Challenges and Pitfalls of CEL and ALE Analyses in Geotechnics

Geotechnical boundary value problems involving soil-structure interaction and large deformations can be simulated well using the Coupled Eulerian-Lagrangian (CEL) method. That method has gained widespread acceptance as a numerical approach to analyze geotechnical processes with large soil deformations such as pile installations. There are two numerical meshes that overlap at least in some parts of the model, one Lagrangian and the other Eulerian. The soil is usually represented by the Eulerian part while the Lagrangian part is the structure. The material within the Eulerian region of the mesh overlap may be considered as ‘virtual’ or ‘ghost’ material, as there is only the material of the structure in reality and no additional material at the same place. It’s worth to investigate how much the existence of ‘ghost’ material in CEL simulations affects the numerical results. The investigation is done here by two different soil-structure interaction examples: (i) a laterally loaded pile in undrained soil, and (ii) the numerical simulation of single pile installation. Both examples are also modeled using the Arbitrary Lagrangian Eulerian (ALE) approach which does not require any ‘ghost’ material definition. A comparison of the modelling and results from these different methods – CEL and ALE – gives insights into pitfalls of the analyses. Further analysis challenges arise from the definition of void material on top of the free ground surface as well as the contact models in the numerical simulations, which will also be considered here.