Erosion tests with the annular flume

To predict the changes of natural systems like estuaries and tidal lagoons as a consequence of human interferences often numerical models are applied. An important aspect is the erosion and transport of sediment. Therefore, these models require formulations that describe the erosion of the sediment bed consisting of sand, mud and organic matter. A physically founded and validated process-based description of these mixed sediments is still missing. Recently, a theoretically derived formulation for erosion of mixed sediments has been proposed. However, a proper validation of this new formula is lacking. Therefore, a systematic study has been undertaken. First, studies have been executed to investigate the individually physical properties in this formula. Next, erosion tests have been executed on a large number of varying sample compositions using a straight, re-circulating flume. The effect of sediment structures and cohesiveness of the sediment bed has been investigated, from which the new formulations could be qualified. However, this experimental set-up was relatively small, with possible large influences of e.g. boundary effects. Therefore, another set of erosion experiments had to be carried out to supplement the results of the mentioned tests.
The objective of this study was to quantify the newly defined theorem for the erosion behavior of sand-mud mixtures. The experiments were carried out on three different artificially generated sand-silt-clay mixtures, using a large annular flume. The mixtures were homogeneously mixed and 100% saturated and, subsequently, placed in the flume. A unidirectional flow was generated by rotating the top-lid and the flume in opposite directions. The flow-induced shear stresses near the bottom were varied by increasing the rotational speeds step by step. An observation on the erosion behavior of the beds was carried out, as well as the measurement of the concentration of suspended fines over the vertical. Due to the occurring secondary currents, the eroded sand accumulated along the outer bend of the flume. By collecting this the amount of eroded sand was quantified. The shear stresses near the bottom were not measured directly, but were determined by means of a large eddy simulation model.
The results show two parts regarding the erosion behavior of the mixtures. Before the complete failure of the bed (part 2), different modes of erosion were identified in which only small amounts of material were eroded (part 1). Mainly due to the placement of the bed, the bed strength was slightly varying in the upper part of the sediment bed concerning both the horizontal as the vertical direction. In this first part floc erosion occurred for the fines. Sand was transported as bed load as well as in the sheet flow regime. For this part, the concentration of eroded fines showed a typically non-linear increase as usually occurred for erosion tests with deposited beds. Several possible explanations for this unexpected behavior are discussed.
For part 2, surface erosion was observed, during which the concentration was increasing linearly with time. The erosion rate as function of bed shear stresses revealed that the threshold of erosion for samples with a sand-silt skeleton was relatively high, while the threshold for the sample with the clay-water matrix was lower. A transition in erosion behavior was observed when regarding the erosion parameters for both fines and sand. This transition occurred for a plasticity index of around seven, which agrees well with the transition in mechanical behavior for sand-dominated to clay-dominated sediment as found in previous empirical geotechnical studies.