Suction caissons in dense sand, part I: installation, limiting capacity and drainage

Abstract

Suction caissons are a promising foundation concept for supporting offshore wind turbines. Compared to applications in the oil and gas industry, where most practical experience exists, significant differences arise in terms of load paths and magnitudes, soil type and caisson aspect ratio (skirt length to diameter). In a set of two companion papers, this contribution investigates the response of suction caissons in dense sand through a series of centrifuge experiments. The caisson was installed using suction, followed by sequences of cyclic loading and then extraction, all steps completed continuously in-flight. This first paper discusses installation, limiting capacities and drainage, whereas the second paper focuses on vertical cyclic loading into tension. The work demonstrates that suction caisson installation behaviour is well described by existing calculation methods. Tests performed at different installation rates demonstrate that careful assessment of the pumping rate is needed to ensure successful installation, with low pumping rates resulting in premature refusal. In the centrifuge tests, full skirt penetration was achieved without apparent loosening of the soil plug. The limiting capacity in tension, measured during the testing at both fast and slow uplift rates, was also well described by existing calculation methods.

Suction caissons in dense sand, part II: vertical cyclic loading into tension

“…(2017), but in brief involved in-flight installation of a suction caisson (with diameter, 157 D = 80 mm, skirt length, L = 40 mm, skirt thickness, t = 0.5 mm) into Baskarp sand with 158 a relative density, Dr = 98±1%, at 100 g. The sand was saturated with pore fluids having 159 a viscosity 100 and 660 times that of water to generate differing drainage responses during 160 installation and loading. All tests considered in this paper had the same installation 161 history, with self-weight installation followed by suction installation at a pumping rate of Suction caissons in dense sand, Part II: Bienen, Klinkvort, O'Loughlin, Zhu, Byrne Vertical cyclic loading into tension _____________________________________________________________________________________ 6 392.7 mm 3 /s (model scale, with the choice of flow rate and installation response discussed 163 in Bienen et al, 2017). 164…”

“…At full skirt embedment, the drained frictional tensile capacity of this caisson with an 176 aspect ratio L/D = 0.5 (where D is the caisson diameter) in very dense Baskarp sand 177 (Dr = 98%) was established to be around 15 kPa for z/L = 1 (Bienen et al, 2017). 178…”

“…The cyclic testing reflected loading on the individual caissons supporting a tripod jacket 132 structure, particularly the windward caisson. Detailed consideration of caisson load paths 133 has identified the most relevant loading scenarios to be explored (with details of the 134 background provided in Bienen et al, 2017), particularly with low average stresses 135 (around zero), comparable in magnitude to the frictional capacity. This is a significant difference to previous testing of caissons in sand (Kelly et al, 2006b).…”

“…As this paper focuses on the effect of cyclic loading into 171 the tensile region, a suitable reference load is the tensile drained frictional capacity of the 172 caisson, which depends on the skirt embedment, the sand relative density and the 173 mobilised friction angle at the skirt-soil interface. The caisson installation and cyclic 174 loading history plays a role in the resulting frictional tensile capacity (Bienen et al, 2017). 175…”

“…Two soils were used: (i) Redhill 191 110 sand at a relative density of 80-82% with a loading frequency of 1 Hz and (ii) 192 Oakamoor HPF5 sand with a relative density of 53-73% at 1 Hz and 10 Hz. The particle 193 size distributions of the Redhill and Oakamoor sands are provided in the companion paper Suction caissons in dense sand, Part II: Bienen, Klinkvort, O'Loughlin, Zhu, Byrne Vertical cyclic loading into tension _____________________________________________________________________________________ 7 (Bienen et al, 2017), with that of Redhill being similar to the Baskarp sand used in the 195 centrifuge tests. The estimated permeability of the two sands differed by three orders of 196 magnitude, with estimates of 5×10 -4 m/s for Redhill 110 and 5×10 -7 m/s for Oakamoor 197 HPF5.…”