Reliability analysis of consolidation settlement in clay subsoil using FOSM and Monte Carlo simulation

Monica Susanne Löfman,Leena Katariina Korkiala-Tanttu

doi:10.1016/j.trgeo.2021.100625

Monica Susanne Löfman, Leena Katariina Korkiala-Tanttu

Open Access

https://doi.org/10.1016/j.trgeo.2021.100625

Copy DOI

Journal: Transportation Geotechnics	Publication Date: Jul 21, 2021
Citations: 6	License type: cc-by

Affiliation: Aalto University

Abstract

The settlement of ground-supported transportation infrastructures may vary greatly depending on the uncertainties related to compressibility properties and overconsolidation of clay. These uncertainties can be considered systematically by means of reliability analysis where the probability distribution of consolidation settlement is evaluated. Hence, this paper applies two well-known reliability methods, first-order second-moment method (FOSM) and Monte Carlo simulation, to various theoretical consolidation scenarios and three case studies of test embankments founded on normally consolidated (NC) or overconsolidated (OC) clay subsoil. The aim is to assess the accuracy and limitations of FOSM when applied to estimate the mean, the standard deviation, or the upper 95th percentile of the settlement. According to the results, the accuracy of FOSM is adequate for most serviceability limit state (SLS) purposes if the clay is normally consolidated. On the other hand, large relative errors may be introduced for OC clays where the relative contribution of OC and NC strains creates a strongly non-linear problem. It is concluded that FOSM, combined with second-order mean, may be applied to preliminary settlement estimation of OC clays. However, in the actual SLS verification with certain target reliability, Monte Carlo simulation should be used to ensure high enough accuracy.

Highlights

When designing transportation infrastructures founded on clay subsoil, the ultimate primary consolidation settlement is crucial when assessing the potential need for soil improvement and alternative foundation methods
Since Sallow and Sresidual are design-dependent, this paper aims to estimate the probability distribution of the ultimate primary consolidation settlement Sp, which is the basis of any soil improvement assessment or preloading design
The relative errors in estimating the mean, the stan dard deviation (Std), or the 95th percentile of strain or settlement by means of first-order secondmoment method (FOSM) were assessed by comparing to the Latin hypercube Monte Carlo

Summary

Introduction

When designing transportation infrastructures founded on clay subsoil, the ultimate primary consolidation settlement is crucial when assessing the potential need for soil improvement and alternative foundation methods. The advantages in using FOSM include that it requires only (2N + 1) realizations where N is the number of random variables. The required number of realizations in Monte Carlo simulation is usually in the scale of 104...106. Regarding consolidation analysis performed with finite element modeling for example, performing all the needed Monte Carlo simulation runs may lead to impractical analysis times. In addi tion, FOSM allows straightforward assessment of sensitivity factors and each variable’s contribution to the performance function [24,31]. FOSM has been suggested for various geotechnical problems (e.g., [7,35]), including consolidation settlement [9]

Objectives

Results

Discussion

Conclusion