Probabilistic Analysis of an Embankment Using the Collocation Stochastic Response Surface Method
摘要
This study presented a coupled deterministic and probabilistic analysis of an embankment founded on compressible soil. The objective was to quantify settlements and assess structural reliability while accounting for geotechnical uncertainties. Deterministic analyses showed that consolidation governed long-term settlement, varying between 14 and 19 cm. Symmetrical displacements and stabilized pore pressures were observed in the results. A Monte Carlo probabilistic analysis was conducted at three monitoring points to evaluate the failure probability associated with maximum displacement. The probability of failure decreased exponentially as the allowable displacement threshold increased. At the critical point, it reached 25.5% for 20 cm, decreasing to 3.3% at 25 cm and 0.67% at 30 cm. Spatial variations in risk highlighted settlement variability. The main contribution of this work was the development of a two-phase methodology that combined a deterministic model with a stochastic response-surface-based Monte Carlo framework. This approach enabled rapid probabilistic evaluations and reliability assessments, significantly reducing computational cost while preserving accuracy for engineering interpretation of settlement behavior under uncertainty. The results indicated high reliability for a 30 cm threshold and increased sensitivity for stricter serviceability limits between 15 and 20 cm. The results emphasized the influence of soil compressibility and load variability on system performance.