Research on the Prediction Model of Vibration Acceleration Based on TBM Tunnelling Parameters

The double-shield TBM has been widely adopted in urban subway tunnel construction in China. TBM excavation inevitably induces ground vibrations, which may pose risks to adjacent structures and adversely affect human comfort. However, current vibration assessment methods rely heavily on field monitoring and lack a quantitative relationship with TBM excavation parameters. To address this gap, this study proposes a novel approach to predict vibrations based on thrust, torque, and monitoring distance. Based on field tests conducted in the double-shield TBM section of the Qingdao Metro, a test segment consisting of 1063 rings was selected for analysis. Excavation parameters were systematically adjusted, and both the parameter values and corresponding vibration accelerations were recorded. Using Spearman’s rank correlation coefficient, the relationship between the excavation parameters and the root mean square (RMS) value of the vertical vibration acceleration was examined. The results indicate that both thrust (r = 0.72, p < 0.01) and torque (r = 0.47, p < 0.01) are significantly correlated with the vertical vibration acceleration. Using the improved GA-BP algorithm, a prediction model for vibration acceleration was established based on tunnelling parameters and monitoring distance. The results indicate that the model relying on thrust and distance yields the best performance, achieving a mean squared error (MSE) of 0.92888. The single-point prediction error ranges from 0.02 to 1.63 mm/s2. Compared with a model that uses distance alone, the proposed model reduces the single-point prediction error by 73.6%, representing a significant improvement in accuracy. The predictive model was further validated in the left tunnel, where the prediction error ranged from 0.03 to 1.07 mm/s2, confirming its feasibility. This study provides a method for real-time prediction of TBM-induced construction vibrations and offers a basis for adjusting tunnelling parameters to mitigate vibrations during tunnel construction.