Probabilistic assessment of supported excavations under demand-based sequential strut loss considering soil spatial variability

Sequential loss of bracing support represents a critical risk scenario in deep excavations, yet its interaction with inherent soil spatial variability and its impact on excavation system reliability have not been systematically quantified. This study proposes a probabilistic framework to investigate the influence of soil spatial variability on demand-based sequential strut loss and overall excavation reliability. A three-dimensional numerical model is constructed, and random fields of the corrected standard penetration resistance ( N 1 ) 60 are generated to represent spatially variable soil stiffness and strength. Sequential strut-removal analyses are performed under deterministic and probabilistic conditions to evaluate the evolution of strut axial-force redistribution, wall deformation, and soil shear strength degradation. The demand-based removal sequence indicates that the most highly loaded struts are generally located near the central region, and the prescribed support loss develops laterally and downward in a center-outward and top-to-bottom pattern in both deterministic and probabilistic analyses. Increasing the coefficient of variation broadens the spread in the number of removed struts required to reach the prescribed system performance limit, while the large scale of fluctuation can induce greater variation in the sequential loss length (i.e., the critical number of removed struts). Comparison between the deterministic and probabilistic analyses shows that the deterministic analysis overestimates the safety of the supported excavation and does not capture the possibility of earlier attainment of the prescribed system performance limit under sequential strut loss in spatially variable soils. These findings highlight the need to incorporate soil spatial variability into reliability-based excavation design and show that probabilistic analysis provides a more realistic basis for assessing excavation performance under demand-based sequential strut loss and improving bracing strategies.