1. Submarine stratum instability is an important marine geohazard affecting subsea tunnels, offshore foundations, pipelines, communication cables, and other marine infrastructure. In fault-influenc...

Surrounding rock deformation in soft rock tunnels is controlled by complex nonlinear interactions among geological conditions, construction parameters, and support measures, making accurate prediction challenging. In this study, a project-scale deformation prediction dataset was constructed using field monitoring data from a sandy shale tunnel project. Eight engineering factors were selected as input variables, including excavation method, initial support strength, closure time, tunnel burial depth, lithology, rock integrity, groundwater condition, and the relative orientation between major structural planes and the tunnel. A hybrid prediction framework integrating a novel whale optimization algorithm (NWOA) and a long short-term memory (LSTM) network was developed. The proposed NWOA improves the standard whale optimization algorithm by introducing a nonlinear convergence strategy, an adaptive weight coefficient, and a dynamic spiral position updating mechanism to enhance the hyperparameter search process of the LSTM model. Model performance and stability were further assessed using repeated and nested cross-validation. The corresponding RMSEs were 0.2294 ± 0.0734 and 0.2219 ± 0.0751 percentage points, the MAEs were 0.1427 ± 0.0350 and 0.1505 ± 0.0585 percentage points, and the R² values were 0.8660 ± 0.0548 and 0.8732 ± 0.0599, respectively. These comparable results support project-specific predictive performance for the investigated tunnel sections.

Water and mud inrush is a major hazard impeding the construction sustainability of tunnel crossings of water-rich active faults, where seismic-induced fracturing may compromise the integrity of water-resisting rock masses. Natural rock masses commonly contain pre-existing cracks and mineral veins, but their influences on dynamic failure under coupled geostress, hydraulic pressure, and seismic loading remain insufficiently understood. Taking the Kangding No. 2 Tunnel of the Sichuan–Tibet Railway crossing the Yunongxi active fault as the engineering background, a coupled hydro-mechanical discrete element model was established to investigate representative pre-existing crack and quartz-vein configurations corresponding to inclinations of 0°, 30°, 60°, and 90°. The model was subjected to principal stresses of 23.7 and 12.7 MPa, a groundwater pressure of 3.33 MPa, and combined horizontal and vertical seismic loading, with the horizontal component prescribed at 0.8 g and 5 Hz. The results show that pre-existing crack tips acted as preferential sites for fracture initiation, whereas quartz veins generally impeded fracture propagation before vein rupture. Shear fractures dominated the dynamic failure process in all investigated cases. The tensile-fracture ratios ranged from 7.90% to 13.15% for the crack-containing models, with the maximum value occurring in the 30° configuration, compared with only 7.01–7.84% for the vein-containing models. The AE b values of the crack-containing models ranged from 2.14 to 2.94 and exhibited a trapezoid-shaped variation among the investigated configurations, whereas the vein-containing models showed an oblique N-shaped variation, with a maximum b value of 2.23 in the 90° configuration. Abrupt increases and pronounced fluctuations in AE counts and energy accompanied intensified fracture propagation and model failure. These results provide numerical indications of potentially fracture-prone structural configurations, which can guide precise directional grouting, early warning and ecological protection in the migration of water and mud inrush, eventually supporting the construction sustainability of the tunnel.