【摘要】 利用测井资料求取岩石的可钻性是一种简便可行的途径。常规的岩石可钻性测井预测模型都是基于回归分析而建立起来的 ,形式简单 ,精度不高。介绍了一种基于BP神经网络利用测井资料求取岩石可钻性的实用方法。该方法从测井信息与岩石可钻性的内在关系出发 ,选用与岩石可钻性密切相关的多个测井参数 ,通过BP神经网络技术建立利用测井资料准确求取岩石可钻性的数学模型 ;将该方法用于SC油田DU 4井等多口井的测井资料处理中 ,为邻井及时提供了较为准确的地层岩石可钻性剖面 ,也为该区的新井钻头选型提供了较好的依据。更多还原

【Abstract】 It is a simple and feasible approach to calculate rock drillability from log data. The normal prediction model of rock drillability is based on regression analysis. As a result, it’s compution is simple and it’s accuracy is also lower. This paper presents a practical approach to calculate rock drillability from logging data based on BP neural network. From log information and rock drillability, we select some log parameters related to rock drillability, and establish a mathematical model to accurately obtain rock drillability by BP neural network. This method is successfully applied to processing log data of BU4 well, SC oilfield. Accordingly, the accurate formation rock drillability section is presented in time, and the good guidance is provided for selecting bit types in a new well.更多还原