基于机器学习的裂缝水驱气藏采收率预测方法

doi:10.13809/j.cnki.cn32-1825/te.2025.05.013

油气藏评价与开发 ›› 2025, Vol. 15 ›› Issue (5): 834-843.doi: 10.13809/j.cnki.cn32-1825/te.2025.05.013

基于机器学习的裂缝水驱气藏采收率预测方法

孙秋分¹(), 秦佳正²(), 冯乔¹, 乔宇²^,³, 刘雅昕², 赵启阳¹, 徐良¹, 闫春¹

1.中国石油杭州地质研究院，浙江杭州 310023
2.西南石油大学油气藏地质及开发工程全国重点实验室，四川成都 610500
3.中国石油冀东油田分公司储气库建设项目部，河北唐山 063000

收稿日期:2024-08-01 发布日期:2025-09-19 出版日期:2025-10-26
通讯作者: 秦佳正（1993—），女，博士，副研究员，从事储气库及CO₂埋存、人工智能在油气领域的应用等方面的工作。地址：四川省成都市新都区新都大道8号，邮政编码：610500。E-mail：jqin_swpu@163.com
作者简介:孙秋分（1981—），男，硕士，高级工程师，从事油气储量评估方面的工作。地址：浙江省杭州市西湖区西溪路920号，邮政编码：310023。E-mail：sunqf_hz@petrochina.com.cn
基金资助:
中国石油科学研究与技术开发项目课题“国内已开发气田储采平衡分析与SEC增储评估技术研究”(2022DJ7902)

A machine learning-based method for recovery rate prediction in fractured water-driven gas reservoirs

SUN Qiufen¹(), QIN Jiazheng²(), FENG Qiao¹, QIAO Yu²^,³, LIU Yaxin², ZHAO Qiyang¹, XU Liang¹, YAN Chun¹

1. PetroChina Hangzhou Research Institute of Geology, Hangzhou, Zhejiang 310023, China
2. State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation, Southwest Petroleum University, Chengdu, Sichuan 610500, China
3. Gas Storage Construction Project Department, PetroChina Jidong Oilfield Company, Tangshan, Hebei 063000, China

Received:2024-08-01 Online:2025-09-19 Published:2025-10-26

摘要/Abstract

摘要：

X气藏为受背斜构造控制的块状裂缝性边水气藏，受边水水侵的影响，气藏见水迅速，严重降低采收率，亟待科学的理论与方法指导开发实践。为此，研究在系统分析关键参数对采收率的影响规律基础上，构建裂缝水驱气藏采收率预测方法，为动态优化开发方案提供科学依据。基于气藏基础地质与生产数据，运用EDFM（嵌入式离散裂缝模型）建立单井机理模型，通过单因素敏感性分析，揭示水体倍数、采气速度、基质渗透率、渗透率各向异性的作用规律：水体倍数与采收率呈负相关，随着水体倍数增加，气井水气比上升速率显著加快，稳产时间大幅缩短；采气速度对气藏稳产时间影响尤为突出，存在特定的最佳采气速度阈值，可使采收率达到最大值；基质渗透率与采收率呈正相关，基质渗透率越低，气藏采收率越低，水气比上升速度越快，稳产时间越短；渗透率各向异性过低时，由于渗流能力较差导致采收率降低，比值增大则加速水侵使采收率降低。在此基础上，设计125组交叉实验方案，通过数值模拟获取基础数据，进而建立裂缝性水驱气藏采收率预测模型。为提升模型预测精度，对原始数据进行离散化处理，并采用决策树算法构建预测模型。经参数优化后，模型预测精度达到96%。基于X气藏2口生产井的实际动态数据，将该模型预测结果与Blasingame产量递减分析法进行对比验证，结果表明：该模型预测结果与实际生产数据高度吻合，具有较高的可靠性与实用性，为裂缝性水驱气藏采收率预测提供了一种高效、精准的技术手段。

关键词: 机器学习, 嵌入式离散裂缝模型, 采收率预测, 裂缝性水驱气藏, 数值模拟

Abstract:

Gas reservoir X is a block-fractured edge-water gas reservoir controlled by anticline structures. Due to edge-water invasion, rapid water breakthrough severely reduced recovery efficiency, highlighting the urgent need for theoretical and methodological guidance for reservoir development. To address this, based on a systematic analysis of the influence patterns of key parameters on recovery rate, a recovery prediction model for fractured water-driven gas reservoirs was established. This model provides a scientific basis for the dynamic optimization of development schemes. Using basic geological and production data, a single-well mechanistic model was constructed with the embedded discrete fracture model (EDFM). Through single-factor sensitivity analysis, the influence patterns of water body multiple, gas production rate, matrix permeability, and permeability anisotropy were revealed. Recovery rate exhibited a negative correlation with water body multiple. As the water body multiple increased, the water-gas ratio of gas wells rose significantly faster, and the stable production time was greatly shortened. Gas production rate had a pronounced impact on the stable production time of the reservoir. An optimal gas production rate threshold existed that maximized the recovery rate. Matrix permeability was positively correlated with recovery rate. Lower matrix permeability led to lower recovery rate. The faster the increase in water-gas ratio, the shorter the stable production time. When permeability anisotropy was too low, poor seepage capacity resulted in a reduced recovery rate. An increased ratio accelerated water invasion, further decreasing the recovery rate. Based on these findings, 125 sets of cross-experimental schemes were designed, and basic data were obtained through numerical simulations. A recovery rate prediction model for fractured water-driven gas reservoirs was established. To improve the prediction accuracy of the model, the original data were discretized, and a decision tree algorithm was used. After parameter optimization, the prediction accuracy of the model reached 96%. The model was validated using actual dynamic data from two production wells in the gas reservoir X. The prediction results were compared with those from the Blasingame production decline analysis method. The results showed high consistency between model predictions and actual production data, indicating high reliability and practicality of the model. This provides an efficient and precise technical method for recovery rate prediction in fractured water-driven gas reservoirs.

Key words: machine learning, embedded discrete fracture model, recovery rate prediction, fractured water-driven gas reservoirs, numerical simulation

中图分类号:

TE319

孙秋分,秦佳正,冯乔, 等. 基于机器学习的裂缝水驱气藏采收率预测方法[J]. 油气藏评价与开发, 2025, 15(5): 834-843.

SUN Qiufen,QIN Jiazheng,FENG Qiao, et al. A machine learning-based method for recovery rate prediction in fractured water-driven gas reservoirs[J]. Petroleum Reservoir Evaluation and Development, 2025, 15(5): 834-843.

图/表 12

图1

图2

图3

图4

图5

图6

表1

表2

表3

表4

图7

图8

参考文献 26

[1]	国家能源局. 天然气可采储量计算方法: [S]. 北京: 石油工业出版社, 2022.
	National Energy Bureau of the People’s Republic of China. Natural gas recoverable reserves calculation method: [S]. Beijing: Petroleum Industry Press, 2022.
[2]	尹涛, 杨屹铭, 靳锁宝, 等. 概率法在岩性气藏储量风险评估中的应用[J]. 西南石油大学学报(自然科学版), 2020, 42(3): 60-68. doi: 10.11885/j.issn.1674-5086.2019.09.16.04
	YIN Tao, YANG Yiming, JIN Suobao, et al. Application of probability method in the reserves risk evaluation of lithologic gas reservoirs[J]. Journal of Southwest Petroleum University (Science & Technology Edition), 2020, 42(3): 60-68.
[3]	孙贺东, 王宏宇, 朱松柏, 等. 基于幂函数形式物质平衡方法的高压、超高压气藏储量评价[J]. 天然气工业, 2019, 39(3): 56-64.
	SUN Hedong, WANG Hongyu, ZHU Songbai, et al. Reserve evaluation of high pressure and ultra high pressure reservoirs with power function material balance method[J]. Natural Gas Industry, 2019, 39(3): 56-64.
[4]	付斌, 李进步, 张晨, 等. 强非均质致密砂岩气藏已开发区井网完善方法[J]. 天然气地球科学, 2020, 31(1): 143-149. doi: 10.11764/j.issn.1672-1926.2019.09.005
	FU Bin, LI Jinbu, ZHANG Chen, et al. Improvement of well pattern in development area of tight sandstone gas reservoir[J]. Natural Gas Geoscience, 2020, 31(1): 143-149. doi: 10.11764/j.issn.1672-1926.2019.09.005
[5]	郑玲丽, 朱冰倩, 张宇豪, 等. 缝洞型碳酸盐岩油藏水驱特征曲线类型及适应性: 以塔河油田为例[J]. 油气藏评价与开发, 2024, 14(6): 899-907.
	ZHENG Lingli, ZHU Bingqian, ZHANG Yuhao, et al. Types and applicability of waterflooding characteristic curves in fractured-cavity carbonate reservoirs: A case study of Tahe Oilfield[J]. Petroleum Reservoir Evaluation and Development, 2024, 14(6): 899-907.
[6]	崔传智, 李怀亮, 吴忠维, 等. 考虑压驱注水诱发裂缝影响的注水井压力分析[J]. 油气藏评价与开发, 2023, 13(5): 686-694.
	CUI Chuanzhi, LI Huailiang, WU Zhongwei, et al. Analysis of pressures in water injection wells considering fracture influence induced by pressure-drive water injection[J]. Petroleum Reservoir Evaluation and Development, 2023, 13(5): 686-694.
[7]	梁运培, 张怀军, 王礼春, 等. 连续加载应力下真实裂缝流场和渗透率演化规律数值研究[J]. 油气藏评价与开发, 2023, 13(6): 834-843.
	LIANG Yunpei, ZHANG Huaijun, WANG Lichun, et al. Numerical simulation of flow fields and permeability evolution in real fractures under continuous loading stress[J]. Petroleum Reservoir Evaluation and Development, 2023, 13(6): 834-843.
[8]	陈祥, 王冠, 刘平礼, 等. 四川盆地灯影组酸压裂缝导流能力实验和模拟研究[J]. 油气藏评价与开发, 2024, 14(4): 569-576.
	CHEN Xiang, WANG Guan, LIU Pingli, et al. Experimental and simulation study on fracture conductivity of acid-fracturing in Dengying Formation of Sichuan Basin[J]. Petroleum Reservoir Evaluation and Development, 2024, 14(4): 569-576.
[9]	SIDLE R E, LEE W J. An update on the use of reservoir analogs for the estimation of oil and gas reserves[C]// SPE Hydrocarbon Economics and Evaluation Symposium. SPE, 2010: 1-9.
[10]	谭晓华, 彭港珍, 李晓平, 等. 考虑水封气影响的有水气藏物质平衡法及非均匀水侵模式划分[J]. 天然气工业, 2021, 41(3): 97-103.
	TAN Xiaohua, PENG Gangzhen, LI Xiaoping, et al. Material balance method and classification of non-uniform water invasion mode for gas reservoirs with water considering the effect of water sealed gas[J]. Natural Gas Industry, 2021, 41(3): 97-103.
[11]	吕志凯, 唐海发, 刘群明, 等. 塔里木盆地库车坳陷超深层裂缝性致密气藏水封气动态评价方法[J]. 天然气地球科学, 2022, 33(11): 1874-1882. doi: 10.11764/j.issn.1672-1926.2022.07.007
	Zhikai LYU, TANG Haifa, LIU Qunming, et al. Dynamic evaluation method of water sealed gas for ultra-deep fractured tight gas reservoir in Kuqa Depression, Tarim Basin[J]. Natural Gas Geoscience, 2022, 33(11): 1874-1882. doi: 10.11764/j.issn.1672-1926.2022.07.007
[12]	王国锋, 周梦飞, 胡勇, 等. 裂缝—孔隙型边底水气藏提高采收率大型物理模拟实验[J]. 天然气地球科学, 2024, 35(1): 96-103. doi: 10.11764/j.issn.1672-1926.2023.07.009
	WANG Guofeng, ZHOU Mengfei, HU Yong, et al. Large-scale physical simulation experiment for enhanced gas recovery in fractured-porous water-drive gas reservoirs[J]. Natural Gas Geoscience, 2024, 35(1): 96-103. doi: 10.11764/j.issn.1672-1926.2023.07.009
[13]	刘念肖, 雷登生, 黄小亮, 等. 考虑水封气的水驱气藏开发因素数值模拟研究[J]. 重庆科技学院学报(自然科学版), 2022, 24(2): 31-36.
	LIU Nianxiao, LEI Dengsheng, HUANG Xiaoliang, et al. Numerical simulation research on development factors of water drive gas reservoir considering water-sealed gas[J]. Journal of Chongqing University of Science and Technology (Natural Sciences Edition), 2022, 24(2): 31-36.
[14]	胡景涛, 王本成, 王勇飞, 等. 基于试井分析方法的元坝气田水侵早期识别[J]. 非常规油气, 2023, 10(3): 89-97.
	HU Jingtao, WANG Bencheng, WANG Yongfei, et al. Early identification of water invasion in Yuanba Gas Field based on well test analysis method[J]. Unconventional Oil & Gas, 2023, 10(3): 89-97.
[15]	李兴娟, 姜应兵, 丁立明. 塔河油田AD4单元流场变化及水侵规律认识[J]. 非常规油气, 2022, 9(5): 123-128.
	LI Xingjuan, JIANG Yingbing, DING Liming. Understanding of flow distribution and water intrusion rule in AD4 well area of Tahe Olifield[J]. Unconventional Oil & Gas, 2022, 9(5): 123-128.
[16]	侯亚伟, 刘超, 徐中波, 等. 多层水驱开发油田采收率快速预测方法[J]. 石油钻探技术, 2022, 50(5): 82-87.
	HOU Yawei, LIU Chao, XU Zhongbo, et al. A method for rapidly predicting recovery of multi-layer oilfields developed by water-flooding[J]. Petroleum Drilling Techniques, 2022, 50(5): 82-87.
[17]	盖建. 基于自动机器学习的采油井压裂效果预测方法[J]. 油气地质与采收率, 2023, 30(1): 161-170.
	GAI Jian. Prediction method for hydraulic fracturing effect of oil production well based on automatic machine learning technology[J]. Petroleum Geology and Recovery Efficiency, 2023, 30(1): 161-170.
[18]	尚福华, 郑伟. 基于决策树的注采连通关系判别研究[J]. 计算机应用研究, 2013, 30(7): 2051-2054.
	SHANG Fuhua, ZHENG Wei. Study on inferring interwell connectivity of injection-production system based on decision tree[J]. Application Research of Computers, 2013, 30(7): 2051-2054.
[19]	张世昆, 陈作. 人工智能在压裂技术中的应用现状及前景展望[J]. 石油钻探技术, 2023, 51(1): 69-77.
	ZHANG Shikun, CHEN Zuo. Status and prospect of artificial intelligence application in fracturing technology[J]. Petroleum Drilling Techniques, 2023, 51(1): 69-77.
[20]	徐帅. 致密油藏注水吞吐数值模拟及开发制度优化[D]. 北京: 中国石油大学(北京), 2022.
	XU Shuai. Numerical simulation and development system optimization of water injection huff and puff in tight oil reservoirs[D]. Beijing: China University of Petroleum (Beijing), 2022.
[21]	窦凯文. 致密油水平井体积压裂参数优化研究[D]. 东营: 中国石油大学(华东), 2018
	DOU Kaiwen. Optimization of volumetric-fracturing parameters for horizontal wells in tight oil reservoirs[D]. Dongying: China University of Petroleum (East China), 2018.
[22]	LEE S H, LOUGH M F, JENSEN C L. Hierarchical modeling of flow in naturally fractured formations with multiple length scales[J]. Water Resources Research, 2001, 37(3): 443-455.
[23]	LI L, LEE S H. Efficient field-scale simulation of black oil in a naturally fractured reservoir through discrete fracture networks and homogenized media[J]. SPE Reservoir Evaluation &Engineering, 2008, 11(4): 750-758.
[24]	XU Y, CAVALCANTE FILHO J S, YU W, et al. Discrete-fracture modeling of complex hydraulic-fracture geometries in reservoir simulators[J]. SPE Reservoir Evaluation & Engineering, 2017, 20(2): 403-422.
[25]	王平, 沈海超. 加拿大M致密砂岩气藏高效开发技术[J]. 石油钻探技术, 2022, 50(1): 97-102.
	WANG Ping, SHEN Haichao. High-efficient development technologies for the M tight sandstone gas reservoir in Canada[J]. Petroleum Drilling Techniques, 2022, 50(1): 97-102.
[26]	BREIMAN L, FRIEDMAN J, OLSHEN R A, et al. Classification and regression trees[M]. Belmont California: Wadsworth, 1984.

井名	M	k/10^-3 μm²	v/%	A	B
水平1	1	0.1	3	0.10	1
水平2	5	0.5	6	0.50	5
水平3	10	1.0	9	1.00	10
水平4	15	5.0	12
水平5	20	10.0

方案	M	k/10^-3 μm²	v/%	A	B	R/%
方案1	1	5.0	3	0.10	1	54.71
方案2	1	1.0	12	0.10	10	47.94
方案3	1	10.0	9	0.50	10	49.00
方案4	1	0.1	12	1.00	1	22.10
方案5	1	10.0	6	0.10	5	53.28
方案6	1	0.5	12	0.50	5	50.26
方案7	1	0.1	3	1.00	5	25.12
方案8	1	0.5	6	1.00	5	43.74
方案9	1	5.0	12	0.10	10	52.77
方案10	5	1.0	3	0.10	1	60.10
方案11	5	0.1	6	1.00	5	23.01
方案12	5	10.0	3	0.50	1	64.19
方案13	5	0.5	9	0.10	10	60.38
方案14	5	5.0	6	1.00	1	52.85
方案15	5	0.1	12	0.10	10	36.59
方案16	5	10.0	12	0.50	5	63.54
方案17	5	10.0	3	1.00	10	61.92
方案18	5	0.1	9	0.50	1	25.33
方案19	10	0.1	3	0.10	5	35.12
方案20	10	5.0	9	1.00	5	67.12
方案21	10	1.0	6	0.50	10	67.11
方案22	10	1.0	12	0.50	5	66.47
方案23	10	0.5	6	0.10	1	56.43
方案24	10	10.0	9	1.00	1	66.84
方案25	10	5.0	9	1.00	10	67.05
…	…	…	…	…		…
方案122	20	0.5	9	0.10	10	61.07
方案123	20	1.0	12	0.50	5	61.65
方案124	20	5.0	3	0.10	1	61.67
方案125	20	0.5	3	0.50	5	57.33

数据类型	参数	分类规则	分类结果
建模特征	M	M<10	1
		10≤M<20	2
		M≥20	3
	k/10^-3μm²	k<1	1
		1≤k≤5	2
		k>5	3
	v/%	v≤6	1
	v/%	v>6	2
	A	A<0.5	1
	A	A≥0.5	2
	B	B<5	1
	B	B≥5	2
建模标签	R/%	R<35	1（低）
		35≤R≤55	2（中）
		R>55	3（高）

参数	最优值
评估指标	基尼系数
决策树最大深度	6
叶节点存在所需的最小样本数	1
节点分枝策略	最佳随机分枝

基于机器学习的裂缝水驱气藏采收率预测方法

A machine learning-based method for recovery rate prediction in fractured water-driven gas reservoirs

RichHTML

PDF (PC)

赞

可视化

摘要/Abstract

引用本文

使用本文

图/表 12

参考文献 26

相关文章 15

Metrics

本文评价

推荐阅读 0

[1]	齐怀彦, 杨国斌, 朱亚娣, 邓茗心, 耿少阳, 田伟超. 基于核磁共振与孔隙尺度模拟的致密油藏渗吸机理研究 [J]. 油气藏评价与开发, 2025, 15(5): 824-833.
[2]	王烁石, 纪强, 郭平, 刘煌, 温连辉, 徐锐锋, 汪周华, 张瑞旭. LARSEN & SKAUGE相渗滞后模型在高温高压CO₂-水互驱实验中的适应性 [J]. 油气藏评价与开发, 2025, 15(4): 613-624.
[3]	孙东升, 张顺康, 王智林, 葛政俊, 林波. 苏北断块型圈闭基于安全性CO₂地质封存能力计算方法研究 [J]. 油气藏评价与开发, 2025, 15(4): 641-645.
[4]	朱苏阳, 彭真, 邸云婷, 彭小龙, 刘东晨, 官文洁. 页岩气产能评价研究进展：内涵、方法和方向 [J]. 油气藏评价与开发, 2025, 15(3): 488-499.
[5]	柴妮娜, 李嘉瑞, 张力文, 王俊杰, 刘亚鹏, 朱伦. 夹层型陆相页岩油储层压裂裂缝扩展实验研究 [J]. 油气藏评价与开发, 2025, 15(1): 124-130.
[6]	张章, 孟鹏, 杨威, 张小龙, 黄奇, 王浩然. 基于地震属性堆叠泛化集成学习的辫状河储层构型表征——以渤海湾盆地C-2油田为例 [J]. 油气藏评价与开发, 2025, 15(1): 64-72.
[7]	盖长城, 李洪达, 任路, 曹伟, 郝杰. 地热数值模拟与油藏数值模拟方法对比分析 [J]. 油气藏评价与开发, 2024, 14(6): 849-856.
[8]	瞿常青, 林千果. 基于通量监测-CFD模拟的CO₂驱油封存地表泄漏大气扩散研究 [J]. 油气藏评价与开发, 2024, 14(6): 885-891.
[9]	张益, 宁崇如, 陈亚舟, 姬玉龙, 赵立阳, 王爱方, 黄晶晶, 于凯怡. 致密油藏大排量注水吞吐技术及参数优化研究 [J]. 油气藏评价与开发, 2024, 14(5): 727-733.
[10]	曹小朋, 刘海成, 李忠新, 陈先超, 江朋宇, 樊浩. 基于EDFM的页岩油水平井注水吞吐优化研究 [J]. 油气藏评价与开发, 2024, 14(5): 734-740.
[11]	廖凯, 张士诚, 谢勃勃. 页岩油体积压裂后合理焖井时间模拟研究 [J]. 油气藏评价与开发, 2024, 14(5): 749-755.
[12]	陈祥, 王冠, 刘平礼, 杜娟, 王铭, 陈伟华, 李金龙, 刘金明, 刘飞. 四川盆地灯影组酸压裂缝导流能力实验和模拟研究 [J]. 油气藏评价与开发, 2024, 14(4): 569-576.
[13]	盖长城, 赵忠新, 任路, 颜艺灿, 侯本峰. 中深层水热型地热资源集群式开发井位部署参数研究与应用——以HTC地热田为例 [J]. 油气藏评价与开发, 2024, 14(4): 638-646.
[14]	陈学忠, 赵慧言, 陈满, 徐华卿, 杨建英, 杨晓敏, 唐慧莹. 海陆过渡相页岩储层合层开采数值模拟研究 [J]. 油气藏评价与开发, 2024, 14(3): 382-390.
[15]	张连锋, 张伊琳, 郭欢欢, 李洪生, 李俊杰, 梁丽梅, 李文静, 胡书奎. 近废弃油藏延长生命周期开发调整技术 [J]. 油气藏评价与开发, 2024, 14(1): 124-132.