神经网络方法拟合核磁共振测井曲线  

Fitting Virtual Magnetic Resonance Logs Using Neural Networks Method

作　　者：李瑜[1] 郑敏娟[1] 程国建[1] 

机构地区：[1]西安石油大学计算机学院,陕西西安710065

出　　处：《微电子学与计算机》2009年第2期66-68,共3页Microelectronics & Computer

基　　金：国家自然科学基金项目(40572082);西安石油大学科技创新基金项目(Z07079)

摘　　要：针对常规测井数据无法准确评价低电阻储层的特点,通过建立基于神经网络的高度非线性映射模型,由常规测井曲线拟合核磁共振测井曲线实现测井解释.通过与实际数据的相关度比较,验证了该方法具有模拟核磁共振描述井下储层特征的能力.将其应用于低电阻油藏描述中,可以降低投资成本,对提高油田的勘探效益具有实际意义.In view of the data obtained by conventional well logs can t appraise low resistivity reservoirs well and truly,basing on the Neural Networks the capability of expressing arbitrary nonlinear mapping,a methodology of fitting virtual magnetic resonance logs using data obtained by conventional well logs is proposed to realize well logging interpretation.The correlation compared with the actual data shows that this methodology can simulate nuclear magnetic resonance providing the capability of measurement of r...

关 键 词：核磁共振测井 神经网络 BP算法 

分 类 号：P631.2[天文地球—地质矿产勘探]