建模和致密碎屑岩储层流体识别.四川盆地XC地区致密碎屑岩储层流体识别结果表明,对致密碎屑岩储层中的气层和气水同层两类线性不可分的样本数据,使用核Fisher判别分析均能实现分离,说明该方法在识别致密碎屑岩储层流体识时准确率高,具有很好的应用前景.

Fluid identification is one of difficulties in tight clastic reservoir prediction. KFDA has been applied to fluid identification with strong nonlinear extraction ab/lity. This paper comprehensively introduces KFDA theory, then logging attribute parameters is used to calculate the elastic attributes, these two kinds of attributes as variables for the fluid identification. Finally kernel Fisher discriminant was introduced to apply to the fluid identification of tight clastic reservoir. Taking Xujiahe formation of western Sichuan for example, the experimental results show that the method of tight clastic rock fluid identification is effective and high accuracy, but also have good wide perspective.