CS29C - Concurrent Session 29C: Using Laboratory Experiments to Explore the Effect of Non-horizontal Layers on Backward Erosion Piping

Dams and levees are critical infrastructure systems that protect over 20 million lives and $2 trillion in property in the United States. There are many important failure modes that must be considered for dams and levees, but internal erosion is one of the most damaging and hard to observe failure modes. Backward erosion piping (BEP) is one form of internal erosion that occurs as flowing water progressively removes soil particles through an unfiltered exit, such as a sand boil. If BEP is allowed to progress it can lead to breaches, as occurred at multiple points along the Mississippi River, including Kaskaskia Island and Bois Brule during the 1993 floods. Current practice evaluates the potential for internal erosion by comparing results of seepage analyses to critical gradient thresholds. Many researchers have used both laboratory tests and analytical models to determine the magnitude of these critical gradients, but these studies almost always consider horizontal layers. Some authors have proposed correction factors to account for inclined bedding, but there is limited data available to validate these corrections. This lack of data adds to the already significant uncertainty in BEP calculations.

This study uses BEP experiments in a medium-scale flume to measure the pore pressure distribution in the vicinity of the pipe tip. These pore pressure measurements are used to determine the gradient needed to cause erosion to progress. The newly designed flume includes a frame that can rotate and lock at a variety of bedding angles to determine the gradient required to progress piping to failure when the foundation layer is inclined. The results demonstrate that positive bedding angles are more resistant to erosion requiring higher gradients for progression, while negative bedding angles can erode more easily. Natural deposits will rarely have perfectly horizontal layering and so these differences in critical gradient need to be accounted for. It is also likely that many deposits will have both positive and negative bedding angles in different locations, which may lead to varying erosion rates as the pipe progresses. The results demonstrate the value in using laboratory tests to validate BEP models for more realistic scenarios.