To address the problems of complex influencing factors of chamber pressure and delayed response in manual regulation during shield construction, this study proposes a chamber pressure prediction and parameter-optimization method integrating SSA-RF and PSO based on field tunneling data from 1082 rings. First, on the basis of conventional tunneling parameters, muck improvement parameters, including the cumulative amount of foam mixture in the current ring, the cumulative amount of cutterhead water spray in the current ring, and the cumulative amount of bentonite in the current ring, are introduced. Nine key input variables are then screened by combining Pearson correlation coefficient analysis and distance correlation coefficient analysis. Subsequently, an SSA-RF chamber pressure prediction model is established. The results show that the model achieves an R² of 0.9968 and a MAPE of only 0.34% on the test set, outperforming the other models. Furthermore, with the objective of minimizing the deviation between the predicted chamber pressure and the target value, the PSO algorithm is used to optimize the propulsion speed and the rotation speed of the screw conveyor. The results indicate that, based on the predicted chamber pressure values, the PSO algorithm can output optimized adjustment suggestions for propulsion speed and screw conveyor rotation speed, thus realizing model-assisted regulation of chamber pressure. This study can provide a reference for chamber pressure prediction and auxiliary regulation of tunneling parameters during earth pressure balance shield tunneling.

The shield tunneling method is widely used for underground space development in the world, where vibratory screening is a critical pretreatment stage for resource utilization of shield muck. However, existing parameters for vibrating screens are mainly based on soft soil and sandy cobblestone strata. Research regarding shield muck from moderately weathered strata remains scarce, leading to poor equipment adaptability and low efficiency. This study focuses on a dry granular system of typical shield muck from moderately weathered strata after washing, drying, and homogenization. By combining laboratory standard sieve tests with EDEM discrete element simulations, a DEM numerical model of double-deck vibratory screening was established. Using single-factor controlled variables and orthogonal experimental designs, the comprehensive influences of vibration frequency, amplitude, and screen inclination angle on a double-deck screen were systematically investigated. Within the investigated parameter ranges, the screen inclination angle produced a relatively large variation in screening efficiency. Based on the preliminary evaluation using the numerical model developed in this study, a vibration frequency of 15 Hz, an amplitude of 2.0 mm, and a screen inclination angle of 12° exhibited favorable overall screening performance among the investigated parameter combinations, yielding upper- and lower-deck screening efficiencies of 95.16% and 82.96%, respectively. The results indicate that changes in the feed state of the lower deck induced by upper-deck pre-classification may be associated with the differentiated responses of the two screen decks to operating parameters, providing a reference for multi-stage screening analysis and parameter selection for shield muck from moderately weathered strata.