Cross-project utilisation of tunnel boring machine (TBM) construction data: a case study using big data from Yin-Song diversion project in China. Issue 1 (2nd January 2023)

• Record Type: Journal Article
• Authors: Li, Xu; Li, Haibo; Du, Saizhao; Jing, Liujie; Li, Pengyu
• Is Part Of: Georisk. Volume 17:Issue 1(2023)
• Page Start: 127
• Page End: 147
• Publication Date: 2023-01-02
• Subjects: TBM performance prediction -- machine learning -- big data -- conversion relationship -- rock tunnel; Environmental risk assessment -- Periodicals; Hazardous geographic environments -- Risk assessment -- Periodicals; Reliability (Engineering) -- Periodicals; Risk management -- Periodicals; 624.151
• DOI: 10.1080/17499518.2023.2184834
• Languages: English
• ISSNs: 1749-9518

Abstract

ABSTRACT: The variation in Tunnelling boring machine (TBM) equipment and geological information of tunnels result in substantial differences in real-time TBM tunnelling data. This variation makes it difficult to apply machine learning models trained by historical engineering data on new projects. To overcome this challenge, a novel data conversion approach from a mechanical analysis perspective has been proposed to normalise TBM tunnelling data, such as cutterhead torque and cutterhead thrust, which help to unify data from different projects under the same framework. Furthermore, the effectiveness of this approach has been verified through analogy analysis and machine learning applications. With the application of these conversion relationships, the machine learning model trained on a completed Yin-Song project with big data (12, 501 boring cycles) is applied to the on-going Yin-Chao Water Diversion Project in China with limited data (777 boring cycles) and gives reliable predictions for each performance parameter (with R 2 for the cutterhead thrust of 0.81 and R 2 for the cutterhead torque of 0.70). This approach enhances the usefulness of TBM intelligence for cross-engineering geophysical prospecting in different geological conditions.