On characteristic values and the reliability-based assessment of dykes

Abstract

A case study involving the assessment and re-design of an existing dyke, founded on a layered soil, has compared deterministic analysis based on 5-percentile property values and a reliability-based random finite element analysis consistent with the requirements of Eurocode 7. The results show that a consideration of the spatial nature of soil variability generally leads to higher computed factors of safety and, for those dyke sections requiring remedial action, to more economic designs. Back-figured characteristic values are shown to be considerably higher than the 5percentile soil properties; hence, a reduction in over-conservatism is achieved.

On characteristic values for calculating factors of safety for dyke stability

“…One solution is to use a simplified approach for calculating the value of X k , which can then, after application of partial factors, be used in deterministic analyses to obtain reliability-based values of F. For example, Dutch engineering practice calculates F for dykes by using the 5 percentile of either the underlying soil property distribution or a distribution which takes some account of spatial variability by using simplified variance reduction. Recently, Hicks et al (2019) showed how RFEM can be used (with or without partial factors) to directly determine reliability-based factors of safety, without having to explicitly derive the characteristic values. Although this method is computationally intensive, it removes the need to determine X k , is completely general, and automatically accounts for both variance reduction and the reduced mean due to weaker zones.…”

“…Figure 2 shows the idealised dyke cross-section analysed by Hicks et al (2019), which is the same as that used previously by Kames (2015) in limit equilibrium slope stability analyses based on 5-percentile characteristic values. Table 2 lists, for each soil layer, the mean, 5 percentile and coefficient of variation (COV) of the shear strength parameters (cohesion c′ and tangent of friction angle ϕ′), which were assumed to follow log-normal distributions, as well as the unit weight γ, which was assumed to be deterministic (Kames, 2015).…”

“…Table 2 lists, for each soil layer, the mean, 5 percentile and coefficient of variation (COV) of the shear strength parameters (cohesion c′ and tangent of friction angle ϕ′), which were assumed to follow log-normal distributions, as well as the unit weight γ, which was assumed to be deterministic (Kames, 2015). Hicks et al (2019) also assumed the vertical and horizontal scales of fluctuation to be θ v ¼ 0:5 m and θ h ¼ 6:0 m, respectively, based on CPT data from a similar site (de Gast et al, 2017). Hicks et al (2019) used RFEM with the strength reduction method to compute the probability distribution of possible values of F, given the soil parameter statistics listed in Table 2.…”

“…However, a more rigorous approach is the random finite-element method (RFEM) (Fenton & Griffiths, 2008), which combines random field theory with the finite-element method within a Monte-Carlo framework (Griffiths et al, 2009;Hicks & Spencer, 2010). In particular, Hicks et al (2019) used RFEM to account for spatial variability of soil properties in the reliability-based assessment and re-design of a dyke in the Netherlands. The assessment revealed that the factor of safety did not meet national safety requirements.…”

“…This paper uses various simplified approaches to determine characteristic soil property values for the dyke crosssection analysed by Hicks et al (2019). These values have then been used in deterministic finite-element slope stability analyses, and the resulting factors of safety (F) compared with F ¼ 0:98, the 5-percentile system response previously computed using RFEM.…”