旋转导向钻井工具系统实时测量的智能粒子滤波方法

盛立; 刘一凡; 高明; 周东华

doi:10.16383/j.aas.c250136

旋转导向钻井工具系统实时测量的智能粒子滤波方法

doi: 10.16383/j.aas.c250136 cstr: 32138.14.j.aas.c250136

盛立^1,,
刘一凡^1,,
高明^1,,
周东华^2,

1.
中国石油大学(华东)控制科学与工程学院青岛 266580
2.
东南大学自动化学院南京 214135

基金项目: 国家自然科学基金(62473379, 62173343, 62033008), 山东省自然科学基金(ZR2024MF072, ZR2022ZD34), 山东省泰山学者项目研究基金资助

详细信息

作者简介:
盛立：中国石油大学(华东)控制科学与工程学院教授. 2010年获得江南大学博士学位. 主要研究方向为随机控制与滤波、网络化控制系统以及现代系统的故障检测与诊断. E-mail: shengli@upc.edu.cn

刘一凡：中国石油大学(华东)控制科学与工程学院博士研究生. 主要研究方向为随机系统的滤波与故障诊断. E-mail: liuyifan202306@163.com

高明：中国石油大学(华东)控制科学与工程学院教授. 2009年获得江南大学博士学位. 主要研究方向为网络化控制系统的鲁棒控制与故障诊断. E-mail: gaoming@upc.edu.cn

周东华：东南大学自动化学院教授. 1990年获得上海交通大学博士学位. 主要研究方向为动态系统的故障诊断与容错控制, 故障预测与最优维护技术. 本文通信作者. E-mail: zdh@tsinghua.edu.cn

计量
- 文章访问数: 6
- HTML全文浏览量: 7
- 被引次数: 0
出版历程
- 收稿日期: 2025-04-02
- 录用日期: 2025-07-13
- 网络出版日期: 2025-09-19

Intelligent Particle Filter for Real-time Measurement of Rotary Steerable Drilling Tool System

1.
College of Control Science and Engineering, China University of Petroleum (East China), Qingdao 266580
2.
School of Automation, Southeast University, Nanjing 214135

Funds: Supported by National Natural Science Foundation of China (62473379, 62173343, 62033008), the Natural Science Foundation of Shandong Province (ZR2024MF072, ZR2022ZD34), and the Research Fund for the Taishan Scholar Project of Shandong Province of China

More Information

Author Bio:
SHENG Li　Professor at the College of Control Science and Engineering, China University of Petroleum (East China). He received his Ph.D. degree from Jiangnan University in 2010. His research interest covers stochastic control and filtering, networked control systems, and fault detection and diagnosis for modern systems

LIU Yi-Fan　Ph.D. candidate at the College of Control Science and Engineering, China University of Petroleum (East China). His research interest covers filtering and fault diagnosis for stochastic systems

GAO Ming　Professor at the College of Control Science and Engineering, China University of Petroleum (East China). She received her Ph.D. degree from Jiangnan University in 2009. Her research interest covers robust control and fault diagnosis for networked control systems

ZHOU Dong-Hua　Professor at the School of Automation, Southeast University. He received his Ph.D. degree from Shanghai Jiao Tong University in 1990. His research interest covers fault diagnosis, fault-tolerant control, fault prediction, and optimal maintenance for dynamic systems. Corresponding author of this paper

摘要

摘要: 针对旋转导向钻井工具系统中工具面角的实时测量问题, 提出了一种基于深度学习的智能粒子滤波算法. 首先, 针对粒子滤波中的粒子短缺与退化问题, 建立了条件生成对抗网络(Conditional generative adversarial network, CGAN)引导的粒子选择机制. 在该机制中, 生成器网络通过对抗训练优化采样分布, 生成高质量粒子集; 判别器则评估生成粒子在真实后验分布中的概率值, 指导粒子权重计算. 其次, 针对井下复杂工况中存在的噪声协方差未知且时变问题, 设计了基于深度残差网络(Residual network, ResNet)的协方差矩阵估计器. 该模块与CGAN引导的粒子滤波以端到端的方式集成, 形成了闭环优化系统. ResNet模块得益于粒子滤波算法中的模型信息, 并为粒子滤波提供协方差矩阵的估计. 最后, 在旋转导向钻井工具平台上进行实验, 结果表明所提算法能够有效解决工具面角的实时测量问题, 与已有算法相比具有更高的精度.
- 智能粒子滤波 /
- 旋转导向钻井工具系统 /
- 实时测量 /
- 深度学习算法 /
- 未知噪声协方差矩阵
Abstract: To address the real-time measurement challenge of Toolface in rotary steerable drilling tool system (RSDTS), this paper proposes an intelligent particle filtering algorithm based on deep learning. Initially, the particle selection mechanism guided by conditional generative adversarial network (CGAN) is established to tackle the issues of particle impoverishment and degeneracy in particle filtering. In this mechanism, the generator network optimizes the sampling distribution through adversarial training, producing a high-quality set of particles; the discriminator evaluates the probability of the generated particles within the true posterior distribution, guiding the particle weight calculation. Subsequently, the covariance matrix estimator is designed based on a deep residual network (ResNet) to address the unknown but time-varying noise covariance matrices in complex downhole conditions. This module is integrated with the CGAN-guided particle filter in an end-to-end manner, forming a closed-loop optimization system. The ResNet module benefits from the model information in the particle filtering algorithm and provides the particle filter with estimates of the covariance matrices. Finally, experiments are conducted on the rotary steerable drilling tool platform. The results demonstrate that the proposed algorithm effectively addresses the real-time measurement issue of Toolface and exhibits higher accuracy compared with existing algorithms.
- Intelligent particle filter /
- rotary steerable drilling tool system /
- real-time measurement /
- deep learning algorithm /
- unknown noise covariance matrix

HTML全文

图 1 ResCGAN-PF算法图示

Fig. 1 The diagram of the ResCGAN-PF algorithm

下载: 全尺寸图片幻灯片

图 2 RSDTS实验平台

Fig. 2 RSDTS experimental platform

下载: 全尺寸图片幻灯片

图 3 RSDTS数据集

Fig. 3 RSDTS dataset

下载: 全尺寸图片幻灯片

图 4 RSDTS测量噪声

Fig. 4 RSDTS measurement noise

下载: 全尺寸图片幻灯片

图 5 粒子分布随时间变化图

Fig. 5 Particle distributions over time

下载: 全尺寸图片幻灯片

图 6 四个时刻粒子分布图

Fig. 6 Particle distributions at four time instances

下载: 全尺寸图片幻灯片

图 7 ResCGAN-PF估计结果

Fig. 7 Estimation results of ResCGAN-PF

下载: 全尺寸图片幻灯片

图 8 不同算法的估计误差(第一组)

Fig. 8 Estimation errors of algorithms in group 1

下载: 全尺寸图片幻灯片

图 9 不同算法的估计误差(第二组)

Fig. 9 Estimation errors of algorithms in group 2

下载: 全尺寸图片幻灯片

表 1 WT9011G4K芯片参数

Table 1 Parameters of chip WT9011GK

参数	加速度计	陀螺仪
测量范围	16 g	± 4 000°/s
误差	0.0005 g	0.061°/s
温漂	+0.00015 g/℃	0.005(°/s)/℃
采样频率	5-256 Hz	5-256 Hz

下载: 导出CSV

表 2 本文涉及的对比算法

Table 2 Comparative algorithms

算法名称	算法特点
PF^[6]	使用重采样技术来缓解粒子退化问题, 是粒子滤波领域的基础方法.
GPF^[8]	引入交叉和变异操作来动态调节粒子状态分布, 以解决粒子短缺问题.
CVAEPF^[14]	将条件变分自编码器引入粒子滤波框架, 利用其生成能力引导粒子状态分布.
VBPF^[17]	假设过程噪声协方差矩阵已知, 利用VB算法对测量噪声协方差矩阵进行自适应估计.
EMEKF^[15]	对系统进行线性化处理, 使用EM算法对过程和测量噪声协方差进行联合估计.

下载: 导出CSV

参考文献(24)

[1]	盛立, 牛艺春, 刘诗洋, 王伟亮, 高明, 周东华. 旋转导向钻井工具装备的微小故障检测. 中国科学: 信息科学, 2022, 52(11): 2106−2120 Sheng Li, Niu Yi-Chun, Liu Shi-Yang, Wang Wei-Liang, Gao Ming, Zhou Dong-Hua. Incipient fault detection of rotary steerable drilling tool equipment. Scientia Sinica Informationis, 2022, 52(11): 2106−2120
[2]	Xue Q L, Wang R H, Sun F, Huang L L, Han L J. Continuous measurement-while-drilling utilizing strap-down multi-model surveying system. IEEE Transactions on Instrumentation and Measurement, 2014, 63(3): 650−657 doi: 10.1109/TIM.2013.2282412
[3]	Sheng L, Niu Y, Wang W L, Gao M, Geng Y F, Zhou D H. Estimation of Toolface for dynamic point-the-bit rotary steerable systems via nonlinear polynomial filtering. IEEE Transactions on Industrial Electronics, 2022, 69(7): 7192−7201 doi: 10.1109/TIE.2021.3097601
[4]	Niu Y, Sheng L, Gao M, Wang Y C, Zhou D H. Variational Bayesian-based moving horizon estimation of Toolface for rotary steerable drilling tool systems. IEEE Transactions on Industrial Electronics, 2023, 70(1): 813−823 doi: 10.1109/TIE.2022.3153829
[5]	Sheng L, Niu Y C, Gao M, Geng Y F, Zhou D H. Particle filter-based fault detection for Toolface measurement of rotary steerable systems. IEEE Transactions on Instrumentation and Measurement, 2023, 72: Article No. 3519511
[6]	Arulampala M S, Maskell S, Gordon N, Clapp T. A tutorial on particle filters for online nonlinear/non-Gaussian Bayesian tracking. IEEE Transactions on Signal Processing, 2002, 50(2): 174−188 doi: 10.1109/78.978374
[7]	Liu J S, Ch en, R. Sequential Monte Carlo methods for dynamic systems. Journal of the American Statistical Association, 1998, 93(443): 1032−1044 doi: 10.1080/01621459.1998.10473765
[8]	Yin S, Zhu X P. Intelligent particle filter and its application to fault detection of nonlinear system. IEEE Transactions on Industrial Electronics, 2015, 62(6): 3852−3861
[9]	田梦楚, 薄煜明, 陈志敏, 吴盘龙, 赵高鹏. 萤火虫算法智能优化粒子滤波. 自动化学报, 2016, 42(1): 89−97 Tian Meng-Chu, Bo Yu-Ming, Chen Zhi-Min, Wu Pan-Long, Zhao Gao-Peng. Firefly algorithm intelligence optimized particle filter. Acta Automatica Sinica, 2016, 42(1): 89−97
[10]	Li T C, Sun S D, Sattar T P, Corchado J M. Fight sample degeneracy and impoverishment in particle filters: A review of intelligent approaches. Expert Systems with Applications, 2014, 41(8): 3944−3954 doi: 10.1016/j.eswa.2013.12.031
[11]	张文安, 林安迪, 杨旭升, 俞立, 杨小牛. 融合深度学习的贝叶斯滤波综述. 自动化学报, 2024, 50(8): 1502−1516 Zhang Wen-An, Lin An-Di, Yang Xu-Sheng, Yu Li, Yang Xiao-Niu. A survey on Bayesian filtering with deep learning. Acta Automatica Sinica, 2024, 50(8): 1502−1516
[12]	Ma M, Fu L, Zhai Z, Sun R B. Transformer based Kalman filter with EM algorithm for time series prediction and anomaly detection of complex systems. Measurement, 2024, 229: Article No. 114378 doi: 10.1016/j.measurement.2024.114378
[13]	Qian Y N, Yan R Q. Remaining useful life prediction of rolling bearings using an enhanced particle filter. IEEE Transactions on Instrumentation and Measurement, 2015, 64(10): 2696−2707 doi: 10.1109/TIM.2015.2427891
[14]	Jiao R H, Peng K X, Dong J. Remaining useful life prediction of lithium-ton batteries based on conditional variational autoencoders-particle filter. IEEE Transactions on Instrumentation and Measurement, 2020, 69(11): 8831−8843 doi: 10.1109/TIM.2020.2996004
[15]	Huang Y, Zhang Y, Xu B, Wu Z, Chambers A. A new adaptive extended Kalman filter for cooperative localization. IEEE Transactions on Aerospace and Electronic Systems, 2018, 54(1): 353−368 doi: 10.1109/TAES.2017.2756763
[16]	Niu Y, Sheng L, Gao M, Zhou D H. Moving horizon estimation for stochastic descriptor systems with inaccurate noise covariance matrices. IEEE Transactions on Industrial Electronics, 2024, 71(8): 9530−9540 doi: 10.1109/TIE.2023.3325555
[17]	Li K, Zhao S Y, Liu F. Joint state estimation for nonlinear state-space model with unknown time-variant noise statistics. International Journal of Adaptive Control and Signal Processing, 2021, 35(4): 498−512 doi: 10.1002/acs.3208
[18]	Bertipaglia A, Alirezaei M, Happee R, Shyrokau B. An unscented Kalman filter-informed neural network for vehicle sideslip angle estimation. IEEE Transactions on Vehicular Technology, 2024, 73(9): 12731−12746 doi: 10.1109/TVT.2024.3389493
[19]	Zhang Y P, Huang Y F, Deng K, Shi B F, Wang X Y, Li L, Song J. Vehicle dynamics estimator utilizing LSTM-ensembled adaptive Kalman filter. IEEE Transactions on Industrial Electronics, 2025, 72(5): 5429−5439 doi: 10.1109/TIE.2024.3482011
[20]	Zeng G X, Wu H, Chen S X, Yang M B. Modified Swin Transformer-based continuous-discrete extended Kalman filtering for bearings-only tracking with unknown noise statistics. IEEE Sensors Journal, 2025, 25(3): 9842−9852
[21]	Ahwiadi M, Wang W. An adaptive particle filter technique for system state estimation and prognosis. IEEE Transactions on Instrumentation and Measurement, 2020, 69(9): 6756−6765 doi: 10.1109/TIM.2020.2973850
[22]	王坤峰, 苟超, 段艳杰, 林懿伦, 郑心湖, 王飞跃. 生成式对抗网络GAN的研究进展与展望. 自动化学报, 2017, 43(3): 321−332 Wang Kun-Feng, Gou Chao, Duan Yan-Jie, Lin Yi-Lun, Zheng Xin-Hu, Wang Fei-Yue. Generative adversarial networks: the state of the art and beyond. Acta Automatica Sinica, 2017, 43(3): 321−332
[23]	刘建伟, 谢浩杰, 罗雄麟. 生成对抗网络在各领域应用研究进展. 自动化学报, 2020, 46(12): 2500−2536 Liu Jian-Wei, Xie Hao-Jie, Luo Xiong-Lin. Research progress on application of generative adversarial networks in various fields. Acta Automatica Sinica, 2020, 46(12): 2500−2536
[24]	胡铭菲, 左信, 刘建伟. 深度生成模型综述. 自动化学报, 2022, 48(1): 40−74 Hu Ming-Fei, Zuo Xin, Liu Jian-Wei. Survey on deep generative model. Acta Automatica Sinica, 2022, 48(1): 40−74