Backward erosion piping: Experiments on erosion and progression rate

Abstract

Backward erosion piping is a form of internal erosion where small pipes are formed below a dike. These pipes are formed in a direction opposite to the flow that transports sand particles. Piping is a very important failure mechanism in the protection of dikes, which can be unpredictable due to the different soil characteristics. Piping is well managed in the Netherlands, but there is a lot to be discovered regarding theerosion processes inside the pipe. For example, how fast the pipe develops and what this is dependent on. The objective of this thesis is to monitor and study the development of the pipe, to extend the knowledge about the different processes that influence the progression rate of the pipe. The main research question is: How do the different parameters and processes influence the progression and the sediment transport rate in laboratory experiments of backward erosion piping? To monitor the development of the pipe, different small-scale laboratory experiments were performed to obtain new data regarding piping and to study the influences of different parameters. The piping experiments were performed in three series: (1) configuration of the setup, (2) effect of grain size and (3) hydraulic loading. These experiments were performed in the previously developed setup of Vera van Beek (Van Beek, 2015). This setup was modified to measure the pore pressures and to guide the pipe through the middle of the setup. While conducting the experiments, different measurements were performed. This included measuring of the pipe length and geometry, collecting the sand boil and a dye injection to follow the flow. The literature study performed for this thesis has shown that in the past many experiments were performed regarding piping, but these studies did not focus on the different processes and the sediment transport rate of the pipe. By focussing on the progression phase of the pipe (continuous transport), the experimental data is compared to existing models (primary and secondary erosion). Sellmeijer’s model (Förster et al., 2012) is the current rule that is applied in the Netherlands for dike safety, but this model does not include time-dependency. This research showed that the development of the pipe is not a stationary process but depends on various conditions, such as
soil characteristics. The hypothesis formed at the beginning of this thesis listed several soil parameters which influence the progression rate of the pipe. Concerning the sediment transport rate of the pipe, Cheng’s model for bedload transport (Cheng, 2004) is evaluated and compared with the measured results. The analysis showed that the adapted formula of Cheng (Equation 2.24) overestimates the sediment transport rate in the pipe. From the analysis of the experimental results, it can be concluded that two parameters play an important role in the progression and sediment transport rate of the pipe: the particle diameter and hydraulic permeability. These parameters showed an influence on the progression rate which can be used to study piping on a larger scale. The most interesting result is the fact that experiments with a larger particle diameter have an overall larger progression rate.