利用混合模型CRBM-PSO-XGBoost识别致密砂岩储层岩性

doi:10.11743/ogg20210518

摘要/Abstract

摘要：

识别储层岩性是开展地层对比和沉积展布等地质基础研究工作的重要前提。致密砂岩储层多为砂、泥岩薄互层，形成的岩性不仅种类较多，而且大部分岩性的测井响应也较为相似，为相关的岩性识别工作带来了困难。基于机器学习技术在模式识别上具有强大的分析性能，选用了稳定性好且计算效率高的XGBoost模型来解决致密砂岩储层岩性识别问题。该模型在建模过程中需要较多的经验参数参与，且计算速度随着自变量的增加而逐渐降低，为此提出了采用CRBM模型和PSO模型对其进行改进。CRBM模型具备数据提取功能，可从源数据中挖掘出更少但更利于建模的新自变量，而PSO模型可通过迭代计算确定XGBoost所有经验参数的最优值。以姬塬油田西部长4+5段部分取心井资料为基础，通过设计两个实验来验证提出的混合模型的预测能力。为加强验证效果，在实验中加入了PNN和SVM的优化模型进行对比。实验后发现所提出模型的预测准确率最高，均在90%以上。实验结果表明所提出的模型不仅能有效解决致密砂岩储层岩性识别问题，且较以往经典预测模型更具推广应用性。

关键词: XGBoost模型, CRBM模型, PSO模型, 岩性识别, 致密砂岩储层, 姬塬油田, 鄂尔多斯盆地

Abstract:

Lithology identification is universally regarded as a prerequisite for some fundamental geological work such as stratigraphic correlation and analysis of sedimentary system distribution. The tight sandstone reservoirs are generally composed of multiple thin sandstone-mud alternating beds, featuring a variety of lithological types, most of which are similar in log responses, posing great difficulty for lithology identification. Given the powerful analytical performance of machine learning techniques shown on pattern recognition, XGBoost, a stable and highly efficient model is adopted to predict lithology of the tight sandstone reservoirs. However, this model needs many empirical parameters to complete its prediction, and its computing efficiency will be reduced when it deals with more independent variables. The CRBM and PSO models are, therefore, introduced to improve the prediction performance. CRBM is functional to extract features from original data so that fewer but more significant independent variables can be created by the CRBM in modeling. While PSO is effective to determine the optimal values via iterative computation for those empirical parameters used by XGBoost. Based on the data derived from some cored wells of Chang 4+5 Members in western Jiyuan oilfield, two experiments are designed to validate the prediction performance of the hybrid model proposed, CRBM-PSO-XGBoost. Furthermore, two optimized models formed by PNN and SVM are selected to make a comparison, as a way to consolidate the validation effect. The experiment results manifest that the prediction accuracy comes to peak by the proposed model, up to over 90%. In all, the hybrid model proposed in the study, capable of identifying lithology of tight sandstone reservoirs, is of wider application to lithology prediction compared with previous models.

Key words: XGBoost model, CRBM model, PSO model, lithology identification, tight sandstone reservoir, Jiyuan oilfield, Ordos Basin

中图分类号:

TE19

谷宇峰,张道勇,鲍志东. 利用混合模型CRBM-PSO-XGBoost识别致密砂岩储层岩性[J]. 石油与天然气地质, 2021, 42(5): 1210-1222. doi: 10.11743/ogg20210518 Yufeng Gu,Daoyong Zhang,Zhidong Bao. Lithology identification in tight sandstone reservoirs using CRBM-PSO-XGBoost[J]. Oil & Gas Geology, 2021, 42(5): 1210-1222. doi: 10.11743/ogg20210518

图/表 10

图1

图2

图3

表1

表2

图4

图5

图6

表3

图7

参考文献 29

1	孔强夫, 杨才, 李浩, 等. 基于图论聚类和最小临近算法的岩性识别方法[J]. 石油与天然气地质, 2020, 41 (4): 884- 890.
	Kong Qiangfu , Yang Cai , Li Hao , et al. A lithology recognition method based on multi-resolution graph-based clustering and K-nearest neighbor: A case study from the Leikoupo formation carbonate reservoirs in western Sichuan Basin[J]. Oil & Gas Geology, 2020, 41 (4): 884- 890.
2	张天付, 黄理力, 倪新锋, 等. 塔里木盆地柯坪地区下寒武统吾松格尔组岩性组合及其成因与勘探意义[J]. 石油与天然气地质, 2020, 41 (5): 928- 940.
	Zhang Tianfu , Huang Lili , Ni Xinfeng , et al. Lithological combination, genesis and exploration significance of the lower Cambrain Wusonggeer formation of Kalpin area in Tarim Basin: Insight through the deepest Asian onshore well-Well Luntan 1[J]. Oil & Gas Geology, 2020, 41 (5): 928- 940.
3	黄彦庆, 刘忠群, 林恬, 等. 川东北元坝地区徐家河组三段基于相控的相对优质储层预测[J]. 石油与天然气地质, 2021, 42 (4): 863- 872.
	Huang Yanqing , Liu Zhongqun , Lin Tian , et al. Lithofacies-based prediction of relatively high-quality reservoirs of the Xu 3 member in Yuanba area, northeastern Sichuan Basin[J]. Oil & Gas Geology, 2021, 42 (4): 863- 872.
4	郑见中. 辽河坳陷兴隆台潜山带中生界储层测井评价方法[J]. 石油地质与工程, 2019, 33 (6): 29- 33. doi: 10.3969/j.issn.1673-8217.2019.06.007
	Zheng Jianzhong . Logging evaluation method of Mesozoic reservoir in Xinlongtai buried hill belt of Liaohe depression[J]. Petroleum Geology and Engineering, 2019, 33 (6): 29- 33. doi: 10.3969/j.issn.1673-8217.2019.06.007
5	闫林, 冉启全, 高阳, 等. 新疆芦苇沟组致密油赋存形式及可动用性评价[J]. 油气藏评价与开发, 2017, 7 (6): 20- 26. doi: 10.3969/j.issn.2095-1426.2017.06.004
	Ran Lin , Ran Qiquan , Gao Yang , et al. Tight oil occurence form and recoverability evaluation of tight oil reservoir in Lucaogou Formation of Xinjiang[J]. Reservoir Evaluation and Development, 2017, 7 (6): 20- 26. doi: 10.3969/j.issn.2095-1426.2017.06.004
6	袁述武, 李想, 史乐, 等. 克拉玛依油田六、七、九区石炭系内幕有利火山岩储层岩性分布预测[J]. 大庆石油地质与开发, 2019, 38 (6): 40- 45.
	Yuan Shuwu , Li Xiang , Shi Le , et al. Prediction of the lithology distribution of favorable volcanic reservoirs in the inside Carboniferous system in Block 6, 7 and 9 of Karamay Oilfield[J]. Petroleum Geology & Oilfield Development in Daqing, 2019, 38 (6): 40- 45.
7	赵彤彤, 张春雷, 张春雨, 等. 基于模糊熵的KNN分类模型在岩性识别中的应用[J]. 计算机工程与应用, 2018, 54 (24): 260- 265. doi: 10.3778/j.issn.1002-8331.1709-0084
	Zhao Tongtong , Zhang Chunlei , Zhang Chunyu , et al. Application of KNN classification model based on fuzzy entropy in lithology recognition[J]. Computer Engineering and Applications, 2018, 54 (24): 260- 265. doi: 10.3778/j.issn.1002-8331.1709-0084
8	张梓童, 张春雷, 张艳, 等. 数据空间结构性及在KNN算法中的应用[J]. 数学的实践与认知, 2019, 49 (1): 195- 202.
	Zhang Zitong , Zhang Chunlei , Zhang Yan , et al. Data space structure and application in KNN algorithm[J]. Mathematics in Practice and Theory, 2019, 49 (1): 195- 202.
9	陈玉林, 李戈理, 杨智新, 等. 基于KNN算法识别合水地区长7储层岩性岩相[J]. 测井技术, 2020, 44 (2): 182- 185.
	Chen Yulin , Li Geli , Yang Zhixin , et al. Identification of lithology and lithofacies of Chang 7 reservoir in Heshui area by KNN algorithm[J]. Well Logging Technology, 2020, 44 (2): 182- 185.
10	赵杰, 李春华. 基于神经网络的两种岩性识别方法的研究[J]. 现代电子技术, 2009, 32 (23): 136- 138. doi: 10.3969/j.issn.1004-373X.2009.23.043
	Zhao Jie , Li Chunhua . Research on two lithology identification methods based on neural network[J]. Modern Electronics Technique, 2009, 32 (23): 136- 139. doi: 10.3969/j.issn.1004-373X.2009.23.043
11	陈刚. PNN在煤田随钻测井岩性判识应用研究[J]. 地质与资源, 2018, 27 (1): 103- 106. doi: 10.3969/j.issn.1671-1947.2018.01.015
	Chen Gang . Application of PNN in the lithology identification of logging while drilling in coal field[J]. Geology and Resources, 2018, 27 (1): 103- 106. doi: 10.3969/j.issn.1671-1947.2018.01.015
12	陈刚, 汪凯斌, 薛必辞, 等. 随钻测井中岩性识别方法的对比及应用[J]. 煤田地质与勘探, 2018, 46 (1): 165- 169. doi: 10.3969/j.issn.1001-1986.2018.01.028
	Chen Gang , Wang Kaibin , Xue Bici , et al. Comparison and application of LWD lithology identification method[J]. Coal Geology & Exploration, 2018, 46 (1): 165- 169. doi: 10.3969/j.issn.1001-1986.2018.01.028
13	李政宏, 刘永福, 张立强, 等. 数据挖掘方法在测井岩性识别中的应用[J]. 断块油气藏, 2019, 26 (6): 713- 718.
	Li Zhenghong , Liu Yongfu , Zhang Liqiang , et al. Application of data mining method in lithology identification using well log[J]. Fault-block Oil & Gas Field, 2019, 26 (6): 713- 718.
14	林香亮, 朱建伟, 刘光涛, 等. 基于PCA-SVM的砂砾岩岩性识别[J]. 长江大学学报(自然科学版), 2020, 17 (1): 21- 26. doi: 10.3969/j.issn.1673-1409.2020.01.005
	Lin Xiangliang , Zhu Jianwei , Liu Guangtao , et al. Lithologic identification of glutenite based on PCA-SVM[J]. Journal of Yangtze University (Natural Science Edition), 2020, 17 (1): 21- 26. doi: 10.3969/j.issn.1673-1409.2020.01.005
15	石锁, 余继峰, 曹慧涛, 等. 基于高斯核SVM的储层岩性识别——以东濮凹陷上古生界碎屑岩为例[J]. 中国科技论文, 2020, 15 (1): 112- 118. doi: 10.3969/j.issn.2095-2783.2020.01.017
	Shi Suo , Yu Jifeng , Cao Huitao , et al. Reservoir lithology identification based on SVM using radial basis function: an example of Upper Paleozoic clastic rocks in Dongpu sag[J]. China Science Paper, 2020, 15 (1): 112- 118. doi: 10.3969/j.issn.2095-2783.2020.01.017
16	Chen T, Guestrin C. XGboost: A scalable tree boosting system[C]//ACM SIGKDD International Conference on Konwledge Discovery and Data Mining, 2016: 785-794.
17	Zhou K , Zhang J , Ren Y , et al. A gradient boosting decision tree algorithm combining synthetic minority over-sampling technique for lithology identification[J]. Geophysics, 2020, 85, 1- 52. doi: 10.1190/geo2020-0711-fe.1
18	Dev V A , Eden M R . Formation lithology classification using scalable gradient boosted decision trees[J]. Computers & Chemical Enginee-ring, 2019, 128, 392- 404.
19	闫星宇, 顾汉明, 肖逸飞, 等. XGBoost算法在致密砂岩气层测井解释中的应用[J]. 石油地球物理勘探, 2019, 54 (2): 447- 455.
	Yan Xingyu , Gu Hanming , Xiao Yifei , et al. XGBoost algorithm applied in the interpretation of tight-sand gas reservoir on well logging data[J]. Oil Geophysical Prospecting, 2019, 54 (2): 447- 455.
20	杨维, 李歧强. 粒子群优化算法综述[J]. 中国工程科学, 2004, 6 (5): 87- 94. doi: 10.3969/j.issn.1009-1742.2004.05.018
	Yang Wei , Li Qiqiang . Survey on particle swarm optimization algorithm[J]. Engineering Science, 2004, 6 (5): 87- 94. doi: 10.3969/j.issn.1009-1742.2004.05.018
21	刘建华. 粒子群算法的基本理论及其改进研究[D]. 长沙: 中南大学, 2009.
	Liu Jianhua. The research of basic theory and improvement on particle swarm optimization[D]. Changsha: Center South University, 2009.
22	温阳东, 李龙剑. 基于LDIW-PSO算法的BP神经网络在压力传感器中的应用[J]. 化工自动化及仪表, 2014, 41 (9): 1031- 1034. doi: 10.3969/j.issn.1000-3932.2014.09.011
	Wen Yangdong , Li Longjian . Application of LDIW-PSO algorithm-based BP neural network in pressure sensor[J]. Control and Instruments in Chemical Industry, 2014, 41 (9): 1031- 1034. doi: 10.3969/j.issn.1000-3932.2014.09.011
23	Tang T B , Murray A F . Adaptive sensor modeling and classification using a continuous restricted Boltzmann machine (CRBM)[J]. Neurocomputing, 2006, 70 (7-9): 1198- 1206.
24	Huang H B , Li R X , Yang M L , et al. Evaluation of vehicle interior sound quality using a continuous restricted Boltzmann machine-based DBN[J]. Mechanical Systems & Signal Processing, 2017, 84, 245- 267.
25	Chen Y , Lu L , Li X . Application of continuous restricted Boltzmann machine to identify multivariate geochemical anomaly[J]. Journal of Geochemical Exploration, 2014, 140 (4): 56- 63.
26	黎盼. 低渗透砂岩储层微观孔隙结构表征及生产特征分析[D]. 西安: 西北大学, 2019.
	Li Pan. Microscopic pore structure characterization and production characteristic analysis of low-permeability sandstone reservoir: a case study on Chang 4+5 and Chang 6 reservoir of T area in Jiyuan Oilfield[D]. Xi'an: Northwest University, 2019.
27	赵冰瑶. 姬塬油田王盘山地区长4+5储层成岩相微观孔隙结构及渗流特征研究[D]. 西安: 西北大学, 2018.
	Zhao Bingyao. The study on microscopic pore structure and percolation characteristics of diagenetic facies in Chang 4+5 reservoir of Wangpanshan area, Jiyuan Oilfield[D]. Xi'an: Northwest University, 2018.
28	翟利华. 姬塬油田西部长4+5油层组沉积微相与单砂体研究[D]. 北京: 中国石油大学(北京), 2018.
	Zhai Lihua. Study on sedimentary microfacies and single sandbodies on Chang 4+5 member in the western Jiyuan Oilfield[D]. Beijing: China University of Petroleum(Beijing), 2018.
29	沈佳男. 鄂尔多斯盆地姬塬油田西部长4+5油层组储层评价[D]. 北京: 中国石油大学(北京), 2018.
	Shen Jianan. Reservoir evaluation of Chang 4+5 member in west Jiyuan Oilfield in Ordos Basin[D]. Beijing: China University of Petroleum(Beijing), 2018.

	PNN	SVM	XGBoost
初始参数(是否需要优化)	α=X(0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1)(是)	c=1 (是) 核函数=RBF(否) gamma=1 (是)	D=100 (是) η=0.01 (是) max_depth=3 (是) λ=0.1 (是) min_chile_weight=0.1 (是) δ=0 (是) 损失函数=交叉熵损失函数(否)
PSO参数设定	q=10 t=100 ω_max，ω_min=0.9, 0.4 c₁, c₂=1.5 r₁和r₂，范围[0, 1]
	各优化参数的σ_{max_1}和σ_{max_2}
	α (10^-2, 2×10⁰)	c (10^-3, 10³) gamma (10^-3, 10³)	D (10²0, 10⁴) η (10^-3, 10⁰) max_depth (3×10⁰, 10¹) λ (10⁰, 10²) min_chile_weight (10⁰, 10²) δ (10⁰, 10¹)
	各优化参数的W_max
	α (0.1)	c (10) gamma (10)	D(50) η(0.1) max_depth(5) λ(10) min_chile_weight(10) δ(5)
优化后参数	X=(0.85, 0.83, 0.91, 0.79, 0.98, 0.89, 1.01, 0.87)	c=15 gamma=2.5	D=730 η=0.2 max_depth=5 λ=1.3 min_chile_weight=1.2 δ=0

参数	值
最大迭代次数(K)	100
迭代精度(e)	0.001
可见层神经元个数(n_v)	11
隐含层神经元个数(n_h)	6
动量系数(ξ)	0.9
学习速率(η)	0.1
噪音控制参数(μ)	0.05
噪音方差(σ)	0.2
S函数上渐近线(ϕ_h)	1
S函数下渐近线(ϕ_l)	0
mini-batch学习样本容量(χ)	100

预测模型	实验1		实验2
预测模型	HA井	HB井	HA井	HB井
CRBM-PSO-PNN	79.00/50.269 2	75.33/50.269 1	83.00/120.348 5	80.67/120.348 6
CRBM-PSO-SVM	83.67/25.487 1	77.00/25.487 2	88.33/43.247 3	84.33/43.247 2
CRBM-PSO-XGBoost	92.67/35.983 7	90.33/35.983 6	96.33/54.453 9	93.67/54.453 7

[1]	吴伟涛, 冯炎松, 费世祥, 王一妃, 吴和源, 杨旭东. 鄂尔多斯盆地神木气田二叠系石千峰组5段致密气富集因素及有利区预测[J]. 石油与天然气地质, 2024, 45(3): 739-751.
[2]	刘成林, 丁振刚, 范立勇, 康锐, 洪思婕, 朱玉新, 陈践发, 王海东, 许诺. 鄂尔多斯盆地含氦天然气地球化学特征与富集影响因素[J]. 石油与天然气地质, 2024, 45(2): 384-392.
[3]	万俊雨, 朱建辉, 姚素平, 张毅, 李春堂, 张威, 姜海健, 王杰. 鄂尔多斯盆地中、东部奥陶系马家沟组成烃生物及烃源岩地球生物学评价[J]. 石油与天然气地质, 2024, 45(2): 393-405.
[4]	杨丽华, 刘池洋, 黄雷, 周义军, 刘永涛, 秦阳. 鄂尔多斯盆地古峰庄地区疑似侵入岩体的发现及其地质意义[J]. 石油与天然气地质, 2024, 45(1): 142-156.
[5]	师良, 范柏江, 李忠厚, 余紫巍, 蔺子瑾, 戴欣洋. 鄂尔多斯盆地中部三叠系延长组7段烃组分的运移分异作用[J]. 石油与天然气地质, 2024, 45(1): 157-168.
[6]	曹江骏, 王继平, 张道锋, 王龙, 李笑天, 李娅, 张园园, 夏辉, 于占海. 深层致密砂岩储层成岩演化对含气性的影响[J]. 石油与天然气地质, 2024, 45(1): 169-184.
[7]	胡宗全, 王濡岳, 路菁, 冯动军, 刘粤蛟, 申宝剑, 刘忠宝, 王冠平, 何建华. 陆相页岩及其夹层储集特征对比与差异演化模式[J]. 石油与天然气地质, 2023, 44(6): 1393-1404.
[8]	刘成林, 丁振刚, 陈践发, 范立勇, 康锐, 王海东, 洪思婕, 田安琦, 陈学勇. 鄂尔多斯盆地氦源岩特征及生氦潜力[J]. 石油与天然气地质, 2023, 44(6): 1546-1554.
[9]	李勇, 朱治同, 吴鹏, 申陈州, 高计县. 鄂尔多斯盆地东缘上古生界致密储层含气系统压力演化[J]. 石油与天然气地质, 2023, 44(6): 1568-1581.
[10]	曾溅辉, 张亚雄, 张在振, 乔俊程, 王茂云, 陈冬霞, 姚泾利, 丁景辰, 熊亮, 刘亚洲, 赵伟波, 任克博. 致密砂岩气藏复杂气-水关系形成和分布主控因素及分布模式[J]. 石油与天然气地质, 2023, 44(5): 1067-1083.
[11]	梁岳立, 赵晓明, 张喜, 李树新, 葛家旺, 聂志宏, 张廷山, 祝海华. 轨道周期约束下海-陆过渡相页岩层系高精度层序界面识别及其地质意义[J]. 石油与天然气地质, 2023, 44(5): 1231-1242.
[12]	李涵, 付金华, 季汉成, 张雷, 佘钰蔚, 官伟, 井向辉, 王红伟, 曹茜, 刘刚, 魏嘉怡. 鄂尔多斯盆地西南部上古生界风化壳型铝土岩系发育过程及优势储层分布规律[J]. 石油与天然气地质, 2023, 44(5): 1243-1255.
[13]	李晓, 郭鹏, 胡彦智, 李士祥, 杨伟伟. 陆相页岩压裂试验与数值模拟[J]. 石油与天然气地质, 2023, 44(4): 1009-1019.
[14]	高嘉洪, 金之钧, 梁新平, 李士祥, 杨伟伟, 朱如凯, 杜晓宇, 刘全有, 李彤, 董琳, 李鹏, 张旺. 火山活动对鄂尔多斯盆地三叠系长7段淡水湖盆富营养化与沉积水体介质环境的影响[J]. 石油与天然气地质, 2023, 44(4): 887-898.
[15]	王梓毅, 付金华, 刘显阳, 李士祥, 张昌虎, 梁新平, 董琳. 鄂尔多斯盆地上三叠统延长组7段埋藏期热液活动对页岩油储层的影响[J]. 石油与天然气地质, 2023, 44(4): 899-909.