摘要
随着油田勘探程度的不断深入,低阻油层的开发逐渐成为油田增产的主要来源之一。低阻层是一种非常规储层,其含油性受多方面因素影响,常规的测井解释方法比较简单且对数据样本要求较高。于是,论文提出了一种基于梯度提升决策树(GBDT)的低阻油层识别方法。通过提取低阻层测井曲线中的数理特征,形成测井曲线特征集。再利用决策树算法,筛选出区分低阻油层和水层的重要特征,作为GBDT模型的输入,构建低阻油层识别模型。论文应用该模型,对**油田的低阻层进行识别,识别结果经多名石油专家论证,准确率高达90%,能够降低人工识别的成本以及因主观判别错误造成的损失,进一步提高油田的经济效益。
With the deepening of oilfield exploration,the development of low resistivity reservoir has gradually become one of the main sources of oilfield production.Low resistivity reservoir is a kind of unconventional reservoir,and its oil-bearing property is affected by many factors.Conventional logging interpretation method is relatively simple and requires high data samples.Therefore,this paper proposes a low resistivity reservoir identification method based on gradient boosting decision tree(GBDT).By extracting the mathematical characteristics of low resistivity logging curve,the logging curve feature set is formed.Then,the decision tree algo-rithm is used to screen out the important characteristics of distinguishing low resistivity oil layer and water layer,which are used as the input of GBDT model to build the identification model of low resistivity oil layer.In this paper,the model is applied to identify the low resistivity layer in**oilfield.The identification result is proved by many petroleum experts,and the accuracy rate is as high as 90%.It can reduce the cost of manual identification and the loss caused by subjective discrimination error,and further improve the economic benefits of the oilfield.
作者
牛庆威
张如玉
白雨昊
穆有德
王宇辰
吴傲
NIU Qingwei;ZHANG Ruyu;BAI Yuhao;MU Youde;WANG Yuchen;WU Ao(School of Computer Science and Technology,China University of Petroleum(East China),Qingdao 266580)
出处
《计算机与数字工程》
2023年第6期1428-1432,共5页
Computer & Digital Engineering
关键词
GBDT
决策树
特征提取
特征筛选
低阻油层识别
GBDT
decision tree
feature extraction
feature screening
identification of low resistivity reservoir
