小波神经网络在隔夹层研究中的应用  

Application of WNN in the Studies of Interbeds

作　　者：瞿子易[1] 罗鑫[1] 谢润成[1] 葛善良[1] 牛会玲[1] 杨露[2] 

机构地区：[1]成都理工大学油气藏地质及开发工程国家重点实验室,四川成都610059 [2]长庆油田分公司采油三厂,宁夏银川750006

出　　处：《石油天然气学报》2009年第4期265-268,432,共4页Journal of Oil and Gas Technology

摘　　要：在碳酸盐岩储层中,隔夹层的稳定程度直接影响储层的垂向连通性、压力系统和油水关系、细分调整时的层系划分以及所要采取的技术措施。因此,隔夹层识别对储层非均质研究具有重要意义。研究以DALEEL油田为例,在综合考虑孔渗等物性参数和测井资料的基础上,引入小波神经网络对测井参数进行建模,对100个样本数据进行训练,36个预测样本进行识别,识别正确率为94.4%。还采用了BP神经网络和多元统计方法对样本数据进行识别研究,识别正确率分别为80.6%和69.4%。结果表明小波神经网络模型对隔夹层的识别效果较好,为储层非均质研究提供了可靠的依据。In carbonate reservoirs,the stability of interbeds had a direct impact on the vertical reservoir connectivity,pressure systems and oil-water relationship between the breakdown of adjustment layers,as well as technical measures to be taken.Identification of interbeds was important to the study of reservoir heterogeneity.By taking DALEEL Oilfield for example,on the base of considering physical parameters of the porosity and permeability and logging data,the interbeds identification model was established by wavelet neural network(WNN)with logging data,100 sample data were trained,36 samples were identified,the right recognition rate is 94.4%,also BP and multivariate statistical methods was used to study the sample data,the right recognition rates are 80.6% and 69.4%.It shows that the WNN is better for identifying interbeds and providing a reliable basis for the further study of reservoir heterogeneity.

关 键 词：碳酸盐岩储层 隔夹层识别 小波神经网络 

分 类 号：P618.13[天文地球—矿床学]