基于决策树方法的砾岩油藏岩性识别——以克拉玛依油田六中区克下组油藏为例  

Lithological Identification of Conglomerate Reservoirs Base on Decision Tree Method

作　　者：李洪奇[1] 谭锋奇[1] 许长福[2] 王晓光[2] 彭寿昌[2] 

机构地区：[1]中国石油大学(北京)资源与信息学院油气资源与探测国家重点实验室(中国石油大学(北京)),北京102249 [2]新疆油田公司勘探开发研究院,新疆克拉玛依834000

出　　处：《石油天然气学报》2010年第3期73-79,共7页Journal of Oil and Gas Technology

基　　金：国家"十五"重大科技攻关项目(2001BA605A09)

摘　　要：砾岩油藏由于近物源、多水系和快速多变的沉积环境导致储层岩性复杂以及测井响应无规律性变化等特点,岩性的准确识别成为该类油藏调整开发的重点和难点。以克拉玛依油田六中区克下组砾岩油藏为研究对象,利用决策树的分析方法,结合密闭取心资料对砾岩油藏8种岩性的测井曲线敏感性进行研究,优选出原状地层电阻率、声波时差和自然伽马3个测井响应值作为砾岩油藏岩性识别的特征参数,建立决策树岩性识别模型,模型的综合判断准确率达到96.97%,在油田的二次调整开发和水淹层的评价中取得了很好的应用效果。并且在决策树方法分析结果的基础上,利用地球物理方法制作了砾岩油藏岩性识别图版,数据挖掘方法与地球物理方法的有效结合为解决油田勘探和开发中的复杂问题提供了一种新思路和新方法。The characteristics of conglomerate reservoir were complicated reservoir lithology and logging response without regular changes because of the special depositional environment of close source,multi-drainage systems and rapidly changed,and accurate identification of lithology became focus and difficult to the adjustment and development of the conglomerate reservoirs.The conglomerate reservoir in Kexia Formation of Sixth District in Karamay Oilfield was taken as an object the method of decision tree was used to study thee sensitivity of 8 lithological logging curves,and three logging response values such as original formation resistivity,acoustic travel time and natural gamma were optimized as characteristic parameters for conglomerate lithology identification.The lithology identification model of decision tree was established with accuracy rate of 96.97% and good effect was achieved in the second oilfield adjustment and development of and evaluation of water flooded layers.The identification chartboard of conglomerate reservoir lithology is produced by using geophysical methods based on analysis results of decision tree.The effective combination of data mining methods with geophysical methods provides a new idea and a new method to solve complex problems in oilfield exploration and development.

关 键 词：岩性识别 决策树 数据挖掘 克拉玛依油田 砾岩油藏 

分 类 号：P618.13[天文地球—矿床学]