Book Description

Internal erosion and piping are main causes of failures and accidents to embankment dams. The risk of internal erosion and piping for new dams can be controlled by better design and construction of the core of the dam and provision of filters to prevent seepage through the embankment and the foundations. However, the mechanism of internal erosion remains understudied due to the complex interactions between soil structures and water flow. This dissertation intends to provide quantitative analysis to address such difficulties through quantitative analysis on numerical simulation. To deeply understand the process of piping, whole process from initiation of erosion, continuation, progression to heave is established. A number of factors on piping resistance are investigated numerically by using coupled computational fluid dynamics and discrete element method. Then a simple conceptual model of piping resistance is developed considering equilibrium of a soil column at different hydraulic critical state, soil specific gravity, initial void ratio, particle size distribution, sample aspect ratio and friction coefficients. This model highlights the effect of friction on critical hydraulic gradient, which is not included in the classic Terzaghi's formulation. Another type of internal erosion, suffusion, is also reported and studied with the same numerical method. The filter criteria of suffusion are reviewed and examined. The features of suffusion and its microscopic mechanism are investigated by examining the influence of diverse initial fines contents, various initial void ratios, vibrational seepage velocities and different coarse particle size distributions. To facilitate the understanding of internal erosion and address a mass of simulation data, a community detection method is adopted to monitor the internal evolution of soil skeleton. This method extracts the particles' positions and contact forces to construct spatially embedded force networks to characterize the influence of varying coarse particle size distribution in suffusion and triaxial test scenarios. This method successfully extracts community structures that resemble force chains. We verify that the community structures have consistent features as force chains and could represent soil properties.