基于Fisher判别的储层分类渗透率预测技术

Classification of Reservoir Permeability Prediction Technology Based on Fisher Discrimination

莺歌海盆地中深层储层岩性复杂、非均质性强,孔渗关系趋势较差,导致利用测井资料进行储层渗透率评价技术出现了精度大幅降低或失效的情况。为有效地解决研究区储层渗透率精细评价的问题,利用取心井段岩心分析孔渗资料、测井资料,选用流动单元指数FZI值进行储层分类,建立各类储层的渗透率精细计算模型。优选反映储层特征变化的测井变量,运用Fisher判别分析法建立储层类型判别标准的新方法应用于莺歌海盆地东方13气田中,结果表明:储层渗透率计算精度有了明显提高,为研究区的复杂储层渗透率评价提供了一种有效途径。

The deep reservoir of Yinggehai Basin is characterized by its complex lithology and high heterogeneity.This results in a poor porosity-permeability relationship,leading to the failure in using logging data for reservoir permeability evaluation technology with precision.To effectively solve the problem of fine evaluation of reservoir permeability in the study area,this study uses core analysis data of cored interval,well logging data,and selects the flow unit index FZI value for reservoir classification.Hence,models of fine permeability evaluation can be built for various types of reservoir.The reservoir logging attribute variables are optimized to reflect the change characteristics,and a new method is established for the discrimination standard to classify reservoirs by using Fisher discriminant technology for training of each type of reservoir.Based on the reservoir classification,the method is applied in DF13 gas field of Yinggehai Basin.The practical application shows that the accuracy of reservoir permeability prediction has been significantly improved compared with those made with the conventional statistical methods.It then proves suitable for the evaluating the mid-deep to deep formations with wireline log data,and the method should be an effective way to evaluate reservoir permeability in the study area.