[1]罗俊如,丁言瑞,徐明华,等.基于深度AUC最大化算法的井漏风险预测[J].常州大学学报(自然科学版),2024,36(03):34-44.[doi:10.3969/j.issn.2095-0411.2024.03.005]
　LUO Junru,DING Yanrui,XU Minghua,et al.Lost circulation prediction based on deep AUC maximization[J].Journal of Changzhou University(Natural Science Edition),2024,36(03):34-44.[doi:10.3969/j.issn.2095-0411.2024.03.005]
点击复制

基于深度AUC最大化算法的井漏风险预测()
分享到：

常州大学学报(自然科学版)[ISSN:2095-0411/CN:32-1822/N]

卷:
第36卷
期数:
2024年03期
页码:
34-44
栏目:
计算机与信息工程
出版日期:
2024-05-28

文章信息/Info

Title:
Lost circulation prediction based on deep AUC maximization
文章编号:
2095-0411(2024)03-0034-11
作者:
罗俊如1;  丁言瑞1;  徐明华1;  胡 超1;  刘炳官2;  孔维军2;  马强维3;  石 林1
1.常州大学 阿里云大数据学院, 江苏 常州 213164; 2.中国石油化工股份有限公司 江苏油田分公司, 江苏 扬州 225009; 3. 江苏如通石油机械股份有限公司, 江苏 南通 226400
Author(s):
LUO Junru1;  DING Yanrui1;  XU Minghua1;  HU Chao1;  LIU Bingguan2;  KONG Weijun2;  MA Qiangwei3;  SHI Lin1
1.Aliyun School of Big Data, Changzhou University, Changzhou 213164, China; 2.Jiangsu Oil Field Branch, China Petroleum and Chemical Corporation, Yangzhou 225009, China; 3.Jiangsu Rutong Petro-Machinery Co., Ltd., Nantong 226400, China
关键词:
井漏;  非均衡分类;  深度学习;  AUC最大化
Keywords:
lost circulation;  imbalanced classification;  deep learning;  AUC maximization
分类号:
TP 391.4
DOI:
10.3969/j.issn.2095-0411.2024.03.005
文献标志码:
A
摘要:
井漏是石油天然气钻井过程中经常面临的一项重要挑战,它的发生会极大的降低钻井效率,增加钻井成本。运用人工智能技术实现井漏风险的精准预测具有重要意义。文章将钻井液泄露分类问题转化为不平衡分类问题。类别间的非均衡性和钻探特征间缺乏高度相关性,对传统的深度学习模型提出了挑战。在这种情况下,传统的准确率度量很难正确评估模型性能。此外,研究引入了一种称为FAUC-S的深度AUC最大化(DAM)算法实现井漏风险预测,该算法是通过关注困难样本的AUC损失来训练组合深度学习模型。实验中还应用了一些经典的深度学习模型实现井漏风险的分类。实验结果表明,与其他3个模型相比,FAUC-S获得了最高的精确度、召回率和F1分数,实验验证了FAUC-S模型优越的分类性能。因此,该深度模型的成功应用可以有效地帮助钻井队解决钻井问题。
Abstract:
Lost circulation is a significant challenge in oil and gas drilling, which can lead to various costly and time-consuming problems. It is of great significance to use artificial intelligence technology to accurately predict the risk of lost circulation. The lost circulation prediction problem was converted into an imbalanced classification problem, which pose challenges to traditional deep learning models due to the imbalance between categories and the lack of high correlation between drilling features. Accuracy is not an appropriate measurement for imbalanced classification algorithms. A deep AUC maximization(DAM)algorithm, which is called FAUC-S, is introduced in this paper. It trains a combination deep learning model by focusing on the AUC loss of hard samples(FAUC-S). Several traditional deep learning methods are also applied to classify lost circulation risk during oil exploration in the experiments. The result shows that the FAUC-S method achieved the highest accuracy, recall, and F1 score among the other three models. This confirms that the FAUC-S model has superior classification performance. Therefore, the successful implementation of this deep model can help drilling teams effectively solve drilling problems.

参考文献/References:

[1] CHEN G H, HE Y L, QIU Z S, et al. Research and application of fine identification method of lost circulation characteristics in drilling[J]. Petroleum Drilling Techniques, 2024, 52(1): 1-6.
[2] SUN J S, BAI Y R, CHENG R C, et al. Research progress and prospect of plugging technologies for fractured formation with severe lost circulation[J]. Petroleum Exploration and Development, 2021, 48(3): 732-743.
[3] 李大奇, 康毅力, 刘修善, 等. 裂缝性地层钻井液漏失动力学模型研究进展[J]. 石油钻探技术, 2013, 41(4): 42-47.
[4] APALEKE A S, AL-MAJED A, HOSSAIN M E. Drilling fluid: state of the art and future trend[C]// SPE North Africa Technical Conference and Exhibition. Cairo: SPE, 2012: SPE-149555-MS.
[5] 陈卓, 耿立军, 岳明, 等. 渤中13-2区块复杂薄弱层高强度承压堵漏技术研究及应用[J].常州大学学报(自然科学版), 2023, 35(4): 77-86.
[6] KALHOR MOHAMMADI M, RIAHI S, BOEK E S. An insight review on formation damage induced by drilling fluids[J]. Reviews in Chemical Engineering, 2023, 39(3): 387-415.
[7] KIRAN R, TEODORIU C, DADMOHAMMADI Y, et al. Identification and evaluation of well integrity and causes of failure of well integrity barriers(a review)[J]. Journal of Natural Gas Science and Engineering, 2017, 45: 511-526.
[8] MARDANIRAD S, WOOD D A, ZAKERI H. The application of deep learning algorithms to classify subsurface drilling lost circulation severity in large oil field datasets[J]. SN Applied Sciences, 2021, 3(9): 785.
[9] SULE I, IMTIAZ S, KHAN F, et al. Risk analysis of well blowout scenarios during managed pressure drilling operation[J]. Journal of Petroleum Science and Engineering, 2019, 182: 106296.
[10] WOOD D A, MARDANIRAD S, ZAKERI H. Effective prediction of lost circulation from multiple drilling variables: a class imbalance problem for machine and deep learning algorithms[J]. Journal of Petroleum Exploration and Production Technology, 2022, 12(1): 83-98.
[11] KUANG L C, LIU H, REN Y L, et al. Application and development trend of artificial intelligence in petroleum exploration and development[J]. Petroleum Exploration and Development, 2021, 48(1): 1-14.
[12] KAMGUE LENWOUE A R, LI Z H, TANG C F, et al. Recent advances and challenges of the application of artificial intelligence to predict wellbore instabilities during drilling operations[J]. SPE Drilling & Completion, 2023, 38(4): 645-662.
[13] ZHU X Y, HAO J, GUO Y H, et al. AUC maximization in imbalanced lifelong learning[C]//Proceedings of the Thirty-Ninth Conference on Uncertainty in Artificial Intelligence. Pittsburgh: ACM, 2023: 2574-2585.
[14] YUAN Z N, YAN Y, SONKA M, et al. Large-scale robust deep AUC maximization: a new surrogate loss and empirical studies on medical image classification[C]//2021 IEEE/CVF International Conference on Computer Vision(ICCV). Montreal: IEEE, 2021: 3020-3029.
[15] YUAN Z N, GUO Z S, CHAWLA N, et al. Compositional training for end-to-end deep AUC maximization[C]//International Conference on Learning Representations. [S.l.: s.n.], 2022.
[16] YANG T B, YING Y M. AUC maximization in the era of big data and AI: a survey[J]. ACM Computing Surveys, 2022, 55(8): 172.
[17] WANG X Y, LYU Y L, JING L P. Deep generative model for robust imbalance classification[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR). Seattle: IEEE, 2020: 14112-14121.
[18] JAN B, FARMAN H, KHAN M, et al. Deep learning in big data analytics: a comparative study[J]. Computers and Electrical Engineering, 2019, 75: 275-287.
[19] CARRINGTON A M, MANUEL D G, FIEGUTH P W, et al. Deep ROC analysis and AUC as balanced average accuracy, for improved classifier selection, audit and explanation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023, 45(1): 329-341.
[20] JOLOUDARI J H, MAREFAT A, ALI NEMATOLLAHI M, et al. Effective class-imbalance learning based on SMOTE and convolutional neural networks[J]. Applied Sciences, 2023, 13(6): 4006.
[21] YANG Z Y, XU Q Q, BAO S L, et al. AUC-oriented domain adaptation: from theory to algorithm[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023, 45(12): 14161-14174.
[22] HANLEY J A, MCNEIL B J. The meaning and use of the area under a receiver operating characteristic(ROC)curve[J]. Radiology, 1982, 143(1): 29-36.
[23] YING Y, WEN L, LYU S. Stochastic online AUC maximization[C]//Neural Information Processing Systems. 2016, 29: 451-459.
[24] XU S K, DING Y R, WANG Y H, et al. FAUC-S: deep AUC maximization by focusing on hard samples[J]. Neurocomputing, 2024, 571: 127172.
[25] SAITO K, SAENKO K. OVANet: one-vs-all network for universal domain adaptation[C]//2021 IEEE/CVF International Conference on Computer Vision(ICCV). Montreal: IEEE, 2021: 8980-8989.
[26] MWANGI B, TIAN T S, SOARES J C. A review of feature reduction techniques in neuroimaging[J]. Neuroinformatics, 2014, 12(2): 229-244.
[27] RAMSUNDAR B, ZADEH R B. TensorFlow for deep learning: from linear regression to reinforcement learning[M]. Sebastopol: O'Reilly Media, Inc., 2018.
[28] GAO Y, GLOWACKA D. Deep gate recurrent neural network[EB/OL].(2016-04-11)[2024-03-04]. http://arxiv.org/abs/1604.02910.pdf.
[29] YU K P, TAN L, MUMTAZ S, et al. Securing critical infrastructures: deep-learning-based threat detection in IIoT[J]. IEEE Communications Magazine, 2021, 59(10): 76-82.
[30] LUO J R, QIAO H, ZHANG B. A minimax probability machine for nondecomposable performance measures[J]. IEEE Transactions on Neural Networks and Learning Systems, 2023, 34(5): 2353-2365.

相似文献/References:

[1]宫闻浩,李朝玮,李 栋,等.基于异质集成的井漏预警模型[J].常州大学学报(自然科学版),2024,36(02):39.[doi:10.3969/j.issn.2095-0411.2024.02.005]
　GONG Wenhao,LI Chaowei,LI Dong,et al.Lost circulation early warning model based on heterogeneous integration[J].Journal of Changzhou University(Natural Science Edition),2024,36(03):39.[doi:10.3969/j.issn.2095-0411.2024.02.005]

备注/Memo

备注/Memo:
收稿日期: 2024-02-19。
基金项目: 中国石油-常州大学创新联合体资助项目(2021DQ06); 江苏省双创博士研究资助项目(JSSCBS20210885); 常州大学阿里云大数据学院研究资助项目(ZMF21020012)。
作者简介: 罗俊如(1987—), 男, 江西吉安人, 博士, 讲师。 通信联系人: 石林(1979—), E-mail: slcczu@cczu.edu.cn
更新日期/Last Update: 1900-01-01