Working mode detection method based on bidirectional LSTM for pipe jacking inertial automatic guidance system

Abstract The pipe-jacking inertial guidance method is a key technology to solve the guidance problems of complex pipe-jacking projects, such as long distances and curves. However, since its guidance information is obtained by gyroscope integration, the accrued error is significant, which limits its...

Full description

Saved in:
Bibliographic Details
Main Authors: Yutong Zu, Lu Wang, Yuanbiao Hu, Gansheng Yang, Boning He, Zheng Zhou
Format: Article
Language:English
Published: Nature Portfolio 2025-08-01
Series:Scientific Reports
Subjects:
Online Access:https://doi.org/10.1038/s41598-025-14304-5
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849343671029202944
author Yutong Zu
Lu Wang
Yuanbiao Hu
Gansheng Yang
Boning He
Zheng Zhou
author_facet Yutong Zu
Lu Wang
Yuanbiao Hu
Gansheng Yang
Boning He
Zheng Zhou
author_sort Yutong Zu
collection DOAJ
description Abstract The pipe-jacking inertial guidance method is a key technology to solve the guidance problems of complex pipe-jacking projects, such as long distances and curves. However, since its guidance information is obtained by gyroscope integration, the accrued error is significant, which limits its application. The pipe-jacking construction process has two working modes of static and jacking. The accumulated errors of static state can be corrected with zero velocity to suppress the system position dispersion. Therefore, zero-velocity detection is required. However, the pipe-jacking velocity is so slow (1–2 m/h) that the traditional threshold-based zero-velocity detection method cannot accurately detect the zero-velocity interval (ZVI). Bidirectional long short term memory (Bi-LSTM) can effectively extract the features of repetitive and regular movements during the long-time pipe-jacking construction process. Therefore, this research proposes a working pattern detection model for pipe-jacking based on Bi-LSTM deep learning framework. Through establishing a data collection system to construct the data set and training the model, the accuracy of the test set reaches 98.54%. In addition, a zero-velocity correction model is established. According to the zero-velocity detection results of the Bi-LSTM model, zero-velocity correction is performed. Subsequently, an experimental platform is established to simulate a curve pipe-jacking and attitude experiments. The attitude experiment proves that the proposed model detects the ZVI accurately. The inclination error is corrected by 0.06 °, and the azimuth error is corrected by 0.18 °. Finally, the proposed model is validated by the crossing project of the China-Russia Eastern Natural Gas Pipeline, and the results show that the proposed model effectively detects the working pattern of pipe-jacking machine with strong robustness and adaptability. In summary, the method can effectively improve the detection accuracy of the pipe-jacking working pattern. It lays the foundation for the application of the inertial guidance system in complex pipe-jacking, such as long-distance and curved projects.
format Article
id doaj-art-c58d03b334e14b2e85f16fba766615a1
institution Kabale University
issn 2045-2322
language English
publishDate 2025-08-01
publisher Nature Portfolio
record_format Article
series Scientific Reports
spelling doaj-art-c58d03b334e14b2e85f16fba766615a12025-08-20T03:42:53ZengNature PortfolioScientific Reports2045-23222025-08-0115111510.1038/s41598-025-14304-5Working mode detection method based on bidirectional LSTM for pipe jacking inertial automatic guidance systemYutong Zu0Lu Wang1Yuanbiao Hu2Gansheng Yang3Boning He4Zheng Zhou5State Key Laboratory of Deep Earth Exploration and Imaging, School of Engineering and Technology, China University of GeosciencesState Key Laboratory of Deep Earth Exploration and Imaging, School of Engineering and Technology, China University of GeosciencesState Key Laboratory of Deep Earth Exploration and Imaging, School of Engineering and Technology, China University of GeosciencesState Key Laboratory of Deep Earth Exploration and Imaging, School of Engineering and Technology, China University of GeosciencesChina Railway Construction Heavy Industry Corporation LimitedState Key Laboratory of Deep Earth Exploration and Imaging, School of Engineering and Technology, China University of GeosciencesAbstract The pipe-jacking inertial guidance method is a key technology to solve the guidance problems of complex pipe-jacking projects, such as long distances and curves. However, since its guidance information is obtained by gyroscope integration, the accrued error is significant, which limits its application. The pipe-jacking construction process has two working modes of static and jacking. The accumulated errors of static state can be corrected with zero velocity to suppress the system position dispersion. Therefore, zero-velocity detection is required. However, the pipe-jacking velocity is so slow (1–2 m/h) that the traditional threshold-based zero-velocity detection method cannot accurately detect the zero-velocity interval (ZVI). Bidirectional long short term memory (Bi-LSTM) can effectively extract the features of repetitive and regular movements during the long-time pipe-jacking construction process. Therefore, this research proposes a working pattern detection model for pipe-jacking based on Bi-LSTM deep learning framework. Through establishing a data collection system to construct the data set and training the model, the accuracy of the test set reaches 98.54%. In addition, a zero-velocity correction model is established. According to the zero-velocity detection results of the Bi-LSTM model, zero-velocity correction is performed. Subsequently, an experimental platform is established to simulate a curve pipe-jacking and attitude experiments. The attitude experiment proves that the proposed model detects the ZVI accurately. The inclination error is corrected by 0.06 °, and the azimuth error is corrected by 0.18 °. Finally, the proposed model is validated by the crossing project of the China-Russia Eastern Natural Gas Pipeline, and the results show that the proposed model effectively detects the working pattern of pipe-jacking machine with strong robustness and adaptability. In summary, the method can effectively improve the detection accuracy of the pipe-jacking working pattern. It lays the foundation for the application of the inertial guidance system in complex pipe-jacking, such as long-distance and curved projects.https://doi.org/10.1038/s41598-025-14304-5Deep learningZero-velocity detectionPipe-jacking guidanceInertial navigation systemKalman filtering
spellingShingle Yutong Zu
Lu Wang
Yuanbiao Hu
Gansheng Yang
Boning He
Zheng Zhou
Working mode detection method based on bidirectional LSTM for pipe jacking inertial automatic guidance system
Scientific Reports
Deep learning
Zero-velocity detection
Pipe-jacking guidance
Inertial navigation system
Kalman filtering
title Working mode detection method based on bidirectional LSTM for pipe jacking inertial automatic guidance system
title_full Working mode detection method based on bidirectional LSTM for pipe jacking inertial automatic guidance system
title_fullStr Working mode detection method based on bidirectional LSTM for pipe jacking inertial automatic guidance system
title_full_unstemmed Working mode detection method based on bidirectional LSTM for pipe jacking inertial automatic guidance system
title_short Working mode detection method based on bidirectional LSTM for pipe jacking inertial automatic guidance system
title_sort working mode detection method based on bidirectional lstm for pipe jacking inertial automatic guidance system
topic Deep learning
Zero-velocity detection
Pipe-jacking guidance
Inertial navigation system
Kalman filtering
url https://doi.org/10.1038/s41598-025-14304-5
work_keys_str_mv AT yutongzu workingmodedetectionmethodbasedonbidirectionallstmforpipejackinginertialautomaticguidancesystem
AT luwang workingmodedetectionmethodbasedonbidirectionallstmforpipejackinginertialautomaticguidancesystem
AT yuanbiaohu workingmodedetectionmethodbasedonbidirectionallstmforpipejackinginertialautomaticguidancesystem
AT ganshengyang workingmodedetectionmethodbasedonbidirectionallstmforpipejackinginertialautomaticguidancesystem
AT boninghe workingmodedetectionmethodbasedonbidirectionallstmforpipejackinginertialautomaticguidancesystem
AT zhengzhou workingmodedetectionmethodbasedonbidirectionallstmforpipejackinginertialautomaticguidancesystem