IOP Conference Series: Earth and Environmental Science | 2021
Prediction of TBM Operational Parameters Using an Integrated Data Mining Framework
Abstract
Tunnel Boring Machine (TBM) is widely used in tunnel construction because of its high safety, less effect on surroundings and rapid excavation speed. However, its performance is highly dependent on the adjustment of operational parameters during tunnelling. In many cases, the project delays even the accidents take place due to the mis-operation of inexperienced drivers. To improve the adaptability of a TBM in complex geological conditions, this study proposes an integrated data mining framework to perform near real-time prediction of two operational parameters including thrust and torque. The integrated framework provides a set of data processing methods including data cleaning, partition of full tunnelling cycles, feature extracting and model establishment. These data mining methods were applied to analyse the in-situ data of a water conveyance project. The results showed that the proposed framework performed well in predicting those two parameters with the determination coefficient R2 all exceeding 0.9, which illustrated the feasibility of using the proposed framework to assist driving in TBM construction.