人工神经网络在储层敏感性预测中的应用

Applications of Artifical Neural Network in Predicting Formation Sensitivity

介绍了改进的ＢＰ网络训练算法。提出了利用人工神经网络快速预测储层潜在敏感性（水敏性、速敏性）的方法步骤。分析表明，该方法受人为因素干扰小（总符合率大于８０％）、所需参数少、适用范围广（特别适于探井），能定量地反映出储层潜在敏感性程度，从而为制定保护油气层技术措施提供较可靠的依据。

The improved BP network training method was introduced.The method and step of using Artifical Neural Network can quickly predict the formation potential sensitivity.The analyses show that the method needs fewer parameters and has a broader applicability(especialy in explortary wells),and quantitavely reflect the level of the reservior potential sensitivily,so can provides reliable foundation for drafting the techniques and mesures to pro tect the reservior.