摘要
藻灰岩在测井曲线图上具有不同于其它岩性的特征,据此确定了7 种测井参数(自然伽玛、自然电位、井径、声波时差、补偿中子、补偿密度及深电阻率),将测井数据归一化后得到了相应的岩性判别样本参数。利用F Means中速度较快的聚类迭代(LBG)算法对研究样本进行分类,优选了一组数据用于判别分析。在进行Q因子分析的基础上建立了用于岩性判别的3 个判别函数,并对花土沟油田某井段进行了岩性识别预测分析。
Algae limestone represents different characteristics in the plot of well logging. Accordingly seven parameters(GR,SP,CAL,AC,CNL,DEN,RT) selected from well logging data are standardized to form sample parameters for lithologic identification. The LBG(Linde-Buzo-Gray) algorithms in F-means cluster is used to classify the samples, a set of data is then selected for the identification and analysis. Based on Q gene analysis, three identification functions have been obtained and used in lithologic identification and prediction in terms of well logging data of Huatugou oilfield. The method is fast and efficient in identifying algae limestone based on well logging data.
出处
《吉林大学学报（地球科学版）》
EI
CAS
CSCD
北大核心
2005年第3期382-388,共7页
Journal of Jilin University：Earth Science Edition
基金
中国石油天然气集团公司青海分公司项目(03370502000126)
关键词
藻灰岩
归一化
聚类迭代
岩性判别
algae limestone
digitalized
LBG algorithms
lithologic identification
