Cobble Sea Defence: Hydraulic Interface Stability of Sand underlying a Single Filter Layer

Abstract

A cobble sea defence appears to be an easy constructible protection, with relatively low total costs of ownership. Flume tests show that sand under such single layered constructions is stable. However, it is not exactly known how the hydraulic loading of a breaking wave is reducing in the filter. Two datasets containing pressure measurements in a revetment were available for identification of this reduction: 1) A dataset of a test performed in the Groser Wellenkanal (Hannover), for improvement of the understanding of all relevant processes in Elastocoast revetments (obtained from the Braunschweig University of Technology). 2) A dataset of tests in the Delta Flume, for the verification and optimization of the cobble shore design of the ‘Maasvlakte 2’ (obtained from PUMA). The dataset of the Groser Wellenkanal test was analysed to increase the insight in the behaviour of pressures in a filter resulting from (breaking) waves and to explain processes and test results of the Delta Flume model tests. Hydraulic loading at the top of a revetment exist out of two types of loads; impact and non-impact loads. The impact load, resulting from plunging waves, can be distinguished in an impact peak and a quasi-static part. The impact peak is very high (>10 kPa), last for only a fraction of time (<0.2s) and reduces completely in the filter. The non-impact load and the quasi-static part of the impact load reduce less and are responsible for the hydraulic loading at the interface of sand underlying a thick, single filter layer. Although large influence was expected for turbulence generated by breaking waves it does not have an important role in the hydraulic loading that eventually reaches the interface. The parallel gradients at the interface are only the result of pressures in- and decreasing with the wave period. The maximal destabilizing perpendicular gradients are frequently accompanied by a maximal parallel gradient. The main loading mechanism at the interface results from the run-down; during run-down the largest parallel and perpendicular gradients are generated. Predicted gradients acting on the interface between gravel and sand in the Delta Flume models are higher than traditional stability criteria. Therefore, an erosion process of the sandy embankment would be expected for the tested Delta Flume models. However, erosion was not observed. The stable interface in the Delta Flume tests can thus not be explained from the pressure reduction resulting from the analysis of the measurements in the Delta Flume models itself. The damping observed in the Groser Wellenkanal and the non-distorted interface in the Delta Flume models give strong indications that the reduction observed in the gravel material of the Delta Flume models is not representative for the performance of the gravel layer, resulting in a conservative prediction of gradients at the interface of gravel and sand. Several factors could have been influencing the pressures in the revetments of the Delta Flume models and the reduction found in the analysis of the measurements.