摘要
根据孔洞和裂缝的发育情况及其在储集和渗滤中所起的作用，结合地质分类方法，将碳酸盐岩储层划分为孔洞型、裂缝—孔洞型、孔洞—裂缝型和裂缝型４种储集类型。总结了用成像测井资料鉴别碳酸盐岩溶洞和裂缝的方法，根据ＦＭＩ分析结果从１６口井中划分出含４种储集类型的储层，描述了４种储集类型储层的常规测井响应特征。从常规测井资料中提取出指示孔洞、裂缝发育程度及指示储层均质性和纵向变化情况的测井响应平均值、均方差、关联度及残差等参数，组成识别储集类型的样本。设计３层ＢＰ网络，对样本进行训练，得到模型参数后对储集类型进行识别。
Carbonate reservoir is classified as four types, i.e., vug type, fracture vug type, vug fracture type and fracture type according to the growth degree of vugs and fractures and their effects on reservoir and infiltration as well as geologic classification method. Summed up here are the methods for indentifying vugs and fractures in carbonate with imaging logging data. Four types of reservoir are sorted out with FMI data from 16 wells and their conventional log response characteristics are also described. Seven characteristic parameters such as mean value, mean square deviation, correction index, residual error of conventional log response are extracted. They indicate the growth degree of vugs and fractures as well as reservoir homogeneity and longitidinal changes, thus, forming the samples for recognizing reservoir types. A three layer BP neural network is designed and used to train the samples. Then, the reservoir types are categorized when the model parameters are obtained. The obtained result agrees well with that of imaging logging and core analysis.
出处
《测井技术》
CAS
CSCD
北大核心
1999年第5期355-360,共6页
Well Logging Technology
关键词
碳酸盐岩
类型
测井资料
神经网络
识别
储集层
carbonate rock [reservoir type] log data imaging logging [Sample] neural network recognization
