Answers are grounded in this paper and cite their sources.
Embedded pile formulations represent piles by beam elements superimposed on an overlapping soil continuum and are widely used in geotechnical finite-element analysis. Recent developments extend the coupling from the pile centreline to an implicit interaction surface at the physical pile circumference, introducing modelling choices whose consequences are not yet fully established. This study formulates and assesses these choices within a contact-consistent embedded beam formulation implemented in the freely available finite-element code numgeo . The mechanical normal traction is distinguished from the pressure entering the Coulomb criterion, complete traction is suppressed upon separation, and a capped generalised-hyperbolic relation is proposed for gradual base mobilisation. Verification and benchmark analyses under axial tension, axial compression and lateral loading are performed against resolved continuum models. The choice of capacity-pressure source is shown to control sensitivity to the normal penalty: host-stress recovery largely decouples terminal shaft resistance from the penalty stiffness, whereas interface pressure introduces a direct dependence. A modal analysis further shows that interface and host stiffnesses act in parallel for uniform radial deformation, rendering series-type harmonic-mean softening inconsistent with the resolved-host formulation. For base loading, elastic-perfectly plastic and Kondner-type relations mobilise resistance too rapidly, whereas the proposed relation reproduces the gradual continuum response. Under lateral loading, refined embedded models reproduce both pile-head resistance and pile-axis deformation, while a tenfold increase in all interaction stiffnesses changes the terminal reaction by only 2.2%. • A configurable interaction-surface embedded beam in free finite-element software. • Contact-consistent shaft and base: open points lose all traction, not only normal. • Capped hyperbolic base relation with a shape exponent and closed-form cap settlement. • Axial and lateral benchmarks against resolved continuum models quantify each choice.
No takes yet. Share an insight, caveat, or question.
Jan Macháček (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: