Federated Dictionary Learning-Based Intelligent Monitoring for Geological Drilling Processes
-
摘要: 针对地质钻进过程中多井场数据分布差异显著、隐私约束严格且人工标注缺失等现实问题，提出一种基于联邦字典学习的智能监测方法. 该方法以稀疏字典表征为核心，通过事件启发式打分与对齐机制实现无人工标签条件下的异常模式识别与可解释区分；同时在数据不出域的前提下，引入多井协同训练与服务器端加权聚合，以提升模型在异质井场间的鲁棒性与跨井泛化能力. 基于多个实际钻井工程的现场测井数据，本文构建多井场联邦数据集并开展验证实验. 结果表明，所提方法在多井场条件下取得更优的监测效果，分离度指标的平均水平达到4.018；弱监督精度的平均水平达到0.804，显示其在无标注条件下仍能稳定识别典型异常事件. 研究结果验证了事件标注机制在提升模型可解释性与跨井泛化能力方面的重要作用，为复杂地质环境下的分布式智能监测提供了一种可行且有效的技术途径.Abstract: To address practical challenges in geological drilling, including pronounced inter-well distribution shifts, stringent privacy constraints, and the lack of manual annotations, this paper proposes an intelligent monitoring method based on federated dictionary learning. Centered on sparse dictionary representations, the proposed approach integrates an event-driven heuristic scoring and alignment mechanism to enable interpretable discrimination and abnormal pattern recognition without requiring human-labeled data. Meanwhile, under a data-locality constraint where raw data never leave local sites, the method incorporates collaborative multi-well dictionary training and a sample-size-weighted server-side aggregation scheme to enhance robustness and cross-well generalization across heterogeneous drilling environments. Experiments are conducted on field logging data collected from multiple real-world drilling projects, where a multi-well federated dataset is constructed for evaluation. The results demonstrate that the proposed method achieves superior monitoring performance under multi-well settings, with an average separability score of 4.018 and an average weakly supervised precision of 0.804, indicating stable identification of typical abnormal events in the absence of annotations. These findings substantiate the effectiveness of the event labeling mechanism in improving model interpretability and cross-well generalization, and provide a feasible and effective technical pathway for distributed intelligent monitoring in complex geological environments.
-
图 3 不同特征子集与方法的重建误差及距离度量对比
a. 本文方法下的井场7监测结果图；b. 本文方法下的井场2监测结果图；c. E-DTWA方法下的井场7监测结果图；d. E-DTWA方法下的井场2监测结果图；e. PCA方法下的井场7监测结果图；f. PCA方法下的井场2监测结果图；g. A-BMIL方法下的井场7监测结果图；h. A-BMIL方法下的井场2监测结果图
Fig. 3. Comparison of reconstruction errors and distance metrics across different feature subsets and methods
表 1 钻进典型异常事件机理与特征对比
Table 1. Comparison of mechanisms and observable characteristics of typical drilling anomalies
异常类型 主要诱因 典型信号变化 粘滑振动 钻柱扭转惯性与井壁摩擦耦合 扭矩、转速周期性波动 卡钻 井壁塌方、差压吸附、井径缩小 钩载、钻压急剧升高，钻柱受限 井漏与井涌 井底与地层压力失衡 立管压力突变、环空流量异常 喷嘴堵塞 岩屑堆积、泥浆固相 泵压上升、流量下降 钻头磨损 硬地层钻进、携屑不畅 钻速下降、扭矩缓慢上升 振动与冲击类异常 非均质地层、钻具共振耦合 钩载、扭矩高频波动 
下载: 导出CSV
表 2 USROP数据集主要监测变量
Table 2. Main monitored variables in the USROP dataset.
变量 缩写 单位 measured depth - m weight on bit WOB kgf average standpipe pressure SPP kPa average surface torque TQ kN/m rate of penetration ROP m/h average rotary speed - rpm mud flow Q L/min mud density p g/cm3 diameter - mm average hookload HKLD kgf
下载: 导出CSV
表 3 各井场下各方法的弱监督精度Pw与分离度d对比
Table 3. Cross-well comparison of $ {P}_{\mathrm{w}} $ and separation $ d $ for the evaluated methods
方法 指标 井场1 井场2 井场3 井场4 井场5 井场6 井场7 平均 本文方法 Pw 0.750 0.704 0.684 0.834 0.887 0.920 0.852 0.804 d 6.876 2.939 3.674 4.538 2.987 3.875 3.239 4.018 E-DTWA
(Kloska et al., 2023)Pw / / / / / / / / d 3.667 3.845 2.321 2.654 2.069 1.964 1.785 2.615 PCA
(Jaffel et al., 2014)Pw / / / / / / / / d 3.067 2.316 3.321 2.458 1.983 1.542 1.061 2.249 A-BMIL
(Kong et al., 2019)Pw 0.687 0.535 0.704 0.654 0.554 0.768 0.329 0.604 d 3.688 2.636 1.644 2.569 4.872 3.065 1.801 2.639
下载: 导出CSV
-
Altindal, M. C., Nivlet, P., Tabib, M., et al., 2024. Anomaly Detection in Multivariate Time Series of Drilling Data. Geoenergy Science and Engineering, 237: 212778. https://doi.org/10.1016/j.geoen.2024.212778 Brito, L. C., Susto, G. A., Brito, J. N., et al., 2023. Fault Diagnosis Using Explainable AI: a Transfer Learning-Based Approach for Rotating Machinery Exploiting Augmented Synthetic Data. Expert Systems with Applications, 232: 120860. https://doi.org/10.1016/j.eswa.2023.120860 Chen, X., He, M., Xu, M. B., et al., 2023. Early Gas Kick Detection-Inversion-Control Integrated System: The Significance of Applications of Managed Pressure Drilling: a Review. Geoenergy Science and Engineering, 229: 212134. https://doi.org/10.1016/j.geoen.2023.212134 Choe, Y. M., Kim, G. S., Kim, I. S., et al., 2023. Influence of Torsional Stick-Slip Vibration on Whirl Behavior in Drill String System. Geoenergy Science and Engineering, 227: 211931. https://doi.org/10.1016/j.geoen.2023.211931 Deng, G. Q., Yao, A. G., Gong, Z., 2017. Real-Time Positioning Method for Horizontal Directional Drilling Based on Ground Magnetic Beacons, Earth Science, 42(12): 2336-2344(in Chinese with English abstract). Faghihi, M. A., Mohammadi, H., Yazdi, E. A., et al., 2024. Distributed Model for the Drill-String System with Multiple Regenerative Effects in the Bit-Rock Interaction. Journal of Sound and Vibration, 571: 118120. https://doi.org/10.1016/j.jsv.2023.118120 He, J. H., Cheng, Z. J., Guo, B., 2024. Anomaly Detection in Telemetry Data Using a Jointly Optimal One-Class Support Vector Machine with Dictionary Learning. Reliability Engineering & System Safety, 242: 109717. https://doi.org/10.1016/j.ress.2023.109717 Huang, K. L., Feng, Y. Q., Zhou, C. X., et al., 2024. Analysis of Bit-Balling Failure and Hydraulic Structure Optimization Design of Annular-Grooved PDC Bits. Geoenergy Science and Engineering, 241: 213134. https://doi.org/10.1016/j.geoen.2024.213134 Huang, K. K., Zhang, L., Wu, D. H., et al., 2023. Nonstationary Industrial Process Monitoring Based on Stationary Projective Dictionary Learning. IEEE Transactions on Control Systems Technology, 31(3): 1122-1132. https://doi.org/10.1109/TCST.2022.3210407 Huang, W. K., Ye, M., Shi, Z. K., et al., 2024. Federated Learning for Generalization, Robustness, Fairness: a Survey and Benchmark. IEEE Transactions on Pattern Analysis and Machine Intelligence, 46(12): 9387-9406. https://doi.org/10.1109/TPAMI.2024.3418862 Jaffel, I., Taouali, O., Elaissi, E., et al., 2014. A New Online Fault Detection Method Based on PCA Technique. IMA Journal of Mathematical Control and Information, 31(4): 487-499. https://doi.org/10.1093/imamci/dnt025 Jia, X. D., Xun, P., Peng, W., et al., 2025. Deep Anomaly Detection for Time Series: a Survey. Computer Science Review, 58: 100787. https://doi.org/10.1016/j.cosrev.2025.100787 Kang, Y. Q., Liu, J., Wu, F., et al., 2024. Deep Convolutional Dictionary Learning Network for Sparse View CT Reconstruction with a Group Sparse Prior. Computer Methods and Programs in Biomedicine, 244: 108010. https://doi.org/10.1016/j.cmpb.2024.108010 Kloska, M., Grmanova, G., Rozinajova, V., 2023. Expert Enhanced Dynamic Time Warping Based Anomaly Detection. Expert Systems with Applications, 225: 120030. https://doi.org/10.1016/j.eswa.2023.120030 Kong, Q. Q., Yu, C. S., Xu, Y., et al., 2019. Weakly Labelled AudioSet Tagging with Attention Neural Networks. IEEE/ACM Transactions on Audio, Speech, and Language Processing, 27(11): 1791-1802. https://doi.org/10.1109/TASLP.2019.2930913 Li, H., Sun, X. Y., 2025. Research on Coal-Rock Interface Identification Technology Based on Measurement-While-Drilling Vibration Signals. Mine Construction Technology, 46(4): 61-65(in Chinese with English abstract). Liang, Z. Y., Liang, Z., Wang, H. Z., et al., 2025. FedDict: Towards Practical Federated Dictionary-Based Time Series Classification. IEEE Transactions on Knowledge and Data Engineering, 37(4): 1785-1803. https://doi.org/10.1109/TKDE.2025.3528023 Liu, H. F., Chen, J. X., Dy, J., et al., 2023. Transforming Complex Problems into K-Means Solutions. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(7): 9149-9168. https://doi.org/10.1109/TPAMI.2023.3237667 Nüsse, P. M., Ambrus, A., Aarsnes, U. J. F., et al., 2023. Evaluation of Distributed Damping Subs with Active Control for Stick-Slip Reduction in Drilling. Geoenergy Science and Engineering, 231: 212255. https://doi.org/10.1016/j.geoen.2023.212255 Qin, Y., Yang, R., He, B., et al., 2024. Multi-Layer Convolutional Dictionary Learning Network for Signal Denoising and Its Application to Explainable Rolling Bearing Fault Diagnosis. ISA Transactions, 147: 55-70. https://doi.org/10.1016/j.isatra.2024.01.027 Qu, F. T., Liao, H. L., Lu, M., et al., 2024. Recognition of Drill String Vibration State Based on WGAN-Div and CNN-IWPSO-SVM. Geoenergy Science and Engineering, 243: 213342. https://doi.org/10.1016/j.geoen.2024.213342 Rakhi, Gupta, B., Lamba, S. S., 2024. An Efficient Local Outlier Detection Approach Using Kernel Density Estimation. Franklin Open, 8: 100162. https://doi.org/10.1016/j.fraope.2024.100162 Shao, C. F., Ge, Z. L., Zhou, Z., et al., 2023. Experimental and Numerical Investigation of Abrasive Water Jet Nozzle Erosion. Powder Technology, 430: 119031. https://doi.org/10.1016/j.powtec.2023.119031 Siqueira, V. S., Cuadros, M. A. S. L., Munaro, C. J., et al., 2024. Expert System for Early Sign Stuck Pipe Detection: Feature Engineering and Fuzzy Logic Approach. Engineering Applications of Artificial Intelligence, 127: 107229. https://doi.org/10.1016/j.engappai.2023.107229 Sorbo, S., Ruocco, M., 2024. Navigating the Metric Maze: a Taxonomy of Evaluation Metrics for Anomaly Detection in Time Series. Data Mining and Knowledge Discovery, 38(3): 1027-1068. https://doi.org/10.1007/s10618-023-00988-8 Xia, T., Cheng, C., Pang, Q. Z., 2023. Safety Risk Early Warning of Deep Foundation Pit Deformation Based on Long Short-Term Memory Network, Earth Science, 48(10): 3925-3931(in Chinese with English abstract). Xiao, P., Wang, C., Lin, Z., et al., 2024. Knowledge-Based Clustering Federated Learning for Fault Diagnosis in Robotic Assembly. Knowledge-Based Systems, 294: 111792. https://doi.org/10.1016/j.knosys.2024.111792 Yao, R. H., Jiang, H. K., Jiang, W. X., et al., 2024. Deep Discriminative Sparse Representation Learning for Machinery Fault Diagnosis. Engineering Applications of Artificial Intelligence, 135: 108836. https://doi.org/10.1016/j.engappai.2024.108836 Zhang, B. H., Xue, J. R., 2023. Weakly-Supervised Anomaly Detection with a Sub-Max Strategy. Neurocomputing, 560: 126770. https://doi.org/10.1016/j.neucom.2023.126770 Zhang, C. G., Liu, T., Du, X. J., 2023. A Deep Neural Network Regularization Measure: The Class-Based Decorrelation Method. Entropy, 26(1): 7. https://doi.org/10.3390/e26010007 邓国庆, 姚爱国, 龚正, 等, 2017. 基于地面磁信标的水平定向钻进实时定位方法, 地球科学, 42(12): 2336-2344. doi: 10.3799/dqkx.2017.567 李恒, 孙晓宇, 2025. 基于随钻振动信号的煤岩边界识别技术研究, 建井技术, 46(4): 61-65. 夏天, 成诚, 庞奇志, 2023. 基于长短时记忆网络的深基坑变形安全风险预警, 地球科学, 48(10): 3925-3931. doi: 10.3799/dqkx.2021.250 -
点击查看大图
图(3) / 表(3)
计量
- 文章访问数: 285
- HTML全文浏览量: 40
- PDF下载量: 56
- 被引次数: 0
