Logging lithology identification method research based on PSO-SVM: a case study of Paleozoic (Pz) reservoir in K oil field, South Turgay Basin, Kazakhstan
-
摘要:
南图尔盖盆地K油田古生界(Pz)岩性多样、孔隙结构复杂，针对常规岩性解释方法对该储层岩性识别准确度未达到预期效果，严重制约了测井储层解释等问题，提出基于粒子群算法优化支持向量机(PSO-SVM)的岩性识别方法.通过岩心资料分析不同岩性的测井响应特征，建立测井相识别图版.选择对研究区岩性敏感的自然伽马、阵列感应电阻率、声波、中子、密度与光电吸收截面指数等七条测井曲线参数作为输入特征值，以粒子群算法优选合适的支持向量机参数(惩罚因子和核函数参数)对研究区4口取心井进行样本学习，建立基于PSO-SVM的岩性识别模型，其识别准确率达到了97%.相对于传统SVM算法，PSO-SVM岩性识别模型预测结果的速度更快，精度更高.通过将该模型应用于取心井与试油井，在正确识别岩性的同时，有效提高了测井储层解释的准确性.结果表明，在K油田复杂岩性识别中应用PSO-SVM方法，可为提高测井储层解释的准确性提供较可靠的岩性依据.
Abstract:The Paleozoic (PZ) reservoir of the K Oilfield in South Turgay Basin is characterized by diverse lithology and complex pore structure. In view of the problems that conventional lithology interpretation methods fail to achieve the expected accuracy of lithology identification, which seriously restricts the logging reservoir interpretation, a lithology identification method based on Particle Swarm Optimization Support Vector Machine (PSO-SVM) is proposed. Analyze the logging response characteristics of different lithologies through core data, and establish a logging facies identification chart. Seven logging curve parameters, which are sensitive to lithology in the study area, such as natural gamma ray, array induction resistivity, acoustic wave, neutron, density and photoelectric absorption cross section index, are selected as input eigenvalues. Select appropriate support vector machine parameters (penalty factors and kernel function parameters) through the particle swarm algorithm to learn samples from 4 core wells in the study area, and establish a lithology recognition model based on PSO-SVM, of which the identification accuracy reached 97%. Compared with traditional SVM algorithm, PSO-SVM lithology identification model can predict results faster and with higher accuracy. By applying this model to coring wells and oil test wells, while correctly identifying lithology, it effectively improves the accuracy of logging reservoir interpretation. The results show that the application of PSO-SVM in K Oilfield complex lithology identification can provide a more reliable lithology basis for improving the accuracy of logging reservoir interpretation.
-
-
图 2
(a) K18井灰岩测井综合图；(b)K30井泥质灰岩测井综合图；(c)K8井变质岩测井综合图
Figure 2.
(a)Comprehensive logging map of limestone in well K18;(b)Comprehensive logging map of argillaceous limestone in well K30;(c)Comprehensive logging map of metamorphic rock in well K8
图 7
粒子群优化算法的适应度曲线
Figure 7.
The fitness curve of particle swarm optimization algorithm
表 1
K区块部分取心井物性特征统计表
Table 1.
Statistical table of physical properties of partial coring wells in block K
岩性 样品数 孔隙度/% 渗透率/(0.987×10-15 m2) 最大值 最小值 平均值 最大值 最小值 平均值 灰岩 41 21.35 2.24 5.19 13.26 0.015 0.829 泥质灰岩 39 15.7 2.24 7.68 6.588 0.006 0.419 变质岩 8 10.74 2.27 4.755 1.017 0.062 0.292 
下载: 导出CSV
表 2
K油田测井相识别图版
Table 2.
K oil field logging facies identification chart
岩性 GR/API Rt/Ω·m AC/(μs/m) DEN/(g/cm3) CNL/% Pe 灰岩 20~45 30~2000 150~280 2.37~2.65 5~22 2.4~5 泥质灰岩 40~110 4~30 190~350 2.39~2.66 10.5~25 2.4~3.5 变质岩 40~80 10~100 180~300 2.47~2.64 6.9~18 2.5~4
下载: 导出CSV
表 3
K油田筛选岩心统计表
Table 3.
Statistical Table of selected core in K oil field
灰岩 泥质灰岩 变质岩 总数 测试样本数 166 239 50 455
下载: 导出CSV
表 4
不同训练样本数的预测结果表
Table 4.
Table of prediction results for different numbers of training samples
SVM PSO-SVM 训练样本数 100 150 200 250 300 100 150 200 250 300 测试样本数 100 100 100 100 100 100 100 100 100 100 预测正确率/% 81 83 83 84 84 84 87 90 90 93
下载: 导出CSV
-
Cao Q, Wang Z Z, Fan T L, et al. 2018. Application of permeability interpretation based on fuzzy support vector machine in sandy conglomerate reservoir. Journal of Xinjiang University (Natural Science Edition) (in Chinese), 35(3): 264-271.
Ding S F, Qi B J, Tan H Y. 2011. An overview on theory and algorithm of support vector machines. Journal of University of Electronic Science and Technology of China (in Chinese), 40(1): 2-10.
Fan X W. 2003. Support vector machine and its applications [Ph. D. thesis](in Chinese). Hangzhou: Zhejiang University.
Fu G M, Yan J Y, Zhang K, et al. 2017. Current status and progress of lithology identification technology. Progress in Geophysics (in Chinese), 32(1): 26-40, doi: 10.6038/pg20170104.
Han X H, Zhi L F, Liu R, et al. 2013. A lithologic identification method in Es4 reservoir of Guangli oilfield with least square support vector machine. Progress in Geophysics (in Chinese), 28(4): 1886-1892, doi: 10.6038/pg20130430.
Kong X Y, Yin J Y, Zhang F Q. 2007. Oil-gas geological features and its exploration potential in South Turgay Basin, Kazakhstan. Lithologic Reservoirs (in Chinese), 19(3): 48-53.
Li J L, Liu G P. 2016. In the application of lithology recognition based on PCA-SVM algorithm. Jiangxi Science (in Chinese), 34(4): 504-510, doi: 10.13990/j.issn1001-3679.2016.04.020.
Li P C, Xu S H. 2005. Support vector machine and kernel function characteristic analysis in pattern recognition. Computer Engineering and Design (in Chinese), 26(2): 302-304.
Li X Y, Zhou J Y, Li H Q, et al. 2012. Computational intelligent methods for predicting complex lithologies and multiphase fluids. Petroleum Exploration and Development (in Chinese), 39(2): 243-248.
Lin X L, Yuan R, Sun Y Q, et al. 2018. The basic theory and research progress of support vector machine. Journal of Yangtze University (Natural Science Edition) (in Chinese), 15(17): 48-53, doi: 10.16772/j.cnki.1673-1409.2018.17.010.
Liu W. 2009. Well logging evaluation methods for complex low permeability oil and gas reservoir [Ph. D. thesis](in Chinese). Qingdao: China University of Petroleum.
Mou D, Wang Z W, Huang Y L, et al. 2015. Lithological identification of volcanic rocks from SVM well logging data: Case study in the eastern depression of Liaohe Basin. Chinese Journal of Geophysics (in Chinese), 58(5): 1785-1793, doi: 10.6038/cjg20150528.
Shao X G, Yang H Z, Chen G. 2006. Parameters selection and application of support vector machines based on particle swarm optimization algorithm. Control Theory & Applications (in Chinese), 23(5): 740-743, 748.
Song Y J, Zhang J F, Yan W L, et al. 2007. A new identification method for complex lithology with support vector machine. Journal of Daqing Petroleum Institute (in Chinese), 31(5): 18-20, 46.
Su F, Ma L, Luo R Z, et al. 2020. Research and application of logging lithology identification based on improve multi-class twin support vector machine. Progress in Geophysics (in Chinese), 35(1): 174-180, doi: 10.6038/pg2020DD0185.
Vapnik V N. 1995. The Nature of Statistical Learning Theory. New York: Springer.
Wang X Y, Chen Z G, Zhong X R, et al. 2018. PSO-SVM based leakage diagnosis method of water supply pipeline. Modern Electronics Technique (in Chinese), 41(7): 156-159, 164.
Wang Z T, Song X Q. 2018. A study on space-borne SAR signal recognition technology based on support vector machine. Modern Radar (in Chinese), 40(11): 31-36.
Xie J. 2015. Based on ACO to optimize the SVM research in logging lithological identification [Master's thesis](in Chinese). Daqing: Northeast Petroleum University.
Yang F Q, Chen K G, Huang C B, et al. 2019. Application of PSO-LIBSVM in recognition of potassium salt deposits. Advances in Earth Science (in Chinese), 34(7): 757-764.
Yin J P. 2016. Research on selection of kernel functions and key parameters in support vector machine [Master's thesis](in Chinese). Harbin: Harbin Institute of Technology.
Zhang Z J, Fang S. 2019. Application of Support Vector Machine in lithology identification based on genetic algorithm optimization. Global Geology (in Chinese), 38(2): 486-491.
Zhao X L, Wang G W, Zhou Z L, et al. 2015. A review of lithology interpretation methods using geophysical well logs. Progress in Geophysics (in Chinese), 30(3): 1278-1287, doi: 10.6038/pg20150338.
曹茜, 王志章, 樊太亮, 等. 2018. 基于FSVM法分岩性解释砂砾岩储层渗透率. 新疆大学学报(自然科学版), 35(3): 264-271. https://www.cnki.com.cn/Article/CJFDTOTAL-XJDZ201803004.htm
丁世飞, 齐丙娟, 谭红艳. 2011. 支持向量机理论与算法研究综述. 电子科技大学学报, 40(1): 2-10. https://www.cnki.com.cn/Article/CJFDTOTAL-DKDX201101003.htm
范昕炜. 2003. 支持向量机算法的研究及其应用[博士论文]. 杭州: 浙江大学.
付光明, 严加永, 张昆, 等. 2017. 岩性识别技术现状与进展. 地球物理学进展, 32(1): 26-40, doi: 10.6038/pg20170104. http://www.progeophys.cn/article/doi/10.6038/pg20170104
韩学辉, 支乐菲, 刘荣, 等. 2013. 应用最小二乘支持向量机识别广利油田沙四段储层岩性. 地球物理学进展, 28(4): 1886-1892, doi: 10.6038/pg20130430. http://www.progeophys.cn/article/doi/10.6038/pg20130430
孔祥宇, 殷进垠, 张发强. 2007. 哈萨克斯坦南图尔盖盆地油气地质特征及勘探潜力分析. 岩性油气藏, 19(3): 48-53. doi: 10.3969/j.issn.1673-8926.2007.03.009
李佳丽, 刘光萍. 2016. 基于PCA-SVM算法在岩性识别中的应用. 江西科学, 34(4): 504-510, doi: 10.13990/j.issn1001-3679.2016.04.020.
李盼池, 许少华. 2005. 支持向量机在模式识别中的核函数特性分析. 计算机工程与设计, 26(2): 302-304. doi: 10.3969/j.issn
李雄炎, 周金昱, 李洪奇, 等. 2012. 复杂岩性及多相流体智能识别方法. 石油勘探与开发, 39(2): 243-248. https://www.cnki.com.cn/Article/CJFDTOTAL-SKYK201202017.htm
林香亮, 袁瑞, 孙玉秋, 等. 2018. 支持向量机的基本理论和研究进展. 长江大学学报(自科版), 15(17): 48-53, doi: 10.16772/j.cnki.1673-1409.2018.17.010.
刘伟. 2009. 复杂低渗透油气储层测井评价[博士论文]. 青岛: 中国石油大学.
牟丹, 王祝文, 黄玉龙, 等. 2015. 基于SVM测井数据的火山岩岩性识别——以辽河盆地东部坳陷为例. 地球物理学报, 58(5): 1785-1793, doi: 10.6038/cjg20150528.
邵信光, 杨慧中, 陈刚. 2006. 基于粒子群优化算法的支持向量机参数选择及其应用. 控制理论与应用, 23(5): 740-743, 748. doi: 10.3969/j.issn.1000-8152.2006.05.014
宋延杰, 张剑风, 闫伟林, 等. 2007. 基于支持向量机的复杂岩性测井识别方法. 大庆石油学院学报, 31(5): 18-20, 46. https://www.cnki.com.cn/Article/CJFDTOTAL-DQSY200705008.htm
苏赋, 马磊, 罗仁泽, 等. 2020. 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. 地球物理学进展, 35(01): 174-180, doi: 10.6038/pg2020DD0185. http://www.progeophys.cn/article/doi/10.6038/pg2020DD0185
王学渊, 陈志刚, 钟新荣, 等. 2018. 基于粒子群优化的SVM供水管道泄漏诊断方法. 现代电子技术, 41(7): 156-159, 164. https://www.cnki.com.cn/Article/CJFDTOTAL-XDDJ201807037.htm
王哲涛, 宋小全. 2018. 基于支持向量机的星载SAR信号识别技术研究. 现代雷达, 40(11): 31-36. https://www.cnki.com.cn/Article/CJFDTOTAL-XDLD201811007.htm
解敬. 2015. 基于ACO优化SVM在测井岩性识别中的研究[硕士论文]. 大庆: 东北石油大学.
杨福强, 陈科贵, 黄长兵, 等. 2019. PSO-LIBSVM在钾盐矿层识别中的应用研究. 地球科学进展, 4(7): 757-764. https://www.cnki.com.cn/Article/CJFDTOTAL-DXJZ201907012.htm
尹嘉鹏. 2016. 支持向量机核函数及关键参数选择研究[硕士论文]. 哈尔滨: 哈尔滨工业大学.
张昭杰, 方石. 2019. 基于遗传算法优化的支持向量机在岩性识别中的应用. 世界地质, 38(2): 486-491. doi: 10.3969/j.issn.1004-5589.2019.02.019
赵显令, 王贵文, 周正龙, 等. 2015. 地球物理测井岩性解释方法综述. 地球物理学进展, 30(3): 1278-1287, doi: 10.6038/pg20150338. http://www.progeophys.cn/article/doi/10.6038/pg20150338
-
图(9)
表(4)
计量
- 文章访问数: 2723
- PDF下载数: 72
- 施引文献: 0