Retaining walls
reliability analysis, reinforced soil structures, geosynthetic, geotextile,
sand.
Deterministic methods are traditionally adopted for reinforced soil structures design, similarly to other types of Geotechnical Engineering structures. This design method does not adequately consider the inherent uncertainties associated to both, the construction work and the variability of the soil and other used materials. Following the deterministic approach, safety factors that promote separation between the mean values of resistance and applied loads are typically used. Nonetheless, reliability based design allows considering uncertainties using statistical tools, which provides a rational way to incorporate uncertainties into geotechnical design. According, the present work intends to evaluate the behavior of geosynthetic-reinforced soil (GRS) wall using reliability analysis, considering sand as backfill and geotextiles, as reinforcement layers. The study was conducted for external and internal failure mechanisms. Different methods of reliability analysis were used, specifically, the first order reliability method (FORM) and point estimate method (PEM). Parametric analyzes were performed in order to evaluate the influence of the variables involved in the problem with respect to the value of the reliability index. The coefficient of variation for the tensile strength of different types of geotextiles was obtained by analyzing the results of tensile tests and creep rupture tests of geotextiles commercially available in Brazil. The results showed that the increase in the coefficient of variation of both the tensile strength and the friction angle provided the decrease of the reliability index related to the internal stability of the structure. In addition, structures using woven geotextiles have resulted in higher reliability index than those designed with non-woven geotextiles. The higher the height of the structure the lower the reliability index. For all failure modes related to external stability and for pullout failure, the FORM method resulted in higher reliability index values than the PEM method, indicating that the point estimates method is more conservative in relation to the first order method. Nevertheless, for the geosynthetic rupture failure mode there was no well-defined trend regarding the more conservative method for calculating the reliability. Overall, reliability based design for GRS walls was found to provide consistent results, corresponding to a conceptually more adequate design approach.