Abstract:  In this study, a tunnel boring machine(TBM) tunneling adaptability prediction model based on the Elman neural network improved by grey relational analysis (GRA) and sparrow search algorithm (SSA) is proposed to accurately evaluate the adaptability of a TBM in a tunnel and ensure safe and efficient TBM tunneling. First, 14 primary influencing factors are preliminarily selected to establish an evaluation index system of tunnel TBM tunneling adaptability based on geological conditions, tunneling parameters, unfavorable geologies, and construction organization. Second, the correlation between the index and tunneling adaptability is analyzed using GRA. Then, SSA is introduced to optimize the Elman neural network to improve the model′s performance, and the leftonecrossvalidation method is used to validate the model′s accuracy and reliability, so that the model is close to the original data′s distribution characteristics. Finally, the model′s applicable results in a bid section of the North Xinjiang water diversion project are compared with those of Elman, PSOElman, BP neural network models, and the field data. The comparative results show that the SSAElman model′s prediction results are the most consistent with the actual engineering results, indicating that the model can correctly and effectively predict and evaluate the TBM tunneling adaptability with high rationality and operability. The results can provide a new method for evaluating tunnel TBM adaptability.

Key words: tunnel construction, driving adaptability of tunnel boring machine, grey correlation analysis, sparrow search algorithm, Elman neural network