Downhole WOB Prediction Method Based on CNN-Bi-LSTM Network Optimized by TDCSO
-
摘要:
为了准确预测井底钻压，提高钻井效率、降低钻井成本，建立了融合双向长短期记忆网络（Bi-LSTM）和卷积神经网络（CNN）的混合模型。采用三角函数驱动的粒子群优化（TDCSO）方法对模型进行超参数优化，以提高预测钻压的精度；采用美国犹他州FORGE 58−32井和FORGE 58−62井的2个公开数据集对建立的模型进行验证，并采用平均绝对误差、均方根误差、决定系数和均方误差等指标进行模型性能评估。研究结果表明，所提出TDCSO-CNN-Bi-LSTM模型平均绝对误差、均方误差和均方根误差等3个关键性能指标较好，其中决定系数大于0.980，明显优于现有的LSTM、GRU、CNN-LSTM、CNN-Bi-LSTM等方法。研究表明，所提出的TDCSO-CNN-Bi-LSTM模型在井底钻压预测方面具有出色的准确性，能够实现实时监测，并与自动钻进系统集成，实现对钻压的精准控制，不仅提高了钻井效率，还降低了钻井成本，对未来的钻井作业具有重要的实际应用价值。
Abstract:In order to accurately predict downhole weight on bit (WOB), improve drilling efficiency, and reduce drilling cost, a hybrid model combining bidirectional long short-term memory network (Bi-LSTM) and convolutional neural network (CNN) was established. The model used the trigonometric function-driven particle swarm optimization (TDCSO) method for hyperparameter optimization, so as to improve the accuracy of WOB prediction. The public data sets of Well FORGE 58−32 and Well FORGE 58−62 in Utah were used to verify the established model, and the model performance was evaluated by the mean absolute error (MAE), root mean square error (RMSE), coefficient of determination, and mean square error (MSE). The results show that the proposed TDCSO-CNN-Bi-LSTM model achieves excellent results in three key indicators of MAE, MSE, and RMSE, and the coefficient of determination was higher than 0.98, which is significantly better than the existing methods such as LSTM, GRU, CNN-LSTM and CNN-Bi-LSTM, etc. The study shows that the proposed TDCSO-CNN-Bi-LSTM model has excellent accuracy in downhole WOB prediction and enables real-time monitoring. It can be integrated with an automated drilling system to achieve precise control of WOB. This not only improves drilling efficiency but also reduces drilling cost, which has important practical application value for future drilling operations.
-
Keywords:
- downhole WOB /
- LSTM /
- neural network /
- optimization algorithm /
- model optimization
-
-
![]()
图 3 基于TDCSO算法的CNN-Bi-LSTM网络流程
Figure 3. CNN-Bi-LSTM network process based on TDCSO algorithm
![]()
图 4 时变系数与适应度的变化曲线
Figure 4. The changing curve graph of time-varying coefficients and fitness values
![]()
图 5 单模型与混合模型的井底钻压拟合结果
Figure 5. Fitting effect of downhole WOB for single models and hybrid models
![]()
图 6 组合模型和TDCSO优化组合模型的钻压拟合结果
Figure 6. Fitting effect of WOB for combined models and TDCSO optimization models
表 1 CNN-Bi-LSTM模型组件和参数设置
Table 1 Components and parameter configuration of CNN-Bi-LSTM model
模型组件 含义 输出维度 卷积核/填充大小 激活函数 Conv1d 一维卷积层 6 559×103×299 3/1 ReLU MaxPool 最大池化层 3/0 Conv1d 一维卷积层 6 559×103×149 3/1 ReLU MaxPool 最大池化层 3/0 Conv1d 一维卷积层 6 559×103×74 3/1 ReLU MaxPool 最大池化层 3/0 Bi-LSTM 双向循环神经网络 74×6 559×206 ReLU Linear 线性层 6 559×1 
下载: 导出CSV
表 2 单模型与混合模型的性能评价结果
Table 2 Performance evaluation results of single models and hybrid models
模型 均方误差 平均绝对误差 决定系数 均方根误差 耗时/s Bi-RNN 0.000 499 0.010 320 0.999 551 0.017 317 1.36 Bi-GRU 0.000 347 0.007 540 0.999 779 0.012 139 1.33 Bi-LSTM 0.000 414 0.012 473 0.999 514 0.018 012 1.38 CNN-Bi-RNN 0.000 381 0.011 461 0.989 398 0.019 535 1.66 CNN-Bi-GRU 0.006 694 0.060 429 0.989 452 0.081 818 1.51 CNN-Bi-LSTM 0.000 322 0.014 398 0.996 958 0.019 691 1.98
下载: 导出CSV
表 3 基于TDCSO算法的模型性能评价结果
Table 3 Evaluation results of model performance based on TDCSO algorithm
模型 均方误差 平均绝对误差 决定系数 均方根误差 耗时/s TDCSO-CNN-Bi-RNN 0.000 321 0.0124 6320.982 398 0.019 932 2.34 TDCSO-CNN-Bi-GRU 0.000 869 0.048 492 0.943 452 0.039 028 2.98 TDCSO-CNN-Bi-LSTM 0.000 292 0.011 398 0.989 953 0.021 698 2.39
下载: 导出CSV
-
[1] FU Jianhong, REN Zhipeng, BAI Jing, et al. The friction-reducing principle and application of the drill string with a hydroos-cillator[J]. Journal of Petroleum Science & Engineering, 2018, 165: 453–461.
[2] XU Baolong, LIU Songyong, LI Hongsheng. Drill string’s axial force transfer law in slide directional drilling in underground coal mine[J]. Tunnelling and Underground Space Technology, 2022, 130: 104701. doi: 10.1016/j.tust.2022.104701
[3] WANG Chao, LIU Gonghui, LI Jun, et al. New methods of eliminating downhole WOB measurement error owing to temperature variation and well pressure differential[J]. Journal of Petroleum Science & Engineering, 2018, 171: 1420–1432.
[4] 吴泽兵,谷亚冰,姜雯,等. 基于遗传优化算法的井底钻压智能预测模型[J]. 石油钻采工艺,2023,45(2):151–159. WU Zebing, GU Yabing, JIANG Wen, et al. Intelligent prediction models of downhole weight on bit based on genetic optimization algorithm[J]. Oil Drilling & Production Technology, 2023, 45(2): 151–159.
[5] 李臻,宋先知,李根生,等. 基于双输入序列到序列模型的井眼轨迹实时智能预测方法[J]. 石油钻采工艺,2023,45(4):393–403. LI Zhen, SONG Xianzhi, LI Gensheng, et al. Real-time intelligent prediction of well trajectory based on dual-input sequence-to-sequence model[J]. Oil Drilling & Production Technology, 2023, 45(4): 393–403.
[6] 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 Haitao, WANG Jianhua, QIU Chen, et al. Recognition method of drilling conditions based on bi-directional long-short term memory recurrent neural network and conditional random field[J]. Oil Drilling & Production Technology, 2023, 45(5): 540–547.
[7] 张矿生,宫臣兴,陆红军,等. 基于集成学习的井漏智能预警模型及智能推理方法[J]. 石油钻采工艺,2023,45(1):47–54. ZHANG Kuangsheng, GONG Chenxing, LU Hongjun, et al. Intelligent early warning model and intelligent reasoning method based on integrated learning for loss circulation[J]. Oil Drilling & Production Technology, 2023, 45(1): 47–54.
[8] OYEDERE M, GRAY K. Torque-on-bit (TOB) prediction and optimization using machine learning algorithms[J]. Journal of Natural Gas Science and Engineering, 2020, 84: 103623. doi: 10.1016/j.jngse.2020.103623
[9] 徐海萍. 基于ARIMA-GRNN模型的油井产量动态预测[J]. 世界石油工业,2022,29(1):64–69. XU Haiping. Dynamic prediction of oil well yield based on the ARIMA-GRNN model[J]. World Oil Industry, 2022, 29(1): 64–69.
[10] 陈宇光. 基于组合模型的天然气管道短期负荷预测[J]. 石油化工自动化,2024,60(2):20–24. CHEN Yuguang. Short-term load forecasting of natural gas pipeline based on the combined model[J]. Petrochemical Industry Automation, 2024, 60(2): 20–24.
[11] 黄小龙,刘东涛,宋吉明,等. 基于大数据及人工智能的钻速实时优化技术[J]. 石油钻采工艺,2021,43(4):442–448. HUANG Xiaolong, LIU Dongtao, SONG Jiming, et al. Real-time ROP optimization technology based on big data and artificial intelligence[J]. Oil Drilling & Production Technology, 2021, 43(4): 442–448.
[12] 樊冬艳,杨灿,孙海,等. 基于时间序列相似性与机器学习方法的页岩气井产量预测[J]. 中国石油大学学报(自然科学版),2024,48(3):119–126. doi: 10.3969/j.issn.1673-5005.2024.03.013 FAN Dongyan, YANG Can, SUN Hai, et al. Shale gas well production forecasting based on time sequence similarity and machine learning methods[J]. Journal of China University of Petroleum (Edition of Natural Science), 2024, 48(3): 119–126. doi: 10.3969/j.issn.1673-5005.2024.03.013
[13] 肖京男,汪志明,魏建光,等. 改进LSSVM在水平井产能预测中的应用[J]. 石油钻探技术,2010,38(6):95–98. doi: 10.3969/j.issn.1001-0890.2010.06.021 XIAO Jingnan, WANG Zhiming, WEI Jianguang, et al. Application of LSSVM improved horizontal well productivity prediction[J]. Petroleum Drilling Techniques, 2010, 38(6): 95–98. doi: 10.3969/j.issn.1001-0890.2010.06.021
[14] PAN Shaowei, BO Yang, WANG Shukai, et al. Oil well production prediction based on CNN-LSTM model with self-attention mechanism[J]. Energy, 2023, 284: 128701. doi: 10.1016/j.energy.2023.128701
[15] HOCHREITER S, SCHMIDHUBER J. Long short-term memory[J]. Neural Computation, 1997, 9(8): 1735–1780. doi: 10.1162/neco.1997.9.8.1735
[16] O’SHEA K, NASH R. An introduction to convolutional neural networks[DB/OL]. （2015-12-02）[2024-06-10]. https://arxiv.org/abs/1511.08458. O’SHEA K,NASH R. An introduction to convolutional neural networks[DB/OL]. (2015-12-02)[2024-06-10]. https://arxiv.org/abs/1511.08458.
[17] ELMAZ F, EYCKERMAN R, CASTEELS W, et al. CNN-LSTM architecture for predictive indoor temperature modeling[J]. Building and Environment, 2021, 206: 108327. doi: 10.1016/j.buildenv.2021.108327
[18] 孙伟峰,冯剑寒,张德志,等. 结合LSTM自编码器与集成学习的井漏智能识别方法[J]. 石油钻探技术,2024,52(3):61–67. doi: 10.11911/syztjs.2024006 SUN Weifeng, FENG Jianhan, ZHANG Dezhi, et al. An intelligent lost circulation recognition method using LSTM-autoencoder and ensemble learning[J]. Petroleum Drilling Techniques, 2024, 52(3): 61–67. doi: 10.11911/syztjs.2024006
[19] 韩克宁,王伟,樊冬艳,等. 基于产量递减与LSTM耦合的常压页岩气井产量预测[J]. 油气藏评价与开发,2023,13(5):647–656. HAN Kening, WANG Wei, FAN Dongyan, et al. Production forecasting for normal pressure shale gas wells based on coupling of production decline method and LSTM model[J]. Petroleum Reservoir Evaluation and Development, 2023, 13(5): 647–656.
[20] CHEN Zhiming, PENG Dong, WANG Tianyi, et al. Pressure transient analysis in shale wells with heterogeneous fractures by using a deep learning based surrogate model[R]. IPTC 22333, 2022.
[21] RHANOUI M, MIKRAM M, YOUSFI S, et al. A CNN-BiLSTM model for document-level sentiment analysis[J]. Machine Learning and Knowledge Extraction, 2019, 1(3): 832–847. doi: 10.3390/make1030048
[22] KENNEDY J, EBERHART R. Particle swarm optimization[C]//Proceedings of ICNN’95-International Conference on Neural Networks. Piscataway, NJ: IEEE Press, 1995: 1942-1948.
[23] JANG J S R, SUN C T, MIZUTANI E. Neuro-fuzzy and soft computing-a computational approach to learning and machine intelligence [J]. IEEE Transactions on Automatic Control, 1997, 42(10): 1482–1484. doi: 10.1109/TAC.1997.633847
[24] MOAYEDI H, RAFTARI M, SHARIFI A, et al. Optimization of ANFIS with GA and PSO estimating α ratio in driven piles[J]. Engineering with Computers, 2020, 36(1): 227–238. doi: 10.1007/s00366-018-00694-w
[25] 周俊,陈璟华,刘国祥,等. 粒子群优化算法中惯性权重综述[J]. 广东电力,2013,26(7):6–12. ZHOU Jun, CHEN Jinghua, LIU Guoxiang, et al. Review of inertia weight in particle swarm optimization algorithm[J]. Guangdong Electric Power, 2013, 26(7): 6–12.
[26] 王鹏飞,任丽佳,高燕. 基于改进收缩因子的粒子群优化算法[J]. 电子科技,2022,35(5):14–18. WANG Pengfei, REN Lijia, GAO Yan. PSO algorithm based on improved contraction factor[J]. Electronic Science and Technology, 2022, 35(5): 14–18.
[27] 王俊伟,汪定伟. 粒子群算法中惯性权重的实验与分析[J]. 系统工程学报,2005,20(2):194–198. WANG Junwei, WANG Dingwei. Experiment and analysis of inertia weight in particle swarm optimization[J]. Journal of Systems Engineering, 2005, 20(2): 194–198.
[28] MIRJALILI S. SCA: a sine cosine algorithm for solving optimization problems[J]. Knowledge-Based Systems, 2016, 96: 120-133.
[29] NASH G, MOORE J. Utah FORGE: logs and data from deep well 58-32 (MU-ESW1): 1006[R]. Salt Lake City: Energy and Geoscience Institute at the University of Utah, 2018.
-
期刊类型引用(19)
1. 王川，陈秋帆，夏勇. 海底泥浆举升钻井系统钻遇天然气水合物时的动态风险分析. 安全与环境学报. 2024(02): 479-487 . 
百度学术
2. 翟诚，吴迪，秦冬冬. 天然气水合物注热分解诱发储层变形破坏的正交数值模拟研究. 特种油气藏. 2024(02): 112-119 . 百度学术
3. 刘芳，冯馨，孙皓宇，张旭辉. 水合物分解中深水基础抗拔性能模型试验研究. 防灾减灾工程学报. 2023(02): 359-365 . 百度学术
4. 贺保卫，马志宇，崔海朋，杜鹏. 基于Unity 3D的可燃冰开采环境监测模拟系统设计. 计算机应用与软件. 2023(08): 121-125+154 . 百度学术
5. 吴艳辉，代锐，张磊，朱志潜，高禹，刘楷，徐鹏，张雨. 深水井筒海水聚合物钻井液水合物生成抑制与堵塞物处理方法. 钻井液与完井液. 2023(04): 415-422 . 百度学术
6. 赵凯，李润森，冯永存，高伟，张振伟，窦亮彬，毕刚. 非均匀地应力场下水合物储层水平井井周塑性区分布. 中南大学学报(自然科学版). 2022(03): 952-962 . 百度学术
7. 王磊，杨进，李莅临，胡志强，柯珂，臧艳彬，孙挺. 深水含水合物地层钻井井口稳定性研究. 岩土工程学报. 2022(12): 2312-2318 . 百度学术
8. 王志刚，李小洋，张永彬，尹浩，胡晨，梁金强，黄伟. 海域非成岩天然气水合物储层改造方法分析. 钻探工程. 2021(06): 32-38 . 百度学术
9. 马永乐，张勇，刘晓栋，侯岳，杨金龙，宋本岭，刘涛，李荔. 海域天然气水合物低温抑制性钻井液体系. 钻井液与完井液. 2021(05): 544-551+559 . 百度学术
10. 王偲，谢文卫，张伟，陈靓，陈浩文. RMR技术在海域天然气水合物钻探中的适应性分析. 探矿工程(岩土钻掘工程). 2020(02): 17-23 . 百度学术
11. 史静怡，樊建春，武胜男，李磊. 深水井筒天然气水合物形成预测及风险评价. 油气储运. 2020(09): 988-996 . 百度学术
12. 李莅临，杨进，路保平，柯珂，王磊，陈柯锦. 深水水合物试采过程中地层沉降及井口稳定性研究. 石油钻探技术. 2020(05): 61-68 . 本站查看
13. 李子丰，韩杰. 海底天然气水合物开采的环境安全性探讨. 石油钻探技术. 2019(03): 127-132 . 本站查看
14. 李庆超，程远方，邵长春. 允许适度坍塌的水合物储层最低钻井液密度. 断块油气田. 2019(05): 657-661 . 百度学术
15. 牛洪波，于政廉，孙菁，徐加放. 天然气水合物动力学抑制剂与水分子相互作用研究. 石油钻探技术. 2019(04): 29-34 . 本站查看
16. 迟咏梅，徐松杰，曹玉廷，董坚. 新型两亲性聚酰胺的合成及性质. 应用化学. 2017(03): 269-275 . 百度学术
17. 庞维新，李清平，程艳，王炳明. 水合物堵塞治理工具和方法研究. 石油机械. 2016(03): 68-72 . 百度学术
18. 光新军，王敏生. 海洋天然气水合物试采关键技术. 石油钻探技术. 2016(05): 45-51 . 本站查看
19. 孙小辉，孙宝江，王志远，王金堂. 超临界CO_2钻井井筒水合物形成区域预测. 石油钻探技术. 2015(06): 13-19 . 本站查看
其他类型引用(13)
计量
- 文章访问数: 192
- HTML全文浏览量: 36
- PDF下载量: 46
- 被引次数: 32
