Dam displacement serves as a critical visual indicator for assessing structural integrity and stability in dam engineering. Data-driven displacement forecasting has become essential for modern dam safety monitoring systems, though conventional approaches—including statistical models and basic machine learning techniques—often fail to capture comprehensive feature representations from multivariate environmental influences. To address these challenges, a bidirectional gated recurrent unit (BiGRU)-enhanced neural network is developed, incorporating sliding window mechanisms to model time-dependent hysteresis characteristics. The BiGRU’s architecture systematically integrates historical temporal patterns through overlapping window segmentation, enabling dual-directional sequence processing via forward–backward gate structures. Validated with four instrumented measurement points from a major concrete gravity dam, the proposed model exhibits significantly better performance against three widely used recurrent neural network benchmarks in displacement prediction tasks. These results confirm the model’s capability to deliver high-fidelity displacement forecasts with operational stability, establishing a robust framework for infrastructure health monitoring applications.

Gully bank instability provides critical sediment supplies for debris flow hazards in alpine canyon regions. The physical processes of unconsolidated bank failure driven by coupled rainfall and flood scouring require further experimental characterization. This study performs flume model experiments under two test scenarios, that is, flood loading with and without antecedent rainfall. Macroscopic failure features and subsurface hydrological responses are monitored via a dual perspective visual system and pore‐water pressure sensors. Results indicate that flood‐driven gully bank failure manifests as typical retrogressive landslides caused by cantilever undercutting. External toe erosion destroys basal support and reshapes bank geometry, whereas flood infiltration increases pore water pressure and reduces matric suction to weaken soil shear strength. Although antecedent rainfall may degrade soil mechanical properties by raising moisture content, earlier landslide initiation and larger failure magnitude are observed under the no‐rainfall condition, with the main failure processes completing within 90 s. Rainfall drives fine particles to migrate downward and fill subsurface voids, increasing soil compactness. This improved antierosion capacity partially offsets the moisture‐induced strength loss, resulting in a smaller overall failure volume during the rainfall present test. Derived from simplified laboratory conditions, these findings offer mechanistic insights into rainfall and flood triggered gully bank instability and support hazard related research for alpine debris flow catchments.

The increasing scarcity of urban spatial resources has led to the growing development of urban subways. The earth pressure balanced (EPB) shield machine has become one of the main tunnel boring machines in metro tunnel construction due to its economic efficiency and versatility. In this study, the spatial heterogeneity of soil properties in the shield tunnel stratum is considered, and the three‐dimensional (3D) random discrete element method (RDEM) is applied to investigate the mechanisms of tunnel face instability and ground deformation for EPB shield tunneling in spatially heterogeneous stratum at both macro and microscopic levels. Numerical triaxial tests with a flexible boundary are carried out to establish the relationship between key macro‐microscopic parameters. Based on the RDEM, a shield tunnel model with spatially heterogeneous stratum and EPB shield machine is built and validated. Then, 16 sets of shield tunnel excavation simulations are conducted, including one set with homogeneous stratum and 15 sets with spatially heterogeneous strata. To consider the influence of excavation disturbance, the EPB shield machine keeps moving and the cutterhead keeps rotating in all simulation processes. Results reveal that overlooking spatial heterogeneity in soil properties may underestimate the risks of shield tunnel face instability and ground deformation.