Critical Pressure during Installation of Suction Caissons in Sand

Suction caissons are used in a variety of bottom founded and floating offshore applications, in shallow and deep waters. The present research focuses on the analysis of suction-assisted installation in sand. Suction pressure measurements from a number of field projects and laboratory centrifuge tests in sandy soils led to find that, in 70% of the studied cases, the critical suction pressure is underestimated by current design methods. The main goal of the present thesis is to improve the understanding of the underlying physical mechanisms. More reliable assessment of the critical suction pressure will result in more cost-effective foundations. Five hypotheses were introduced as possible driving mechanisms of the attainment of critical suction: H1: Change in void ratio (?e) and permeability (?k) of the sand plug. H2: Soil arching caused by friction at the inner caisson wall. H3: Uplifted soil particles of the plug surface at the final phase of installation. H4: Soil internal stability (internal rearrangement of soil particles). H5: Transient versus steady-state flow pattern during installation in sand. A physical model was used to investigate the validity of hypotheses H1 through H4, namely the Laboratory Upward Flow Test (LUFT). LUFT is a normal ground gravity (1g) experiment based on ASTM standards for measuring the permeability of granular soils, using a constant head method. It constitutes a simplification of the suction caisson installation process. Twenty five LUFT experiments were performed. An experimental parametric study was carried out by varying: relative density (Dr), sample gradation (sand grain size, coefficients of uniformity (Cu) and curvature (Cc)), sample height over sample diameter or aspect ratio (?) and hydraulic head difference (?h). The term “experimental critical hydraulic gradient” was introduced in order to express the mean value of the hydraulic gradient along the sample when the critical condition occurs. It is proven that the assumption of steady-state conditions during suction-assisted caisson installation in sand (hypothesis H5) is valid, based on Finite Element (FE) analyses. The experimental approach of the hypotheses H1 though H4 is well based. It is concluded that the aspect ratio (?) and the coefficient of uniformity (Cu) influence the experimental critical hydraulic gradient. The hypotheses which are linked to those two parameters are H2 and H4. Hypotheses H1 and H3 do not affect the experimental critical hydraulic gradient, based on LUFT results. Experimental critical hydraulic gradient values 15% higher than the “theoretical critical hydraulic gradient” were observed for non-uniform samples with ? of 1.0 and 1.5 (most common in practice). This can be explained by presence of fine particles in between the coarser ones, increasing the soil structure stability that results in higher pressure differences. Higher experimental critical hydraulic gradient values are obtained with greater ? in dense samples. Interface (soil-wall) shear resistance increases with sample height and greater seepage forces can be applied before failure occurs. The observed failure mechanism for uniform samples is piping. Absence of finer particles facilitates pipe creation through the soil structure. Non-uniform samples with ? > 0.5 fail with sample separation, which is thought to be a shear failure mechanism. Voids between coarse particles are obstructed by finer ones and water cannot escape through. The internal soil resistance is higher than the shear resistance at the wall and shear failure occurs.

Subject

Suction
Caisson
Installation
Sand
Critical hydraulic gradient
Critical pressure

To reference this document use:

http://resolver.tudelft.nl/uuid:4346855e-c3e6-4bab-8ad9-3afb71a12c14

Part of collection

Student theses

Document type

master thesis

Rights

Files

PDF

Panagoulias2015.pdf

459.56 MB

Close viewer