Intelligent recognition method for drilling conditions based on 1dCNN-BiGRU and attention mechanism

doi:10.3969/j.issn.2096-1693.2025.03.010

Abstract

Abstract:

This study addresses the challenges of poor real-time performance and low accuracy in drilling condition identification by introducing an innovative intelligent recognition method. The proposed approach integrates a one-dimensional convolutional neural network (1dCNN) for local feature extraction, a bidirectional gated recurrent unit (BiGRU) to capture sequential dependencies, and a multi-head attention mechanism to emphasize critical information. This fusion enables efficient discrimination among 13 drilling conditions, including rotary drilling, slide drilling, whipstocking, and reverse whipstocking. In the model design phase, comprehensive ablation studies were conducted to evaluate the contributions of each module—1dCNN, BiGRU, self-attention, and multi-head attention—as well as their serial and parallel configurations. The performance was further optimized using the Optuna framework for automatic hyperparameter tuning. Experimental results demonstrated that the model achieved an accuracy of 96.22% on time-domain data from a single well. Additionally, in both intra- and inter-block transfer tests, the overall accuracy ranged from 94% to 97%, with each drilling condition exceeding an 80% recognition rate. Real-time testing on field data also showed a high degree of consistency with actual operational conditions. Overall, the proposed method provides a robust technical framework for real-time monitoring and optimization of drilling operations.

Key words: drilling, drilling conditions, artificial intelligence, neural networks, self-attention

摘要：

本研究针对钻井工况识别中实时性差、准确率低的问题，提出了一种创新的智能识别方法。该方法融合一维卷积神经网络(1dCNN)用于局部特征提取、双向门控循环单元(BiGRU)捕捉时序依赖以及多头注意力机制对齐时间域并强化关键信息，从而实现对旋转钻进、滑动钻进、划眼、倒划眼等13种工况的高效区分。在模型设计上，通过消融实验系统评估了各模块(1dCNN、BiGRU、自注意力与多头注意力机制)及其串并联结构的贡献，并借助Optuna自动调参进一步优化了性能。实验结果表明：在单井时间域数据测试中，工况识别准确率达96.22%；在同区块及跨区块井数据迁移测试中，整体准确率保持在94%~97%，且各工况识别率均超过80%；此外，实时数据测试结果与现场实际操作高度吻合。该方法为钻井工况的实时监控与作业优化提供了有力技术支撑。

关键词: 油气钻井, 钻井工况, 人工智能, 神经网络, 注意力机制

CLC Number:

TE2
TP18

WANG Zheng, SONG Xianzhi, LI Hongsong, YU Jiawei, WANG Yifan, ZHANG Chongyuan. Intelligent recognition method for drilling conditions based on 1dCNN-BiGRU and attention mechanism[J]. Petroleum Science Bulletin, 2025, 10(5): 926-940.

王正, 宋先知, 李洪松, 于佳伟, 王一帆, 张重愿. 基于1dCNN-BiGRU和注意力机制的钻井工况智能识别方法[J]. 石油科学通报, 2025, 10(5): 926-940.

/ / Recommend / Download Citations

URL: http://sykxtb.dgqinyehang.com/EN/10.3969/j.issn.2096-1693.2025.03.010

http://sykxtb.dgqinyehang.com/EN/Y2025/V10/I5/926

Figures/Tables 21

References 31

[1]	钟仪华, 刘雨鑫, 林旭旭, 等. 基于马尔科夫链和贝叶斯网络的钻井风险预测[J]. 石油钻采工艺, 2016, 38(3): 5.
	[ZHONG Y H, LIU Y X, LIN X X, et al. Drilling risk prediction based on Markov chain and Bayesian network[J]. Petroleum Drilling and Production Technology, 2016, 38(3): 5.
[2]	YANG A X, WU M, HU J, et al. Discrimination and correction of abnormal data for condition monitoring of drilling process[J]. Neurocomputing, 2021, 433: 275-286.
[3]	李根生, 宋先知, 祝兆鹏, 等. 智能钻完井技术研究进展与前景展望[J]. 石油钻探技术, 2023, 51(4): 35-47.
	[LI G S, SONG X Z, ZHU Z P, et al. Research progress and prospects of intelligent drilling and completion technology[J]. Petroleum Drilling Technology, 2023, 51(4): 35-47]
[4]	李根生, 宋先知, 田守嶒, 等. 智能钻井技术研究现状及发展趋势[J]. 石油钻探技术, 2020, 48(1): 1-8.
	[LI G S, SONG X Z, ZHU Z P, et al. Research status and development trends of intelligent drilling technology[J] Petroleum Drilling Technology, 2020, 48(1): 1-8.]
[5]	张菲菲, 崔亚辉, 于琛, 等. 基于机器学习的钻井工况识别技术现状及发展[J]. 长江大学学报: 自然科学版, 2023, 20(4): 53-65.
	[ZHANG F F, CUI Y H, YU C, et al. The current status and development of drilling condition recognition technology based on machine learning[J]. Journal of Changjiang University: Natural Science Edition, 2023, 20(4): 53-65]
[6]	孙挺, 赵颖, 杨进, 等. 基于支持向量机的钻井工况实时智能识别方法[J]. 石油钻探技术, 2019, 47(5): 28-33.
	[SUN T, ZHAO Y, YANG J, et al. Real time intelligent recognition method for drilling conditions based on support vector machine[J]. Petroleum Drilling Technology, 2019, 47(5): 28-33.]
[7]	TRIPATHI A M, RASHMI D, SUBBIAH S, et al. Oil well drilling activities recognition using a hierarchical classifier[J]. Journal of Petroleum Science and Engineering, 2021, 196: 107883.
[8]	毛光黔, 宋先知, 丁燕, 等. 基于梯度提升决策树算法的钻井工况识别方法[J]. 石油钻采工艺, 2023, 45(5): 532-539.
	[MAO G Q, SONG X Z, DING Y, et al. A drilling condition recognition method based on gradient boosting decision tree algorithm[J]. Petroleum Drilling and Production Technology, 2023, 45(5): 532-539.]
[9]	BEN Y X, CHRIS J, CAO D Z. Development and Application of a Real-Time Drilling State Classification Algorithm with Machine Learning[C]// SPE/AAPG/SEG Unconventional Resources Technology Conference (2019). Denver, Colorado: Society of Petroleum Engineers, 2019: 3053-3066.
[10]	侯欣欣. 基于录井数据的钻井工况识别与时效分析[D]. 北京: 中国免费靠逼视频(北京), 2020.
	[HOU X X. Identification and timeliness analysis of drilling conditions based on logging data[D]. Beijing: China University of Petroleum, 2020.]
[11]	谯英, 林慧, 周文俊, 等. 基于CNN-BiGRU并行混合网络的钻井工况智能识别[C]// 2022年中国油气智能科技大会论文集. 北京: 中国免费靠逼视频出版社, 2022: 123-133.
	[QIAO Y, LIN H, ZHOU W J, et al. Intelligent recognition of drilling conditions based on CNN BiGRU parallel hybrid network[C]//. Proceedings of the 2022 China Oil and Gas Intelligent Technology Conference. Beijing: China University of Petroleum Press, 2022: 123-133.]
[12]	WANG C, LIU G H, YANG Z R, et al. Downhole working conditions analysis and drilling complications detection method based on deep learning[J]. Journal of Natural Gas Science and Engineering, 2020, 81: 103485.
[13]	王海涛, 王建华, 邱晨, 等. 基于双向长短期记忆循环神经网络和条件随机场的钻井工况识别方法[J]. 石油钻采工艺, 2023, 45(5): 540-547.
	[WANG H T, WANG J H, QIU C, et al. A drilling condition recognition method based on bidirectional long short-term memory recurrent neural network and conditional random field[J]. Petroleum Drilling and Production Technology, 2023, 45(5): 540-547.]
[14]	YANN L, LEON B. Gradient-based learning applied to document recognition[J]. Proceedings of the IEEE, 1998, 86(11): 2278-2324.
[15]	胡越, 罗东阳, 花奎, 等. 关于深度学习的综述与讨论[J]. 智能系统学报, 2019, 14(1): 19.
	[HU Y, LUO D Y, HUA K, et al. A review and discussion on deep learning[J]. Journal of Intelligent Systems, 2019, 14(1): 19]
[16]	WEI Z, ZHANG G G, ZHAO C C, et al. Multichannel consecutive data cross-extraction with 1DCNN-attention for diagnosis of power transformer[J]. International Journal of Electrical Power & Energy Systems, 2024, 158: 109951.
[17]	DAVID O, TEMIDAYO O O, EMMANUEL G D. Consumer complaints of consumer financial protection bureau via two-stage residual one-dimensional convolutional neural network (TSR1DCNN)[J]. Data and Information Management, 2023, 7(4): 100046.
[18]	CHO K, DZMITRY B, BOUGARES F, et al. Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translation[C]// Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP). Doha, Qatar: Association for Computational Linguistics, 2014: 1724-1734.
[19]	WANG J T, WEN X T, HE Y L, et al. Logging curve prediction based on a CNN-GRU neural network[J]. Geophysical Prospecting for Petroleum, 2022. 61(2): 276-285. doi: 10.3969/j.issn.1000-1441.2022.02.009
[20]	ASHISH V, NOAM S, NIKI P, et al. Attention is all you need[C]//Advances In Neural Information Processing Systems 30: Annual Conference On Neural Information Processing Systems 2017, CA, USA.
[21]	HU Z F, CHEN L B J, LOU Y, et al. EEG-Based Emotion Recognition Using Convolutional Recurrent Neural Network with Multi-Head Self-Attention[J]. Applied Sciences, 2022, 12(21): 11255.
[22]	LIAN Z, LIU B, TAO J H. CTNet: Conversational Transformer Network for Emotion Recognition. IEEE/ACM Trans. Audio, Speech and Lang. IEEE/ACM Transactions on Audio, Speech, and Language Processing[J], 2021. 29: 985-1000.
[23]	LIU T Y, ZOU B H, HE M M, et al. LncReader: identification of dual functional long noncoding RNAs using a multi-head self-attention mechanism[J]. Briefings in Bioinformatics, 2023, 24(1): 579.
[24]	朱张莉, 饶元, 吴渊, 等. 注意力机制在深度学习中的研究进展[J]. 中文信息学报, 2019, 33(6): 1-11.
	[ZHU Z L, RAO Y, WU Y, et al. Research progress on attention mechanisms in deep learning[J]. Chinese Journal of Information Science, 2019, 33(6): 1-11]
[25]	刘建伟, 刘俊文, 罗雄麟, 等. 深度学习中注意力机制研究进展[J]. 工程科学学报, 2021, 43(11): 1499-1511.
	[LIU J W, LIU J W, LUO X L, et al. Research progress on attention mechanisms in deep learning[J]. Journal of Engineering Science, 2021, 43 (11): 1499-1511]
[26]	ALAA A H, RAFID S, RAAID A, et al. Oil spill detection based on machine learning and deep learning: A review[C]//In 2022 5th International Conference on Engineering Technology and its Applications (IICETA). 2022.
[27]	肖立志. 机器学习数据驱动与机理模型融合及可解释性问题[J]. 石油物探, 2022, 61(2): 205-212. doi: 10.3969/j.issn.1000-1441.2022.02.002
	[XIAO L Z. Machine learning data-driven and mechanistic model fusion and interpretability issues[J]. Petroleum Geophysical Exploration, 2022, 61(2): 205-212]
[28]	CHEN Y T, ZHANG D X. Physics-constrained deep learning of geomechanical logs[J]. IEEE Transactions on Geoscience and Remote Sensing, 2020. 58(8): 5932-5943.
[29]	DU M G, CHEN Y T, ZHANG D X, et al. AutoKE: An automatic knowledge embedding framework for scientific machine learning[J]. IEEE Transactions on Artificial Intelligence, 2023, 4(6): 1564-1578.
[30]	BIRD J J, CHLOE M, MANSO L, et al. Fruit quality and defect image classification with conditional GAN data augmentation[J]. Scientia Horticulturae, 2022, 293: 110684.
[31]	周志华. 机器学习[M]. 北京: 清华大学出版社, 2016: 38-39.
	[ZHOU Z H. Machine learning[M]. Beijing: Tsinghua University Press, 2016: 38-39.]

钻井工况	工况描述
旋转钻进	钻杆带动钻头旋转破碎岩石
滑动钻进	用于定向钻井，钻杆不旋转，依靠井下动力钻具带动钻头旋转
接卸单根	在原有的钻柱上增加或减少单根、立柱的过程
坐卡	钻杆钻具坐在卡瓦上，此时大钩载荷很低
循环	通过钻井泵泵送钻井液，并且钻柱在循环过程中不转动
转动循环	通过钻井泵泵送钻井液，并且钻柱在循环过程中转动
向上洗井	通过钻井泵泵送钻井液，钻杆不转动，钻头位置向上移动，清除岩屑
向下洗井	通过钻井泵泵送钻井液，钻杆不转动，钻头位置向下移动，清除岩屑
划眼	通过钻井泵泵送钻井液，钻杆转动，钻头位置向下移动，修整下部井壁
倒划眼	通过钻井泵泵送钻井液，钻杆转动，钻头位置向上移动，修整上部井壁
起钻	从井筒中移出钻杆，通常是进行钻头钻具的更换，或者设备的维护
下钻	将钻杆下入井筒中，通常是需要进行钻进操作
井下施工	通常需要将全部的钻杆、钻具移出井筒，进行固井、完井、测井施工作业

钻井工况	工况描述
旋转钻进	钻杆带动钻头旋转破碎岩石
滑动钻进	用于定向钻井，钻杆不旋转，依靠井下动力钻具带动钻头旋转
接卸单根	在原有的钻柱上增加或减少单根、立柱的过程
坐卡	钻杆钻具坐在卡瓦上，此时大钩载荷很低
循环	通过钻井泵泵送钻井液，并且钻柱在循环过程中不转动
转动循环	通过钻井泵泵送钻井液，并且钻柱在循环过程中转动
向上洗井	通过钻井泵泵送钻井液，钻杆不转动，钻头位置向上移动，清除岩屑
向下洗井	通过钻井泵泵送钻井液，钻杆不转动，钻头位置向下移动，清除岩屑
划眼	通过钻井泵泵送钻井液，钻杆转动，钻头位置向下移动，修整下部井壁
倒划眼	通过钻井泵泵送钻井液，钻杆转动，钻头位置向上移动，修整上部井壁
起钻	从井筒中移出钻杆，通常是进行钻头钻具的更换，或者设备的维护
下钻	将钻杆下入井筒中，通常是需要进行钻进操作
井下施工	通常需要将全部的钻杆、钻具移出井筒，进行固井、完井、测井施工作业

输入参数	旋转钻进	滑动钻进	接卸单根	坐卡	循环	转动循环	向上洗井	向下洗井	划眼	倒划眼	起钻	下钻	井下施工
钻压	>0	>0	0	—	0	0	0	0	0	0	0	0	0
扭矩	>0	0	0	—	0	>0	0	0	>0	>0	0	0	0
转速	>0	0	0	0	0	>0	0	0	>0	>0	0	0	0
立管压力	>0	>0	0	0	>0	>0	>0	>0	>0	>0	0	0	0
大钩高度	↓	↓	—	—	—	—	↑	↓	↓	↑	↑	↓	/
大钩载荷	↑	↑	—	—	—	—	↑	↓	↓	↑	↑	↓	/
入口流量	>0	>0	0	0	>0	>0	>0	>0	>0	>0	0	0	0
钻头位置	↓	↓	—	—	—	—	↑	↓	↓	↑	↑	↓	/
钻头位置/井深	≥1	≥1	<1	<1	<1	<1	<1	<1	<1	<1	<1	<1	<1

输入参数	旋转钻进	滑动钻进	接卸单根	坐卡	循环	转动循环	向上洗井	向下洗井	划眼	倒划眼	起钻	下钻	井下施工
钻压	>0	>0	0	—	0	0	0	0	0	0	0	0	0
扭矩	>0	0	0	—	0	>0	0	0	>0	>0	0	0	0
转速	>0	0	0	0	0	>0	0	0	>0	>0	0	0	0
立管压力	>0	>0	0	0	>0	>0	>0	>0	>0	>0	0	0	0
大钩高度	↓	↓	—	—	—	—	↑	↓	↓	↑	↑	↓	/
大钩载荷	↑	↑	—	—	—	—	↑	↓	↓	↑	↑	↓	/
入口流量	>0	>0	0	0	>0	>0	>0	>0	>0	>0	0	0	0
钻头位置	↓	↓	—	—	—	—	↑	↓	↓	↑	↑	↓	/
钻头位置/井深	≥1	≥1	<1	<1	<1	<1	<1	<1	<1	<1	<1	<1	<1

神经网络模块	层定义	层参数
BiGRU	GRU层	隐藏节点数64，层数2，双向网络
1dCNN	卷积层1	核数量9，核尺寸1，步长1，无填充
注意力层	全连接层1	隐藏节点数8
	全连接层2	隐藏节点数8
	全连接层3	隐藏节点数8，Softmax激活函数

Please choose a citation manager

Content to export