Document Type

Conference Paper


Available under a Creative Commons Attribution Non-Commercial Share Alike 4.0 International Licence


Geology, 2.1 CIVIL ENGINEERING, Civil engineering, Environmental and geological engineering, Geotechnics

Publication Details

This conference paper was published on 2019 by Editorial Universidad Santo Tomas in the Proceedings of the 1er CONGRESO INTERNACIONAL en Innovación, Desarrollo y Aplicaciones en Ingeniería Civil CI-IDeA. Full proceedings available online at


In geotechnical engineering the design of bored- pile walls is commonly performed in terms of the embedment length (D) defined according to the limit equilibrium methods (also known as balance methods) formulated by Krey, Blum, Rowe, Hansen and the empirical method of Look in sandy soils. The limit equilibrium methods are calculations associated with the shape of the earth pressures distributions in the shaft, earth pressures theories, geotechnical properties and safety factors, while the empirical criteria establish that the relationship between embedment and the free height (H) is of the order of D/H = 1.5. These design methods are complemented by reliability-based analyzes to evaluate the uncertainties associated with soil variability and the use of deterministic concepts. The Monte Carlo method is applied to evaluate the inherent variability of soil friction angle and the soil-wall interface angle in the calculation of the stability of adjacent bored-pile walls.

The results show that Hansen’s method requires a lower D/H ratio to obtain the failure probabilities of the USACE (PF = 0.001) in comparison with the other methods. The above is because the method considers that the passive earth pressure increases linearly from the excavation level to the bottom of the embedment depth, however, the other methods considers an earth pressure balance around the pivot point in the shaft. The results show that for a probability of failure established at 0.001, the obtained D/H by the Krey, Blum, Rowe and Hansen methods are 2.13, 1.97, 1.97 and 1.93, respectively. These values represent differences between 28.7% and 42.0% with respect to the deterministic design criteria of Look. Thus, the Look criterion, from the probabilistic point of view, can be considered as an inadmissible risk method, especially since it does not consider surcharge and water table.


Administrative Department of Science, Technology and Innovation of Colombia – Colciencias