On the transient process of contact erosion (instabilities during granular erosion)

Internal erosion is one of the major threats for water retaining structures like embankment dams or levees. Numerous design criteria were developed in the last decades for this reason. As the main concern of the geotechnical engineer is to prevent erosion, it is not astonishing that a vast majority of the criteria are focussed on the onset or initiation of erosion, or more precise, to avoid it. However, many existing structures have already experienced the one or other event of internal erosion. Thus, it is beneficial to understand not only the initiation, but also the progression of the erosion process. This process is though still not completely understood.Among the different types of internal erosion, contact erosion is characterised by two soils (base and filter) of different grain size distributions (PSD) forming an interface to each other. The erosion process is triggered by a water flow, which can be either parallel or perpendicular to the interface. The latter configuration is sometimes referred to as filtration and is in the focus of this research. If erosion occurs, base particles are transported into the pores of the filter by a water flow, forming a mixture zone with a lower porosity and permeability. For a better understanding of the contact erosion process, the formation of this zone must be understood. Two main aspects can be identified: The behaviour of the base material under the hydraulic load and the arrangement of the particles in the mixture zone, which is expressed in the porosity as a macroscopic parameter.Porosity alterations during erosion tests can be determined in different ways. These are among others: Direct observations of changes in layer heights, Computed Tomography (CT) and radiometric methods. Electromagnetic methods take advantage of the different dielectric permittivity of the solid, liquid and gaseous phases of a soil, which interact with an electric pulse travelling along a sensor surrounded by the material under test. An average value can be obtained with Time Domain Reflectometry (TDR). With an inversion algorithm, the spatial distribution of the parameter of interest, e.g. porosity or moisture, can be computed out of the TDR-trace. This method is called Spatial TDR (STDR), which was chosen for this research work. The three step inverse model was established and refined for this setup and its accuracy proven in calibration and verification measurements.In order to study the behaviour of the base material alone under hydraulic load, a test stand called Fluidisation Setup was developed. Different base materials were tested under increasing hydraulic potentials up to the point of hydraulic heave. The inner stress conditions during this tests were checked with vane-shear-tests. A clear correlation of shear resistance and hydraulic load was found. Additionally, the dielectric permittivity of the material was measured up to high porosities obtained during the fluidisation.A Coaxial Erosion Cell (CEC) was developed to act as a sensor by itself for the erosion tests. The coaxial arrangement ensures an even and defined distribution of the electromagnetic field, which is beneficial for the accuracy of the measurements. The material under test is positioned in the annulus between inner and outer conductor. An inspection window allows visual observations and pressure transducers the reading of the distribution of the hydraulic head along the cell. With this cell and the STDR-measurements, the longitudinal distribution of the porosity during the tests can be determined in nearly real time in short intervals. A second cell of a more conventional setup was used for verifying the results of the CEC. Additionally, superimposed loads were applied in this tests in order to study the influence of different effective stresses on onset and progress of contact erosion.It was found that the size ratio of base and filter pores has a considerable influence on the initiation and progress of the erosion process. It not only influences the hydraulic gradient at which the process may initiate, but also the porosity and the distribution of the hydraulic gradient in the formed MZ as well as the elevation of the MZ for a given gradient. It further has an influence on the effect of effective stresses on the onset of erosion