Shield Tail Seal System Operation State Detection Research Based on Data Segmentation of Slope and Shield Tail Seal Cavity Pressure Change Laws

Zhenggang Guo, Min Wang, Qingchao Sun, Wanxiu Hao, Zhixuan Zhou, Fuwen Zhang · 2025

In recent years, with the continuous advancement of various underground engineering construction projects in China, shield technology as an important technology in underground tunnel construction,Shield machine as an indispensable construction equipment in shield technology, shield machine systems can be safe and reliable operation is more and more important. Because of the working environment of the shield tail seal, it is difficult to verify its working status through experiments, because there is still no perfect shield tail seal status detection method. In this paper, a shield tail sealing system state detection scheme based on the slope and shield tail sealing system operation law is proposed for the operational state detection of shield tail sealing system in shield tunneling machine. The main research content is as follows: (1) according to the shield tail sealing system pressure change rule and slope change rule put forward a shield tail sealing system sealing cavity data for data segmentation, segmentation for different cycle data of three different stages. (2) According to the segmentation algorithm to segment the data features of the first two phases of each cycle respectively, the state detection algorithm based on dynamic time regularization and system operation law is proposed for different phases.(3) Aiming at the data features of the third stage segmented by the segmentation algorithm, a system operation state detection method based on sliding window outlier detection is proposed. (4) The accuracy of the comprehensive validation algorithm for shield tail sealing system operation anomaly detection is verified based on real data from real shield sealing system operation. The final results show that the proposed algorithm consists of good stability and accuracy.

Read the paper · More papers on PaperTik