Thrust force assessment model for avoiding TBM jamming using improved RIME algorithm

Tunnel Boring Machine (TBM) has prevailed in tunnel construction. However, the poor geological adaptability severely limits the tunneling speed and application of TBM in long and deep-buried tunnel projects. Therefore, a preliminary assessment of the required TBM’s thrust force can lower the risk of a TBM shield jamming disaster induced by the high ground pressure acting on the TBM shield. Before building the assessment model, a novel optimization algorithm, SLORIME, improved on the relatively newly developed RIME, is proposed. Then, based on many numerical results, a database of 8118 data is carefully established. Next, RF, SVM, and XGBoost are introduced and used to build a convenient and portable machine learning model instead of manually looking up the paper-based simulation results. Finally, SLORIME-XGBoost is determined after thorough comparisons. It can achieve the best performance (training set: R 2 =0.9996; validation set: R 2 =0.9978; testing set: R 2 =0.9976), followed by RIME-XGBoost, XGBoost, SLORIME-RF, RIME-RF, RF, SLORIME-SVM, RIME-SVM, and SVM in sequence. SLORIME, as proposed in this research, has a more promising application in the future in the face of a similar engineering dataset.