Random Forest-Based Quantitative Evaluation Method for Reservoir Heterogeneity: A Case Study of Ngs4 and Ngs5 Reservoirs of Upper Guantao Formation in Chengbei 84 Well Area

doi:10.7657/XJPG20260414

Abstract

Abstract:

The sand groups 4 and 5 of the upper Guantao formation (Ngs4 and Ngs5) in the Chengdao oilfield, Bohai Bay Basin, represent meandering river deposits, and exhibit rapid reservoir sedimentary facies transition. Reservoirs in the upper Guantao formation are found with pronounced interlayer and planar heterogeneities. Based on two dimensions of reservoir quality and spatial distribution, four evaluation index systems comprising basic parameters, intralayer heterogeneity parameters, interlayer heterogeneity parameters, and planar heterogeneity parameters respectively were established. Employing the Random Forest algorithm, the importance of each parameter to heterogeneity was quantified. Subsequently, basic composite index and intralayer/interlayer/planar heterogeneity composite indexes were calculated, a formula for calculating reservoir heterogeneity composite index was designed, and the reservoir heterogeneity composite index was yielded. The results indicate that a higher heterogeneity composite index reflects a weaker reservoir heterogeneity.

Key words: Chengdao oilfield, Guantao formation, reservoir heterogeneity, quantitative evaluation, Random Forest, sedimentary microfacies

CLC Number:

TE19

SUN Tingbin, WANG Ling, WANG Kaipeng, LI Ensheng, LIU Changhui, WANG Zichuan, ZHAO Rongqi. Random Forest-Based Quantitative Evaluation Method for Reservoir Heterogeneity: A Case Study of Ngs4 and Ngs5 Reservoirs of Upper Guantao Formation in Chengbei 84 Well Area[J]. Xinjiang Petroleum Geology, 2026, 47(4): 496-505.

Figures/Tables 15

Fig. 1.

Fig. 2.

Fig. 3.

Table 1.

Table 2.

Table 3.

Fig. 4.

Table 4.

Fig. 5.

Table 5.

Fig. 6.

Fig. 7.

Fig. 8.

Fig. 9.

Fig. 10.

References 32

[1]	杨少春. 储层非均质性定量研究的新方法[J]. 石油大学学报(自然科学版), 2000, 24(1):53-56.
	YANG Shaochun. A new method for quantitative study of reservoir heterogeneiy[J]. Journal of the University of Petroleum,China(Edition of Natural Science), 2000, 24(1):53-56.
[2]	高树新, 杨少春, 胡红波, 等. 胜坨油田坨21断块沙二段储层非均质性定量表征[J]. 油气地质与采收率, 2004, 11(5):10-13.
	GAO Shuxin, YANG Shaochun, HU Hongbo, et al. Quantitative characterization on reservoir heterogeneity of the second member of the Shahejie formation in Tuo21 fault block,Shengtuo oilfield[J]. Petroleum Geology and Recovery Efficiency, 2004, 11(5):10-13.
[3]	岳大力, 林承焰, 吴胜和, 等. 储层非均质定量表征方法在礁灰岩油田开发中的应用[J]. 石油学报, 2004, 25(5):75-79. doi: 10.7623/syxb200405015
	YUE Dali, LIN Chengyan, WU Shenghe, et al. Application of quantitative method for characterizing reservoir heterogeneity to the development of reef limestone reservoir[J]. Acta Petrolei Sinica, 2004, 25(5):75-79. doi: 10.7623/syxb200405015
[4]	梁力文, 旷理雄, 敬小军, 等. 用非均质性综合指数研究大路沟A区储层[J]. 新疆石油地质, 2013, 34(2):215-218.
	LIANG Liwen, KUANG Lixiong, JING Xiaojun, et al. Application of heterogeneity index to study of Block-A reservoir in Dalugou area of Jing’an oilfield,Ordos Basin[J]. Xinjiang Petroleum Geology, 2013, 34(2):215-218.
[5]	张航, 李治平, 赖枫鹏, 等. 利用动态生产数据识别储层非均质性[J]. 科学技术与工程, 2015, 15(5):76-80.
	ZHANG Hang, LI Zhiping, LAI Fengpeng, et al. The use of dynamic production data to determine the reservoir heterogeneity[J]. Science Technology and Engineering, 2015, 15(5):76-80.
[6]	王喜鑫, 侯加根, 刘钰铭, 等. 基于层次分析与模糊数学的河口坝非均质性定量表征:以王官屯油田官195断块为例[J]. 天然气地球科学, 2017, 28(12):1914-1924. doi: 10.11764/j.issn.1672-1926.2017.11.002
	WANG Xixin, HOU Jiagen, LIU Yuming, et al. Quantitative characterization of vertical heterogeneity of mouth bar based on the analytic hierarchy process and the fuzzy mathematics:A case study of Guan195 fault block of Wangguantun oilfield[J]. Natural Gas Geoscience, 2017, 28(12):1914-1924. doi: 10.11764/j.issn.1672-1926.2017.11.002
[7]	王长胜, 石玉江, 周金昱, 等. 致密油储层砂体宏观非均质性测井定量表征[J]. 测井技术, 2020, 44(2):204-208.
	WANG Changsheng, SHI Yujiang, ZHOU Jinyu, et al. Logging characterization of macroscopic heterogeneity of tight oil reservoirs[J]. Well Logging Technology, 2020, 44(2):204-208.
[8]	周志杰, 张小莉, 杨振, 等. 基于优化熵权的储层非均质性定量表征新方法研究:以鄂尔多斯盆地志丹油区F井区长61 1储层为例[J]. 地球物理学进展, 2023, 38(4):1713-1721.
	ZHOU Zhijie, ZHANG Xiaoli, YANG Zhen, et al. Study on a new method for quantitative characterization of reservoir heterogeneity based on optimized entropy weight:An example of the Chang 61 1 reservoir in the F well area of Zhidan oifield in Ordos Basin[J]. Progress in Geophysics, 2023, 38(4):1713-1721.
[9]	KADRI A, BENZAGOUTA M S. Heterogeneous oil reservoirs characterization using artificial intelligence techniques:Application to the Hassi Messaoud oil field in the Algerian-Saharan platform[J]. Journal of Applied Geophysics, 2025, 242:105878. doi: 10.1016/j.jappgeo.2025.105878
[10]	肖寿恒. KL6-C区块古近系碎屑岩储层非均质性研究[D]. 北京: 中国石油大学(北京), 2017.
	XIAO Shouheng. Study on the heterogeneity of clastic reservoir of Paleogene formation of KL6-C block[D]. Beijing: China University of Petroleum(Beijing), 2017.
[11]	严科, 杨少春, 任怀强. 储层宏观非均质性定量表征研究[J]. 石油学报, 2008, 29(6):870-874. doi: 10.7623/syxb200806015
	YAN Ke, YANG Shaochun, REN Huaiqiang. Research on quantitative characterization of macroscopic heterogeneity of reservoir[J]. Acta Petrolei Sinica, 2008, 29(6):870-874. doi: 10.7623/syxb200806015
[12]	于翠玲, 林承焰. 储层非均质性研究进展[J]. 油气地质与采收率, 2007, 14(4):15-18.
	YU Cuiling, LIN Chengyan. Advancement of reservoir heterogeneity research[J]. Petroleum Geology and Recovery Efficiency, 2007, 14(4):15-18.
[13]	张海娜, 季雅新, 许彦群, 等. 海上埕岛油田河流相砂体内部结构研究[J]. 石油工程建设, 2021, 47(增刊2):156-162.
	ZHANG Haina, JI Yaxin, XU Yanqun, et al. Internal structure of fluvial facies sand body in offshore Chengdao oilfield[J]. Petroleum Engineering Construction, 2021, 47(Supp.2):156-162.
[14]	蒋思华. 埕岛油田馆上段河道砂储层连通关系研究[D]. 北京: 中国石油大学(北京), 2019.
	JIANG Sihua. Study on the connectivity of channel sand reservoirs in the upper Guantao formation of the Chengdao oilfield[D]. Beijing: China University of Petroleum(Beijing), 2019.
[15]	杨鹏飞, 张磊, 李大伟. 埕岛油田馆上段储层非均质性研究[J]. 西南石油学院学报, 2003, 25(5):19-22.
	YANG Pengfei, ZHANG Lei, LI Dawei. The study of the reservoir anisotropy in the upper Guantao formation in Chengdao oil field[J]. Journal of Southwest Petroleum Institute, 2003, 25(5):19-22.
[16]	姚帮. 埕岛北部断裂带古近系东营组沉积体系特征研究[D]. 山东青岛: 中国石油大学(华东), 2018.
	YAO Bang. Sedimentary system of the Paleogene Dongying formation in the northern fault zone of Chengdao[D]. Qingdao, Shandong: China University of Petroleum (East China), 2018.
[17]	郑和荣, 林会喜, 王永诗. 埕岛油田勘探实践与认识[J]. 石油勘探与开发, 2000, 27(6):1-3.
	ZHENG Herong, LIN Huixi, WANG Yongshi. Practice and knowledge of exploration on Chengdao oil field[J]. Petroleum Exploration and Development, 2000, 27(6):1-3.
[18]	常文会, 赵永刚, 卢松. 曲流河环境沉积微相和测井相特征分析[J]. 天然气工业, 2010, 30(2):48-51.
	CHANG Wenhui, ZHAO Yonggang, LU Song. Features of sedimentary microfacies and electrofacies of meandering river deposits[J]. Natural Gas Industry, 2010, 30(2):48-51.
[19]	刘伟. 曲流河砂体流动单元空间分布规律研究:以埕岛油田馆陶组为例[J]. 断块油气田, 2005, 12(3):28-31.
	LIU Wei. Study on dimensional distribution of flow units in meandering sandstone reservoir:Talking about Guantao formation in Chengdao oilfield as an example[J]. Fault-Block Oil & Gas Field, 2005, 12(3):28-31.
[20]	马志欣, 李进步, 付斌, 等. 苏里格气田二叠系山西组曲流河储层岩相与构型[J]. 新疆石油地质, 2025, 46(3):280-287.
	MA Zhixin, LI Jinbu, FU Bin, et al. Lithofacies and architecture of meandering river reservoirs in Permian Shanxi formation,Sulige gas field[J]. Xinjiang Petroleum Geology, 2025, 46(3):280-287.
[21]	郭子熙, 马骉, 张帅, 等. 深度学习在油气产量预测中的研究进展与技术展望[J]. 天然气工业, 2024, 44(9):88-98.
	GUO Zixi, MA Biao, ZHANG Shuai, et al. Research status and prospects of deep learning in oil and gas production prediction[J]. Natural Gas Industry, 2024, 44(9):88-98.
[22]	匡立春, 刘合, 任义丽, 等. 人工智能在石油勘探开发领域的应用现状与发展趋势[J]. 石油勘探与开发, 2021, 48(1):1-11. doi: 10.11698/PED.2021.01.01
	KUANG Lichun, LIU He, REN Yili, et al. Application and development trend of artificial intelligence in petroleum exploration and development[J]. Petroleum Exploration and Development, 2021, 48(1):1-11. doi: 10.1016/S1876-3804(21)60001-0
[23]	姜世一, 孙盼科, 张林, 等. 基于径向基—多层感知器神经网络联合的复杂岩相智能识别与表征[J]. 天然气工业, 2022, 42(9):47-62.
	JIANG Shiyi, SUN Panke, ZHANG Lin, et al. Intelligent identification and characterization of complex lithofacies based on radial basis-multilayer perception neural network joint model[J]. Natural Gas Industry, 2022, 42(9):47-62.
[24]	郑晓梅, 何桂平, 王子龙. 基于决策树对致密储层流体识别的方法研究:以盐池地区长8储层为例[J]. 北京石油化工学院学报, 2023, 31(3):21-27.
	ZHENG Xiaomei, HE Guiping, WANG Zilong. Study on fluid identification method of tight reservoir:A case study of Chang 8 reservoir in Yanchi area[J]. Journal of Beijing Institute of Petrochemical Technology, 2023, 31(3):21-27.
[25]	鲁春华, 姜汉桥, 李杰, 等. 采用改进的FCM聚类算法进行储层优势通道分级[J]. 石油科学通报, 2020, 5(4):520-530.
	LU Chunhua, JIANG Hanqiao, LI Jie, et al. A classification method for reservoir thief zones based on an improved FCM clustering algorithm[J]. Petroleum Science Bulletin, 2020, 5(4):520-530.
[26]	杜昕, 范廷恩, 董建华, 等. 基于多层感知机网络的薄储层预测[J]. 石油地球物理勘探, 2020, 55(6):1178-1187. doi: 10.13810/j.cnki.issn.1000-7210.2020.06.002
	DU Xin, FAN Ting’en, DONG Jianhua, et al. Characterization of thin sand reservoirs based on a multi-layer perception deep neural network[J]. Oil Geophysical Prospecting, 2020, 55(6):1178-1187. doi: 10.13810/j.cnki.issn.1000-7210.2020.06.002
[27]	干磊, 何东博, 郭建林, 等. 机器学习方法在储层分类中的应用[J]. 数学的实践与认识, 2019, 49(13):138-144.
	GAN Lei, HE Dongbo, GUO Jianlin, et al. Application of machine learning method in reservoir classification[J]. Mathematics in Practice and Theory, 2019, 49(13):138-144.
[28]	马陇飞, 萧汉敏, 陶敬伟, 等. 基于梯度提升决策树算法的岩性智能分类方法[J]. 油气地质与采收率, 2022, 29(1):21-29.
	MA Longfei, XIAO Hanmin, TAO Jingwei, et al. Intelligent lithology classification method based on GBDT algorithm[J]. Petroleum Geology and Recovery Efficiency, 2022, 29(1):21-29.
[29]	杨威, 龙华. 基于加权距离进行密度计算的聚类方法研究[J]. 数据通信, 2019, 4:11-15.
	YANG Wei, LONG Hua. Research on clustering method based on weighted distance for density calculation[J]. Data Communications, 2019, 4:11-15.
[30]	付君豪, 陈再贺, 付宪, 等. 利用随机森林算法预测束鹿凹陷束21井区馆陶组储层砂地比[J]. 大庆石油地质与开发, 2023, 42(6):34-41.
	FU Junhao, CHEN Zaihe, FU Xian, et al. Prediction of sandstone-strata thickness ratio of Guantao formation reservoir in Shu 21 well area of Shulu sag via random forest algorithm[J]. Petroleum Geology & Oilfield Development in Daqing, 2023, 42(6):34-41.
[31]	秦志军, 操应长, 冯程. 基于改进型随机森林算法的页岩岩性识别:以准噶尔盆地芦草沟组为例[J]. 新疆石油地质, 2024, 45(5):595-603.
	QIN Zhijun, CAO Yingchang, FENG Cheng. Shale lithology identification based on improved random forest algorithm:A case of Lucaogou formation in Junggar Basin[J]. Xinjiang Petroleum Geology, 2024, 45(5):595-603.
[32]	赵磊, 任怀强, 徐守余, 等. 基于层次分析法和熵权法的储层非均质性定量表征:以胜坨油田坨142块沙二段15砂组为例[J]. 甘肃科学学报, 2020, 32(3):10-16.
	ZHAO Lei, REN Huaiqiang, XU Shouyu, et al. The quantitative characterization of reservoir inhomogeneity based on analytic hierarchy process and entropy method:Take sand group 15 in Tuo142 block Es2 of Shengtuo oilfield as an Example[J]. Journal of Gansu Sciences, 2020, 32(3):10-16.

评价指标体系	参数	符号及计算公式	符号注释
基本参数	砂体厚度	H_s	H_s——砂体厚度，m
	孔隙度	φ	φ——孔隙度，%
	渗透率	k	k——渗透率，mD
层内非均质参数	渗透率变异系数	${V}_{k}=\sqrt{\sum _{i=1}^{n}({k}_{i}{-\stackrel{-}{k})}^{2}/n}/\stackrel{-}{k}$	k_i——层内第i个砂体渗透率，mD $\stackrel{-}{k}$——层内砂体平均渗透率，mD n——层内砂体数，个 i——层内第i个样品
	渗透率突进系数	${T}_{k}={k}_{max}/\stackrel{-}{k}$	k_max——层内砂体最大渗透率，mD
	孔隙度变异系数	${V}_{\varphi }=\sqrt{\sum _{i=1}^{n}({\varphi }_{i}{-\overline{\varphi })}^{2}/n}/\overline{\varphi }$	${\varphi }_{i}$——层内第i个砂体孔隙度，% $\overline{\varphi }$——层内砂体平均孔隙度，%
	孔隙度突进系数	${T}_{\varphi }={\varphi }_{max}/\overline{\varphi }$	${\varphi }_{max}$——层内砂体最大孔隙度，%
	夹层频数	tP_j=N_j/H_s	N_j——夹层个数，个
	夹层密度	D_j=H_j/H_s	H_j——夹层厚度，m
层间非均质参数	砂地比	S=H_s/H	H——地层厚度，m
	分层系数	F=N_s/H	N_s——小层内钻遇砂层数，个
	砂厚变异系数	${V}_{s}=\sqrt{\sum _{i=1}^{n}({s}_{i}{-\stackrel{-}{s})}^{2}/n}/\stackrel{-}{s}$	s_i——层内第i个砂体厚度，m $\stackrel{-}{s}$——层内砂体平均厚度，m
平面非均质参数	河道宽度	W_i	W_i——河道宽度，km
	河道弯曲度	K_a=L/I	K_a——河道弯曲度 L——河道实际长度，km I——河道直线长度，km
	河道弯曲振幅	A_m	A_m——河道弯曲振幅，km
	河道弯曲波长	W_a	W_a——河道弯曲波长，km
	薄层砂体宽度	W_k	W_k——薄层砂体宽度，km
	薄层砂体长度	W_c	W_c——薄层砂体长度，km

算法	方法简介	优点	结果验证	特征重要性评估能力
多层感知机算法	前馈神经网络，含多个隐含层，通过神经元将多个输入非线性映射到单一输出^[26-27]	处理复杂非线性关系，映射多特征到综合指数	结果分散，地质意义不明显	弱
决策树算法	以树形结构展示决策规则和分类结果的归纳学习模型^[28]	用于特征重要性评估与分类	预测结果与实际分类对比，效果不佳	中等，稳定性差
聚类算法	无监督学习方法，用于将一组数据根据特征相似性分成若干个类别或簇^[29]	直观展示数据分组（散点图）	难以直接用于参数预测及特征重要性评估	弱，无监督学习限制
随机森林算法	集成学习方法，构建多棵决策树并综合其结果进行预测^[30-31]	有效评估特征重要性及贡献	结果基本符合预期	强

非均质性参数		相关性
孔隙度突进系数	孔隙度极差	0.887
渗透率突进系数	渗透率极差	0.882
夹层频数	夹层密度	0.991
砂地比	砂体厚度	0.945

非均质性	基本综合指数	层内非均质综合指数	层间非均质综合指数	平面非均质综合指数
弱	[0.25，1.00）	[0，0.4）	[0.2，1.0）	[0.2，1.0）
中等	[0.15，0.25）	[0.4，1.0）	[0，0.2）	[0，0.2）
强	[0，0.15）	[0.4，1.0）	[0，0.2）	[0，0.2）

砂体编号	基本综合指数	层内非均质综合指数	层间非均质综合指数	平面非均质综合指数	非均质综合指数	砂体构型样式
③	1.00	0.49	1.00	0.67	3.38	河道砂体强侵蚀叠加
⑨	0.42	0.57	0.47	0.79	1.14	河道砂体强侵蚀叠加
⑧	0.33	0.49	0.55	0.77	1.06	孤立状河道砂体
⑤	0.11	0.09	0.12	0.46	1.07	薄互层砂体
⑥	0.14	0.15	0.18	0.46	1.05	薄互层砂体
④	0.13	0.16	0.13	0.46	1.02	薄互层砂体
⑦	0.02	0.21	0.12	0	0.75	孤立状薄层砂体
②	0.12	1.00	0.25	0.61	0.65	孤立状薄层砂体
①	0.01	0.47	0.12	0.61	0.63	孤立状薄层砂体