利用贝叶斯判别法识别岩性基础上的孔隙度评价

Porosity Evaluation Based on Lithology Indentification Using Bayesian Classifier

伊拉克M油田具有岩性复杂、储集空间多样、非均质性强等特征,常规孔隙度计算模型不能较好地反映其储层孔隙度。为了提高其孔隙度计算精度,在计算孔隙度前先对储层作了岩性判别。以最小错误率的贝叶斯判别法为基础,选取研究区测井资料齐全和岩性分布较均衡的5口井作为样本,选择合适的测井曲线作为输入曲线,利用贝叶斯公式求得每个深度点的各类后验概率值,后验概率值最大的即为该点所属的类别,以此为依据对所有井作了岩性识别。在岩性识别的基础上对经过岩心分析的岩样分类拟合以获得各类岩性的拟合模型以及在该地区的骨架值,根据所输出的岩性代码,可选择该深度点所对应的测井解释模型。识别岩性后计算的孔隙度与岩心分析孔隙度的误差很小,为储层的解释评价提供了理论基础。

Oilfield M in Irag is characterized by complex lithology,various reservoir space,strongly anisotropy and so on,and the common porosity calculation model couldn't be a good response to the reservoir porosity,so the lithologic discrimination has been done before porosity calculation.Based on the minimum error rate training method,five wells are chosen as the sample which has complete logging data and balanced lithology distribution.Then each posterior probability of every depth is acquried based on the Bayesian formula after chosing appropriate logging curves as input curves.The classes which every depth point belongs to would be decided according to the maximum posterior probability,then proceed with these steps above to all wells to finish lithology identification.All kinds of lithologic fitting model and skeleton value in research area can be obtained based on lithology identification,the output lithology code can choose corresponding logging interpretation model.Calculation of porosity after lithology identification turns out to be accurate,it provides a strong foundation for the interpretation of reservier.