声波曲线重构技术在火山岩储层预测中的应用  

Application of Log Reconstruction Technique in Volcanic Reservoir Prediction

作　　者：朱超[1,2] 高先志[2] 杨德相 崔俊峰[3] 

机构地区：[1]杭州地质研究院,浙江杭州310023 [2]中国石油大学(北京)资源与信息学院,北京102249 [3]中国石油华北油田勘探开发研究院,河北任丘062550

出　　处：《石油天然气学报》2010年第3期258-262,共5页Journal of Oil and Gas Technology

摘　　要：火山岩储层具有较强的非均质性,常规的声波测井曲线约束反演由于受多种因素的影响,不能有效地反映火山岩储层特征。针对火山岩这种特殊的岩性体,探索性的采用深、浅双侧向测井曲线重构声波曲线,在此基础上,以神经网络反演为手段,对二连盆地洪浩尔舒特凹陷火山岩储层分布进行预测,取得了较好的应用效果。Volcanic reservoir had a stronger heterogeneity,conventional sonic log constrained inversion was usually influenced by many uncertain factors,therefore it could not be used to effectively reflect the characters of volcanic reservoirs.In allusion to the peculiar volcanic lithology,shallow and deep dual laterolog was used to reconstruct sonic logging curve.On the basis stated by means of BP neural network reversion,the distribution of volcanic reservoirs in Honghaoershute Sag of Erlian Basin is predicted and good effect is obtained.

关 键 词：火山岩储层 储层预测 非均质性 曲线重构 神经网络 二连盆地 

分 类 号：P618.13[天文地球—矿床学]