|本期目录/Table of Contents|

[1]韩 建,刘 梦,曹志民,等.一种基于多模型融合的测井曲线复原方法[J].工业仪表与自动化装置,2021,(06):106-111.[doi:10.19950/j.cnki.cn61-1121/th.2021.06.021]
　HAN Jian,LIU Meng,CAO Zhimin,et al.Multi-model fusion-based approach of well logging curve restoration[J].Industrial Instrumentation & Automation,2021,(06):106-111.[doi:10.19950/j.cnki.cn61-1121/th.2021.06.021]
点击复制

一种基于多模型融合的测井曲线复原方法

《工业仪表与自动化装置》[ISSN:1000-0682/CN:61-1121/TH]

卷:
期数:
2021年06期
页码:
106-111
栏目:
出版日期:
2021-12-15

文章信息/Info

Title:
Multi-model fusion-based approach of well logging curve restoration
作者:
韩 建1; 2; 刘 梦1; 2; 曹志民1; 2; 3; 李 婧1; 2; 李佳露1; 2; 王思捷1; 2
(1.东北石油大学 物理与电子工程学院；2.东北石油大学 黑龙江省共建测试计量技术及仪器仪表研发中心；3.大庆油田勘探开发研究院，黑龙江 大庆 163318)
Author(s):
HAN Jian1; 2; LIU Meng1; 2; CAO Zhimin1; 2; 3; LI Jing1; 2; LI Jialu1; 2; WANG Sijie1; 2
(1.School of Physics and Electronic Engineering, Northeast Petroleum University; 2.Research and Development Center for Testing and Measurement Technology and Instrumentation, Heilongjiang Province Universities, Northeast Petroleum University;3.Daqing oilf
关键词:
测井曲线; 模型融合; GBDT; LightGBM; XGBoost
Keywords:
log;  model fusion;  GBDT;  LightGBM;  XGBoost
分类号:
TP391
DOI:
10.19950/j.cnki.cn61-1121/th.2021.06.021
文献标志码:
A
摘要:
利用测井数据进行井内物理地质描述是测井勘探时常用的手段，但是测井数据经常会出现缺失的问题。所以测井曲线的复原一直是石油勘探领域迫切解决的问题。该文提出了一种基于多种树模型融合的方法，利用Stacking集成方法融合模型，对井内的多条测井曲线进行复原。实验结果表明，通过多模型融合方法的复原结果优于单一树模型，该方法具有更小的误差和更高的精度。
Abstract:
A Method of Logging Curve Recovery Based on Multi-model Fusion Using logging data to describe physical geology in well is often used in well logging exploration, but the problem of missing logging data often occurs.So the restoration of well logging curve has always been an urgent problem in petroleum exploration.In this paper, a method based on the fusion of multiple tree models is proposed, and the multiple logging curves in the well are restored by using the fusion model of Stacking integration method.The experimental results show that the recovery results of the multi-model fusion method are better than that of the single tree model, and the proposed method has smaller errors and higher precision.

参考文献/References:

[1] 陈云天.基于机器学习的测井曲线补全与生成研究[D].北京：北京大学,2020.

[2] HA H,OLSON J R,BIAN L,Analysis of heavy metal sources in soil using kriging interpolation on principal components[J].Environmental Science&Technology,2014, 48(9):4999-5007.
[3] BATEMAN R M.Openhole log analysis and formation evaluation[M].D.ReidelPub.Co.International Human Resources Development Corp. 1986.
[4] ASQUITH G B , KRYGOWSKI D . Basic Well Log Analysis[M]. 2004.
[5] 张晶玉,张会来,范廷恩,等.时移测井曲线重构方法及其在A油田中的应用[J]. 物探化探计算技术,2018, 40(03):318-323.
[6] ROLON L ,MOHAGHEGH S D,AMERI S,et al.Using artificial neural networks to generate synthetic well logs[J]. Journal of Natural Gas Science & Engineering, 2009, 1(4-5):118-133.
[7] 杨志力,周路, 彭文利,等. BP神经网络技术在声波测井曲线重构中的运用[J]. 西南石油大学学报(自然科学版), 2008(01):63-66+13.
[8] 张东晓,陈云天,孟晋.基于循环神经网络的测井曲线生成方法[J].石油勘探与开发,2018,45(04):598-607.
[9] 王俊,曹俊兴,尤加春.基于GRU神经网络的测井曲线重构[J].石油地球物理勘探, 2020,55(03):510-520+468.
[10] 柯国霖.梯度提升决策树(GBDT)并行学习算法研究[D]. 福建:厦门大学,2016.
[11] CHEN T ,TONG H , BENEST M . xgboost: Extreme Gradient Boosting[J]. 2016.
[12] 黄卿,谢合亮.机器学习方法在股指期货预测中的应用研究——基于BP神经网络、SVM和XGBoost的比较分析[J].数学的实践与认识, 2018(8):297-307.
[13] LIANG W,LUO S,ZHAO G,et al. Predicting Hard Rock Pillar Stability Using GBDT, XGBoost, and LightGBM Algorithms. 2020,8(5).
[14] 魏志强,张浩,陈龙.一种采用SmoteTomek和LightGBM算法的Web异常检测模型[J].小型微型计算机系统,2020,41(03):587-592.

相似文献/References:

备注/Memo

备注/Memo:
收稿日期：2021-06-21
作者简介：韩建（1976），男（汉族），黑龙江大庆人，博士，教授，主要从事油井信号检测，机器学习，模式识别研究。
通信作者：曹志民.
更新日期/Last Update: 1900-01-01