基于信息熵识别油气层和水层的聚类方法

简要地介绍了不确定性、信息熵、联合熵、条件熵、互信息的基本概念。将地层含有某种流体属性(如油层)的特征参数的概率分布应用于联合熵中,提出其概率分布是特征参数X与流体的特征参数Y的距离函数,再用熵与条件熵之差计算得到的互信息去识别地层的流体属性,即提出了基于信息熵识别油气层和水层的聚类方法。利用实际测井资料提取了油气层和水层的特征参数,并利用聚类方法识别出油气层和水层。识别结果与试油结论相符。该聚类方法不仅可用于识别油气层和水层,还可推广应用于划分地层岩性等领域。

Clustering method for identification of oil (gas)-bearing formation and water-bearing formation based on information entropy

The basic concepts of uncertainty, information entropy, united entropy, term entropy and mutual information were introduced briefly. The probability distribution of the fluid attribution in the oil (gas)-bearing formation or water-bearing formation where characteristic parameters occurred, was applied to united entropy. The probability distribution is the distance function between characteristic parameter X and the characteristic parameter Y of fluid. A method for identification of the fluid type by using the mutual information of the difference between entropy and term entropy was proposed, namely the clustering method for identification of the oil (gas)-bearing formation and water-bearing formation on the basis of information entropy. The characteristic parameters of the oil (gas), water-bearing formation were withdrawn by using well logging data. The oil (gas)-bearing formation and water-bearing formation were identified by the clustering method. The identifying results agree with the test oil conclusion. The clustering method can be used not only in the identification of the oil (gas)-bearing formation and water-bearing formation, but also in the demarcation of rock.