Assessment of soil spatial variability for linear infrastructure using cone penetration tests

Abstract

Soil spatial variability has a significant impact on the reliability of geotechnical structures. In particular, the horizontal variability is important for linear infrastructure, which has only limited vertical height and width, but extensive length. Due to depositional and geological processes, the variability is often substantially different in the vertical and lateral directions. This variability can be characterised by a spatial correlation length, or scale of fluctuation, which is a measure of how significantly soil properties are correlated in space. An analysis of the reliability of such a measure has been undertaken using synthetic data, leading to a design chart which quantifies the statistical uncertainty in the scale of fluctuation for specific site investigation designs, which can be an important input for probabilistic analyses of the structure response. Moreover, practical guidance for site investigation design is proposed which can reduce the statistical uncertainty. The method has been applied to a real site investigation comprising a row of 29 closely spaced cone penetration tests (CPTs), within a larger site investigation of 100 CPTs, and applied to a simple design calculation for a long embankment to illustrate the impact on slope stability assessment. The site investigation data are made available to add to the limited amount of detailed data in this field.

On the reliability assessment of a controlled dyke failure

“…with respect to the actual depth-dependent mean). Table 2 shows the scales of fluctuation to be small (0•3-0•8 m) in the vertical direction, as has generally been found by other researchers for other soils, and as found for the same soil layers at the adjacent test site at Leendert de Boerspolder (de Gast, 2020;de Gast et al, 2020). The scales of fluctuation in the horizontal direction were found to be 4-8 times larger than the vertical scales of fluctuation.…”

“…2), with the CPTs in each row positioned at 2•5 m centres on average. As the derived values of θ h were similar in magnitude to the CPT spacing, there is some doubt as to their accuracy, although they are generally consistent with values obtained at the adjacent test site, which used CPT spacings of only 1•25 m (de Gast et al, 2017(de Gast et al, , 2019(de Gast et al, , 2020. Moreover, it seems reasonable that θ h would be at least several times greater than θ v , due to the natural process of deposition, and there is confidence in the derived values of θ v owing to the large number of closely spaced data available to construct the experimental autocorrelation functions in the vertical direction.…”

“…Two site investigations were conducted at Leendert de Boerspolder as part of the Reliable dykes project: at one location, 100 cone penetration tests with pore pressure measurement (CPTus) and six borings were undertaken as part of a detailed study on soil heterogeneity not reported in this paper (de Gast et al, 2017(de Gast et al, , 2019(de Gast et al, , 2020; the second series of tests was conducted prior to, and at the location of, the failure test and the tests were used in the analyses presented herein. Fig.…”

“…A possible cause of the difference between the general tendency predicted by the RFEM results and the actual location of the failure is the spatial variability of soil properties along the length of the dyke, coupled with the narrow probability distributions being sensitive to moderately small changes in soil strength. Although some account has been taken of the spatial variability in the third dimension through the three rows of CPTs, these rows were at a spacing of 20 m, which is much greater than the horizontal scales of fluctuation listed in Table 2 and determined at the adjacent test site (de Gast et al, 2017(de Gast et al, , 2019(de Gast et al, , 2020. Hence, it seems likely that, while the whole slope was near failure at the maximum drawdown, the actual location of the failure was determined by the spatial variation of soil properties along the dyke and, in particular, by the spatially variable soil profile at a location for which little prior information was available.…”

Assessment of the spatial variability of a Croatian flood embankment using the cone penetration test

“…Soil properties vary as a function of space and time. Quantifying exactly how they vary is of upmost importance in determining the capacity of geotech nical structures (De Gast, Vardon and Hicks, 2020). This is particularly important for aged linear infra structure, which underwent less rigorous design than its modern counterparts and has a lower safety margin as a result (Reale et al, 2016;Reale, Xue and Gavin, 2017).…”