Determination of Spatial Variability in d70 Grain Size Values Using High Density Site Measurements

More Info
expand_more

Abstract

Research to the failure mechanism piping (internal, backward erosion) in the Netherlands shows that the dikes require costly berms to meet the required safety standards. According to the Sellmeijer piping-formula, the d70 (70%-quantile of the grain size distribution by weight) is an important parameter in the resistance against piping. Therefore a better insight in the d70 variability in the naturally deposited sand layers beneath dikes is of great importance. High density site measurements (5x5 meter spacing) give insight in local heterogeneity of the d70 and the implications for representative design values. Three test sites were developed at locations with distinct geomorphological deposits. The spatial variability in the d70 appears to be large. Despite the large amount of data at a very dense measurement grid, there is no obvious spatial correlation in the measurements. The sites have a similar or larger coefficient of variation than the much sparser regional datasets that are normally used to derive the representative d70. Current regulations prescribe the characteristic (5% lower bound) value as the design parameter in the piping formula. Considering the large variation in d70 values this approach may not be appropriate. Given the influence of the d70 on dike dimensions, hence investment costs, establishing a method for determining a representative d70 value is important. We recommend to develop more sites at distinct geomorphological deposits to gain insight in the spatial variability in grain size distribution and to determine design values with matching site sampling requirements. An important aspect of further research is the interaction between d70 variability and the physical development of the piping channel.