3D modelling of backward erosion piping experiments. (September 2022)

• Record Type: Journal Article
• Authors: van Beek, Vera; Robbins, Bryant; Rosenbrand, Esther; van Esch, John
• Is Part Of: Geomechanics for energy and the environment. Volume 31(2022)
• Publication Date: 2022-09
• Subjects: Internal erosion -- Embankments -- Piping -- 3D seepage flow; Engineering geology -- Periodicals; Power resources -- Periodicals; Energy development -- Technological innovations -- Periodicals; Engineering geology -- Environmental aspects -- Periodicals; Energy development -- Technological innovations; Engineering geology; Engineering geology -- Environmental aspects; Power resources; Geology -- Periodicals; Energy-Generating Resources -- Periodicals; Periodicals; Electronic journals; 621.042
• DOI: 10.1016/j.gete.2022.100375
• Languages: English
• ISSNs: 2352-3808

Abstract

Abstract: Backward erosion piping poses a threat to levees and dams. Prediction of this phenomenon is generally conducted using models that are based on a two-dimensional pipe and flow pattern. However, the concentration of flow towards the pipe is essentially a three-dimensional phenomenon, causing model width and exit geometry to affect its progression. This paper presents the 3D simulation of backward erosion experiments on uniform sands using the Sellmeijer model, as a first step towards the development of a prediction model. The paper confirms that two-dimensional outlet configurations, such as a ditch or plane, which tend to result in the development of multiple pipes, can best be modelled using the existing two-dimensional model. For configurations in which flow converges to a single location, as often occurs in the field, three-dimensional modelling will be required. The study illustrates that three-dimensional modelling of backward erosion piping still presents challenges, such as the understanding of the erosion mechanism at the pipe tip, the numerical simulation of the pipe and the validation of models at a larger scale. Highlights: For 2D-configurations the 2D Sellmeijer model is more suitable for BEP prediction than the 3D model. For 3D-configurations the 2D Sellmeijer model overestimates, and the 3D model underestimates the critical head. A 3D situation in the field cannot be captured by a constant factor applied to the 2D Sellmeijer rule.