A pile load test is a reliable method to validate pile design and assess uncertainties in the in-situ soil conditions and parameters. One particularly innovative approach is the Osterberg Cell (O-cell) test. Unlike conventional methods that apply load from the top, the O-cell technique applies load from the bottom of the pile. A hydraulically actuated cell, installed near the base, expands under pressure, pushing upward against the shaft and downward against the pile base. The upward resistance mobilizes shaft friction, while the downward force mobilizes end bearing resistance. The pressure is gradually increased until the pile reaches its ultimate capacity, either through side shear, end bearing, or a combination of both.

Figure 1: Pile-load test configuration - Stratigraphy and cell locations

The pile configuration under investigation is shown in Figure 1. Concrete was placed only between elevations -14.0 m and -51.3 m. A temporary steel casing was installed above -14.0 m to prevent shaft friction between the pile and the surrounding soil. The load testing was performed on a 1000 mm × 2800 mm barrette using two distinct O-cell assemblies positioned at different depths:

• The upper O-cell was installed between elevations -34.70 m and -35.20 m,
• The lower O-cell was positioned between -46.70 m and -47.20 m.

Displacements were recorded at both the upper and lower bearing plates for each O-cell assembly.

The load test was conducted in two stages:

• First, the lower O-cell assembly was pressurized to 53.20 MN and then unloaded.
• Next, the upper assembly was pressurized to 58.63 MN and subsequently relaxed. During this phase, the lower assembly was initially drained and then locked.

The experimental results from these tests are presented in Figure 2.

Figure 2: Pile-load test experimental results

This example focuses on modelling an O-cell pile load test in PLAXIS 3D using volume (continuum) elements. It highlights several modelling techniques aimed at:

• enhancing the computational efficiency of the numerical model, and
• achieving accurate calibration against the target load–displacement curves.

Taking advantage of the project’s geometric symmetry, only one quarter of the barrette is modelled. Furthermore, the current analysis is limited to the first loading and unloading stage.