物理约束多源数据驱动机械钻速智能预测方法

doi:10.7623/syxb202609010

石油学报 ›› 2026, Vol. 47 ›› Issue (9): 1874-1887.DOI: 10.7623/syxb202609010

• 石油工程 • 上一篇

物理约束多源数据驱动机械钻速智能预测方法

廖华林^1,2, 刘建胜¹, 裴学良³, 许军富³, 程丙方³, 王华健⁴

1. 中国石油大学（华东）石油工程学院山东青岛 266580;
2. 怀柔实验室新疆研究院新疆乌鲁木齐 830000;
3. 中石化胜利石油工程有限公司山东东营 257000;
4. 中石化经纬有限公司山东青岛 266000

收稿日期:2025-12-06 修回日期:2026-07-01 发布日期:2026-10-14
通讯作者: 廖华林，男，1974年8月生，2005年获石油大学（华东）博士学位，现为中国石油大学（华东）石油工程学院教授、博士生导师，主要从事油气井工程方面的教学与科研工作。Email：liaohualin2003@126.com
作者简介:廖华林，男，1974年8月生，2005年获石油大学（华东）博士学位，现为中国石油大学（华东）石油工程学院教授、博士生导师，主要从事油气井工程方面的教学与科研工作。Email：liaohualin2003@126.com
基金资助:
国家科技重大专项（2025ZD1401900，2025ZD1401902）和中央高校基本科研业务费专项资金项目（26CX05024A）资助。

An intelligent prediction method for rate of penetration driven by the physical constraint and multi-source data

Liao Hualin^1,2, Liu Jiansheng¹, Pei Xueliang³, Xu Junfu³, Cheng Bingfang³, Wang Huajian⁴

1. School of Petroleum Engineering, China University of Petroleum, Shandong Qingdao 266580, China;
2. Xinjiang Research Institute of Huairou Laboratory, Xinjiang Urumqi 830000, China;
3. Sinopec Shengli Petroleum Engineering Co., Ltd.,Shandong Dongying 257000, China;
4. Sinopec Matrix Corporation, Shandong Qingdao 266000, China

Received:2025-12-06 Revised:2026-07-01 Published:2026-10-14

摘要/Abstract

摘要： 钻井机械钻速智能预测是油气井工程领域关键研究方向，但现有数据驱动模型多基于单一来源数据，未能有效融合多源数据与物理先验知识，导致模型在不同井间的泛化能力有限。结合深度学习理论和经典钻速方程，提出了一种融合多源信息和物理规律的数据—物理耦合驱动机械钻速预测方法。该方法采用并联的双通道多层感知机模块，分别提取地层信息和钻井工程参数特征，并利用Transformer模型注意力机制实现地质—工程多源数据融合。通过在损失函数中引入修正杨格钻速方程作为物理约束，增强模型的物理一致性，有效抑制了过拟合。结果表明:引入物理约束后，模型的泛化性能和预测精度均显著提升，相较于纯数据驱动模型，平均绝对误差降低约14.6%，均方误差降低约61.9%，均方根误差降低约38.3%，R²提高约19.8%；双通道多层感知机特征提取模块显著改善了预测精度与稳定性，与单通道框架相比，R²提升约5.5%，均方根误差降低约30.2%；所提出的物理约束多源数据驱动钻速智能预测方法在不同区块井中均表现出良好的稳定性和泛化能力，相较于传统数据驱动模型具有明显优势。

关键词: 智能钻速预测, 地质—工程多源数据, 物理约束, 双通道神经网络, 注意力机制

Abstract: Intelligent prediction of rate of penetration (ROP) is a critical research direction in oil and gas well engineering. However, existing methods are predominantly based on single-source data-driven models, which fail to effectively integrate multi-source data with physical knowledge, resulting in limited generalization capability across different wells. In this study, a data-physics-coupled ROP prediction method integrating multi-source information and drilling physics is proposed by combining deep learning theory with the classical ROP equation. The proposed method employs a parallel dual-channel multilayer perceptron (MLP) architecture to independently extract features from formation information and drilling engineering parameters, while the attention mechanism of the Transformer model is utilized to achieve the fusion of geological and engineering multi-source data. Furthermore, a modified ROP equation of the F. S. Young model is embedded in the loss function as a physical constraint, enhancing the physical consistency and effectively mitigating overfitting of the model. The results show that the incorporation of the physical constraint significantly enhances both the generalization capability and prediction accuracy of the model. Compared with the purely data-driven model, the proposed method reduces mean absolute error (MAE), mean squared error (MSE) and root mean square error (RMSE) by approximately 14.6%, 61.9% and 38.3%, respectively, while increasing R² by approximately 19.8%. The dual-channel MLP feature extraction module further improves prediction accuracy and model stability; compared with the single-channel framework, it increases R² by approximately 5.5% and reduces RMSE by approximately 30.2%. The proposed ROP prediction method, which integrates multi-source data with the physical constraint, exhibits excellent stability and generalization capability across wells from different blocks and demonstrates clear advantages over conventional data-driven models.

Key words: intelligent rate of penetration prediction, geological-engineering multi-source data, physical constraint, dual-channel neural network, attention mechanism

中图分类号:

TE242

廖华林, 刘建胜, 裴学良, 许军富, 程丙方, 王华健. 物理约束多源数据驱动机械钻速智能预测方法[J]. 石油学报, 2026, 47(9): 1874-1887.

Liao Hualin, Liu Jiansheng, Pei Xueliang, Xu Junfu, Cheng Bingfang, Wang Huajian. An intelligent prediction method for rate of penetration driven by the physical constraint and multi-source data[J]. Acta Petrolei Sinica, 2026, 47(9): 1874-1887.

参考文献

[1] 石祥超,王宇鸣,刘越豪,等. 关于人工智能方法用于钻井机械钻速预测的探讨[J].石油钻采工艺,2022,44(1):105-111.
SHI Xiangchao,WANG Yuming,LIU Yuehao,et al.Discussion on the application of artificial intelligence method to the prediction of drilling machinery ROP[J].Oil Drilling & Production Technology,2022,44(1):105-111.
[2] LI Zhong,XIE Renjun,WU Yi,et al.Progress and prospect of CNOOC’s oil and gas well drilling and completion technologies[J].Natural Gas Industry B,2022,9(2):209-217.
[3] 李根生,熊超,黄中伟,等.干热岩钻井技术发展现状与建议[J].石油学报,2026,47(1):279-293.
LI Gensheng,XIONG Chao,HUANG Zhongwei,et al.Development status and recommendations for hot dry rock drilling technology[J].Acta Petrolei Sinica,2026,47(1):279-293.
[4] 孙金声,杨景斌,吕开河,等.致密油气钻井液技术研究现状与展望[J].石油学报,2025,46(1):279-288.
SUN Jinsheng,YANG Jingbin,Lü Kaihe,et al.Research status and prospects of drilling fluid technology for tight oil and gas[J].Acta Petrolei Sinica,2025,46(1):279-288.
[5] 龚斌,王虹雅,王红娜,等.基于大数据分析算法的深部煤层气地质—工程一体化智能决策技术[J].石油学报,2023,44(11):1949-1958.
GONG Bin,WANG Hongya,WANG Hongna,et al.Integrated intelligent decision-making technology for deep coalbed methane geology and engineering based on big data analysis algorithms[J].Acta Petrolei Sinica,2023,44(11):1949-1958.
[6] 李宁,徐彬森,武宏亮,等.人工智能在测井地层评价中的应用现状及前景[J].石油学报,2021,42(4):508-522.
LI Ning,XU Binsen,WU Hongliang,et al.Application status and prospects of artificial intelligence in well logging and formation evaluation[J].Acta Petrolei Sinica,2021,42(4):508-522.
[7] 孙宝江,周梓强,孙骞.油气钻采工程中的人工智能方法研究进展[J].石油学报,2025,46(11):2141-2173.
SUN Baojiang,ZHOU Ziqiang,SUN Qian.Progress on artificial intelligence methods in oil and gas drilling and production[J].Acta Petrolei Sinica,2025,46(11):2141-2173.
[8] 林元华,宗玉宇,梁政,等.石油钻井机械钻速预测研究进展[J].石油钻探技术,2004,32(1):10-13.
LIN Yuanhua,ZONG Yuyu,LIANG Zheng,et al.The developments of ROP prediction for oil drilling[J].Petroleum Drilling Techniques,2004,32(1):10-13.
[9] 阎铁,毕雪亮,刘春天,等.一种预测深井钻井速度的人工神经网络法[J].石油钻探技术,2001,29(6):9-10.
YAN Tie,BI Xueliang,LIU Chuntian,et al.A manual nerve network for prodiction of deep drilling rate[J].Petroleum Drilling Techniques,2001,29(6):9-10.
[10] SABAH M,TALEBKEIKHAH M,WOOD D A,et al.A machine learning approach to predict drilling rate using petrophysical and mud logging data[J].Earth Science Informatics,2019,12(3):319-339.
[11] ASHRAFI S B,ANEMANGELY M,SABAH M,et al.Application of hybrid artificial neural networks for predicting rate of penetration (ROP):a case study from Marun oil field[J].Journal of Petroleum Science and Engineering,2019,175:604-623.
[12] 李琪,屈峰涛,何璟彬,等.基于PSO-BP的钻井机械钻速预测模型[J].科学技术与工程,2021,21(19):7984-7990.
LI Qi,QU Fengtao,HE Jingbin,et al.Rate of penetration for drilling prediction model based on PSO-BP[J].Science Technology and Engineering,2021,21(19):7984-7990.
[13] 宋宇,彭福康,孟卓然,等.基于BO-LSTM的海洋浅层钻井机械钻速预测方法[J].自动化与仪表,2024,39(10):14-17.
SONG Yu,PENG Fukang,MENG Zhuoran,et al.Prediction method for the drilling speed of marine shallow seismic drill rig based on BO-LSTM[J].Automation & Instrumentation,2024,39(10):14-17.
[14] SOBHI I,DOBBI A,HACHANA O.Prediction and analysis of penetration rate in drilling operation using deterministic and metaheuristic optimization methods[J].Journal of Petroleum Exploration and Production Technology,2022,12(5):1341-1352.
[15] SALUBI V,MAHON R,OLUYEMI G.The combined effect of fluid rheology,inner pipe rotation and eccentricity on the flow of Newtonian and non-Newtonian fluid through the annuli[J].Journal of Petroleum Science and Engineering,2022,211:110018.
[16] HEGDE C,DAIGLE H,MILLWATER H,et al.Analysis of rate of penetration (ROP) prediction in drilling using physics-based and data-driven models[J].Journal of Petroleum Science and Engineering,2017,159:295-306.
[17] CAO Tong,CHEN Xuyue,YU Kaian,et al.Hydraulic modeling and optimization of jet mill bit considering the characteristics of depressurization and cuttings cleaning[J].Petroleum Science,2023,20(5):3085-3099.
[18] 李博,王鲁朝.基于GA-BPANN的钻井机械钻速预测模型[J].西部探矿工程,2024,36(2):56-61.
LI Bo,WANG Luchao.A prediction model for drilling speed of drilling machinery based on GA-BPANN[J].West-China Exploration Engineering,2024,36(2):56-61.
[19] LI Gensheng,SONG Xianzhi,TIAN Shouceng,et al.Intelligent drilling and completion:a review[J].Engineering,2022,18:33-48.
[20] 张立刚,苗振华,黄小刚,等.基于MEA-BP神经网络的钻井机械钻速预测[J].自动化与仪表,2022,37(11):87-92.
ZHANG Ligang,MIAO Zhenhua,HUANG Xiaogang,et al.Prediction of drilling ROP based on MEA-BP neural network[J].Automation & Instrumentation,2022,37(11):87-92.
[21] SU Kanhua,DA Wenhao,LI Meng,et al.Research on a drilling rate of penetration prediction model based on the improved chaos whale optimization and back propagation algorithm[J].Geoenergy Science and Engineering,2024,240:213017.
[22] 徐英卓,王若禹,王六鹏.基于遗传算法优化的随机森林钻井机械钻速预测模型研究[J].智能计算机与应用,2023,13(3):210-214.
XU Yingzhuo,WANG Ruoyu,WANG Liupeng.Research on ROP prediction model of random forest drilling machinery based on genetic algorithm optimization[J].Intelligent Computer and Applications,2023,13(3):210-214.
[23] GHANITOOS H,GOHARIMANESH M,AKBARI A.Prediction of drill penetration rate in drilling oil wells using mathematical and neurofuzzy modeling methods[J].Energy Reports,2024,11:145-152.
[24] YANG Xiao,WU Min,LU Chengda,et al.Prediction of rate of penetration based on drilling conditions identification for drilling process[J].Neurocomputing,2024,579:127439.
[25] 李琪,屈峰涛,何璟彬,等.基于BAS-BP的钻井机械钻速预测模型[J].西安石油大学学报:自然科学版,2021,36(6):89-95.
LI Qi,QU Fengtao,HE Jingbin,et al.Prediction model of mechanical ROP during drilling based on BAS-BP[J].Journal of Xi’an Shiyou University:Natural Science Edition,2021,36(6):89-95.
[26] 苏兴华,孙俊明,高翔,等.基于GBDT算法的钻井机械钻速预测方法研究[J].计算机应用与软件,2019,36(12):87-92.
SU Xinghua,SUN Junming,GAO Xiang,et al.Prediction method of drilling rate of penetration based on GBDT algorithm[J].Computer Applications and Software,2019,36(12):87-92.
[27] HEGDE C,GRAY K.Evaluation of coupled machine learning models for drilling optimization[J].Journal of Natural Gas Science and Engineering,2018,56:397-407.
[28] ZHANG Chengkai,SONG Xianzhi,LIU Zihao,et al.Real-time and multi-objective optimization of rate-of-penetration using machine learning methods[J].Geoenergy Science and Engineering,2023,223:211568.
[29] 陈宗琦,陈修平,李轲,等.基于KMeans-RF算法的顺北区块钻井机械钻速预测模型[J].石化技术,2022,29(6):117-119.
CHEN Zongqi,CHEN Xiuping,LI Ke,et al.Prediction model of drilling ROP in Shunbei block based on KMeans-RF algorithm[J].Petrochemical Industry Technology,2022,29(6):117-119.
[30] GUPTA I,TRAN N,DEVEGOWDA D,et al.Looking ahead of the bit using surface drilling and petrophysical data:machine-learning-based real-time geosteering in Volve field[J].SPE Journal,2020,25(2):990-1006.
[31] ELKATATNY S.Real-time prediction of rate of penetration while drilling complex lithologies using artificial intelligence techniques[J].Ain Shams Engineering Journal,2021,12(1):917-926.
[32] CHEN Gang,YANG Henglin,YANG Dianru,et al.A new rate of penetration prediction model based on the XGBoost and explanation method[C]∥Proceedings of the 56th U.S.Rock Mechanics/Geomechanics Symposium.Santa Fe,USA:ARMA,2022.
[33] FENG Zhipeng,GANI H,DAMAYANTI A D,et al.An explainable ensemble machine learning model to elucidate the influential drilling parameters based on rate of penetration prediction[J].Geoenergy Science and Engineering,2023,231:212231.
[34] CHEN Xuyue,WENG Chengkai,DU Xu,et al.Prediction of the rate of penetration in offshore large-scale cluster extended reach wells drilling based on machine learning and big-data techniques[J].Ocean Engineering,2023,285:115404.
[35] 祝兆鹏,朱林,宋先知,等.机理约束下钻井机械钻速智能预测泛化方法[J].天然气工业,2024,44(9):179-189.
ZHU Zhaopeng,ZHU Lin,SONG Xianzhi,et al.A generalization method of intelligent ROP prediction under mechanism constraints[J].Natural Gas Industry,2024,44(9):179-189.
[36] 夏冰.基于GBDT算法的空气冲旋钻井机械速度预测方法[J].制造业自动化,2022,44(3):185-188.
XIA Bing.Mechanical speed prediction method of air percussion rotary drilling based on GBDT algorithm[J].Manufacturing Automation,2022,44(3):185-188.
[37] 胥知画,姜杰,周长春,等.基于粒子群优化的融合特征选择钻速预测模型研究[J].钻探工程,2025,52(2):134-143.
XU Zhihua,JIANG Jie,ZHOU Changchun,et al.Research on a rate of penetration (ROP) prediction model based on feature selection integrated with particle swarm optimization (PSO)[J].Drilling Engineering,2025,52(2):134-143.
[38] HASSAN A,ELKATATNY S,AL-MAJED A.Coupling rate of penetration and mechanical specific energy to improve the efficiency of drilling gas wells[J].Journal of Natural Gas Science and Engineering,2020,83:103558.
[39] ALALI A M,ABUGHABAN M F,AMAN B M,et al.Hybrid data driven drilling and rate of penetration optimization[J].Journal of Petroleum Science and Engineering,2021,200:108075.
[40] AMADI K,IYALLA I,PRABHU R.Modeling and predicting performance of autonomous rotary drilling system using machine learning techniques[R].SPE 208450,2021.
[41] LI Yuanjun,SAMUEL R.Prediction of penetration rate ahead of the bit through real-time updated machine learning models[R].SPE 194105,2019.
[42] ENCINAS M A,TUNKIEL A T,SUI Dan.Downhole data correction for data-driven rate of penetration prediction modeling[J].Journal of Petroleum Science and Engineering,2022,210:109904.
[43] BIZHANI M,KURU E.Towards drilling rate of penetration prediction:Bayesian neural networks for uncertainty quantification[J].Journal of Petroleum Science and Engineering,2022,219:111068.
[44] BOUKREDERA F S,YOUCEFI M R,HADJADJ A,et al.Enhancing the drilling efficiency through the application of machine learning and optimization algorithm[J].Engineering Applications of Artificial Intelligence,2023,126:107035.
[45] MAHMOODZADEH A,NEJATI H R,MOHAMMADI M,et al.Forecasting tunnel boring machine penetration rate using LSTM deep neural network optimized by grey wolf optimization algorithm[J].Expert Systems with Applications,2022,209:118303.
[46] 樊永东,金衍,林伯韬,等.地质-工程驱动的邻井辅助同井钻速预测与优化方法[J].石油钻探技术,2025,53(1):31-40.
FAN Yongdong,JIN Yan,LIN Botao,et al.Prediction and optimization of ROP assisted by adjacent well data based on geological and engineering driving[J].Petroleum Drilling Techniques,2025,53(1):31-40.
[47] LAWAL A I,KWON S,ONIFADE M.Prediction of rock penetration rate using a novel antlion optimized ANN and statistical modelling[J].Journal of African Earth Sciences,2021,182:104287.
[48] GAN Chao,CAO Weihua,WU Min,et al.Two-level intelligent modeling method for the rate of penetration in complex geological drilling process[J].Applied Soft Computing,2019,80:592-602.
[49] ELMGERBI A M,ETTINGER C P,TEKUM P M,et al.Application of machine learning techniques for real time rate of penetration optimization[R].SPE 202184,2021.
[50] HARELAND G,WU A,RASHIDI B.A drilling rate model for roller cone bits and its application[R].SPE 129592,2010.
[51] VASWANI A,SHAZEER N,PARMAR N,et al.Attention is all you need[C]∥Proceedings of the 31st International Conference on Neural Information Processing Systems.Long Beach,USA:Curran Associates Inc.,2017:6000-6010.

物理约束多源数据驱动机械钻速智能预测方法

An intelligent prediction method for rate of penetration driven by the physical constraint and multi-source data

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 2

编辑推荐

Metrics

本文评价

[1]	骆正山, 张景奇, 骆济豪, 王小完. 基于IMA-AmMLP模型的CO₂驱最小混相压力预测[J]. 石油学报, 2024, 45(10): 1522-1528.
[2]	杨培杰, 穆星, 印兴耀. 叠前三参数同步反演方法及其应用[J]. 石油学报, 2009, 30(2): 232-236.