Incorporating Spatial Variability into FEM Analyses of Anchored Retaining Walls

Abstract

In a finite element-based reliability analysis, distributions of geotechnical input parameters from soil properties measured at the site may overlook spatial variability, especially across sparsely investigated areas. In doing so, soil parameters from weak zones are often considered as representative values across entire soil layers, potentially leading to overly conservative safety assessments. This issue becomes critical when analysing geotechnical failures in anchors, where their tensile resistance relies heavily on the averaged soil properties along the anchors. To address these challenges, this paper presents a method that incorporates spatial averaging into a reliability analysis. To demonstrate its application, a quay wall at the Port of Rotterdam is presented where the parameter distributions and their spatial variability were derived from the Cone Penetration Test (CPT). Comparing the reliability index before and after the inclusion of spatial variability shows that ignoring spatial variability may be overly conservative when assessing shaft capacity of piles or anchors.