| 基于改进多分类孪生支持向量机的测井岩性识别方法研究与应用 |
| |
| 引用本文： | 苏赋,马磊,罗仁泽,倪艳.基于改进多分类孪生支持向量机的测井岩性识别方法研究与应用[J].地球物理学进展,2020(1):174-180. |
| |
| 作者姓名： | 苏赋 马磊 罗仁泽 倪艳 |
| |
| 作者单位： | 西南石油大学;新疆油田公司数据公司 |
| |
| 基金项目： | 国家重点研发计划深地专项项目(2016YFC0601100);四川省科技计划项目(2019CXRC0027)赞助. |
| |
| 摘 要： | 岩性识别是认识地层及求解储层参数的基础,受地质环境复杂性和非均质性影响,测井曲线间存在着大量的信息冗余,数据集类间分布不平衡,常用的分类算法无法满足实际需求.针对常用分类算法容错性差,识别岩性单一和无法有效解决类间不平衡的问题,本文改进合成少数过采样技术(Synthetic Minority Over Sampling Technique,SMOTE)来处理数据集,可得到类间平衡的新数据集,并提出一种新的模糊隶属度函数改进模糊孪生支持向量机,在北美Hugoton油气田实际测井数据的基础上,用改进多分类孪生支持向量(Improve Multi Class Twin Support Vector Machine,IMCTSVM)综合自然伽马(GR)、电阻率(RL)、光电效应(PE)、中子密度孔隙度差异(DPHI)和平均中子密度孔隙度(PHIND)五种测井参数,以及相对位置(RELPOS)和非海洋/海洋指标(NM_M)两种地质约束变量,识别出9种岩性.将识别结果与传统支持向量机、深度神经网络等方法进行对比与分析,发现IMCTSVM算法优于上述两种分类算法,取得了较好的识别效果.
|
| 关 键 词： | 模糊孪生支持向量机 不平衡数据集 岩性识别 合成少数过采样技术(SMOTE) |
Research and application of logging lithology identification based on improve multi-class twin support vector machine |
| |
| SU Fu,MA Lei,LUO Ren-ze,NI Yan.Research and application of logging lithology identification based on improve multi-class twin support vector machine[J].Progress in Geophysics,2020(1):174-180. |
| |
| Authors: | SU Fu MA Lei LUO Ren-ze NI Yan |
| |
| Institution: | (Southwest Petroleum University,Chengdu 610500,China;Xinjiang Oilfield Company PetroChina,Karamay 834000,China) |
| |
| Abstract: | Lithology identification is the basis for understanding the stratum and solving the reservoir parameters.Logging curve dataset is affected by geological environment complexity and heterogeneity.There are two problems when using logging data for lithology classification.One is that there is a large amount of information redundancy between the logging curves,and the other is the imbalance between the data sets.Common classification algorithm cannot meet actual needs.In order to solve the problem that the existing common classification algorithm has poor fault tolerance,identify single lithology and cannot effectively overcome the imbalance between classes,this paper proposes an improved algorithm.This paper improves Synthetic Minority Over Sampling Technique(SMOTE)to process data sets.Using ISMOTE to process data sets,we can get a new data set balanced between classes.This paper proposes a new fuzzy membership function to improve the fuzzy twin support vector machine.IMTSVM using gamma ray,resistivity,photoelectric effect,neutron density porosity difference,average neutron density porosity 5 logging parameters and relative position,nonmarine-marine indicator 2 geologic constraining variables of Hugoton oil and gas field real logging data in North America.It identifies nine lithologies.The recognition results are compared with traditional support vector machines and deep neural networks.It is found that the IMCTSVM algorithm is better than the above two classification algorithms and has achieved good recognition results. |
| |
| Keywords: | Fuzzy twin support vector machine Unbalanced data set Lithology identification Synthetic Minority Over Sampling Technique(SMOTE) |
| 本文献已被 CNKI 维普 等数据库收录！ |
|
