Reliability Analysis of Anchored and Cantilevered Flexible Retaining Structures

Abstract

The primary objectives of this paper are to describe the details of a reliability analysis procedure for the geotechnical design of anchored and cantilevered flexible retaining structures using the 2002 AASHTO Bridge Design Specifications and to apply the procedure to evaluate the relationship among the major design parameters and the reliability index (β). The load models employed in the analyses consist of the apparent earth pressure (AEP) concept for anchored walls and the Rankine earth pressure distribution for cantilevered walls. The load statistics are derived primarily from the bias observed in the AEP model, in addition to the statistical variation in the primary input parameters, namely the Rankine coefficient of active earth pressure (Ka) and the soil unit weight (γs). The resistance statistics for anchor pullout and the passive resistance of discrete embedded vertical elements are based solely on field test data. The statistical variation in the passive resistance of continuous embedded vertical elements is evaluated by considering the variation of the primary input parameters, namely the Rankine coefficient of passive earth pressure (Kp) and γs. For anchored walls, the geotechnical limit states of anchor pullout and passive resistance are considered. Only overturning (passive resistance) is considered for cantilevered walls. The analyses demonstrate that the most important parameter in the reliability analysis for these types of walls is the statistical variability of the geotechnical resistance. Reliability indices are calculated using Monte Carlo simulations for anchored and cantilevered flexible retaining walls of typical dimensions. The results of the analyses also indicate that the reliability index of such structures designed according to existing geotechnical practice can vary significantly as functions of the degree to which engineering judgment is permitted and the type of soil strength data employed in resistance predictions.