A Novel Hybrid Transfer Learning Framework for Dynamic Cutterhead Torque Prediction of the Tunnel Boring Machine

A tunnel boring machine (TBM) is an important large-scale engineering machine, which widely applied in construction. Precise cutterhead torque prediction plays essential role the cost estimation of energy consumption and safety operation tunneling process, since it directly influences adaptable adjustment excavation parameters. Complicated variable geological conditions, leading to operational status parameters TBM, usually exhibit some spatio-temporally varying characteristic, poses a serious challenge conventional data-based methods for dynamic prediction. In this study, novel hybrid transfer learning framework, namely TRLS-SVR, proposed knowledge from historical dataset that may contain multiple working patterns alleviate fresh data noise interference when addressing issues. Compared with data-driven algorithms, TRLS-SVR considers long-ago data, can effectively extract leverage public latent implied datasets current collection situ TBM project located China utilized evaluate performance framework.