Utilizing machine learning to predict the uplift of shield tunnel linings ahead of the cutterhead during construction enables timely adjustments of control parameters, mitigating lining uplift issues. Nevertheless, existing models exhibit limited performance in long sequence time-series forecasting (LSTF) and face challenges in accurately predicting the uplift of multiple lining rings ahead of the shield cutterhead. Considering the impact of shield control, attitude parameters, and geological condition, and utilizing the Boruta algorithm to determine model input features, a shield tunnel segment uplift prediction model based on LightGBM-Informer was proposed. This model incorporates a wavelet transform filter and a complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) method to eliminate noise in time series data. The accuracy and applicability of the proposed model were validated using the monitoring data from subway shield tunnel projects in Nanjing and Xiamen. The results demonstrate that the model exhibits enhanced prediction accuracy in comparison to other models, including recurrent neural network (RNN), long short-term memory (LSTM), gated recurrent unit (GRU), and Transformer. Additionally, it demonstrates robust generalization capabilities across diverse geological condition datasets. As the length of the prediction sequence increases, the performance advantages of the model become more pronounced, accurately predicting the uplift of 1−2 rings of linings ahead of the shield cutterhead. Feature importance analysis based on Shapley additive explanations (SHAP) method indicates that earth chamber pressure and vertical displacement at the shield head and tail have significant impacts on lining uplift. The model provides theoretical guidance for intelligent control of shield tunnel lining construction in complex, water-rich environments.

In tunnel boring machine (TBM) tunnels, the pea gravel as a filling layer between the tunnel lining segments and surrounding rock is of significant importance for the load-bearing capacity and impermeability of the segments. Due to the poor flowability of cement slurry, it fails to adequately fill the backfill layer, resulting in defects such as voids behind the walls and inadequate grouting. Enzyme-induced calcium carbonate precipitation technology (EICP) has emerged as an environmentally friendly and efficient reinforcement method. The grouting material is liquid, exhibiting excellent fluidity and diffusivity, making it a promising solution for grouting in pea gravel backfill layers. To optimize the effectiveness of EICP grouting in pea gravel, an attempt was made to use standard sand and pea gravel as backfill aggregates. In order to quantitatively analyze the optimal mixing ratio, experiments were conducted with different ratios of pea gravel to sand (0.5, 0.75, 1.0, 1.25, 1.5) and varying grouting frequencies (9, 12, 15 times) in sand column solidification tests. Through unconfined compressive strength tests, permeability tests, determination of calcium carbonate content, ultrasonic velocity measurements, and scanning electron microscopy (SEM) microscopic analysis, the impact of different ratios of pea gravel to sand on the solidification effectiveness of EICP was analyzed from both macro and micro perspectives. The results indicate that the optimal ratio for EICP reinforcement of mixed pea gravel and sand is 1:1.5. After 15 grouting cycles, the uniaxial compressive strength of the specimens can reach up to 4.55 MPa, and the permeability coefficient is 1.72×10⁻⁵ m/s. Samples with a higher sand content exhibit a notable phenomenon where interparticle voids are readily filled and compacted by calcium carbonate crystals. This process results in a higher proportion of effective bonding among calcium carbonate crystals, consequently contributing to an elevated unconfined compressive strength of the stone body. The findings of this study can provide a theoretical basis for the engineering application of EICP technology in reinforcing TBM backfilled pea gravel.