Transient Vertical Loading of Model Suction Caissons in a

Pressure Chamber

by

R.B. Kelly, G.T. Houlsby and B.W. Byrne

Report No. OUEL 2291/06

University of Oxford Department of Engineering Science Parks Road, Oxford, OX1 3PJ, U.K.

Tel. 01865 273162/283300

Fax. 01865 283301 Email Civil@eng.ox.ac.uk

http://www-civil.eng.ox.ac.uk/

Transient Vertical Loading of Model Suction Caissons in a Pressure Chamber

R.B. Kelly, G.T. Houlsby and B.W. Byrne

This report consists of four papers and two technical reports that have resulted from a study on vertical loading of model suction caissons in a pressurised chamber. The work was carried out as part of a joint industry project investigating the application of suction caissons to offshore wind turbines. The four papers describe in varying detail aspects of the work whilst the technical reports contain greater detail on the tests carried out (in two phases) and provide records of the test data. The work consists of: a) “Transient vertical loading of model suction caissons in a pressure chamber.” Kelly, R.B., Houlsby, G.T. and Byrne, B.W. b) “Pressure chamber testing of model caisson foundations in sand.” Presented at BGA International Conference on Foundations, Dundee, September 2003. Kelly, R.B., Houlsby, G.T., Byrne, B.W and Martin, C.M. c) “Tensile loading of model caisson foundations for structures on sand.” Presented at International Symposium of Offshore and Polar Engineering, Toulon, France, 2004. Kelly, R.B., Houlsby, G.T., Byrne, B.W. and Martin, C.M. d) “The tensile capacity of suction caissons in sand under rapid loading.” Presented at International Symposium for Frontiers in Geotechnics, Perth, Australia, 2005. Kelly, R.B., Houlsby, G.T. and Byrne, B.W. d) “Vertical loading tests in a pressurised chamber: Phase one experimental data.” Kelly, R.B., Byrne, B.W., Houlsby, G.T. and Martin, C.M. e) “Vertical loading tests in a pressurised chamber: Phase two experimental data.” Kelly, R.B., Byrne, B.W., Houlsby, G.T. and Martin, C.M.

1

Transient vertical loading of model suction caissons in a pressure chamber

R.B. Kelly1, G.T. Houlsby2 and B.W. Byrne2 Keywords: serviceability loading, ultimate loading, foundations, offshore

ABSTRACT

Vertical monotonic and cyclic loading tests on model suction caissons have been carried out,

to explore conditions relevant to the design of multiple-footing foundations for offshore wind

turbines. The tests were conducted in a pressurised chamber, which simulate up to 20m of water

depth, to investigate the effect of cavitation on the ultimate tensile response of the caisson. Cyclic

loading tests were also conducted at elevated pressures. Data from cyclic tests are presented to

compare the response of caissons subjected to varying ambient pressures and rates of loading. The

ambient pressure had little effect on the cyclic response. Faster rates of loading generated larger

pressures beneath the lid of the caisson and increased the load-displacement stiffness. The ultimate

tensile capacity was significantly affected by the ambient pressure and rate of loading. Implications

for the design of offshore caisson foundations are discussed.

INTRODUCTION

The offshore wind energy industry is a rapidly expanding sector in the UK, with over 500

wind turbines planned for construction within the next few years and up to 2000 turbines by the

year 2010. The first developments are being constructed near-shore, in water up to about 15m deep,

and will mainly be founded on large “mono-piles”. The cost of mono-pile foundations is a

significant part of the total cost of an offshore wind turbine installation. Mono-piles are the

preferred technology as they have been extensively proven offshore. However, when wind farms

are built in greater water depths, alternative designs may be necessary. The stiffness of the sub-sea

support structure and foundation system will be critical to the overall design, as the turbine

superstructure is dynamically sensitive. To increase the stiffness of the support structure it is likely

that multiple-footing designs will be used, consisting of piles or alternatives. It is likely that

multiple pile foundations may not be the most economic solution in some instances. An alternative

is to use a tetrapod jacket structure incorporating “suction caisson” footings at each corner (Byrne

and Houlsby, 2003). Figure 1 shows a typical tetrapod, with the salient dimensions. Similar

arrangements have been used offshore as foundations for a small number of fixed platforms (Bye et

al., 1995; Tjelta 1994, 1995).

1 Coffey Geotechnics Pty Ltd, Sydney, NSW, Australia (formerly Research Assistant, University of Oxford). 2 Department of Engineering Science, University of Oxford.

2

A key feature of offshore wind turbine structures is that, for their size, they are relatively

light, yet they are subjected to large horizontal forces and overturning moments from wind and

waves. In the case of a tetrapod foundation, the overturning moment is principally carried by “push-

pull” action by opposing footings, and it is the variation of vertical load, and in particular the

possibility of tension on a footing, that is most important (Byrne and Houlsby, 2003). The design

needs are to select an appropriate diameter D and depth h of the caisson, as well as the spacing s of

the foundations (Figure 1). Quantifying the allowable tension on the caisson is vital for designing

the size of the structure. A research project, Byrne et al. (2002), has been carried out to explore the

design of suction caissons under various loading conditions, especially the vertical loading relevant

to multiple-footing structures. Design guidelines for vertical loading are developed principally

based on laboratory tests on small-scale model caissons (Byrne and Houlsby, 2002, 2004; Byrne et

al., 2003; Kelly et al., 2003, 2004, 2006a), verified by field scale testing (Houlsby et al., 2005b).

This paper reports selected vertical loading data from laboratory tests, with the following features:

(a) A computer-controlled hydraulic actuator has been used, so that cyclic loading paths can

be applied at high loading rates. The model used here is significantly larger than in tests reported by

Byrne and Houlsby (2002, 2004), and the loads are correspondingly larger too.

(b) Two sands of different gradings have been used, so that a range of behaviour, from

drained to approaching undrained, could be obtained. This compares with the oil-saturated fine sand

used by Byrne and Houlsby (2002, 2004) in which the foundation response was closer to undrained

conditions than to drained.

(c) Tests were carried out in a pressurised container, so that the effect of the water depth could

be simulated. This is important as the pressure (relative to seabed pressure) at which cavitation of

the pore fluid occurs will be dependent on the ambient water pressure. The principle of effective

stress guarantees that altering the water pressure should of course have no impact on the response of

the soil unless, when the footing is under tension and/or the soil is dilating rapidly, the suction in

the pore fluid becomes sufficiently large for cavitation to occur. The development of suctions in

dilative sand under undrained conditions is discussed by McManus and Davis (1997). The suctions

will be limited by the cavitation pressure of the fluid, which may therefore influence the overall

foundation response. Tests were carried out at atmospheric pressure and at 200kPa above

atmospheric (i.e. simulating a water depth of 20m) to examine the effect of the cavitation pressure.

For further details of the testing programme, testing equipment and procedures, and

comprehensive records of individual tests see Kelly et al. (2006b).

3

EQUIPMENT AND TESTING PROCEDURES

The pressure chamber, loading apparatus and model caisson are shown in Figures 2 and 3.

The chamber is a watertight cylinder, 1m diameter and 1m high. A hydraulically powered actuator,

fixed to the lid of the chamber, provided either load or displacement controlled vertical loading on

the model caisson. The caisson was located beneath the lid of the pressure chamber and fixed to the

actuator via a ram passing through a gland in the chamber lid. The model caisson was 280mm in

diameter, had a skirt length 180mm and wall thickness 3.125mm. A custom built water-proof load

cell was fixed on top of the caisson to record the vertical load. A second load cell was located

outside the pressure chamber, recording the loads applied by the actuator and used for feedback for

actuator control. A comparison of the loads from the two cells provided an indication of the friction

on the ram passing through the gland. Frictional loads of several kilonewtons were commonly

observed. Discrepancies between the quoted nominal values of the mean and cyclic loads (measured

and controlled by the external load cell) and the load data presented in the figures (measured by the

internal load cell) are due to friction on the ram.

Pressure sensors were attached to the caisson to record water pressures beneath the caisson lid

and at the skirt tip. The latter was measured with an Entran micro-miniature sensor placed flush

with the base of the skirt, in a 3mm diameter aluminium tube which was glued into a groove milled

into the inside of the skirt wall. A pressure sensor was also fixed in the chamber wall above water

level to record the ambient pressure. The caisson displacement was measured by an LVDT fixed to

the ram outside the pressure chamber. The actuator also incorporated an integral displacement

transducer. The data were recorded using PC-based data acquisition systems, and the actuator was

controlled using software supplied by Instron.

Two test beds of saturated sand were constructed in the chamber, so that the effects of

drainage on foundation response could be explored. One used Redhill 110, a fine-grained silica

sand. The other used Oakamoor HPF5, a sandy-silt artificially created by crushing, which creates

highly angular particles leading to high frictional strengths. The particle size distributions of the

sands are given in Figure 4. Maximum and minimum void ratios, and permeabilities estimated by

Houlsby et al. (2005a) for each material are shown in Table 1. The permeabilities were estimated

using Hazen’s empirical correlation, and therefore indicate only approximate relative values. Peak

friction angles were estimated from drained triaxial compression tests as 48.4º for the HPF5 (RD =

78%) and 43.9º for the Redhill 110 sand (RD = 75%). Estimated critical state angles are 41º for

HPF5 and 36º for Redhill 110.

The Redhill 110 test bed was constructed in two stages. Initially a filter layer, approximately

4

80mm thick, of Leighton Buzzard 16-30 silica sand was placed. Redhill 110 was then placed over

the filter layer. The sample was saturated by flushing water through it from the base of the chamber.

Any remaining air was removed by applying a vacuum to the top of the pressure chamber and

flushing carbon dioxide through the bed from the bottom. The HPF5 test bed was constructed in a

different manner, to minimise the hazard from fine silica dust raised during placement of this sand.

The water was drained from the pressure chamber and most of the Redhill 110 removed. A 10mm

layer of Redhill 110 was left to provide an additional filter above the 80mm filter of Leighton

Buzzard 16-30 sand. The tank was then filled with water and the HPF5 sand was placed through the

water into the tank. Repeated vibration of the pressure chamber was used to remove air from the

sample.

Each test bed was prepared to the required density by first loosening the sand by applying an

upward hydraulic gradient. The bed was then densified by placing a flat steel plate on top of the

soil, to provide a surcharge, and applying vibration to the walls of the chamber. The relative density

of the Redhill 110 samples was 80% to 82%, corresponding to submerged unit weights of 9.8

kN/m3 to 9.9kN/m3. The HPF5 samples had relative densities from 53% to 73%, corresponding to

submerged unit weights of 9.4 kN/m3 to 10.0 kN/m3.

The caissons were installed by slowly pushing them into the test bed. A vent valve in the

caisson lid was open to prevent water pressure building up inside the caisson (which might cause

piping in the sand). The caisson was installed at a rate of 0.1mm/s in Redhill 110 and at 0.05mm/s

in HPF5, until the lid of the caisson came in contact with the water, when the rate was reduced to

0.02mm/s. Typically the lid of the caisson made contact with the sand at a vertical load of about

2kN (vertical stress of 32kPa), and the installation was continued until a load of 5kN (81kPa) was

reached, to ensure firm contact with the sand surface. At this stage, for tests in Redhill 110, the

installation was paused and a slight increase in ambient pressure was applied to the chamber to

purge air from within the caisson through a drainage line. This was not necessary for tests in HPF5,

as there was a slight flow of water continuously through the drainage line during installation.

Installation was then continued until a load of 35kN (568kPa) was applied. Data from the

installation of the caissons in Redhill 110 sand have been presented by Kelly et al. (2003), who also

present equations predicting the loads required for installation, taking into account the enhancement

of skin friction due to down-drag of the sand adjacent to the caisson (Houlsby and Byrne, 2005).

The vent valve to the caisson was then closed, and cyclic loading and/or ultimate tensile loading

was applied to the caisson. Brief details of the tests presented in this paper are given in Table 2, and

comprehensive descriptions of all 27 tests are given by Kelly et al. (2006b). The cyclic loads

applied to the caisson during the multiple-amplitude cyclic tests are given in Table 3.

5

TEST RESULTS

Multiple-amplitude cyclic loading

The cyclic test data are presented in terms of vertical stress, computed as the applied vertical

load divided by the cross-sectional area of the caisson. Unless stated otherwise, the water pressure

refers to that measured beneath the caisson lid.

Vertical stress and water pressure during a multiple-amplitude cyclic loading test at 1Hz in

Redhill 110 sand are shown in Figure 5. The data show that the displacement per cycle increases

with stress amplitude and is greatest in the first cycle of each set at any stress amplitude (details are

discussed further below). During the last two cyclic load packets, the stiffness reduces significantly

as the minimum load approaches 0kN, and becomes very small once the vertical load becomes

tensile. During the last packet of cycles the target loads were 35kN ± 40kN (569kPa ± 650kPa),

indicating a target tensile load of -5kN, corresponding to an average tensile stress of –81kPa. In fact

this could not be attained, with the maximum tensile stress recorded in any test being –32kPa. Note

particularly that, although tensile stresses are accompanied by large upward displacements, over the

course of an entire load cycle the net displacement was downwards. These observations are

consistent with those reported by Byrne and Houlsby (2002).

Comparison of Figures 5 and 6 shows that the increase in pressure in Test 6 by 200kPa has

had virtually no effect on the vertical stress-deformation response, but has just increased all water

pressures by 200kPa. This is exactly as one would expect from the principle of effective stress, in

the absence of cavitation, and is therefore indirect evidence that cavitation did not occur during

these cyclic tests. Minor differences between the two tests are mainly attributable to the fact that (i)

in Test 6, ten rather than five cycles were carried out in each of the last two cycling packages, (ii)

the vertical loads in Test 2 are slightly higher and, (iii) the relative density of the sand in Test 6 is

slightly lower.

The measured water pressure under the caisson lid (Figures 5(b) and 6(b)) shows that the

pressure amplitude increased with the load amplitude. The changes in pressure (the difference

between the maximum and minimum pressures in a cycle), are plotted against the corresponding

values for change in vertical stress in Figure 7. The data show a near identical pressure-stress

response for the two tests, indicating that the magnitude of the ambient pressure had little influence,

as noted above. The pressures generated during the tests were small, typically about 4.4% of the

applied vertical stress, although there is some nonlinearity in the pressure-stress relationship, with

marginally higher pressure ratios at larger amplitudes.

6

The water pressure measured at the skirt tip in Test 6 is compared with the lid pressure in

Figure 8. The excess pressure at the skirt tip is a small proportion of that measured at the lid. This

was typical of all tests where the skirt tip water pressure was recorded.

Data from multiple-amplitude cyclic tests in HPF5 sand at different loading frequencies are

presented in Figures 9 to 11, showing tests at 0.1Hz, 1Hz and 10Hz respectively. Differences of up

to 15% in the load amplitudes applied are linked to the performance of the load control system at

different frequencies: at the higher rates it is difficult to track the loads, especially for the large

amplitude cycles. In particular in Test 21 (Figure 11) the loading does not go into tension, and there

is a consequent impact on the displacement response for this test. However, for the cyclic packages

with positive loads, there is evidence that as the rate of loading increased the overall penetration of

the caisson decreased: the accumulated deformations were much smaller. The stiffness during

individual cycles is examined in more detail later. At higher frequencies the pressure response

increased. The recorded pressures shown in Figure 11(b) are cut off at about 400kPa, which was the

maximum range of the sensor.

The change in pressure is plotted against change in vertical stress in Figure 12. Approximately

linear trends between pressure and stress could be identified for the tests conducted at 0.1Hz and

1Hz. The pressure is about 15% of the vertical stress in the 0.1Hz test and 25% for the 1Hz test. The

pressure is much higher in these tests than the 4.4% recorded in the coarser Redhill 110 sand,

because the permeability of the HPF5 sand is about three orders of magnitude less than that of

Redhill 110. For the test at 10Hz the relationship between pressure and stress is strongly nonlinear,

but it is just possible that this result has been affected by the response time of the pressure

transducer. There is, however, a consistent trend that the measured pore pressures are larger at

higher rates of cycling and for a less permeable sand.

This may seem, at first sight, an entirely expected observation. However, it is useful to

consider the mechanisms occurring within the soil around the foundation. The problem is one of

transient loading to a large body of soil where, importantly, partial drainage occurs (i.e. the soil is

neither fully undrained nor drained). Load application will increase the mean stress immediately

below the footing, leading to an increase in excess pore water pressure. Clearly the faster the load is

applied, relative to the permeability of the sand, the greater the ratio of vu σ∆∆ will be expected.

However, the soil will also shear to sustain the applied load. In dense sand this shearing leads to

dilation and, if this is suppressed due to insufficient inflow of water, negative excess pressures will

occur. The faster the load application the higher the negative excess pressures in dilating zones. The

different excess pressures at different locations will lead to flow, and the pressures measured at

7

discrete points will reflect in some way a balance between conflicting mechanisms. Higher loads

will lead both to a higher mean stress and also greater shearing of the soil, and therefore to higher

negative excess pressures. To assess these interactions fully would require measuring water

pressures at a number of points in a radial plane through the soil.

Bearing in mind the above comments, it is possible to assess qualitatively the degree of

drainage. For drained sands the pressure can be related to the velocity of the caisson through

Darcy’s law, e.g. Houlsby et al. (2005a), whilst for undrained conditions the pressure would be

expected to be related to the displacement. By examining the relative phases of the displacement,

velocity, stress and pressure in the cycles, it is possible to gauge the degree of drainage. The phase

has been obtained by fitting sinusoids to the relevant signals for the 4th cyclic load packet in Tests 6,

14, 15 and 21, and the results are shown in Table 4. The phase angles in Table 4 are calculated

relative to the displacement, i.e. the phase angle for displacement is taken as zero. The computed

phase for the velocity ranged from 86o to 88o, which is close to the theoretical 90o (the small

difference being due to the fact that neither the displacement nor velocity varies purely sinusoidally

with time). The pressure was found to be approximately in phase with the velocity in Test 6, in

Redhill 110 sand, but closer to the phase of the displacement in the tests conducted in the finer

HPF5 sand. Furthermore in HPF5 sand, as the cyclic loading rate increased the pressure was more

closely in phase with the displacement. These results suggest that the tests in Redhill 110 were in

the drained to partially drained range whilst the tests in HPF5 approached undrained behaviour.

Sinusoidal curves were also fitted to plots of effective load (defined as the applied load minus the

load due to the pore pressure, uAVV −=′ ) with time. The results are also presented in Table 4. The

effective load was broadly in phase with the displacement for all tests, as would be expected.

Differences in phase between effective load and displacement could be due to viscous damping and

dynamic loading, and the faster test in HPF5 show more phase difference. This may be related to

the fact that the change in pressure was not linear with the change in vertical stress (Figure 12).

Definitions of cyclic unloading stiffness and incremental displacement are shown in Figure

13. The variation of the unloading stiffness is plotted in Figure 14 against the number of cycles for

each amplitude of loading applied in Test 6. The unloading stiffness increased with the number of

cycles, although at decreasing rate, during the ±5kN (±81kPa), ±10kN (±162kPa) and ±20kN

(±325kPa) load packets. The unloading stiffness was almost constant during cyclic load packets in

excess of ±20kN (±325kPa), with the stiffness decreasing as the load amplitude increased. Figure

15 shows the unloading stiffness over 1000 cycles for Test 12, conducted at 35kN ± 15kN (568kPa

± 244kPa). The stiffness increased at a very slow, and decreasing, rate throughout the test and could

be treated simply as constant without significant error. These observations are important, as they

8

provide no evidence of degradation of stiffness during cycling, but on the contrary indicate constant

or increasing stiffness.

Figure 16 shows the incremental and cumulative cyclic displacements during Test 15. The

incremental displacement per cycle decreased with the number of cycles and became small, but

positive, after about 200 cycles. About half of the cumulative displacement occurred during the first

200 cycles. Further “ratchetting” of the caisson into the sand then occurred at an almost constant

rate per cycle. Such a response would be expected to result ultimately in a hardening of the response

(as opposed to upward ratchetting movement, which would be of much greater concern to a

designer). There is insufficient data, however, to be able to quantify cumulative movements after

many cycles.

Ultimate tensile loading

Figure 17 shows vertical stress data from three pullout tests in Redhill 110 sand. In two tests

the caissons were pulled out of the sand at rates of 5mm/s and 100mm/s respectively at atmospheric

pressure. The third test was at an ambient pressure of 200kPa above atmospheric, and pulled out at

100mm/s. The arrows on the figure show the direction of movement with time. The data show that

the ultimate load increased with the pullout rate and (in stark contrast with the results for cyclic

loading) also increased with the magnitude of the ambient pressure. The pore pressures during these

tests are shown in Figure 18, showing that the pressure beneath the caisson lid was also dependent

on the rate of pullout. A small suction pressure was generated in the test at 5mm/s. In contrast the

cavitation pressure of –100kPa was reached in the test at 100mm/s. When the ambient pressure was

increased to 200kPa the cavitation limit (expected to be about -300kPa) was not reached, but a

suction of –250kPa was generated. Figure 19 shows suctions beneath the caisson lid and at the skirt

tip, along with the net vertical stress on the caisson, and Figure 20 a detail of the same data plotted

against time. Note the distinct change in stiffness of the load-deformation response as the net stress

on the caisson changes sign. As the caisson is pulled from the soil a rapid response is initially

observed at the lid but little response at the skirt tip ((a) to (b)). A short period later, less than 100

ms (at (b)), a small pressure drop is observed at the skirt tip, indicating that a pressure front

propagates downwards in the caisson. As the load on the caisson changes sign the pressure at the lid

sensor suddenly increases (at (c)) but the skirt pressure continues reducing, suggesting a change of

mechanism (although the detailed mechanics are not clear). Steadier conditions are attained after

(d), when the ratio of the skirt to lid excess pressure varies from about 0.8 at (d) to approximately

0.95 at (e).

In Figure 21, data from Tests 6 and 12, where cycling was carried out prior to pullout, are

9

compared with Test 10, where the caisson was pushed into the sand and immediately pulled out.

Because of the different loading histories, the caisson had penetrated different amounts prior to

pull-out (Test 6 to 267 mm, Test 10 to 211 mm and Test 12 to 240 mm). For comparison the start of

each pull-out has been rebased to 250 mm. The tests were all conducted at an ambient pressure of

200kPa in Redhill 110 sand. The ultimate tensile vertical stress after each of the cyclic tests was less

than in Test 10, but this ultimate tensile vertical stress was, however, directly related to the ultimate

suction pressure in each test. The ultimate suction pressures were smaller in the tests where cycling

was applied than in Test 10. The softened response from the caissons with the significant cyclic

loading histories indicates that cycling may have caused loosening on potential failure planes. This

is supported by the fact that the ultimate capacity in Test 10 (without cycling) was mobilised at a

smaller tensile displacement than for the tests with cycling. No direct measurement of the soil

density during the test was possible, nor was any assessment of heave around the foundation

performed. It is therefore not possible at this stage to state whether the soil became globally denser

or looser during cycling, in spite of the fact that net downward movement of the caisson had

occurred.

IMPLICATIONS FOR PROTOTYPE FOUNDATIONS

Serviceability loading

The transient tensile capacity of a skirted foundation depends on a complex interaction

between the permeability of the soil, the length of the drainage path and the rate of loading. A non-

dimensional parameter that incorporates these variables is Tv, the time parameter used in one-

dimensional consolidation analysis:

2HtcT vv = (1)

where ( )vwv mkc γ= is the co-efficient of consolidation, k is the permeability of the sand, mv is the

inverse of the constrained modulus of the sand, γw is the unit weight of water, t is the time taken for

a specified degree of pore pressure dissipation and H is the length of the drainage path. If it is

assumed that the constrained modulus is proportional to the square root of the mean vertical stress

in the sand beneath the caisson, and that the mean vertical stress and H are both proportional to the

diameter of the caisson then it follows that, for comparable degrees of consolidation in model and

prototype:

10

23

⎟⎟⎠

⎞⎜⎜⎝

⎛=

p

m

m

p

p

mDD

tt

kk

(2)

Where km and kp are the permeability of the model and prototype sands, Dm and Dp are the diameter

of the model and prototype caissons and tm and tp are the period of loading on the model and

prototype caissons. Equation 2 can be used to infer the behaviour of prototype-scale caissons from

the model-scale test data.

The diameter and skirt length of caissons in a full-scale tetrapod foundation supporting an

offshore wind turbine might be of the order of 6m and 4m respectively. The relevant period of wave

loading in UK coastal waters is in the order of 7s (~0.14Hz). A typical permeability of a sand in

offshore locations around the UK may be of the order of 0.001m/s (but clearly with a rather wide

variation). From Equation 2, the rate of loading in the laboratory, required to simulate these field

conditions, was 7 Hz for Redhill 110 sand and 0.007 Hz for HPF5. Based on these values, the tests

in Redhill 110 are the more relevant than the tests in HPF5 for this case. In the field caissons are

therefore likely to be loaded in partially-drained to drained conditions.

The laboratory tests conducted in HPF5 sand showed that large transient pore water pressures

were generated beneath the caisson lid when the caisson was loaded at rates of 1Hz and above.

There was no evidence of pressure accumulation during these tests, but only a few cycles were

applied. The limited model data suggest that accumulation of pore pressure within the caissons will

not occur at prototype scale. If, however, there is an accumulation of pore pressure, greater

deformations of the caissons can be expected. A method, such as that proposed by Bye et al (1995),

could be employed to estimate the accumulation of pore pressure, and design the caisson so that the

total vertical load minus the load due to positive pore pressure does not approach 0kN on the up-

wind foundation of the tetrapod.

Recent analyses suggest that foundations for offshore wind turbines will need to be designed

to exceed a stiffness criterion, rather than to resist an ultimate load case. This requirement allows

the structural stiffness of the turbine structure to be tuned to keep it within a narrow frequency

range. The data from the cyclic tests suggest that, as long as the total vertical load does not

approach 0kN, then the stiffness of the foundation will increase with the number of cycles. The

largest deformations are likely to occur during the first significant loading event after installation.

The size and spacing of the caissons in a tetrapod foundation should be designed to limit the load on

the upwind leg to 0kN at worst, and preferably a slight compressive load, to minimise vertical

deformations of the caissons under serviceability loading conditions. Kelly et al. (2006a) discuss in

11

detail the issue of scaling of stiffness between model and prototype, using evidence from larger

scale field trials.

Design for an ultimate tensile load

One might wish to design the spacing and diameter of caissons to resist an ultimate tensile

load during an extreme storm event with a long return period. The model data suggest that large

tensile loads can be attained under certain conditions. Furthermore, over the period of a cycle, if the

mean load is compressive and the caisson does not fail in tension, the net deformation will be

downward. The model data indicate that the ultimate tensile load is dependent on the suction

pressure that can be generated beneath the lid of the caisson during tensile loading. This will be

greatest in deeper water, in fine-grained sand and when the rate of tensile loading is fast. However,

the data also indicate that large tensions are only possible after significant displacements have

occurred, which may render the structure unserviceable after loading.

Houlsby et al. (2005a) present equations that allow the ultimate tensile load and suction

pressures in sand to be predicted. For example, when a caisson is pulled out sufficiently rapidly to

cause the pore water to cavitate, the ultimate tensile load can be computed using case (d) in Houlsby

et al. (2005a). The ultimate tensile load in this case is given by:

⎟⎠⎞

⎜⎝⎛ δ+−= tan21 oK

DhsAV (3)

where V is the ultimate tensile load, s is the suction pressure, A is the cross-sectional area of the

caisson, h is the length of the caissons skirts, D is the diameter of the caisson, Ko is the lateral earth

pressure co-efficient and δ is the sand-skirt interface friction angle. The variation with time of the

ultimate vertical stress and suction pressure for Test 16 (in HPF5 sand, at atmospheric pressure, and

pulled out at a rate of 100mm/s), and an estimate of the tensile stress using Equation 3, are shown in

Figure 22. The cavitation limit was reached in this test and the suction pressure was about –100kPa.

The vertical stress was calculated using a value 5.0tan =δK . Aside from minor differences prior to

the full cavitation pressure being reached, the ultimate vertical stress is predicted reasonably well.

Houlsby et al. (2005a) show that this approach is also successful for caissons pulled out at different

rates, in different sands, and subject to different ambient pressures. These analyses provide a

framework for the predicting the ultimate tensile load. However, note that this was shown to reduce

after the caisson has been cycled (see Figure 21). Thus there is considerable uncertainty involved in

designing caissons to withstand tensile loads (other than drained friction) and such a design could

not be recommended at this time.

12

CONCLUSIONS

Data have been presented from vertical loading tests of model caissons in sand. The tests were

conducted in a pressure vessel to simulate the effects of water depth, and were designed to

investigate the serviceability and ultimate tensile loading of vertically loaded caisson foundations.

The results indicate that:

(a) The ultimate tensile load can be estimated if the suction pressures beneath the caisson are

known. However, these suction pressures result from a complex interaction between the rate of

loading, the ambient pressure conditions, the soil type and the load history.

(b) The magnitude of the ambient pressure has little or no effect on the deformations of the

caisson during cyclic loading, as expected from an effective stress analysis. In contrast, the

magnitude of the ambient pressure increases the ultimate tensile load available, due to its effect on

the relative pressure at which cavitation occurs.

(c) Under cyclic loading the magnitudes of the pore pressures within the soil are relatively

small, and indicative of partial drainage conditions. As the rate of cycling increases the positive

pore pressures beneath the caisson increase, as does the vertical stiffness of the caisson.

ACKNOWLEDGEMENTS

This research was sponsored by the EPSRC, DTI and a consortium of companies (Fugro Ltd,

SLP Engineering Ltd, Garrad Hassan, GE Wind Ltd, NEG Micon and Shell Renewables Ltd).

REFERENCES

Bye, A., Erbrich, C., Rognlien, B. and Tjelta, T.I. (1995) “Geotechnical design of bucket foundations”, Proc. Offshore Technology Conference, OTC 7793.

Byrne, B.W. and Houlsby, G.T. (2002) “Experimental Investigations of the Response of Suction Caissons to Transient Vertical Loading”, Proc. ASCE, Journal of Geotechnical Engineering, Vol. 128, No. 11, Nov., pp 926-939

Byrne, B.W. and Houlsby, G.T. (2003) “Foundations for Offshore Wind Turbines”, Phil. Trans. of the Royal Society of London, Series A, Vol. 361, December, pp 2909-2930

Byrne, B.W. and Houlsby, G.T. (2004) “Experimental Investigations of the Response of Suction Caissons to Transient Combined Loading”, Proc. ASCE, Jour. of Geotechnical and Geoenvironmental Engineering, Vol. 130, No. 3, pp 240-253

Byrne, B.W., Houlsby, G.T., Martin, C.M. and Fish, P.M. (2002) “Suction caisson foundations for offshore wind turbines.” Wind Engineering, Vol. 26, No 3

Byrne, B.W., Villalobos, F. Houlsby, G.T. and Martin, C.M. (2003) “Laboratory Testing of Shallow Skirted Foundations in Sand”, Proc. Int. Conf. on Foundations, Dundee, 2-5 September, Thomas Telford, pp 161-173

13

Houlsby, G.T., Kelly, R.B. and Byrne, B.W. (2005a), “The tensile capacity of suction caissons in sand under rapid loading”, Proc. Int. Symp. on Frontiers in Offshore Geomechanics, Perth, Australia, pp 405-410

Houlsby, G.T., Kelly, R.B., Huxtable, J. and Byrne, B.W. (2005b), “Field trials of suction caissons in sand for offshore wind turbine foundations”, Géotechnique 56, pp 3-10

Kelly, R.B., Byrne, B.W., Houlsby, G.T. and Martin, C.M. (2003) “Pressure Chamber Testing of Model Caisson Foundations in Sand”, Proc. Int. Conf. Foundations, Dundee, 2-5 September, Thomas Telford, pp 421-431.

Kelly, R.B., Byrne, B.W., Houlsby, G.T. and Martin, C.M. (2004) “Tensile Loading of Model Caisson Foundations for Structures on Sand”, Proc. ISOPE, Toulon, Vol. 2, pp 638-641.

Kelly, R.B., Houlsby, G.T. and Byrne, B.W. (2006a) “A Comparison of Field and Laboratory Tests of Caisson Foundations in Sand and Clay”, Géotechnique, accepted.

Kelly, R.B., Houlsby, G.T. and Byrne, B.W. (2006b) “Transient vertical loading of model suction caissons in a pressure chamber”, Report OUEL 2291/06, Department of Engineering Science, University of Oxford

McManus, K.J. and Davis, R.O. (1997) “Dilation induced pore fluid cavitation in sands”, Géotechnique, Vol 47, pp 173-178

Tjelta, T.I. (1994) “Geotechnical aspects of bucket foundations replacing piles for the Europipe 16/11-E Jacket” Paper OTC 7379, Offshore Technology Conference, Houston, Texas

Tjelta, T.I. (1995) “Geotechnical experience from the installation of the Europipe Jacket with bucket foundations” Paper OTC 7795, Offshore Technology Conference, Houston, Texas

14

Material Maximum Void Ratio Minimum Void Ratio Estimated Permeability(m/s)

Redhill 110 1.037 0.547 5.0 x 10-4 Oakamoor HPF5 1.014 0.467 5.0 x 10-7 Table 1: Properties of Redhill 110 and Oakamoor HPF5 sands

Test Sand RD (%) γ′

(kN/m3) Pressure

(kPa) Frequency

(Hz) Pullout rate

(mm/s) Comments

2 Redhill 81 9.8 0 1 5 Multiple amplitude cyclic test5 Redhill 89 10.1 0 10 100 Multiple amplitude cyclic test6 Redhill 80 9.8 200 1 100 Multiple amplitude cyclic test9 Redhill 81 9.9 0 - 100 Stepped push in then pull out10 Redhill 82 9.9 200 - 100 Stepped push in then pull out11 Redhill 81 9.9 0 - 5 Push in then pull out 12 Redhill 82 9.9 200 0.5 100 1000 cycles 1535 ± kN 14 HPF5 53 9.4 0 1 5 Multiple amplitude cyclic test15 HPF5 55 9.5 0 0.1 10 Multiple amplitude cyclic test16 HPF5 67 9.8 0 - 100 Push in then pull out 21 HPF5 73 10.0 0 10 25 Multiple amplitude cyclic test

Table 2: Test details

Cycle Packet Number of Cycles Amplitude (kN) Amplitude (kPa)1 10 10 162 2 10 20 325 3 10 40 650 4 10 60 974 5* 5 70 1137 6* 5 80 1299

*Tests 5 and 6 have 10 cycles in packets 5 and 6

Table 3: Nominal cyclic loads applied to the model caisson

Displacement Velocity Vertical stress Lid pressure Effective Stress Test Sand Freq.

(Hz) Amp. (mm) Phase* Amp.

(mm/s) Phase Amp. (kPa) Phase Amp.

(kPa) Phase Amp. (kPa) Phase

6 Redhill 1.0 2.7 0o 17 86o 906 25o 35 102o 898 23o 15 HPF5 0.1 5.3 0o 3 88o 1096 14o 123 26o 976 12o 14 HPF5 1.0 5.1 0o 32 87o 1090 18o 322 10o 772 21o 21 HPF5 10.0 2.1 0o 130 86o 930 25o 206 6o 741 30o

* zero by definition: reference phase

Table 4: Amplitude and phase for the 4th cyclic load packet

15

Figure 1: Configuration and salient dimensions of a tetrapod suction caisson foundation for an offshore wind turbine

Figure 2: Pressure vessel, loading apparatus and instrumentation

16

(a) (b)

Figure 3: (a) pressure vessel, (b) model caisson

0

10

20

30

40

50

60

70

80

90

100

0.001 0.01 0.1 1

Particle Size (mm)

Per

cent

Pas

sing

Oakamoor HPF5Redhill 110

Figure 4: Particle size distribution for Redhill 110 and Oakamoor HPF5 sands

17

-200

0

200

400

600

800

1000

1200

1400

200 210 220 230 240 250 260 270

Displacement (mm)

Verti

cal s

tress

(kPa

)

Figure 5(a): Test 2 in Redhill 110 sand @ 1Hz and 0kPa ambient pressure: vertical stress

-200

0

200

400

600

800

1000

1200

1400

200 210 220 230 240 250 260 270

Displacement (mm)

Ver

tical

stre

ss (k

Pa)

Figure 6(a): Test 6 in Redhill 110 sand @ 1Hz and 200kPa ambient pressure: vertical stress

-40

-20

0

20

40

60

80

100

200 210 220 230 240 250 260 270

Displacement (mm)

Pre

ssur

e (k

Pa)

Figure 5(b): Test 2 in Redhill 110 sand @ 1Hz and 0kPa ambient pressure: lid water pressure

160

180

200

220

240

260

280

300

200 210 220 230 240 250 260 270Displacement (mm)

Pre

ssur

e (k

Pa)

Figure 6(b): Test 6 in Redhill 110 sand @ 1Hz and 200kPa ambient pressure: lid water pressure

18

0

10

20

30

40

50

60

70

80

0 200 400 600 800 1000 1200 1400

Change in vertical stress (kPa)

Cha

nge

in p

ress

ure

(kP

a)

Test 2 - 0kPaTest 6 - 200kPa

∆u/∆σv=0.044

Figure 7: Change in pore pressure during loading for Tests 2 and 6

160

180

200

220

240

260

280

300

210 220 230 240 250 260 270

Displacement (mm)

Pre

ssur

e (k

Pa)

Lid PPTSkirt PPT

Figure 8: Pressure at skirt tip compared to pressure beneath lid of caisson in Test 6

19

-200

0

200

400

600

800

1000

1200

1400

200 210 220 230 240 250 260 270

Displacement (mm)

Ver

tical

stre

ss (k

Pa)

Figure 9(a): Test 15 in HPF5 sand @ 0.1Hz and 0kPa ambient pressure: vertical stress

-200

0

200

400

600

800

1000

1200

1400

200 210 220 230 240 250 260 270

Displacement (mm)

Ver

tical

stre

ss (k

Pa)

Figure 10(a): Test 14 in HPF5 sand @ 1Hz and 0kPa ambient pressure: vertical stress

-200

0

200

400

600

800

1000

1200

1400

200 210 220 230 240 250 260 270

Displacement (mm)

Ver

tical

stre

ss (k

Pa)

Figure 11(a): Test 21 in HPF5 sand @ 10Hz and 0kPa ambient pressure: vertical stress

-100

-50

0

50

100

150

200

250

300

350

400

200 210 220 230 240 250 260 270

Displacement (mm)

Pre

ssur

e (k

Pa)

Figure 9(b): Test 15 in HPF5 sand @ 0.1Hz and 0kPa ambient pressure: lid water pressure

-100

-50

0

50

100

150

200

250

300

350

400

200 210 220 230 240 250 260 270

Displacement (mm)

Pre

ssur

e (k

Pa)

Figure 10(b): Test 14 in HPF5 sand @ 1Hz and 0kPa ambient pressure: lid water pressure

-100

-50

0

50

100

150

200

250

300

350

400

200 210 220 230 240 250 260 270

Displacement (mm)

Pre

ssur

e (k

Pa)

Figure 11(b) Test 21 in HPF5 sand @ 10Hz and 0kPa ambient pressure: lid water pressure

20

0

50

100

150

200

250

300

350

400

450

500

0 200 400 600 800 1000 1200 1400

Change vertical stress (kPa)

Cha

nge

in p

ress

ure

(kP

a)

Test 15 - 0.1HzTest 14 - 1HzTest 21 - 10Hz

∆u/∆σv=0.14

∆u/∆σv=0.25

Figure 12: Change in pore pressure during loading for Tests 14, 15 and 21

w

Unload stiffness

Displacement

Axi

al lo

ad Incremental displacement

Figure 13: Definitions of unload stiffness and incremental displacement

21

0

5

10

15

20

25

30

35

40

45

0 1 2 3 4 5 6 7 8 9 10

Number of cycles

Stif

fnes

s (k

N/m

m)

+/- 5kN+/- 10kN+/- 20kN+/- 30kN+/- 35kN+/- 40kN

Figure 14: Variation of unloading stiffness with number of cycles during Test 6 in Redhill 110 sand

0

10

20

30

40

50

60

70

1 10 100 1000

Number of cycles

Stif

fnes

s (k

N/m

m)

Figure 15: Unloading stiffness during cycling at 35kN ± 15kN in Redhill 110 sand

22

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

1 10 100 1000

Number of cycles

Incr

emen

tal d

ispl

acem

ent (

mm

)

0

10

20

30

40

50

60

70

Cum

ulat

ive

disp

lace

men

t (m

m)

IncrementalCumulative

Figure 16: Incremental and cumulative cyclic displacements

-450

-400

-350

-300

-250

-200

-150

-100

-50

0150 160 170 180 190 200 210

Displacement (mm)

Ver

tical

stre

ss (k

Pa)

5mm/s - 0kPa100mm/s - 0kPa100mm/s - 200kPa

Direction of movement

Figure 17: Ultimate tensile loading in Redhill 110 sand

23

-150

-100

-50

0

50

100

150

200

150 160 170 180 190 200 210

Displacement (mm)

Pre

ssur

e (k

Pa)

5mm/s - 0kPa100mm/s - 0kPa100mm/s - 200kPa

Figure 18: Pressures beneath the lid of the caisson during ultimate tensile loading

-400

-200

0

200

400

600

800

50 100 150 200 250Displacement (mm)

Ver

tical

Stre

ss (k

Pa)

-100

-50

0

50

100

150

200

Wat

er P

ress

ure

(kP

a)

Vertical StressLid Pressure SensorSkirt Pressure Sensor

Figure 19: Pressure at skirt tip and beneath lid of caisson in Test 6 during pullout

24

-480

-320

-160

0

160

320

480

640

800

4897.0 4897.2 4897.4 4897.6 4897.8 4898.0 4898.2 4898.4 4898.6

Time (sec)

Ver

tical

Stre

ss (k

Pa)

-120

-80

-40

0

40

80

120

160

200

Wat

er P

ress

ure

(kP

a)

Vertical StressLid Pressure SensorSkirt Pressure Sensor

(a)

(c)

(b)

(d)

(e)

Figure 20: Water pressures and vertical stress against time for Test 6 during pullout.

-450

-400

-350

-300

-250

-200

-150

-100

-50

0100 120 140 160 180 200 220 240 260

Displacement (mm)

Ver

tical

stre

ss (k

Pa)

1000 cycles: 35kN +/-15kN (Test 12)Multi-amplitude cycling (Test 6)Push-pull (Test 10)

Figure 21: Degradation of ultimate tensile capacity after cyclic loading

25

-180

-160

-140

-120

-100

-80

-60

-40

-20

013047 13047.5 13048 13048.5 13049 13049.5

Time (sec)

Stre

ss o

r Pre

ssur

e (k

Pa)

Vertical StressSuctionStress Prediction

Figure 22: Prediction of components of the ultimate tensile load in Test 16

Conference title. Thomas Telford, London, 2003

Pressure Chamber Testing of Model Caisson Foundations in Sand

R.B.Kelly, B.W.Byrne, G.T.Houlsby and C.M.Martin Department of Engineering Science University of Oxford Introduction Within the next few years a number of wind farms will be constructed around the coast of the United Kingdom. In the first instance many of the wind turbine structures will be founded on piles. These foundations, although simple to design as they are a well-established technology, are a significant proportion of the overall installed cost for these structures; of the order of 30%. Various options are being investigated that may reduce the installed cost, and therefore increase the economic viability of these wind-farm developments. One possibility is to use skirted shallow foundations installed by suction either as a single foundation (i.e. monopod structure) or as a multi-foundation system (i.e. quadruped/tripod structure). For the monopod, the key issue is the performance of the foundation under the large moments applied by the wind and wave loads on the structure. For the multi-footing case the applied moment loads from the wind and waves will largely be reacted as vertical compression and tension loads on the individual foundations. Both these loading cases are being investigated at the University of Oxford in laboratory scale tests. This paper concentrates on the tensile vertical capacity of the caisson foundation. A companion paper by Byrne et al. (2003) covers the moment loading case.

There is little previous research on the vertical cyclic loading of skirted foundations in sand. Most recent research has been carried out at the University of Oxford (Byrne, 2000; Johnson, 1999). The research carried out by Byrne (2000) involved vertical loading experiments on skirted foundations in an oil-saturated sand. These indicated that significant tensile capacities are possible and are limited by cavitation of the pore fluid. There is also evidence that the stiffness of the response changes when the load applied to the foundation changes from compression to tension (Byrne and Houlsby, 2002; Johnson, 1999). These experiments were carried out on the laboratory floor, where the

Pressure Chamber Testing of Model Caisson Foundations in Sand 2

ambient fluid pressure is very low. A central issue is therefore that, since the available suction pressure is limited by cavitation of the pore fluid, the maximum available suction is much lower on the laboratory floor as compared with the offshore case. To investigate the effect of water pressure, and therefore cavitation, on the behaviour of suction caissons, a pressure chamber has been developed. This can simulate water pressures corresponding to depths of up to 20m. This paper describes the pressure chamber and presents data from a commissioning test conducted at atmospheric pressure.

A model caisson having a diameter of 280mm and a skirt depth of 180mm was used in the experiment. A waterproof load cell was constructed in order to measure the axial load on the caisson and any moment loads that may have occurred during loading. The displacement of the caisson and the water pressure beneath the caisson base plate were recorded. A fast-acting 100kN Instron actuator applied the loads.

Pressure Chamber A diagram of the pressure chamber is presented in Figure 1. The pressure chamber comprises a watertight cylindrical vessel one metre in diameter and one metre in height. A saturated sand sample, described below, was placed inside the pressure chamber to a height of about 560mm above a perforated false floor. The false floor was constructed to provide a void beneath the sample. The water in the void space could be pressurised in order to fluidise the sand and thus loosen it. The water level inside the pressure chamber was about 100mm above the top of the sand sample. The Instron actuator was fixed to the lid of the pressure chamber to provide vertical load to the model caisson. The model caisson and a waterproof load cell were suspended beneath the lid of the pressure vessel and attached to the actuator via a stainless steel rod, which penetrated through the lid of the pressure chamber via a gland. Sensors near the top and the base of the cylinder monitored pressure within the chamber. A pore pressure sensor was also placed within the base plate of the model caisson. The waterproof load cell measured the axial and moment loads applied to the caisson. The load applied by the actuator was also measured by a 100kN capacity load cell fixed to the ram of the actuator. The displacement of the caisson during loading was measured by an LVDT attached to the rod fixing the actuator to the caisson. During installation, the air within the caisson was vented through a tube from the caisson to the outside of the pressure vessel. Once the caisson had been installed, a slight pressure was applied within the chamber in order to force water through this tube and remove all of the air within it. Signals from the instrumentation were filtered and amplified. The data was recorded by a PC-based system incorporating a Computerboards data acquisition card.

Pressure Chamber Testing of Model Caisson Foundations in Sand 3

Sand Properties and Sample Preparation The sand used to create the test bed in these tests was Redhill 110. This is a commercially available fine-grained, poorly graded silica sand. The average particle size of the sand was about 0.1mm.

The test bed was prepared by first placing a filter layer of Leighton Buzzard 16-30 sand to a thickness of about 80mm. About 700kg of the Redhill 110 was then placed dry on top of the filter layer. The pressure chamber was then sealed and a vacuum applied within it. Carbon dioxide gas was introduced to the base of the chamber and allowed to flush through the sand for a period of time. De-aired water was then introduced to the base of the evacuated chamber until the test bed was entirely submerged. Carbon dioxide gas was again flushed through the sample then the vacuum was gradually released. At this stage any small amounts of carbon dioxide were expected to have dissolved in the water.

The test bed was made dense through the use of a vibrating motor strapped to the side of the pressure chamber.

Test Results The sand bed used in the caisson test was prepared to a relative density of 80%. The initial penetration of the caisson into the sand was conducted at a rate of 0.2mm per second, then vertical load cycles of increasing amplitude were applied at a rate of 1Hz. The final pullout of the caisson from the sand was conducted at a rate of 5mm per second.

The method for conducting the test was as follows:

• Calibrate the Instron load cell • Position the test caisson above the soil • Penetrate the test caisson into the sample until a load of 10kN was

reached. The feedback control response of the actuator was then optimised by cycling at a rate of 5Hz with a load amplitude of ±1kN

• Continue penetration of the caisson until a load of about 30kN was reached

• Apply 10 cycles, at 1Hz, with amplitudes of ±5kN, ±10kN, ±20kN and ±30kN. The pore pressure beneath the base plate of the caisson was allowed to dissipate between each set of cycles.

• Apply 5 cycles, at 1 Hz, with amplitudes of ±35kN and ±40kN. The pore pressure was again allowed to dissipate between tests.

• Pull the caisson out of the sand. The Instron load cell was used to control the applied loads, whereas the

waterproof load cell inside the chamber measured the load data reported below.

Pressure Chamber Testing of Model Caisson Foundations in Sand 4

Caisson Installation Load-penetration data during the installation of the caisson are presented in Figure 2. The load increased non-linearly with depth while the skirts were being penetrated. The shear stress mobilised on the skirts of the caisson can be estimated as σ’v.Ktanδ. This can be integrated over the depth of the skirt, h, to produce an expression for the total friction on the skirt. The bearing on the annulus of the caisson skirt, of width t, is given by a standard bearing capacity expression where Nq and Nγ are bearing capacity factors for a strip footing. The expression for the penetration resistance is therefore given by:

( ) Dtt

NhNDKh

V q πγ

γπδγ

γ

++

=

2'

'tan2'

22

The estimated penetration resistance from this expression is shown on Figure 2 superimposed over the data from the test. The resistance was computed taking Ktanδ = 1.15 and the friction angle to be 45°. During the first 70mm of penetration, the expression above closely predicted the penetration resistance. The resistance was under-estimated on further penetration. Once the base of the caisson came into contact with the sand the load increased rapidly. The stiffness of the foundation increased from about 10N/mm during skirt penetration to about 1,700N/mm on contact of the base plate with the sand.

Cyclic Loading of the Model Caisson Load-displacement data measured during the cyclic phase of the model caisson test are shown in Figure 3. The mean vertical load was about 35kN. The data show that as the cyclic load amplitude increased so did the displacement of the caisson into the sand. The data also showed that the caisson was unable to sustain significant tensile loads during the ±40kN cyclic load set. The maximum tensile force recorded during the ±40kN cyclic load set was about –1kN.

The load, pressure beneath the base plate of the caisson, displacement of the caisson and velocity of the caisson during the 4th cycle of the ±20kN load set are shown in Figure 4. The data show that the load and the displacement of the caisson are related to each other, as are the pressure and velocity of the caisson. The load/displacement data are about a quarter of a cycle out-of-phase with the pressure/velocity.

The loads during the 4th cycle in each set are plotted against re-zeroed displacement in Figure 5. As per Figure 3, the loads increased with penetration into the soil. It can also be seen that the stiffness of the foundation reduced rapidly towards zero as the load moved from compression to tension and remained low upon reloading the caisson for a significant displacement prior to a rapid increase in stiffness as the compressive load increased.

Although the positive and negative excess pore pressures immediately below the base plate of the model caisson increased in magnitude as the cyclic load amplitude increased, the excess pressures did not increase as the load increased. The excess pore pressure increased with load amplitude because the Instron

Pressure Chamber Testing of Model Caisson Foundations in Sand 5

actuator had to displace the caisson at an increasing rate to achieve the target load amplitude within the load period of 1 second. Excess pore pressures are plotted against the velocity of the caisson during the 4th load cycle in each set in Figure 6. While the loads were compressive, during cyclic amplitudes of ±5kN to ±30kN, the excess pore pressures were linearly related to the velocity of the caisson, at least to a first approximation. After the loads became tensile, the excess pressures were unable to be directly related to the velocity of the caisson.

At no time during the test did the excess water pressures beneath the base plate of the caisson approach the cavitation pressure of –100kPa.

Pullout of the Model Caisson The maximum tensile load on the model caisson during its pullout is shown in Figure 3 and was –2.1kN. The corresponding suction pressure beneath the base plate of the model caisson was –17kPa and is shown in Figure 4. A soil plug was observed to remain inside the skirts of the caisson after the caisson had been fully extracted from the sand. The computed load equivalent to the suction pressure, measured under the base plate, was –1.1kN and the saturated weight of the soil plug was estimated to be 0.2kN. It was inferred that the remaining tensile load was generated on the outer surface of the skirt of the caisson.

As the maximum suction pressure during pullout of the caisson was less than that observed during cyclic loading it was inferred that the pullout rate of 5mm per second was insufficient to generate an undrained response in the sand. The rate of loading required to produce cavitation beneath the base plate of the caisson can be estimated from extrapolation of a trend line fitted through the ‘linear’ pressure-velocity data in Figure 6. It was estimated that the caisson would have to be displaced at a rate of 55mm/s for cavitation beneath the base plate to occur. Discussion It has been argued that structures built on shallow foundations resting on the seabed can resist uplift forces greater than their own self weight (Tjelta, 1994). This idea led to caissons replacing piles as foundations for the Europipe 16/11-E jacket. In contrast to jacket structures, offshore wind turbines are extremely lightweight, must resist large overturning loads and are founded in relatively shallow water. If a multi-caisson foundation were to be used to support the wind turbine it would be a great advantage if the tensile capacity of the upwind leg(s) were significantly greater than the drained frictional capacity of the foundation i.e. that a suction could be relied upon for capacity calculations. Two important questions are therefore:

1. What is the ultimate tensile capacity of a caisson foundation in sand? 2. At what displacement is the ultimate tensile capacity mobilized?

Pressure Chamber Testing of Model Caisson Foundations in Sand 6

There is a contrast between the data in Byrne (2000) and the results presented here regarding the ultimate tensile capacity of a caisson foundation. Byrne’s data, obtained from tests conducted in an oil-saturated fine silica sand, suggest that the ultimate tensile capacity is quite large and is governed by cavitation of the pore fluid. The results presented above, for a test conducted in water-saturated silica sand, suggest the ultimate tensile capacity is low and that cavitation of the pore fluid does not occur. The contrasting results occur because the suction pressure beneath the caisson depends on a complex interaction between the permeability of the soil, the length of the drainage path and the rate of loading. A faster rate of loading than that in the test presented here is likely to produce results more reminiscent of those observed by Byrne (2000). The effect of loading rate on stiffness and tensile capacity can be found in Bye et al. (1995), and in particular presented in their Figures 2.6 and 2.7. Loading rates applied to a 550mm diameter caisson with 210mm skirts were varied from 0.1 mm/s to 50 mm/s and the tensile load response varied accordingly. In addition, the small-scale laboratory test data need to be appropriately scaled in order to assess the ultimate tensile capacity of a field scale caisson. Work in this area is ongoing and plans are being made to conduct field trials using a 1.5m diameter caisson.

The data from Byrne (2000) and the test data reported in this paper suggest that the vertical stiffness of a caisson foundation reduces significantly as the loading changes from compression into tension, and that displacements in the order of 10-20% of the caisson diameter are required to mobilize the ultimate tensile capacity. The magnitude of these deformations would be unacceptable for the operation of a wind turbine. The low stiffness as the loads become tensile may limit the serviceability design of an upwind leg of a multi-caisson foundation to zero tensile load. This condition may be required in order to prevent the caisson from ratcheting into or out of the ground, depending on the mean vertical load applied to each caisson. The laboratory-scale data suggests that the maximum tensile load possible before a caisson undergoes significant upward displacement is limited to the self-weight of the caisson, the weight of the soil plug and the external skin friction acting on the caisson’s skirts.

The effect of increasing the ambient pressure in the chamber on the load capacity of the caisson foundation may depend on whether the minimum excess pore pressure of –27.7kPa measured in this test can be exceeded or not. The pore pressure of –27.7kPa is sufficiently remote from the cavitation pressure of –100kPa (relative to atmospheric pressure) to suggest that cavitation did not occur in any part of the sand. It would be expected that increasing the ambient water pressure would only affect the capacity of the caisson foundation if cavitation occurred in the sand during tests at atmospheric pressure. Otherwise, the rules of effective stress will govern the foundation response. However, the suction pressure beneath the base of the caisson may be dependent on the rate of loading and may approach the cavitation limit more closely as the rate of

Pressure Chamber Testing of Model Caisson Foundations in Sand 7

loading is increased. Further tests are planned to investigate the effect of loading rate on the pressure response. Acknowledgments The authors are grateful to the DTI and EPSRC for the funding of this research. The authors would also like to acknowledge the industrial participants to this research project: SLP Engineering Ltd, Shell Renewables Ltd, Enron Wind Overseas Development Ltd, Fugro Ltd, Aerolaminates Ltd, HR Wallingford and Garrad Hassan. The second author is also grateful for the support provided by Magdalen College, Oxford. References 1. Bye, A., Erbrich, C. and Rognlien, B. (1995), “Geotechnical Design of

Bucket Foundations”, OTC Offshore Technology Conference, Paper OTC7793

2. Byrne, B.W. (2000) “Investigations of suction caissons in dense sand”, DPhil thesis, University of Oxford

3. Byrne, B.W. and Houlsby, G.T. (2002) “Experimental investigations of the response of suction caissons to transient vertical loading.” Proc. ASCE, Journal of Geotechnical Engineering, Vol. 128, No. 11, Nov., pp 926-939.

4. Johnson, K. 1999. “Partially drained loading of shallow foundations”. Fourth Year Project. Department of Engineering Science, The University of Oxford.

5. Tjelta, T.J. (1994), “Geotechnical aspects of bucket foundations replacing piles for the Europipe 16/11-E jacket”, OTC Offshore Technology Conference, Paper OTC7379

6. Byrne, B.W., Villalobos, F., Houlsby, G.T. and Martin, C.M. (2003), “Laboratory Testing of Shallow Skirted Foundations in Sand”, ICOF’03

Pressure Chamber Testing of Model Caisson Foundations in Sand 8

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9) (10)

PS

PS

PS

(1) Pressure Chamber(2) Void Beneath Sample(3) Saturated Sand Sample(4) Water Level(5) Instron Actuator(6) Instron Support Structure(7) Test Caisson(8) Waterproof Load Cell(9) Instron Load Cell(10) LVDT Displacement TransducerPS Pressure Sensor

Figure 1 Pressure chamber apparatus

Pressure Chamber Testing of Model Caisson Foundations in Sand 9

Figure 2 Caisson installation

Figure 3 Cyclic loading of caisson

-10

0

10

20

30

40

50

60

70

80

90

200 205 210 215 220 225 230 235 240 245 250

Displacement (mm)

Load

(kN

)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

0 20 40 60 80 100 120 140 160 180 200

Displacement (mm)

Load

(kN

)

Initial PenetrationInferred Load

Pressure Chamber Testing of Model Caisson Foundations in Sand 10

Figure 4 Force, pressure, displacement and velocity during 4th cycle of

±20kN load set

Figure 5 Displacement during the 4th load cycle of each set

-10

0

10

20

30

40

50

60

70

80

90

-8.0 -6.0 -4.0 -2.0 0.0 2.0 4.0 6.0

Displacement (mm)

Load

(kN

)

+/- 5kN+/- 10kN+/- 20kN+/- 30kN+/- 35kN+/- 40kN

-20

-10

0

10

20

30

40

50

60

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Time (s)

For

ce/P

ress

ure

-6

-4

-2

0

2

4

6

8

10

12

14

16

18

Dis

pla

cem

ent/

Vel

oci

ty

Pressure (kPa)Force (kN)Displacement (mm)Velocity (mm/s)

Pressure Chamber Testing of Model Caisson Foundations in Sand 11

Figure 6 Excess pressure vs rate of caisson displacement during the 4th cycle of each set

-40

-20

0

20

40

60

80

100

-60 -40 -20 0 20 40 60 80 100 120

Velocity (mm/s)

Pre

ssur

e (k

Pa)

+/- 5kN

+/- 10kN+/- 20kN+/- 30kN

+/- 35kN+/- 40kN

Paper No. 2004-mfb-03 Kelly Page: 1 of 4

Tensile Loading of Model Caisson Foundations for Structures on Sand

R.B. Kelly, B.W. Byrne, G.T. Houlsby and C.M. Martin

Department of Engineering-Science, Oxford University Oxford, United Kingdom

ABSTRACT The viability of multiple footing structures that use suction caisson foundations would be improved if the up-wind leg(s) could resist significant tensile loads. In the particular case of offshore wind turbines, large moments are applied at foundation level by the action of wind and waves on the structure. For a multiple footing structure the applied moment is resisted primarily by vertical reactions on opposing foundations. If a significant tension can be allowed at the upwind side, then the spacing of the foundations, and therefore the overall size of the structure, can be greatly reduced. Bye et al (1995) suggest that the tensile capacity of caissons is dependent on the rate of loading, compared to the rate of drainage of excess pore pressure beneath the caisson. To investigate this, model caisson tests were carried out in test beds made of two different gradings of sand. Whilst the ultimate capacity is large in both cases (and controlled by the cavitation of the pore fluid) the displacements required to mobilise the loads are large compared to the diameter of the footing. These displacements are of a magnitude that would cause serviceability problems. The experimental results suggest that the upwind footing should be designed for tensile loading no greater than (at most) the drained friction on the skirt. INTRODUCTION Within the next few years a number of wind farms will be constructed around the coast of the UK. In the first instance many of the wind turbine structures will be founded on piles. These foundations, although simple to design, contribute a significant proportion (about 30%) of the overall installed cost for these structures. Various options are being investigated to reduce the costs, and therefore increase the economic viability of offshore wind-farm developments. One possibility is use of skirted shallow foundations, installed by suction either as a single foundation (i.e. monopod structure) or as a multiple foundation system (i.e. quadruped/tripod structure) (see Houlsby and Byrne, 2000; Byrne and Houlsby, 2002b; Byrne et al, 2002; Byrne and Houlsby, 2003). For the monopod, the key issue is the performance of the foundation under the large moments applied by the wind and wave loads on the structure (for a discussion of this problem see Byrne et al, 2003). For the multiple footing case the applied moment loads from the wind and

waves will largely be reacted as vertical compression and tension loads on the individual foundations. Both these loading cases are being investigated at the University of Oxford in laboratory scale and field scale tests funded by the Department of Trade and Industry (DTI), the Engineering and Physical Sciences Research Council (EPSRC) and industrial participants (SLP Engineering Ltd, Shell Renewables Ltd, General Electric Wind Ltd, Fugro Ltd, Aerolaminates Ltd, HR Wallingford and Garrad Hassan) (see Byrne et al, 2002; Byrne et al, 2003; Kelly et al, 2003). This paper concentrates on the tensile vertical capacity of a caisson foundation embedded in a saturated sand, relevant to the design of multiple footing foundations. Previous research on the transient tensile capacity of skirted foundations in saturated sand has been conducted by Bye et al (1995), Johnson (1999), Byrne (2000), Byrne and Houlsby (2002a) and Kelly et al (2003). These studies have shown that significant tensile capacities are possible, under the appropriate loading conditions, and are limited by cavitation of the pore fluid. The displacement of the caissons required to generate the maximum tensile capacity was in the order of 5-10% of their diameter Byrne and Houlsby (2002a), and the stiffness of the response was observed to reduce when the load applied to the foundation changes from compression to tension. Clearly the economic viability of multiple caisson foundations would improve if the upwind leg(s) of the structure can resist significant tensile loads at moderate displacements. The experimental evidence, as reported above, suggests that large upward movements of the caissons are required to develop significant tensile loads. These movements are likely to be of an order sufficient to prevent a wind turbine from operating. For a caisson foundation on sand the transient loading condition will be critical to understanding the load displacement response. The transient response of a foundation on sand is dependent on the rate of loading, compared with the rate of drainage of the excess pore pressures beneath the caisson. In particular it is necessary to understand which variables control aspects of the load-displacement behaviour, such as the stiffness changes at the tension/compression transition. Model caisson tests have been conducted on test beds made of very fine sand, and are compared with tests conducted using a coarser material. The loading rates have been chosen so that the response spans a range of partially drained conditions.

Paper No. 2004-mfb-03 Kelly Page: 2 of 4

TEST APPARATUS A model caisson of diameter 280mm, skirt depth of 180mm and wall thickness of 3.125mm was used in these experiments. The test chamber is shown in Figure 1, and comprises a watertight cylindrical vessel 1m in diameter and 1m in height. A saturated sand sample was constructed inside the test chamber. An Instron actuator was fixed to the lid of the test chamber to provide vertical load to the model caisson. The model caisson and a waterproof load cell were suspended beneath the lid of the test vessel and attached to the actuator via a stainless steel rod, which penetrated through the lid of the chamber via a watertight gland. The load applied by the actuator, outside the test chamber, was also measured by a 100kN capacity load cell fixed to the ram of the actuator. Pressures beneath the lid of the caisson and at the tip of the skirt were measured. Sensors near the top and the base of the cylinder monitored pressure within the chamber. The displacement of the caisson during loading was measured by an LVDT attached to the rod fixing the actuator to the caisson. During installation, the air within the caisson was vented through a tube from the caisson to the outside of the vessel. Once the caisson had been installed, a slight pressure was applied within the chamber in order to force water through this tube and remove all of the air within it. Signals from the instrumentation were filtered and amplified. The data was recorded by a PC-based system.

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9) (10)

PSPS

PS

(1) Pressure Chamber(2) Void Beneath Sample(3) Saturated Sand Sample(4) Water Level(5) Instron Actuator(6) Instron Support Structure(7) Test Caisson(8) Waterproof Load Cell(9) Instron Load Cell(10) LVDT Displacement TransducerPS Pressure Sensor

PS

Fig. 1 Test Chamber SAMPLE PROPERTIES AND PREPARATION The sands used were Oakamoor HPF5 and Redhill 110. HPF5 is a crushed silica sand with d50 of 50µm whereas Redhill 110 is a sieved silica sand (or sand silt) with a d50 of about 120µm. For consolidation and drainage dominated problems, as in this paper, the d10 size is more critical. In the case of Redhill 110 the value for this parameter is about 75µm whilst the HPF5 is an order of magnitude finer at 7µm. A particle size diagram showing the grading of both sands is presented in Figure 2. The Redhill 110 sample was prepared by first placing a filter layer of Leighton Buzzard 16-30 sand to a thickness of about 80mm. About 700kg of the Redhill 110 was placed dry on top of the filter layer. The pressure chamber was sealed and a vacuum applied within it. Carbon dioxide gas was introduced to the base of the chamber and allowed to

flush through the sand for a period of time. De-aired water was then introduced to the base of the evacuated chamber until the test bed was entirely submerged. Carbon dioxide gas was again flushed through the sample, and the vacuum was gradually released. At this stage any small amount of carbon dioxide was expected to have dissolved in the water. The HPF5 sample was constructed differently, to minimise hazard from fine silica particles if a dry preparation method was used. The Leighton Buzzard filter layer was again used together with a thin layer of Redhill 110. The chamber was then filled with water and the HPF5 sand pluviated through water into the chamber. The sand was introduced in stages and vibration applied to the walls of the test chamber to remove air trapped within the sand. Both sands were densified prior to the tests by use of a vibrating motor strapped to the side of the test chamber. The Redhill 110 sample was prepared to a relative density of 80% while the HPF5 test bed had a relative density of about 75%. TEST PROCEDURE The initial penetration of the caisson into the Redhill 110 was conducted at a rate of 0.2mm/s whilst in the HPF5 the rate was reduced to 0.05mm/s to prevent piping failures from occurring. This procedure was applied until the vertical load on the caisson was 10kN. Subsequently, step loads of 5kN were applied to investigate the rate of pore pressure dissipation beneath the caisson. The step loads were applied until the Instron load cell recorded 35kN. Load cycles were then applied with amplitudes of ±5kN, ±10kN, ±20kN, ±30kN, ±35kN and ±40kN. The cycles were applied at rate of 1Hz in the test reported here using the Redhill 110 sand and at rates of 0.1Hz and 1Hz in the tests using the HPF5 sand. (Other tests where the caissons were pushed into the sand until a load of 35kN was reached, and then rapidly pulled out of the sand at a rate of 100mm/s are not reported here). The average times for the dissipation of half of the initial excess pore pressures beneath the lid of the caisson during the step changes was 0.04s in the finer Redhill 110 sand and 18.6s in the coarser HPF5 sand. The permeabilities of the samples were estimated from these data using the equation 2HtcT vv = with 196.0=vT as for 50% drainage in one-dimensional consolidation, and H taken as the depth of the caisson skirt. The permeability is then obtained from vwv mkc γ= , with vm being estimated from the deflection during the load increment. The permeability of the HPF5 sand was estimated as 4x10-7 m/s and that of Redhill 110 sand 1.5x10-4 m/s.

0

10

20

30

40

50

60

70

80

90

100

0.001 0.01 0.1 1

Particle Size (mm)

Per

cent

Pas

sing

Oakamoor HPF5

Redhill 110

Fig. 2 Particle size distribution

Paper No. 2004-mfb-03 Kelly Page: 3 of 4

RESULTS Data from three cyclic tests are shown in Figures 3, 4 and 5. All of the test data show that as the cyclic load amplitude increased so did the displacement of the caisson into the sand. In the coarser Redhill 110 sand during the attempted ±40kN cyclic load set the caisson was able to sustain a tensile load of only about 1kN, whereas in the finer HPF5 sand, loaded at a rate of 0.1Hz, a tensile load of about 7kN could be sustained. The stiffness of the load-displacement response in tension was, however, much lower than that in compression in both tests. These results are consistent with tests conducted by Byrne (2000), and reported by Byrne and Houlsby (2002a), in extremely low permeability oil saturated fine silica sand. The tests were carried out using a footing of 150mm diameter and 50mm depth. An example of a cyclic loading test is shown in Figure 5. Note that the magnitude of the loads is significantly lower with a mean vertical load of 200N. A feature of all the cyclic tests is the accumulation of vertical (downward) displacement during the cycles, as well as the hysteresis in each cycle.

-10

0

10

20

30

40

50

60

70

80

200 210 220 230 240 250 260 270

Vertical Displacement (mm)

Axia

l Loa

d (k

N)

Fig. 3 Load-displacement data: Redhill 110 loaded at a rate of 1Hz

-20

-10

0

10

20

30

40

50

60

70

80

200 210 220 230 240 250 260 270

Vertical Displacement (mm)

Axia

l Loa

d (k

N)

Fig. 4 Load-displacement data: HPF5 loaded at a rate of 0.1Hz Another method of exploring the limits of the caissons behaviour, other than reducing the permeability of the soil, is to increase the rate of loading. The test conducted in HPF5 sand was repeated with the rate of loading increased to 1Hz. Data from this test is shown in Figure 6. The response of the caisson was not as stable as the other two tests. The mean compressive load appeared to change throughout the test and tensile load was apparently not mobilized although the stiffness of the load-displacement response during the ±40kN cyclic load set reduced significantly as the load approached zero. Extremely high pore pressures beneath the lid of the caisson were recorded during this test. The maximum pore pressure was 372kPa during the compressive half-

cycle of the ±40kN load set. This pressure is equivalent to about 23kN, which is about one-third of the total applied load.

-400

-200

0

200

400

600

800

-1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2Vertical Displacement, w (mm)

Axi

al L

oad,

V (N

)

Fig. 5 Load-displacement plot from Byrne (2000).

-20

0

20

40

60

80

200 210 220 230 240 250 260 270

Vertical Displacement (mm)

Axia

l Loa

d (k

N)

Fig. 6 Load-displacement data: HPF5 loaded at 1Hz The cyclic tests, followed by pullout, were conducted to assess whether disturbance to the sand during cycling affected its tensile behaviour. A comparison of the results from the two load-unload tests is shown in Figure 7. Tensile loads of about 10kN were achieved in each test. The pore pressure recorded beneath the lid of the caissons in each test, at the maximum tensile load, was about –100kPa. This is the pressure at which cavitation of the pore fluid occurs in water. The tensile load due to the cavitation pressure would correspond to approximately 6.2kN. The difference between this value and the 10kN measured represents an enhanced friction due to the very large hydraulic gradient at the outside of the caisson skirt. The displacement required to mobilise the maximum tensile load was about 10mm in the Redhill 110 sand and about 20mm in the HPF5 sand. These displacements correspond to about 3.5% and 7% of the caisson diameter respectively. It is significant that the stiffness of the load-displacement response remained much less in tension than compression, even when the caissons were pulled out rapidly from low permeability sand, and prior to any disturbance caused by cyclic loading. DISCUSSION The transient tensile capacity of a skirted foundation depends on an interaction between the permeability of the soil, the length of the drainage path and the rate of loading. The non-dimensional parameter that incorporates these variables is Tv, the time parameter used in one-dimensional consolidation analysis:

Paper No. 2004-mfb-03 Kelly Page: 4 of 4

2HtcT vv = ………………………………………………… (1)

where cv is the coefficient of consolidation, t is the time taken for a certain degree of pore pressure dissipation and H is the length of the drainage path. This equation can be used to assess whether the loading rates applied in these tests are applicable to a prototype situation. The prototype caisson is assumed to have a diameter of 6m and a skirt length of 5m. It is assumed that the time t in the above equation represents the rise-time during a loading cycle (i.e. a quarter cycle), and that the length of the drainage path is equivalent to the length of the caisson skirt. A comparison of calculated Tv values is presented in Table 1. The values of cv in Table 1 were estimated from the step load data in the model caisson tests. It has been assumed that cv for the prototype sand is that as deduced for the sand at the Draupner platform site by Bye et al (1995), where caisson foundations have been used. Table 1 Comparison of Tv values Sand cv (m2/s) t (s) H (m) Tv Redhill 110 0.19 0.25 0.18 1.47 HPF5 at 0.1Hz 0.00039 2.5 0.18 0.03 HPF5 at 1Hz 0.00039 0.25 0.18 0.003 Prototype 0.5 2.5 5.0 0.05 The Tv value in Table 1 for the prototype caisson indicates that it might behave in a similar manner to the model caisson test in HPF5 sand, loaded at a rate of 0.1Hz. Note that in this test the caisson was able to sustain a modest tension during cyclic loading. Note, however, as shown in Figure 4, that the stiffness in tension is significantly lower than in compression. Also note that the footing penetrates into the ground substantially after each tensile excursion. If the loads do not pass into tension the displacement response is less severe. If it is assumed that the tensile load mobilised is proportional to the cube of the dimension of the caisson (as would be usual for drained capacity problems in sand), then the maximum mobilised tensile load would be about 70MN. If, however, the capacity is proportional only to the square of the dimension (as would be true if cavitation were the dominant factor) then the capacity would only be about 3MN for a 6m caisson. However, the low stiffness as the loads become tensile may impose a serviceability design limit, that the load on an upwind leg of a multiple caisson foundation must be limited to zero tensile load, or at most the drained friction value on the skirt. This condition may be required in order to prevent the caisson from ratchetting into or out of the ground, depending on the mean vertical load applied to each caisson. The

laboratory-scale data suggests that the maximum tensile load possible before a caisson undergoes significant upward displacement is limited to the self-weight of the caisson, the weight of the soil plug and the external skin friction acting on the caisson’s skirts. Note that Figures 3, 4 and 6 indicate that, as long as tensile loading is avoided, there is little difference in the vertical load-displacement response across the wide range of Tv values shown in Table 1. CONCLUSIONS Cyclic loading tests of model caissons in dense sand are presented. These data are relevant to the design of multiple caisson foundation systems for offshore wind turbines, in which the design is likely to be limited by the tensile capacity of the foundations. Loading rates were carefully selected so that, when scaled to prototype scale, they represent realistic degrees of partial drainage under wave loading conditions. The results indicate that only small tensions can be permitted on the foundations if excessive vertical movement is to be avoided. ACKNOWLEDGEMENTS The authors are grateful to the DTI and EPSRC for the funding of this research. The authors would also like to acknowledge the industrial participants to this research project: SLP Engineering Ltd, Shell Renewables Ltd, General Electric Wind Ltd, Fugro Ltd, Aerolaminates Ltd, HR Wallingford and Garrad Hassan. The second author is also grateful for the support provided by Magdalen College, Oxford. REFERENCES Bye, A., Erbrich, C. and Rognlien, B. (1995), “Geotechnical Design of

Bucket Foundations”, OTC Offshore Technology Conference, Houston, Paper OTC7793.

Byrne, B.W. (2000) “Investigations of suction caissons in dense sand”, DPhil thesis, Oxford University.

Byrne, B.W. and Houlsby, G.T. (2002a) “Experimental investigations of the response of suction caissons to transient vertical loading.” Proc. ASCE, Jour. of Geotech. Eng. 128, No. 11, Nov., pp 926-939.

Byrne, B.W. and Houlsby, G.T. (2002b) “Investigating novel foundations for offshore wind turbines” Proc. 21st Int. Conf. on Offshore Mechanics and Arctic Engineering OMAE’02, Oslo, Paper OMAE2002-28423.

Byrne, B.W. and Houlsby, G.T. (2003) “Foundations for offshore wind turbines.” Phil. Trans. Roy. Soc. of London, Series A, 231, pp 2909-2300.

Byrne, B.W., Houlsby, G.T., Martin, C.M. and Fish, P.M. (2002) “Suction caisson foundations for offshore wind turbines”, Wind Engineering, 26, No 3.

Byrne, B.W., Villalobos, F., Houlsby, G.T. and Martin, C.M. (2003) “Laboratory testing of shallow skirted foundations in sand”, Proc. BGA Int. Conf. on Foundations, Dundee, Sept., pp 161-173

Houlsby, G.T. and Byrne, B.W. (2000) “Suction caisson foundations for offshore wind turbines and anemometer masts”, Wind Engineering 24, No 4, pp 249-255.

Johnson, K. 1999. “Partially drained loading of shallow foundations”. 4th Year Project. Department of Engineering Science, Oxford University.

Kelly, R.B., Byrne, B.W., Houlsby, G.T. and Martin, C.M. (2003), “Pressure chamber testing of model caisson foundations in sand”, Proc. BGA Int. Conf. on Foundations, Dundee, Sept., pp 421-432.

-20

-10

0

10

20

30

40

0 50 100 150 200

Vertical Displacement (mm)

Axia

l Loa

d (k

N)

Redhill 110HPF5

Fig. 7 Pullout of caissons from Redhill 110 and HPF5 sand

1. INTRODUCTION Suction caissons are an option for the foundations for offshore structures. Under large environmental loads the upwind foundations of a multiple-caisson foundation might be subjected to tensile loads. Recent research indicates that serviceability requirements will often dictate that, under working and frequently encountered storm loads, tensile loads on caissons should be avoided, as they are accompanied by large displacements. However, it may be appropriate to design structures so that under certain extreme conditions the caissons are allowed to undergo tension. It is therefore necessary to have a means of estimating the tensile capacity of a caisson foundation, whilst recognising that large displacements may be necessary to mobilise this capacity. The calculations are also relevant to the holding capacity of caisson anchors subjected to pure vertical load, and to calculation of forces necessary to extract a caisson rapidly (for whatever reason).

Under rapid tensile loading, a suction caisson in sand will exhibit a limiting load which will typically consist of a suction developed within the caisson, and friction on the outer wall. However, a number of different possible modes of failure exist. The purpose of this paper note is to set out simple calculations for capacities under various failure modes, and to compare these with experimental results.

2. TENSILE CAPACITY CALCULATIONS

2.1 Drained capacity If the tensile load is applied very slowly, then pore pressures will be small, and a fully drained calculation is applicable for calculating the capacity. For the purposes of calculation an idealised case of a foundation on a homogeneous deposit of sand is considered here.

The resistance on the caisson is calculated as the sum of friction on the outside and the inside of the skirt. The effective stresses on the annular rim are likely to be sufficiently small that they can be neglected, and it is assumed that the soil breaks contact with the lid of the caisson. The frictional terms are calculated in the same way as for the installation calculation (Houlsby and Byrne, 2005), by calculating the vertical effective stress adjacent to the caisson, then assuming that horizontal effective stress is a factor K times the vertical effective stress. Assuming that the mobilised angle of friction between the caisson wall and the soil is δ then we obtain the result that the shear stress acting on the caisson is δσ′ tanKv . Note that in the subsequent analysis the values of K and δ never appear separately, but only in the combination δtanK , so it is not possible to separate out the effects of these two variables. Allowance is made, however, for the possibility of different values of δtanK acting on the outside and inside of the caisson. A difference

The tensile capacity of suction caissons in sand under rapid loading

Guy T. Houlsby, Richard B. Kelly & Byron W. Byrne Department of Engineering Science, Oxford University

ABSTRACT: We develop here a simplified theory for predicting the capacity of a suction caisson in sand, when it is subjected to rapid tensile loading. The capacity is found to be determined principally by the rate of pullout (relative to the permeability of the sand), and by the ambient pore pressure (which determines whether or not the water cavitates beneath the caisson). The calculation procedure depends on first predicting the suction beneath the caisson lid, and then further calculating the tensile load. The method is based on similar principles to a previously published method for suction-assisted caisson installation (Houlsby and Byrne, 2005). In the analysis a number of different cases are identified, and successful comparisons with experimental data are achieved for cases in which the pore water either does or does not cavitate.

between this analysis and conventional pile design is that the contribution of friction in reducing the vertical stress further down the caisson is taken into account.

If, as a preliminary, no account is taken of the reduction of vertical stress close to the caisson due to the frictional forces further up the caisson, then the tensile vertical load on the caisson for penetration to depth h is given by:

( ) ( ) ( ) ( )iioo DKhDKhV πδγ′−πδγ′

−=′ tan2

tan2

22 (1)

friction on outside friction on inside

where the dimensions are as in Figure 1, and γ′ is the effective weight of the soil. V ′ is the buoyant weight of the caisson and structure.

A check should always be made that the friction calculated inside the caisson does not exceed the weight of the trapped soil plug 42

iDhπγ′ . Ignoring the reduction of the stress in this case

proves unconservative (i.e. it overestimates the force that can be developed), so we develop here a theory which takes this effect into account. Consider first the soil within the caisson. Assuming that the vertical effective stress is constant across the section of the caisson, the vertical equilibrium equation for a disc of soil within the caisson (Figure 2) leads to:

( ) ( ) ( )i

iv

i

iivvDK

D

DKdz

d δσ′−γ′=

π

πδσ′−γ′=

σ′ tan4

4

tan2

(2)

Writing ( )( ) iii ZKD =δtan4 , Eq. (2) becomes

γ′=σ′

+σ′

i

vvZdz

d, which has the solution

( )( )iiv ZzZ −−γ′=σ′ exp1 for 0=σ′v at 0=z . The total frictional terms depend on the integral of the vertical effective stress with depth, and we can also

obtain ( ) ( )( )iii

h

v ZhZhZdz +−−γ′=σ′∫ 1exp2

0. For

small iZh the integral simplifies to 22hγ′ as in Eq. (1). For brevity in the following we shall write the function ( ) ( )( )xxxy +−−= 1exp , so that in the

above ( )ii

h

v ZhyZdz 2

0γ′=σ′∫ .

A similar analysis follows for the stress on the outside of the caisson. We assume that (a) there is a zone between diameters oD and om mDD = in which the vertical stress is reduced through the action of the upward friction from the caisson, (b) within this zone the vertical stress does not vary with radial coordinate and (c) there is no shear stress on vertical planes at diameter mD . We then obtain the same results as for the inside of the caisson, but with

iZ replaced by ( ) ( )( )ooo KmDZ δ−= tan412 . Alternative assumptions could be made for the

variation of mD with depth, but at present there is little evidence to justify any more sophisticated approach. If mD is taken as a variable, then the differential equation for vertical stress will usually need to be integrated numerically.

Accounting for the effects of stress enhancement, Eq. (1) becomes modified to:

( ) ( )

( ) ( )ii

h

vi

oo

h

vo

DKdz

DKdzV

πδσ′

+πδσ′=′

∫

∫

tan

tan

0

0

(3)

In the special case where m is taken as a constant and uniform stress is assumed within the caisson this becomes:

( ) ( )

( ) ( )iii

i

ooo

o

DKZhyZ

DKZhyZV

πδ

γ′−

πδ

γ′−=′

tan

tan

2

2

(4)

σ'v(Ktanδ)iπDidz

σ'vπDi2/4

(σ'v + dσ'v)πDi2/4

γ'(πDi2/4)dz

Figure 2: Vertical equilibrium of a slice of soil within the caisson

z

Do

Di

h

V' Mudline

Figure 1: Caisson geometry

The calculation accounting for stress reduction obviates the need to check that the internal friction does not exceed the soil plug weight, as the capacity asymptotically approaches that value at large h.

2.2 Tensile capacity in the presence of suction If the caisson is extracted more rapidly, then transient excess pore pressures will occur, and the suction within the caisson will need to be taken into account. We return later to the calculation of the relationship between the rate of movement and the suction, but first address the calculation of load in terms of the suction. If the pressure in the caisson is s with respect to the ambient seabed water pressure, i.e. the absolute pressure in the caisson is

shp wwa −γ+ (where ap is atmospheric pressure,

wγ is the unit weight of water and wh the water depth), then we at first assume that the excess pore pressure at the tip of the caisson is as , i.e. the absolute pressure is ( ) ashhp wwa −+γ+ . There is therefore an average downward hydraulic gradient of

has wγ on the outside of the caisson and upward hydraulic gradient of ( ) hsa wγ−1 on the inside.

We assume that the distribution of pore pressure on the inside and outside of the caisson is linear with depth. A detailed flow net analysis shows that this approximation is reasonable. The solutions for the vertical stresses inside and outside the caisson are exactly as before, except that γ′ is replaced by

has+γ′ outside the caisson and by ( ) hsa−−γ′ 1 inside the caisson. The capacity, accounting for the pressure differential across the top of the caisson and pore pressure on the rim (only relevant for a thick caisson), is again calculated as the sum of the external and internal frictional terms:

( )

( ) ( ) ( ) ( )ii

h

vioo

h

vo

ioi

DKdzDKdz

DDas

DsV

πδσ′+πδσ′

=

−π+

π+′

∫∫ tantan

44

00

222

(5)

In the special case of m constant and a uniform stress assumed within the caisson, this gives:

( )

( ) ( )

( ) ( ) ( )iii

i

ooo

o

ioi

DKZhyZ

hsa

DKZhyZ

has

DDas

DsV

πδ

−−γ′−

πδ

+γ′−

=

−π+

π+′

tan1

tan

44

2

2

222

(6)

We can often make a further simplifying assumption, that the suction is sufficiently large that the soil within the caisson liquefies and therefore

( ) 01 =−−γ′h

sa . For a large suction this means that

1≈a and almost all of the suction appears at the caisson tip. The above rearranges to give

hs

has =+γ′ , and equation (6) can be simplified to:

( )

( ) ( )ooo

o

ioi

DKZhyZ

hs

DDas

DsV

πδ

−

=

−π+

π+′

tan

44

2

222

(7)

In the case either that the thickness of the caisson is small, or that 1≈a this simplifies to the following (writing the outer diameter as D, and the caisson area AD =π 42 ):

( )( )

( )

δ

+−=

−πδ

−=′

tan41

tan

2

2

KZhy

DhZsA

sADKZhyZ

hsV

(8)

where ( ) ( )δ−= tan412 KmDZ . Neglecting the effects of stress reduction would give:

( )

δ

+−=′ tan21 KDhsAV (9)

which means that the capacity is simply calculated by applying a linearly varying factor to the suction force beneath the lid.

2.3 Undrained failure A further condition should be considered: that of “undrained failure” of the sand. In any dilative sand, however, the pore pressures developed under undrained conditions are potentially so large that invariable (except in very deep water) the cavitation mechanism would intervene first. Since the undrained strength of sand is in any case very difficult to determine, we do not pursue this case here.

3. RELATIONSHIP BETWEEN SUCTION AND DISPLACEMENT RATE At low displacement rates, the rate of influx of water q to the caisson can be calculated by Darcy’s law, and equated to the rate of displacement times the

area of the caisson. Flow calculations were presented by Houlsby and Byrne (2005), and yield:

dtdhD

FDsk

q i

w

o4

2π−=

γ= (10)

where F is a dimensionless factor as determined by the procedures in Houlsby and Byrne (2005), which may be fitted approximately by the equation

( )DhF 516.3 += for 8.01.0 ≤≤ Dh . If the displacement rate is increased, the above

condition is interrupted by one of two conditions (a) the suction becomes large enough for liquefaction of the sand within the caisson to occur or (b) cavitation occurs within the caisson.

When liquefaction occurs, the permeability of the liquefied sand increases to a large value, with the result that the a factor in the calculation of the load changes (as noted above) to near unity. The displacement rate may still be estimated from a flow calculation, but the appropriate boundary condition now becomes one of the suction applied at the base rather than top of the caisson. Modified values of F (termed LF for this case) are given in Figure 3, and may be fitted approximately by the equation

( )DhF 5exp9.175.1 −+= for 0.11.0 ≤≤ Dh . When cavitation occurs, either before or after

liquefaction, the displacement rate becomes unlimited and (assuming that cavitation occurs at an absolute pressure afp where f is a constant), the suction will be constant and determined by

awwa fpshp =−γ+ , or ( ) wwa hfps γ+−= 1 . In practice it appears that the factor f is near zero.

4. SUMMARY OF ANALYSIS CASES The following summary presents equations for the above cases, for a thin-walled caisson. To simplify the equations we neglecting here the stress reduction effect, although this should be included in more

accurate calculations:

(a) Small dtdh−

dtdh

kD

Fs

o

wγπ−=4

(from Eq. (10)) and:

( ) ( ) ( )2

tan1tan2DhK

hsaK

has

sAV

ioπ

δ

−−γ′+δ

+γ′

−−=′

(for 0=− dtdh , 0=s and these reduce to the equations for the fully drained case).

(b) Liquefaction without cavitation

Onset of liquefaction occurs at ( )ahs

−γ′

=1

, after that

dtdh

kD

Fs

o

w

L

γπ−=4

and:

( )

δ

+−=′ oKDhsAV tan21

Note that this will imply a sudden jump in s and V ′ at the onset of liquefaction.

(c) Cavitation without liquefaction Onset of cavitation occurs at ( ) wwa hfps γ+−= 1 . After that dtdh− is unbounded, s is constant and:

( ) ( ) ( )2

tan1tan2DhK

hsaK

has

sAV

ioπ

δ

−−γ′+δ

+γ′

−−=′

as in case(a).

(d) Cavitation with liquefaction Since s is constant once cavitation occurs, this condition can only occur when liquefaction occurs before cavitation. Onset of cavitation is at

( ) wwa hfps γ+−= 1 , after which dtdh− is unbounded, s is constant and:

( )

δ

+−=′ oKDhsAV tan21 as in case (b).

Note that the above cases only occur in order (a), (b), (d) or (a), (c). When several possibilities exist for calculating load capacity it is often true that the correct case is simply found by calculating all cases and then taking the lowest value. Note in this analysis that this simple approach cannot be adopted as the onset of some states can preclude other cases occurring, and the calculated load is not necessarily the lowest of the cases.

5. COMPARISONS WITH DATA

We present here a number of pullout tests conducted two sands and at different pullout rates. The tests

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

0.0 0.2 0.4 0.6 0.8 1.0Aspect ratio h/D

Dim

ensi

onle

ss fl

ow fa

ctor

FL

Figure 3: Dimensionless flow factor for liquefaction case

were conducted in a pressure chamber: some tests at an ambient (mudline) water pressure equal to atmospheric, and some at atmospheric plus 200kPa. The model caisson was 280mm diameter, 180mm skirt length. In the following the loads presented include the caisson weight.

The first test reported here (Test 9) was conducted on Redhill Sand, at a pullout rate of 100mm/s and atmospheric pressure. Figure 4 shows the record of suction developed beneath the lid of the caisson against time, and Figure 5 shows the corresponding vertical load. It can be seen that (with a minor initial fluctuation) the suction rapidly approaches 100kPa, at which stage cavitation occurs. At around 3044.5s there is a sudden loss of both suction and vertical load, but this is of little practical

interest since by then the displacements are enormous and about three-quarters of the caisson had been pulled out of the soil.

Figure 6 shows the ratio sAV / , showing that this ratio remains approximately constant during most of the pullout.

It can readily be shown that the suction in this case rapidly increased to sufficient value to cause liquefaction (which would occur at a suction of only about 3kPa), and that the relevant case for analysis here is case (d). The predicted values from the theory described above (including stress reduction) are also shown on each of Figures 4 to 6, and it is clear that the theory (whilst not capturing some of the detail at the beginning of the pullout) predicts the broad trends of the test correctly.

Figures 7 and 8 show corresponding results for Test 10 (at the same pullout rate) but at an ambient pressure of atmospheric plus 200kPa. The suctions developed at this rate of loading are insufficient to cause cavitation, which would occur at -300kPa relative to ambient. It can be seen that again the theory predicts the overall pattern of behaviour well. This time it is case (b) that applies. The fluctuations in predicted suction (and hence load) are due to minor variations in the calculated velocity of extraction.

Figures 9 and 10 show the results from Test 11, which is directly comparable to Test 9, but this time

-120.0

-100.0

-80.0

-60.0

-40.0

-20.0

0.0

20.0

3042.8 3043.3 3043.8 3044.3 3044.8 3045.3

t (s)

pres

sure

(kPa

)

ExperimentTheory

Figure 4: Pressure v. time for Test 9

-12.0

-10.0

-8.0

-6.0

-4.0

-2.0

0.0

2.0

3042.8 3043.3 3043.8 3044.3 3044.8 3045.3

t (s)

V (k

N)

ExperimentTheory

Figure 5: Vertical load v. time for Test 9

0.0

0.5

1.0

1.5

2.0

3042.8 3043.3 3043.8 3044.3 3044.8 3045.3t (s)

V / s

A

ExperimentTheory

Figure 6: V/sA v. time for Test 9

-300

-250

-200

-150

-100

-50

04257 4257.5 4258 4258.5 4259 4259.5

t (s)

pres

sure

(kPa

) ExperimentTheory

Figure 7: Pressure v. time for Test 10

-30.0

-25.0

-20.0

-15.0

-10.0

-5.0

0.0

5.0

4257 4257.5 4258 4258.5 4259 4259.5

t (s)

V (k

N)

ExperimentTheory

Figure 8: Vertical load v. time for Test 10

at a pullout rate of only 5mm/s. Although the suctions are sufficient to cause liquefaction, the pullout rate is such that the suction is sufficiently small so that cavitation does not occur, and the vertical loads are correspondingly lower too. The predicted suction and load are also shown on the Figures. The match to the data could be improved by adjusting the permeability, but the value used in the predictions were deliberately kept the same for all three tests discussed. The permeability value used was 3105.0 −×=k m/s, which is somewhat higher than estimated previously for this sand (Kelly et al. 2004). The other parameters used are 7.0tan =δK and 5.1=m .

Finally, Figures 11 and 12 present equivalent data for a test on HP5 sand, which is much finer that Redhill Sand, and has an estimated permeability of

4102.0 −×=k m/s. The extraction rate was 25mm/s, and in this case, although the extraction rate is lower, the pore pressures are sufficient to cause cavitation even with the ambient pressure of atmospheric plus 200kPa.

The predicted and measured values of maximum tensile load for the three tests on Redhill sand and one on HP5 are shown in Table 1. The order of magnitude of the tensile load is correctly predicted in all cases, even though the actual capacity of the caisson varies greatly in the different tests.

6. CONCLUSIONS

In this paper we develop a simplified theory for predicting the maximum tensile capacity of a caisson foundation in sand. The calculated capacity depends critically on the rate of pullout (in relation to the permeability) and the ambient water pressure (which determines whether cavitation occurs). The theory is used successfully to explain widely differing experimental results for caissons pulled out under different conditions.

REFERENCES Houlsby, G.T. and Byrne, B.W., 2005. Design procedures for

installation of suction caissons in sand, Proc. ICE, Geotechnical Engineering, in press.

Kelly, R.B., Byrne, B.W., Houlsby, G.T. and Martin, C.M., 2004. Tensile loading of model caisson foundations for structures on sand, Proc. ISOPE, Toulon, Vol. 2, 638-641

-20

-15

-10

-5

0

5

2980 2990 3000 3010 3020 3030

t (s)

pres

sure

(kPa

)

ExperimentTheory

Figure 9: Pressure v. time for Test 11

-3.0

-2.5

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

2980 2990 3000 3010 3020 3030

t (s)

V (k

N)

ExperimentTheory

Figure 10: Vertical load v. time for Test 11

-350

-300

-250

-200

-150

-100

-50

0

50

8890 8892 8894 8896 8898 8900 8902

t (s)

pres

sure

(kPa

)

ExperimentTheory

Figure 11: Pressure v. time for Test 23

-40

-30

-20

-10

0

10

8890 8892 8894 8896 8898 8900 8902

t (s)

V (k

N)

ExperimentTheory

Figure 12: Vertical load v. time for Test 23

Table 1: Predicted and measured tensile loadsMax. tensile load (kN)Test Predicted Measured

Test 11 (5mm/s, 0kPa) 1.1 2.4 Test 9 (100mm/s, 0kPa) 10.1 11.1 Test 10 (100mm/s, 200kPa) 25.6 24.2 HP5 sand: Test 23 (25mm/s, 200kPa) 30.6 33.2

Vertical Loading Tests in a Pressurised Chamber: Phase One Experimental Data

by

R.B. Kelly, B.W. Byrne, G.T. Houlsby and C.M. Martin

Report No. FOT016/1 “Foundations for Offshore Wind Turbines Research Project”

University of Oxford Department of Engineering Science Parks Road, Oxford, OX1 3PJ, U.K.

Tel. 01865 273162/283300

Fax. 01865 283301 Email Civil@eng.ox.ac.uk

http://www-civil.eng.ox.ac.uk/

TABLE OF CONTENTS

1.0 INTRODUCTION....................................................................................................................1 2.0 PRESSURE VESSEL AND TEST APPARATUS..................................................................2

2.1 Pressure chamber .......................................................................................................................2 2.2 Hydraulic Actuator .....................................................................................................................2 2.3 Selection and Construction of Sand Sample ..............................................................................3 2.4 Model caisson.............................................................................................................................4 2.5 Pressure control system..............................................................................................................4 2.6 Instrumentation ..........................................................................................................................4 2.7 Data acquisition system..............................................................................................................4

3.0 TEST METHODOLOGY ........................................................................................................5 4.0 SUMMARY OF TESTS CONDUCTED.................................................................................6

4.1 Test 1..........................................................................................................................................6 4.2 Test 2..........................................................................................................................................6 4.3 Tests 3 and 4...............................................................................................................................7 4.4 Test 5..........................................................................................................................................7 4.5 Test 6..........................................................................................................................................7 4.6 Test 7..........................................................................................................................................7 4.7 Test 8..........................................................................................................................................8 4.8 Test 9..........................................................................................................................................8 4.9 Test 10........................................................................................................................................8 4.10 Test 11......................................................................................................................................8 4.11 Test 12......................................................................................................................................8 4.13 Test 13......................................................................................................................................9

5.0 CONCLUDING REMARKS ...................................................................................................9 6.0 REFERENCES.........................................................................................................................9 7.0 TABLES.......................................................................................................................................11

1

VERTICAL LOADING TESTS IN A PRESSURISED CHAMBER: PHASE ONE EXPERIMENTAL DATA.

R.B. Kelly, B.W. Byrne, G.T. Houlsby and C.M. Martin

Department of Engineering Science Parks Rd, Oxford, OX1 3PJ

1.0 INTRODUCTION Within the next few years a number of wind farms will be constructed around the coast of the United Kingdom. In the first instance many of the wind turbine structures will be founded on piles. These foundations are simple to design, as they are a well-established technology, and represent a significant proportion of the overall installed cost for these structures. Various options are being investigated that may reduce the installed cost, and therefore increase the economic viability of the wind-farm developments. One possibility is to use skirted shallow foundations installed by suction as a multi-foundation system such as a tripod or quadruped (Byrne & Houlsby, 2002). The applied moment loads from the wind and waves will largely be reacted as vertical compression and tension loads on the individual foundations. There is little previous research on the vertical cyclic loading of skirted foundations in sand. Most recent research has been carried out at the University of Oxford (Byrne (2000); Johnson (1999)). The research carried out by Byrne (2000) involved vertical loading experiments on skirted foundations in an oil-saturated sand. These indicated that significant tensile capacities are possible and are limited by cavitation of the pore fluid. There is also evidence that the stiffness of the response changes when the load applied to the foundation changes from compression to tension (Byrne and Houlsby, 2002; Johnson, 1999). Both sets of experiments were carried out on the laboratory floor, where the ambient fluid pressure is very low. A central issue is that, since the available suction pressure is limited by cavitation of the pore fluid, the maximum available suction is much lower on the laboratory floor as compared with the offshore case. To investigate the effect of water pressure, and therefore cavitation, on the behaviour of suction caissons, a pressure chamber has been developed. This can simulate water pressures corresponding to depths of up to 20m. Bye et al (1995) have shown that the rate of loading, the permeability of the soil and the length of the drainage path affect suction pressures generated beneath the lid of the caisson. However, Johnson (1999), Mangal (2000) and Byrne (2000) present evidence that the rate of loading does not significantly affect the load-displacement response of a vertically loaded model shallow footing. These results would appear to be counter-intuitive, as the pressures beneath the caisson would be expected to affect its load-displacement response. To investigate the effects that the rate of loading has on the pressures generated beneath the caisson and the displacement of the caisson tests were conducted with a water-proof load cell fixed to the top of the caisson and pressure sensors placed to measure pore water pressures beneath the lid of the caisson and at the tip of its skirts. The tests are being conducted in two phases. Phase 1 tests were conducted using Redhill 110 silica sand. Phase 2 tests will be conducted using Oakamoor HPF5 silica sand. Redhill 110 sand is more permeable than Oakamoor HPF5 sand and was used to investigate the response of the model caisson to loads resulting in partially drained to drained conditions in the sand underlying the caisson. The HPF5 sand will be used to investigate the response of the model caisson to loads resulting in partially drained to undrained conditions in the sand underlying the caisson.

2

2.0 PRESSURE VESSEL AND TEST APPARATUS The test rig for the pressurised chamber tests comprises:

- a pressure chamber capable of withstanding 200kPa positive internal pressure; - a fast acting hydraulic actuator to provide vertical static and cyclic loads; - a saturated sand sample; - a test caisson; - a pressure control system; - instrumentation; and - a data acquisition system.

The arrangement of the pressure chamber, hydraulic actuator, sand sample and test caisson is shown in Figure 1. 2.1 Pressure chamber The pressure chamber was modified from an existing calibration chamber used to investigate the behaviour of penetrometers in clay. The calibration chamber was stripped of its bladders, various metal spacing rings and instrumentation fittings to leave a base, a flanged cylinder and a lid. The chamber was made airtight by the provision of rubber gaskets between the base, lid and flanged cylinder as well as providing blanking plugs for the various outlets from the pressure chamber. The chamber was designed to accommodate a working pressure of 200kPa. A false floor was constructed at the base of the pressure chamber for the sand sample to sit upon. The false floor was provided so a reservoir of water could exist beneath the sample. The water could then be pressurised, creating a uniform upward pressure gradient across the sand sample, which has the effect of loosening the sample. Once the sample has been loosened it can then be compacted down to the target test density by applying a vibration to it. The false floor was constructed from a 6mm thick sheet of aluminium sitting on top of a 20mm high spacer ring. A grid of 20mm diameter holes was bored into the aluminium and a sheet of 300 micron aperture wire mesh was fixed to the top of the sheet. A water/gas inlet/outlet line was fixed to the base of the pressure chamber. A support structure for the hydraulic actuator was constructed on top of the pressure chamber. This structure consisted of 4 threaded rods supporting steel plate members. The actuator was fixed to the steel plate. The actuator could then be positioned vertically by screwing the steel plate up or down the threaded rods. A gland plug was provided through the centre of the lid to allow the test caisson to be connected to the actuator. The gland plug consisted of a circular aluminium block with a hole bored through the centre of it to suit the rod fixing the test caisson to the actuator. An Oilite seal was placed within the hole in the plug to provide guidance to the rod fixing the caisson to the actuator. A lip seal was placed around the hole at the base of the plug to minimise pressure release through the gland. A vacuum pump line was fitted to the lid of the chamber to supply vacuum to the chamber. A pressure relief valve was also fitted for safety and set to a pressure of 350kPa. 2.2 Hydraulic Actuator The structural dynamics laboratory at Oxford University possesses several Instron hydraulic actuators. The actuators have the capability to provide fast feedback controlled loading, which allows wide range of load paths and loading rates to be applied to the model caisson. A 100kN capacity actuator was used to apply vertical static and cyclic loads to the test caisson. The

3

maximum rate of load-controlled cycling the actuator can accurately apply to the caisson was found to be about 10Hz. The hydraulic actuator is operated using a dedicated PC based control system. 2.3 Selection and Construction of Sand Sample Phase 1 tests were designed to span partially drained to drained loading conditions. The drainage characteristics of the sand beneath the caisson during loading can be estimated using Equation 1.

2

=

p

m

m

p

m

p

p

m

H

H

t

t

D

D

k

k (1)

where km and kp are the permeability of the model and prototype sands, Dm and Dp are the diameter of the model and prototype caissons, Hm and Hp are the length of the caisson skirts and tm and tp are the period of loading of the model and prototype caissons. Equation 1 has been derived in a similar manner to Equation 2 in Kelly (2002). The permeability of the Redhill 110 sand can be estimated using Equation 2 (Craig, 1987).

100

210d

km = m/s (2)

where d10 is the diameter, in millimetres, of the sand fraction corresponding to 10% by weight passing through a standard sieve test. A particle size diagram showing the results of a standard sieve tests is presented in Figure 2. The d10 value for Redhill 110 sand is about 0.075mm. From Equation 2, the permeability of the sand was estimated as 5.6x10-5m/s. The permeability of the prototype sand was estimated as 1x10-3m/s (Craig, 1987). The diameters of the model and prototype caissons were 0.28m and 6m respectively and their aspect ratio (H/D) was 2/3. The period of wave loading in the field was taken to be 10s. From Equation 1, the rate of loading required to simulate the behaviour of the prototype caisson was 1.8s/cycle or 0.56Hz. Maximum and minimum density tests were conducted on the Redhill 110. The maximum void ratio was found to be 1.037 while the minimum void ratio was 0.547. The sand sample was installed in two stages. Initially a filter layer, approximately 80mm thick, of Leighton Buzzard 16-30 silica sand was placed. The diameter of the particles in Leighton Buzzard 16-30 grade between 1mm and 0.4mm and have a similar particle grading to Redhill 110. The filter layer prevents the fine-grained Redhill 110 sand migrating through the false floor. The Redhill 110 was then installed on top of the filter layer. Approximately 706kg of dry Redhill 110 was placed. This mass of sand was chosen to allow both loose and dense samples to be created while leaving sufficient space between the top of the chamber and the surface of the sand to enable the test caisson to be suspended above the sand prior to testing. The sample was then saturated with water. The saturation procedure was quite complex. It was important to remove as much air as possible from the sand sample to allow the excess pore water pressure response in the sand during testing to simulate offshore conditions. The following procedure was used to saturate the sand sample:

1. The lid was fitted to the pressure chamber and a vacuum applied within the chamber.

4

2. Carbon dioxide gas was applied to the base of the chamber, flushed through the sample and removed through the vacuum line in the lid of the chamber.

3. 30 litres of water was boiled to de-air it and then sucked into the pressure chamber from the base using vacuum pressure. This procedure was repeated until the water level was about 150mm beneath the top of the pressure chamber. This required about 500 litres of water.

4. Carbon dioxide was again flushed through the sample under vacuum pressure. 5. The vacuum was then allowed to slowly dissipate to atmospheric pressure.

2.4 Model caisson The test caisson was constructed at Oxford University from a length of aluminium pipe. The caisson has a diameter of 280mm and a skirt depth of about 180mm giving a depth to diameter ratio of about 0.64. The caisson’s skirt is 3mm thick and its lid is 28mm thick. A pressure port was machined into the lid of the caisson to allow the excess pore water pressure beneath the lid of the caisson to be monitored. A groove was milled into the skirts of the caisson and a 3mm diameter aluminium tube with an Entran micro-miniature pressure sensor inside it was glued into the groove. A bleed hole was provided in the lid of the caisson to allow any air trapped beneath the lid of the caisson to be removed prior to testing. 2.5 Pressure control system Positive air pressure was introduced on top of the water within the pressure chamber via an existing compressed air line. A PC monitored the pressure sensors in the wall of the pressure chamber and the applied pressure was controlled using an air pressure regulator. 2.6 Instrumentation Two load cells, four pressure transducers and two displacement transducers were used to monitor events in the pressure vessel. The instrumentation and their uses are described in Table 1. A purpose built waterproof 100kN capacity load cell was constructed at Oxford University and placed on top of the caisson inside the pressure vessel. The load cell was designed to measure vertical and two orthogonal components of moment load. The load cell underwent significant development in order for it to operate in a pressurised water environment. The load cell consisted of a circular adaptor between the loading rods and the caisson. A 140mm diameter recess was milled into the circular adaptor and strain gauges stuck to the underside of the horizontal metal surface within the recess. A diagram of the load cell is provided in Figure 3. Waterproofing was achieved using an O-ring clamped between the base of the load cell and the top of the caisson. During calibration of this load cell it was found that the applied loads caused a small but significant deflection of the load cell. The deflection was calibrated against the applied load and used to correct the test displacement data. It was also found that output from the load cell was dependent on the ambient pressure. The apparent load was found to decrease in a linear and consistent manner by about 0.45kN per 100kPa. At the end of tests pressurised to 200kPa, when the excess pressure had been removed from the vessel, the load cell data was offset from zero load in the order of +0.9kN. It was inferred that air had seeped into the strain-gauged recess within the load cell during the test through the data cables. Thus the test data required correction for a gradual increase of pressure into the load cell. 2.7 Data acquisition system The data acquisition system comprises a RDP Modular 600 power/filter/amplifier source for the instrumentation, a Computerboards PCM-DAS-16D/16 PC based data acquisition board, junction

5

boxes and cabling. All of the data acquisition system was procured or constructed with the exception of the PCM-DAS-16D/16 PC based data acquisition board. The instrumentation within the pressure chamber was transferred out of the chamber via waterproof Lemo connectors and couplers. Cables from the pressure vessel were then attached to the RDP Modular 600 along with the PDCR 810 pressure sensors and 250mm stroke LVDT. The output signals from the RDP Modular 600 were fed into a second junction box, which then transferred the signals to the PCM-DAS-16D/16 PC based data acquisition board. The data from the Instron load cell and built in displacement transducer were recorded by the dedicated PC based hydraulic control system. 3.0 TEST METHODOLOGY The following methodology was used to prepare the test bed and apply the loads.

1. Approximately 50 litres of water, raised about 6m above the base of the sample, was allowed to pass from the bottom of the sample to the top of the sample in order to fluidise the sand and reduce its density.

2. A circular metal plate, of approximately 1m diameter, was placed as surcharge on top of the sample. The plate had a length of circular hollow section fixed to its top surface to act as a guide through a gland to maintain its vertical alignment. An eccentric mass vibrator, attached to the side of the pressure vessel was activated until the circular plate had sunk a pre-determined distance.

3. The circular plate was removed. Suction pressures were generated beneath the plate as it was lifted off the sand, which caused the sand to loosen. The eccentric mass vibrator was then activated to bring the sand to its target density. The target density was achieved when the top of the sand reached a targeted average depth beneath the top of the pressure vessel. The top of the sand was generally convex, with the highest point being at the centre of the pressure vessel and the lowest points around the edge of the pressure vessel.

4. The void in the lid of the model caisson beneath the pressure sensor was filled with glycerol to limit the potential for air bubbles to be trapped within the void. A protective cap covering the end of the pressure sensor at the tip of the caissons skirts was removed.

5. The lid of the pressure vessel, incorporating the model caisson and Instron actuator, was fixed to the pressure vessel.

6. The instrumentation was connected to the control and data acquisition systems. 7. The hydraulic pumps were activated. 8. The PC based hydraulic controller was turned on. 9. The control program was run and the active and dummy Instron load cells calibrated. 10. Hydraulic pressure was applied to the active Instron actuator. 11. Zero readings were taken for the instrumentation within the pressure vessel. 12. The caisson was pushed slowly into the ground using displacement control. Various rates

of penetration were tried and a rate of 0.1mm/second was found to be optimal. Greater rates of penetration caused significant water pressures to form inside the caisson, above the sand, during penetration. These pressures had the potential to fluidise the sand beneath the caisson during penetration, ruining the test.

13. The caisson was penetrated into the sand until the underside of its lid touched the sand. Penetration of the caisson into the sand was continued until a load of about 10kN was reached. The displacement ramp was then stopped. The soil-structure system then relaxed a little and the loads dropped by 1-2kN.

14. An air pressure ranging between 5-10kPa was applied to the pressure vessel in order to flush out air from beneath the caisson and from the drainage tube leading from the lid of the caisson to the outside of the pressure vessel. It was critical not to provide more than about 5-10kPa as the soil within the caisson could fluidise.

6

15. Optimisation of the Instron actuators load control system occurred during tests 1-4 by applying cyclic square waves and tuning the actuators response until it became close to the square wave input. Optimisation was not conducted during tests 5-13.

16. The ambient pressure within the vessel was set. The ambient pressure was either atmospheric pressure or 200kPa above atmospheric pressure.

17. The caisson was again penetrated into the sand at a rate of 0.1mm/s until the target mean load for the test was registered on the Instron load cell, for most of the tests. The ramp was stopped and the loads again reduced by 1-2kN. The mean target load was reached via a series of step changes in load during tests 9 and 10.

18. The target mean load was set by applying load control to the actuator. The target mean load was 35kN for most of the tests and was 15kN for test 13.

19. Various packets of sinusoidal cyclic loads were applied to the caisson, if required. 20. The caisson was pulled out of the sand. 21. Elevated ambient pressure within the vessel was reduced to atmospheric pressure.

4.0 SUMMARY OF TESTS CONDUCTED Thirteen tests have been conducted in the pressure vessel during Phase 1. A summary of the tests is presented in Table 2. These tests include preliminary tests as well as production tests. A description of each test and figures showing the load-displacement and pressure-displacement data from each test is presented in the following. 4.1 Test 1 This was a commissioning test conducted to ascertain the performance of the pressure vessel rig. The test consisted of penetrating the caisson into the soil to a mean load of 35kN and then applying packets of cyclic loads detailed in Table 3. The time history for this test is shown in Figure 4. Axial load and pore pressure beneath the lid of the caisson are plotted against axial displacement in Figure 5. The test data showed that:

• the permanent cyclic displacement increased as the cyclic load amplitude increased; • there was little tensile capacity during cyclic loading; and • the stiffness of the foundation reduced as the caisson was unloaded.

The data were affected by signal noise during this test. A ‘cracking’ noise was heard during cycle packet 4. At the end of the test, the lid of the pressure vessel was removed and it was observed that the skirt of the caisson had split down the groove where the Entran pressure sensor was fixed to the caisson. The Entran pressure sensor was not operational during this test. The caissons skirts were later welded together and the aluminium tube containing the Entran pressure sensor glued to the outside of the skirts. 4.2 Test 2 Test 2 was an exact replica of test 1, except that cyclic loads were applied at a rate of 1Hz instead of 0.5Hz. The time history for this test is shown in Figure 6. Axial load and pore pressure beneath the lid of the caisson are plotted against axial displacement in Figure 7. The gap in the data shown in Figure 7 between initial penetration and cyclic loading was when optimisation of the load control system for the Instron actuator occurred. The Entran pressure sensor was not operational during this test. Data from this test confirmed the findings from Test 1. The pore pressures measured beneath the lid of the caisson were greater in Test 2 than Test 1 because the caisson remained intact throughout the

7

test. The magnitude of the pore pressures recorded in Test 2 were not sufficient to affect the load-displacement response of the caisson. 4.3 Tests 3 and 4 Test 3 and test 4 were unsuccessful due to excess pressure being applied to the vessel during the stage where air was removed from within the caisson. This caused the soil beneath the caisson to liquefy and reduce in strength. As a result, on the application of cyclic loading, the Instron actuator rapidly reached its displacement limit. The caisson was pulled out of the soil at 100mm/s in each test in order to determine whether cavitation pressures could be generated beneath the caisson. Data from Tests 3 and 4 are presented in Figures 8 and 9. The Entran pressure sensor was not operational during these tests. 4.4 Test 5 The design for the waterproof load cell was modified between Test 4 and Test 5. The application of packets of cyclic loading, similar to previous tests, proceeded smoothly. Prior to pulling the caisson from the soil a large compressive load was applied. The caisson was then pulled from the sand. The time history for Test 5 is presented in Figure 10. The axial load is plotted against the displacement in Figure 11 and the pressures recorded beneath the lid of the caisson and at its skirts are shown in Figure 12. Test 5 was a repeat of test 2, with the exception that the rate of cyclic loading was 10Hz rather than 1Hz. Cyclic load packets 5 and 6 consisted of 10 cycles whereas the equivalent load packets in Test 2 were comprised of 5 cycles. The actuator was unable to attain the target minimum cyclic load during cyclic load packets 5 and 6. The pore pressures recorded beneath the lid of the caisson were less than those recorded in Test 2, where the rate of loading was 1Hz. The loads and pressures recorded as the caisson was pulled out of the sand were greater in Test 5 than Test 2 because the rate of displacement was 100mm/s compared with 5mm/s in the respective tests. The pressure beneath the lid of the caisson increased prior to the pressure at the caissons skirts but the pressures became equal after a displacement of about 30mm. The pressure generated beneath the caisson as it was extracted from the sand was limited by cavitation of the pore fluid. 4.5 Test 6 Test 6 was a repeat of Test 2 with the exception that the ambient pressure was 200kPa rather than 0kPa above atmospheric pressure. The time history for Test 6 is shown in Figure 13 and load-displacement and pressure-displacement data are shown in Figures 14 and 15 respectively. Elevated ambient pressure appeared to have little effect on the load-displacement data during cyclic loading, compared with data from Test 3. The excess pore pressures (recorded pressure – ambient pressure) were also similar to those recorded during Test 2. The axial load and excess pore pressure were both greater as the caisson was extracted from the sand than in Test 5 because the pressure beneath the caisson was not limited by cavitation of the pore fluid. 4.6 Test 7 Cyclic loading during Test 7 was conducted with multiple frequencies and multiple load amplitudes. The cyclic frequencies and load amplitudes are outlined in Table 4. A step change in load was applied between 20-25kN during initial loading in order to assess the time required for excess pore water pressures to dissipate from beneath lid of the caisson. Further step changes from 35-55kN and from 55-35kN were made cyclic load packets 12 and 13. The time history for Test 7 is shown in Figure 16.

8

Load-displacement and pressure-displacement data are presented in Figures 17 and 18 respectively. The pressures beneath the lid of the caisson were observed to increase as the rate of cycling increased for all cyclic load amplitudes. However, the pressures recorded did not appear to affect the load-displacement response of the caisson. The load and pressure response of the caisson as it was pulled out of the sand was similar to that recorded during Test 5. 4.7 Test 8 Test 8 was a repeat of Test 7 with the exception that it was conducted under an ambient pressure of 200kPa rather than 0kPa. Load packet 13 in Table 4 was repeated as load packet 14 in this test. Load packet 14 in Table 4 became load packet 15 in this test. Load packets 15 and 16 in Table 4 were not applied because the maximum stroke of the Instron actuator had been reached. The time history for this test is shown in Figure 19 with load-displacement and pressure-displacement data shown in Figures 20 and 21. The data showed that the elevated pressure did not significantly affect the load-displacement or excess pressure response during cyclic loading. The rate of loading did not affect the data either. The load and pressure response as the caisson was pulled out of the sand was similar to that recorded in Test 6. 4.8 Test 9 Test 9 consisted of pushing the caisson into the sand until a load of 10kN was reached. The load was then increased in a series of steps to further investigate the rate of dissipation of pore pressures beneath the caisson. The loads were stepped in 5kN increments from 10kN to 35kN. During pullout of the caisson from the sand the Entran pressure sensor failed at a pressure of about –100kPa. The time history for Test 9 is shown in Figure 22 with the load and pressure data plotted against displacement in Figure 23. Data from Test 9 can be compared with that from Test 7. The difference between the tests is that the caisson was pulled out of the sand prior to cyclic loading in Test 9 whereas it was extracted after cyclic loading in Test 7. This comparison showed that the maximum ultimate loads and pressures were similar in each test but the displacement required to reach the ultimate state was greater in Test 7 than in Test 9. The initial unloading response in Test 9 was much more stiff than in Test 7. 4.9 Test 10 Test 10 was a repeat of Test 9 with an ambient pressure of 200kPa. The time history of this test is shown in Figure 24 with the load and pressure data shown in Figure 25. Test 10 can be compared with Test 8 in a similar manner to that described in Section 4.8. This comparison produced a similar result to that discussed in Section 4.8. 4.10 Test 11 Test 11 consisted of installing the caisson until a mean load of 35kN was applied by the Instron actuator to the caisson and then pulling the caisson out at 5mm/second. This test was conducted to compare with test 2. The time history for this test is shown in Figure 26 with the load and pressure response presented in Figure 27. The data from Test 11 can be compared with that from Test 2 and similar conclusions to those described in Section 4.8 were made. 4.11 Test 12 Test 12 consisted the application of 1000 cycles once the mean load of 35kN had been applied. The cyclic amplitude was ±15kN and the frequency was 0.5Hz. The test was conducted with an ambient

9

pressure of 200kPa. The time history for this test is shown in Figure 28 and the corresponding load/pressure/displacement data is presented in Figure 29. The load-displacement data showed that the permanent displacement during a single cycle reduced as the number of cycles increased. This phenomenon is known as ‘shakedown’ and is commonly observed in cyclic loading tests on sand. The test data showed no accumulation of pore pressures beneath the lid of the caisson during cycling. Consequently the load-displacement data was not significantly affected by the large number of cycles applied compared with the previous tests. 4.13 Test 13 Test 13 consisted the application of 1000 cycles once the mean load of 15kN had been applied. The cyclic amplitude was ±15kN and the frequency was 0.5Hz. The test was conducted with an ambient pressure of 200kPa. The time history for this test is shown in Figure 30 and the load/pressure/displacement data is shown in Figure 31. This test differed from Test 12 in that the minimum cyclic load was 0kN, where the stiffness of the foundation had been observed to decrease in previous tests. The load-displacement data showed that the total displacement during the 1000 cycles applied in Test 13 was greater than the total displacement recorded in Test 12. However, the ‘shakedown’ process was again observed during Test 13 as was the lack of pore pressure accumulation. 5.0 CONCLUDING REMARKS Tests constituting Phase 1 of the test programme have been completed to investigate the behaviour of a model caisson subject to partially drained to drained conditions. Phase 2 will comprise tests to investigate the behaviour of the model caisson under partially drained to undrained conditions. Data from a series of 13 tests have been presented. The data constitutes time histories and load/pressure plotted against displacement for each of the tests. In general the data showed that:

• The tensile capacity of the model caisson was small; • The axial stiffness of the foundation reduced significantly as the caisson was unloaded to

zero load; • The rate of loading had little effect on the load-displacement response of the caissons during

cyclic loading; • Elevated ambient pressure did not increase the tensile capacity or foundation stiffness during

cyclic loading; • The rate of loading increased the tensile capacity as the caisson was pulled out of the

ground; • Elevated ambient pressure increased the tensile capacity of the caisson as it was pulled out

of the ground; and • The ultimate tensile load was mobilised after a displacement of about 10% of the caissons

diameter as it was pulled out of the ground. The data will be interpreted in a later report. 6.0 REFERENCES Bye, A., Erbrich, C. and Rognlien, B. (1995), Geotechnical Design of Bucket Foundations, OTC

paper 7793, pp 869-883.

10

Byrne, B.W. (2000), Investigations of Suction Caissons in Dense Sand, DPhil Thesis, University of Oxford.

Byrne, B.W. and Houlsby, G.T. (2002), Experimental Investigations of the Response of Suction Caissons to Transient Vertical Loading, Proc. ASCE Journal of Geotechnical Engineering 128 No 11, pp 926-939.

Craig, R.F. (1987), Soil Mechanics 4th Edition, Van Nostrand Reinhold (International).

Johnson, K. (1999), The Behaviour of Partially Drained Footings under Axial Load, 4th year project report, University of Oxford.

Kelly, R.B. (2002), Proposal for Large Scale Field Trials of Suction Caissons: Phase 1: Tests In Sand, Report FOT002 for DTI project “The application of suction caisson foundations to offshore wind turbines”.

Mangal, J. (1999), Partially Drained Loading of Shallow Foundations on Sand, DPhil Thesis, University of Oxford.

11

7.0 TABLES Instrument Use Water-proof 100kN Capacity Load Cell Instron 100kN Capacity Load Cell 2 x PDCR 810 Pressure Sensors PDCR 830 Pressure Sensor Entran Miniature Pressure Sensor D5 250mm stroke LVDT Displacement Transducer Instron Displacement Trandsucer

Placed on top of the test caisson within the pressure chamber to record the vertical loads applied to the test caisson. Fitted to the hydraulic actuator external to the pressure vessel to control the applied loads, to provide a backup to the waterproof load cell and to provide a measure of the friction between the driving rods and the pressure gland. One sensor fitted at the top of the pressure vessel to monitor the applied pressure. The other sensor fitted at the base of the sample to compare with the applied pressure and to record any hydraulic gradients within the sample. Records the pressure beneath the lid of the test caisson Records the pressure at the tip of the test caissons skirt Mounted external to the pressure chamber to record the axial movement of the test caisson Built in to the actuator. Allows feedback control to be provided and is a back up to the LVDT displacement transducer

Table 1 – Instrumentation Used in Pressure Chamber Tests

12

Test Date Ave. Sample Height

Dry Unit Weight

Relative Density

Pressure Cyclic Frequency

Pullout Rate

Comments

(m) (kN/m3) (kPa) Hz (mm/s) 1 03/04/2003 0.57 15.7 0.79 0 0.5 5 Commissioning test. Caisson wall split during test along

pressure tube. 2 11/04/2003 0.56 15.8 0.81 0 1 5 3 28/04/2003 0.56 15.9 0.82 0 - 100 Soil failed during evacuation of air from caisson. Fast pullout

conducted to gain some data 4 30/04/2003 0.55 16.2 0.88 0 - 100 Soil failed during evacuation of air from caisson. Fast pullout

conducted to gain some data 5 0.55 16.3 0.89 0 10 100 Multi amplitude cyclic test 6 25/07/2003 0.56 15.8 0.80 200 1 100 Multi amplitude cyclic test 7 30/07/2003 0.56 15.8 0.80 0 Multi 100 Multi frequency / mult amplitude cyclic test with step changes

to find t50 8 01/08/2003 0.56 15.8 0.80 200 Multi 100 Multi frequency / mult amplitude cyclic test with step changes

to find t50 9 04/08/2003 0.56 15.9 0.81 0 - 100 Stepped Pushin then Pullout 10 05/08/2003 0.56 15.9 0.82 200 - 100 Stepped Pushin then Pullout 11 06/08/2003 0.56 15.9 0.81 0 - 5 Push / Pull 12 07/08/2003 0.56 15.9 0.82 200 0.5 100 1000 cycles 35kN ave load +/- 15kN amplitude 13 08/08/2003 0.56 15.8 0.80 200 0.5 100 1000 cycles 15kN ave load +/- 15kN amplitude

Table 2 Summary of Tests Conducted in the Pressure Vessel to 8 August 2003

13

Cycle Packet Number of Cycles Amplitude 1 10 ±5kN 2 10 ±10kN 3 10 ±20kN 4 10 ±30kN 5 5 ±35kN 6 5 ±40kN

Table 3 – Cyclic loads applied to the caisson during test 1

Cyclic Load Packet No: of Cycles Frequency (Hz) Load Amplitude (kN) 1 2 3 4 5 6 7 8 9

10 11 12 13 14 15 16

10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10

0.1 1 5

10 0.1 1 5

10 0.1 1 5

10 0.1 1 5

10

+/-5 +/-5 +/-5 +/-5 +/-10 +/-10 +/-10 +/-10 +/-20 +/-20 +/-20 +/-20 +/-30 +/-30 +/-30 +/-30

Table 4 – Cyclic frequencies and load amplitudes applied in test 7

14

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9) (10)

PSPS

PS

(1) Pressure Chamber(2) Void Beneath Sample(3) Saturated Sand Sample(4) Water Level(5) Instron Actuator(6) Instron Support Structure(7) Test Caisson(8) Waterproof Load Cell(9) Instron Load Cell(10) LVDT Displacement TransducerPS Pressure Sensor

PS

Figure 1 – Pressure Vessel

0

10

20

30

40

50

60

70

80

90

100

0.001 0.01 0.1 1

Particle Size (mm)

Perc

ent P

assi

ng

Oakamoor HPF5Redhill 110

Figure 2 – Particle size distribution for HPF5 and Redhill 110 sands

15

200

140

50

50

Holes for M10 bolts,60 degrees apart85mm centre tocentre from centroid.

M12 CoarseThread

Cable Gland

20

20Flat Surface fordowty seal 20mmwide. Threadedhole M12 finethread 1.5 pitchhalf way betweenbolt holes

20

20

O-Ring Groove to suit257 series oring148.8ID, 3.53mmthick

30

O-Ring Groove to suit257 series oring148.8ID, 3.53mmthick

35

Flat Surface fordowty seal 20mmwide. Threadedhole M12 finethread 1.5 pitch

Flat Surface fordowty seal 20mmwide. Threadedhole M12 finethread 1.5 pitchhalf way betweenbolt holes

Flat Surface fordowty seal 20mmwide. Threadedhole M12 finethread 1.5 pitch

Plan

Elevation

Figure 3 –Water-proof load cell design.

16

-20

0

20

40

60

80

100

0 500 1000 1500 2000 2500 3000 3500

Time (s)

Load

(kN

)

-100

-80

-60

-40

-20

0

20

Pre

ssur

e (k

Pa)

Load

Lid PPT

Figure 4 – Time History for Test 1

-10

0

10

20

30

40

50

60

70

80

150 160 170 180 190 200 210 220 230 240 250

Displacement (mm)

Load

(kN

)

-20

0

20

40

60

80

100

120

140

160

Pre

ssur

e (k

Pa)

Axial LoadLid PPT

Figure 5 – Load and Pressure plotted against Displacement for Test 1

17

-10

0

10

20

30

40

50

60

70

80

90

0 500 1000 1500 2000 2500 3000 3500 4000

Time (s)

Load

(kN

)

-300

-250

-200

-150

-100

-50

0

50

100

Pre

ssur

e (k

Pa)

Initial Penetration

Cyclic Loading

Instron Optimisation

Lid PPT

Figure 6 – Time History for Test 2

-10

0

10

20

30

40

50

60

70

80

150 170 190 210 230 250

Displacement (mm)

Load

(kN

)

-30

0

30

60

90

120

150

180

210

240

Pre

ssur

e (k

Pa)

Axial LoadLid PPT

Figure 7 – Load and Pressure plotted against Displacement for Test 2

18

-15

-10

-5

0

5

10

15

0 1000 2000 3000 4000 5000 6000

Time (s)

Load

(kN

)

-50

-40

-30

-20

-10

0

10

Pre

ssu

re (

kPa)

Axial LoadLid PPT

Figure 8 – Time History for Test 3

-15

-10

-5

0

5

10

15

20

0 1000 2000 3000 4000 5000 6000

Time (s)

Load

(kN

)

-800

-700

-600

-500

-400

-300

-200

-100

0

100

200

Pre

ssur

e (k

Pa)

Axial Load

Lid PPT

Figure 9 – Time History for Test 4

19

0

20

40

60

80

100

120

140

160

180

200

0 1000 2000 3000 4000 5000 6000 7000

Time (s)

Load

(kN

)

-110

-90

-70

-50

-30

-10

10

30

Pre

ssur

e (k

Pa)

Axial LoadLid PPT

Figure 10 – Time History for Test 5

-20

0

20

40

60

80

100

120

140

160

0 50 100 150 200 250

Displacement (mm)

Axi

al L

oad

(kN

)

installationloading

instron pullout

Figure 11 – Load plotted against Displacement for Test 5

20

-100

-80

-60

-40

-20

0

20

50 70 90 110 130 150 170 190 210 230 250

Displacement (mm)

Pre

ssur

e (k

Pa)

Lid PPTSkirt PPT

Figure 12 – Lid and skirt pressure plotted against Displacement for Test 5

-40

-20

0

20

40

60

80

100

120

0 1000 2000 3000 4000 5000 6000

Time (s)

Load

(kN

)

-500

-400

-300

-200

-100

0

100

200

300

Pre

ssur

e (k

Pa)

Axial LoadLid PPT

Figure 13 – Time History for Test 6

21

-40

-20

0

20

40

60

80

50 100 150 200 250

Displacement (mm)

Load

(kN

)

Figure 14 – Load plotted against Displacement for Test 6

0

50

100

150

200

250

300

50 100 150 200 250

Displacement (mm)

Pre

ssur

e (k

Pa)

Lid PPTSkirt PPT

Figure 15 – Pressure plotted against Displacement for Test 6

22

-20

0

20

40

60

80

100

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Time (s)

Load

(kN

)

-200

-150

-100

-50

0

50

Pre

ssur

e (k

Pa)

Axial Load

Lid PPT

Figure 16 – Time History for Test 7

-20

-10

0

10

20

30

40

50

60

70

150 170 190 210 230 250

Displacement (mm)

Load

(kN

)

Figure 17 – Load plotted against Displacement for Test 7

23

-120

-100

-80

-60

-40

-20

0

20

40

60

150 170 190 210 230 250

Displacement (mm)

Pre

ssur

e (k

Pa)

Lid PPTSkirt PPT

Figure 18 – Pressure plotted against Displacement for Test 7

-40

-20

0

20

40

60

80

100

0 1000 2000 3000 4000 5000 6000

Time (s)

Load

(kN

)

-500

-400

-300

-200

-100

0

100

200

Pre

ssur

e (k

Pa)

Axial LoadLid PPT

Figure 19 – Time History for Test 8

24

-30

-20

-10

0

10

20

30

40

50

60

70

150 170 190 210 230 250

Displacement (mm)

Load

(kN

)

Figure 20 – Load plotted against Displacement for Test 8

0

50

100

150

200

250

100 120 140 160 180 200 220 240 260

Displacement (mm)

Pre

ssur

e (k

Pa)

Lid PPT

Skirt PPT

Figure 21 – Pressure plotted against Displacement for Test 8

25

-20

0

20

40

60

80

100

0 500 1000 1500 2000 2500 3000 3500

Time (s)

Axi

al L

oad

(kN

)

-120

-100

-80

-60

-40

-20

0

20

40

Pre

ssur

e (k

Pa)

Axial Load

Lid PPT

Figure 22 – Time History for Test 9

-20

-10

0

10

20

30

40

0 50 100 150 200

Displacement (mm)

Load

(kN

)

-100

-50

0

50

100

150

200

Pre

ssur

e (k

Pa)

Axial LoadLid PPT

Figure 23 – Load and Pressure plotted against Displacement for Test 9

26

-40

-20

0

20

40

60

80

1000 1500 2000 2500 3000 3500 4000 4500 5000

Time (s)

Load

(kN

)

-500

-400

-300

-200

-100

0

100

200

300

Pre

ssur

e (k

Pa)

Axial LoadLid PPT

Figure 24 – Time History for Test 10

-25

0

25

50

0 50 100 150 200

Displacement (mm)

Load

(kN

)

-100

-50

0

50

100

150

200

Pre

ssur

e (k

Pa)

Axial LoadLid PPT

Figure 25 – Load and Pressure plotted against Displacement for Test 10

27

-10

0

10

20

30

40

50

0 500 1000 1500 2000 2500 3000 3500

Time (s)

Load

(kN

)

-100

-80

-60

-40

-20

0

20

Pre

ssur

e (k

Pa)

Axial LoadLid PPT

Figure 26 – Time History for Test 11

-5

0

5

10

15

20

25

30

35

0 50 100 150 200

Displacement (mm)

Load

(kN

)

-20

0

20

40

60

80

100

120

140

Pre

ssur

e (k

Pa)

Axial LoadLid PPT

Figure 27 – Load and Pressure plotted against Displacement for Test 11

28

-20

-10

0

10

20

30

40

50

60

70

80

0 1000 2000 3000 4000 5000 6000

Time (s)

Load

(kN

)

-50

0

50

100

150

200

250

Pre

ssur

e (k

Pa)

Axial LoadLid PPT

Figure 28 – Time History for Test 12

-20

-10

0

10

20

30

40

50

0 50 100 150 200 250

Displacement (mm)

Load

(kN

)

-80

-40

0

40

80

120

160

200

Pre

ssur

e (k

Pa)

Axial LoadLid PPT

Figure 29 – Load and Pressure plotted against Displacement for Test 12

29

-30

-20

-10

0

10

20

30

40

50

60

0 1000 2000 3000 4000 5000 6000

Time (s)

Axi

al L

oad

(kN

)

-50

0

50

100

150

200

250

Pre

ssur

e (k

Pa)

Axial Load

Lid PPT

Figure 30 – Time History for Test 13

-25

0

25

50

0 50 100 150 200 250

Displacement (mm)

Load

(kN

)

-125

-100

-75

-50

-25

0

25

50

75

100

125

150

175

200

225

250

Pre

ssur

e (k

Pa)

Axial LoadLid PPT

Figure 31 – Load and Pressure plotted against Displacement for Test 13

Vertical Loading Tests in a Pressurised Chamber: Phase Two Experimental Data

by

R.B. Kelly, B.W. Byrne, G.T. Houlsby and C.M. Martin

Report No. FOT019/1 “Foundations for Offshore Wind Turbines Research Project”

University of Oxford Department of Engineering Science Parks Road, Oxford, OX1 3PJ, U.K.

Tel. 01865 273162/283300

Fax. 01865 283301 Email Civil@eng.ox.ac.uk

http://www-civil.eng.ox.ac.uk/

TABLE OF CONTENTS

1.0 INTRODUCTION ...................................................................................................................1 2.0 PROPERTIES AND CONSTRUCTION OF THE HPF5 SAND TEST BED........................1 3.0 TEST METHODOLOGY........................................................................................................2 4.0 SUMMARY OF TESTS CONDUCTED ................................................................................3

4.1 Test 14........................................................................................................................................3 4.2 Test 15........................................................................................................................................3 4.3 Test 16........................................................................................................................................4 4.4 Test 19........................................................................................................................................4 4.5 Test 20........................................................................................................................................4 4.6 Test 21........................................................................................................................................4 4.7 Test 22........................................................................................................................................5 4.8 Test 23........................................................................................................................................5 4.9 Test 24........................................................................................................................................5 4.10 Test 25......................................................................................................................................5

5.0 CONCLUDING REMARKS...................................................................................................6 6.0 REFERENCES.........................................................................................................................6 7.0 TABLES.........................................................................................................................................7 8.0 FIGURES...............................................................................................................................10

VERTICAL LOADING TESTS IN A PRESSURISED CHAMBER: PHASE TWO EXPERIMENTAL DATA.

R.B. Kelly, B.W. Byrne, G.T. Houlsby and C.M. Martin

Department of Engineering Science Parks Rd, Oxford, OX1 3PJ

1.0 INTRODUCTION Vertical loading tests conducted in a pressurised chamber have been carried out in two phases. Phase 1 tests were conducted using Redhill 110 silica sand and were reported in the document FOT016/1 (Kelly et al, 2003) entitled ‘Vertical loading tests in a pressurised chamber: phase one experimental data’. Phase 2 tests have been conducted using Oakamoor HPF5 silica sandy silt. Redhill 110 sand is more permeable than the Oakamoor HPF5 material and was used to investigate the response of the model caisson to loads resulting in partially drained to drained conditions in the sand underlying the caisson. The HPF5 was used to investigate the response of the model caisson to loads resulting in partially drained to undrained conditions in the sand underlying the caisson. The apparatus used in the Phase 2 tests was identical to that used in the Phase 1 tests and is fully described in Kelly et al (2003) 2.0 PROPERTIES AND CONSTRUCTION OF THE HPF5 SAND TEST BED A particle size diagram showing the results of standard sieve tests on the Redhill 110 and the Oakamoor HPF5 material is presented in Figure 1. According to the Unified Soil Classification System Oakamoor HPF5 is a low plasticity sandy silt, while Redhill 110 is a poorly graded, fine grained sand. The HPF5 particles are more angular than the Redhill 110 particles as it is produced artificially by crushing naturally occurring silica sand. Maximum and minimum density tests were conducted on the HPF5. The maximum void ratio was found to be 1.014 while the minimum void ratio was 0.467. In comparison, the Redhill 110 sand had a maximum void ratio of 1.037 and a minimum void ratio of 0.547. The relative density of the HPF5 test beds prepared for the work reported here ranged between 0.53 and 0.73. The angularity of the HPF5 particles causes the material to have a higher strength than the Redhill 110 sand. Data from consolidated-undrained triaxial compression tests on HPF5 with a relative density of 0.78, and Redhill 110 having a relative density of 0.75 are presented in Figure 2 and show that the peak friction angle for the HPF5 was 48.4° while the peak friction angle for the Redhill 110 was 43.9°. The permeability of HPF5 can be estimated using Equation 1 (Craig, 1987).

100

210dkm = m/s (1)

where d10 is the diameter, in millimetres, of the sand fraction corresponding to 10% by weight passing through a standard sieve test. The data shown in Figure 1 indicates that the d10 value for HPF5 is about 0.007mm. From Equation 1, the permeability of HPF5 was estimated as 4.9x10-7m/s.

1

In comparison, the permeability of Redhill 110 was previously estimated as 5.6x10-5m/s (Kelly et al, 2003). The test bed for the Phase 2 tests was constructed in a different manner to the Phase 1 tests in order to minimise the environmental hazard of fine silica dust clouds being created during placement of the HPF5. The water was drained from the pressure chamber and most of the Redhill 110 removed. A 40mm thick layer of Redhill 110 was left to provide an additional filter to the existing 80mm thick filter of Leighton Buzzard 16-30 silica sand. The tank was then filled with water to suppress dust and about 680kg of HPF5 was placed into the tank. As much air as possible was removed from the test bed by repeatedly vibrating the pressure chamber. 3.0 TEST METHODOLOGY The following methodology was used to prepare the test bed and apply the loads.

1. Water was introduced to the bottom of the sample from an elevated position in order to fluidise the sand and reduce its density.

2. A circular metal plate, of approximately 1m diameter, was placed as surcharge on top of the sample. The plate had a length of circular hollow section fixed to its top surface to act as a guide through a gland to maintain its vertical alignment. An eccentric mass vibrator, attached to the side of the pressure vessel was activated until the circular plate had settled a pre-determined distance.

3. The circular plate was removed. The depth from the top of the pressure chamber to the top of the test bed was recorded at 5 locations and the average used to compute the density of the sample.

4. The void in the lid of the model caisson beneath the pressure sensor was filled with glycerol to limit the potential for air bubbles to be trapped within the void.

5. The lid of the pressure vessel, incorporating the model caisson and Instron actuator, was fixed to the pressure vessel.

6. The instrumentation was connected to the control and data acquisition systems. 7. The hydraulic pumps were activated. 8. The PC based hydraulic controller was turned on. 9. The control program was run and the active and dummy Instron load cells calibrated. 10. Hydraulic pressure was applied to the active Instron actuator. 11. Zero readings were taken for the instrumentation within the pressure vessel. 12. The caisson was pushed slowly into the ground using displacement control. The initial rate

of penetration was 0.05mm/second and was maintained until the lid of the caisson came in contact with the surface of the water. The rate of penetration was then reduced of 0.02mm/s in order to minimise the water pressures generated beneath the lid of the caisson while maintaining a rate of penetration suited to allow the test to be conducted in a reasonable time span. Greater rates of penetration caused significant water pressures to form inside the caisson, above the sand, during penetration. These pressures had the potential to fluidise the sand beneath the caisson during penetration, ruining the test.

13. The caisson was penetrated into the sand until the underside of its lid touched the sand. Penetration of the caisson into the sand was continued until a load of about 35kN was reached. The valve on the drainage line leading from the top of the caisson to outside the pressure chamber was closed when a load of 5kN had been applied to the caisson. At this

2

stage only water was being extruded from the caisson indicating that any air trapped beneath the lid of the caisson had been flushed out.

14. The ambient pressure within the vessel was set. The ambient pressure was either atmospheric pressure or 200kPa above atmospheric pressure.

15. Various packets of sinusoidal cyclic loads were applied to the caisson, if required. 16. The caisson was pulled out of the sand. 17. Elevated ambient pressure within the vessel was reduced to atmospheric pressure.

4.0 SUMMARY OF TESTS CONDUCTED Ten tests have been successfully conducted in the pressure vessel during Phase 2. A summary of the tests is presented in Table 1. A description of each test and figures showing the load-displacement and pressure-displacement data from each test is presented in the following. All of the load data presented in the figures were recorded by the waterproof load cell inside the pressure chamber. In the text, all of the loads applied to the caisson refer to loads applied by the Instron actuator and measured by its load cell outside of the pressure chamber. The applied loads can differ from the loads recorded inside the pressure chamber as a result of friction in the gland in the lid of the caisson through which the ram penetrates. 4.1 Test 14 This test was conducted for comparison with Test 2 on Redhill 110 sand (Kelly et al, 2003). The test consisted of penetrating the caisson into the soil to a mean load of 10kN and then applying step loads in 5kN increments to 35kN. The time taken for the excess pore pressure beneath the lid of the caisson to drain during each load increment was recorded for use in estimating the permeability of the soil. Cyclic loads were then applied in addition to the mean load of 35kN in packets described in Table 2. The time history for this test is shown in Figure 3. Axial load and pore pressure beneath the lid of the caisson are plotted against axial displacement in Figure 4. The test data showed that:

• the permeability of the sand beneath the caisson was in the order of 4x10-7m/s, which is in good agreement with the estimate made using Equation 1;

• the permanent cyclic displacement increased as the cyclic load amplitude increased; • there was little tensile capacity during cyclic loading; and • the stiffness of the foundation reduced as the caisson was unloaded.

Positive and negative excess pore pressures beneath the lid of the caisson were significantly greater in this test than the corresponding Test 2 on Redhill 110 sand. This manifested itself most clearly during the ±40kN load packet where up to half of the applied peak cyclic load was resisted by pore water pressure. In Test 2 about 15% of the applied load was resisted by the pore water pressure. 4.2 Test 15 Test 15 was similar to Test 14, except that cyclic loads were applied at a rate of 0.1Hz instead of 1.0Hz and the caisson was pulled out of the sand at a rate of 10mm/s instead of 5mm/s. The time history for this test is shown in Figure 5. Axial load and pore pressure beneath the lid of the caisson are plotted against axial displacement in Figure 6. Data from the cyclic stage of this test confirmed the findings from Test 14. The pore pressures measured beneath the lid of the caisson during cyclic loading were about 25% greater in Test 14 than Test 15 due to the rate of cycling. Similarly, larger pressures and loads were measured while

3

pulling the caisson out of the sand during Test 15 than during Test 14 because the rate of pullout was greater. 4.3 Test 16 The caisson was pushed into the sand until a load of 35kN was reached and then pulled out of the sand at a rate of 100mm/s in this test. The time history for this test is shown in Figure 7. Axial load and pore pressure beneath the lid of the caisson are plotted against axial displacement in Figure 8. The pressure beneath the lid of the caisson reached –95kPa during the pullout, which indicates that cavitation of the pore fluid may have occurred. However, the maximum load recorded during the pullout was less than that recorded during Test 15 where the magnitude of the excess pore pressure beneath the lid of the caisson was smaller during the pullout stage of the test. As the component of the axial load due to suction was greater in Test 16 than in Test 15 this result suggests that the skin friction was smaller in Test 16 than in Test 15. This suggests that either the rate of pullout was greater than the time required for the skin friction to be enhanced by the large hydraulic gradients formed adjacent to the skirt of the caisson during the pullout or that the strength of the sand had reduced due to disturbance arising from extra settlements of the caisson during this test. 4.4 Test 19 Test 19 was conducted in a similar manner to Test 14 except that the chamber was pressurised to 200kPa prior to cycling, the rate of cycling was 0.1Hz and the rate of pullout was 25mm/s. The time history for this test is shown in Figure 9. Axial load and pore pressure beneath the lid of the caisson are plotted against axial displacement in Figure 10. Not all of the load packets were applied as the maximum stroke of the actuators hydraulic ram was almost reached during the ±30kN set of cyclic loads. There were no significant differences between Test 19 and Test 15 during cyclic loading. Cavitation of the pore water pressure beneath the lid of the caisson occurred during pullout in Test 19 and was accompanied by an audible ‘pop’ as the seal created by the sand adjacent to the caisson was broken. The maximum load during pullout was about 30kN, which was about 3 times greater than the load recorded during tests conducted at atmospheric pressure. The excess pore pressure beneath the lid of the caisson was also 3 times greater during this test than in the previous tests. 4.5 Test 20 Test 20 repeated Test 14 with the exceptions that the sample was more dense and the rate of pullout was 25mm/s rather than 5mm/s. The time history for this test is shown in Figure 11. Axial load and pore pressure beneath the lid of the caisson are plotted against axial displacement in Figure 12. The data from inside the pressure chamber was recorded at too slow a rate to obtain good detail hence the load data presented in Figure 12 was taken from the external load cell. There was no significant differences between Tests 14 and 20 during cyclic loading. As per the previous tests, the maximum load during the pullout stage of the test was related to the rate of pullout and was larger in this test than in Test 14. 4.6 Test 21 Test 21 repeated Test 14 using a rate of cycling of 10Hz and a rate of pullout of 25mm/s. The time history for this test is shown in Figure 13. Axial load and pore pressure beneath the lid of the caisson are plotted against axial displacement in Figure 14. The pore pressure readings during the ±40kN cyclic load packet became greater than the capacity of the pressure sensor (350kPa) and went off scale. The pressure sensor was subsequently found to be damaged.

4

The total displacements at the end of the cyclic loading stage were less in this test than in the previous tests while the excess pore water pressures beneath the lid of the caisson were greater than in any of the other tests. 4.7 Test 22 Test 22 was a repeat of Test 20 with a pressure of 200kPa applied to the chamber prior to cyclic loading. The time history for this test is shown in Figure 15. Axial load and pore pressure beneath the lid of the caisson are plotted against axial displacement in Figure 16. The 200kPa pressure was applied to the chamber at about 10000s in Figure 15. The pressure sensor in the lid of the caisson initially recorded an increase in pressure and subsequently showed the pressure to decrease even though the pressure was constant in the chamber. As discussed in the previous section, this indicated that the sensor was faulty and the data from this test should be discarded even though the pressure data recorded during cyclic load shows similar patterns to the other data. The total displacement during cycling was less in this test than in Test 20. This is most likely due to the greater density of the sand in this test and is not caused by the elevated chamber pressure. The maximum load recorded during pullout in this test was similar to those recorded in the other tests where the chamber pressure was set to 200kPa. 4.8 Test 23 Test 23 consisted of pushing the caisson into the sand until a load of 35kN was reached, applying an ambient pressure of 200kPa to the chamber and then pulling the caisson out of the sand at a rate of 25mm/s. The time history for this test is shown in Figure 17. Axial load and pore pressure beneath the lid of the caisson are plotted against axial displacement in Figure 18. The pressure sensor in the lid of the caisson was replaced for this test. Data from Test 23 can be compared with that from Test 16. The pore water pressures recorded during both tests indicated that the pore fluid had cavitated in both tests. The load recorded during Test 23 was greater than that in Test 16 because the change in pressure was greater. 4.9 Test 24 Test 24 consisted of installing the caisson until a load of 35kN was reached, then applying 500 load cycles of ±20kN amplitude and then pulling the caisson out of the sand at a rate of 25mm/s. This test was designed to investigate whether pore pressure would accumulate beneath the lid of the caisson over time. The time history for this test is shown in Figure 17. Axial load and pore pressure beneath the lid of the caisson are plotted against axial displacement in Figure 18. The data presented in Figure 18 show that the intended cyclic loads were applied during the initial stages of the test and then the mean load increased to about 50kN. The mean load increased due to friction between the gland seal in the lid of the caisson and the shaft loading the caisson. As the mean load increased the pore pressures generated in each cycle beneath the lid of the caisson decreased. 4.10 Test 25 Test 25 was designed to compare tests conducted at different mean loads with those conducted with a mean load of 35kN. In this test cyclic loads were applied about mean loads of 5kN, 15kN and 50kN. These mean loads were chosen to be greater than the previous maximum load applied to the sample. The mean loads and cyclic load packets applied during this test are described in Table 3. The caisson was pulled out of the sand at a rate of 25mm/s. The time history for this test is shown in Figure 17. Axial load and pore pressure beneath the lid of the caisson are plotted against axial displacement in Figure 18.

5

The trend of the data from this test is entirely consistent will all of the data from tests conducted at a mean load of 35kN. The displacement during each cycle decreased during each load packet and the amplitudes of the cycles increased with increasing cyclic load. Load cycles where the total load approached zero displaced further than load cycles where the total load did not approach zero. 5.0 CONCLUDING REMARKS Tests constituting Phase 2 of the test programme have been completed to investigate the behaviour of a model caisson subject to partially drained to undrained conditions. Data from a series of 10 tests have been presented. The data constitutes time histories and load/pressure plotted against displacement for each of the tests. In general the data showed that:

• The tensile capacity of the model caisson was small; • The axial stiffness of the foundation reduced significantly as the caisson was unloaded to

zero load; • Elevated ambient pressure did not increase the tensile capacity or foundation stiffness during

cyclic loading; • The rate of loading increased the tensile capacity as the caisson was pulled out of the

ground; • Elevated ambient pressure increased the tensile capacity of the caisson as it was pulled out

of the ground; and • The ultimate tensile load was mobilised after a displacement of about 10% of the caissons

diameter as it was pulled out of the ground. These conclusions are similar to the conclusions in the Phase 1 test report. Although the strength of the HPF5 soil was greater than the Redhill 110 sand and the permeability about 100 times less than the Redhill 110 the overall behaviour of the caisson subject to axial loads was similar. The data will be interpreted in more detail in a later report. 6.0 REFERENCES

Kelly, R.B., Byrne, B.W., Houlsby, G.T. and Martin, C.M. (2003), Vertical loading tests in a pressurised chamber: Phase 1 Experimental Data, Report FOT016/1 for DTI project “The application of suction caisson foundations to offshore wind turbines”.

6

7.0 TABLES

7

Test Date Dry UnitWeight

Relative Density

Pressure CyclicFrequency

Pullout Rate

Comments

(kN/m3) (kPa) Hz (mm/s)14 25/11/2003 15.1 0.53 0 1.0 5 Repeat test 2 using HPF5 sand 15 26/11/2003 15.2 0.55 0 0.1 10 Repeat test 14 with slower cycling and faster pullout 16 28/11/2003 15.8 0.67 0 - 100 Push / Pull 19 16/03/2004 15.8 0.68 200 0.1 25 Pressurised Multi amplitude cyclic test 20 24/03/2004 15.8 0.67 0 1.0 25 Multi amplitude cyclic test 21 26/03/2004 16.1 0.73 0 10.0 25 Multi amplitude cyclic test 22 29/03/2004 16.1 0.72 200 1.0 25 Repeat test 14 under pressure 23 31/03/2004 16.1 0.72 200 - 25 Pressurised push-pull 24 01/04/2004 16.0 0.72 0 10.0 25 500 cycle test 25 13/04/2004 16.0 0.72 0 0.1 25 Multi mean load multi amplitude test

Table 1 Summary of Phase 2 Tests Conducted in the Pressure Vessel

8

Cycle Packet Number of Cycles Amplitude 1 10 ±5kN 2 10 ±10kN 3 10 ±20kN 4 10 ±30kN 5 5 ±35kN 6 5 ±40kN

Table 2 – Cyclic loads applied to the caisson during Test 14

Cyclic Load Packet

No: of Cycles Frequency (Hz) Mean Load (kN)

Cyclic Load Amplitude (kN)

1 2 3 4 5 6 7 8

10 10 10 10 10 10 10 10

0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1

5 5

15 15 15 50 50 50

+/-2.5 +/-5 +/-5 +/-10 +/-15 +/-10 +/-20 +/-30

Table 3 – Cyclic frequencies and load amplitudes applied in Test 25

9

8.0 FIGURES

0

10

20

30

40

50

60

70

80

90

100

0.001 0.01 0.1 1

Particle Size (mm)

Perc

ent P

assi

ng

Oakamoor HPF5Redhill 110

Figure 1 Particle size diagram for HPF5 and Redhill 110

0

10

20

30

40

50

60

0 0.02 0.04 0.06 0.08 0.1 0.12

Axial Strain

Fric

tion

Ang

le (d

egre

es)

HPF5Redhill 110

Figure 2 Triaxial test data for HPF5 and Redhill 110

10

-20

0

20

40

60

80

100

0 2000 4000 6000 8000 10000 12000 14000

Time (s)

Load

(kN

)

-100

0

100

200

300

400

500

Pres

sure

(kPa

)

Axial LoadLid PPT

Figure 3 Time history for Test 14

-20

0

20

40

60

80

100

180 190 200 210 220 230 240 250

Displacement (mm)

Load

(kN

)

-100

0

100

200

300

400

500

Pres

sure

(kPa

)

Axial LoadLid PPT

Figure 4 Axial load and pressure beneath the caisson lid vs displacement

11

-20

0

20

40

60

80

100

0 2000 4000 6000 8000 10000 12000 14000

Time (s)

Load

(kN

)

-100

0

100

200

300

400

500

Pres

sure

(kPa

)

Axial LoadLid PPT

Figure 5 Time history for Test 15

-20

0

20

40

60

80

100

180 190 200 210 220 230 240 250 260 270

Displacement (mm)

Axi

al L

oad

(kN

)

-100

0

100

200

300

400

500

Pres

sure

(kPa

)

Axial LoadLid PPT

Figure 6 Axial load and pressure beneath the caisson lid vs displacement

12

-20

-10

0

10

20

30

40

0 2000 4000 6000 8000 10000 12000 14000

Time (s)

Load

(kN

)

-100

-50

0

50

100

150

200

Pres

sure

(kPa

)

Axial LoadLid PPT

Figure 7 Time history for Test 16

-20

-10

0

10

20

30

40

180 185 190 195 200 205 210 215 220 225

Displacement (mm)

Load

(kN

)

-100

-50

0

50

100

150

200

Pres

sure

(kPa

)

Axial LoadLid PPT

Figure 8 Axial load and pressure beneath the caisson lid vs displacement

13

-40

-20

0

20

40

60

80

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Time (s)

Load

(kN

)

-200

-100

0

100

200

300

400

Pres

sure

(kPa

)

Axial LoadLid PPT

Figure 9 Time history for Test 19

-40

-20

0

20

40

60

80

180 190 200 210 220 230 240

Displacement (mm)

Load

(kN

)

-200

-100

0

100

200

300

400

Pres

sure

(kPa

)

Axial LoadLid PPT

Figure 10 Axial load and pressure beneath the caisson lid vs displacement

14

-20

-10

0

10

20

30

40

50

60

70

80

1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Time (s)

Load

(kN

)

-100

-50

0

50

100

150

200

250

300

350

400

Pres

sure

(kPa

)

Axial LoadLid PPT

Figure 11 Time history for Test 20

-20

-10

0

10

20

30

40

50

60

70

80

180 190 200 210 220 230 240 250

Displacement (mm)

Load

(kN

)

-100

-50

0

50

100

150

200

250

300

350

400Axial LoadLid PPT

Figure 12 Axial load and pressure beneath the caisson lid vs displacement

15

-20

-10

0

10

20

30

40

50

60

70

80

0 2000 4000 6000 8000 10000 12000

Time (s)

Load

(kN

)

-100

-50

0

50

100

150

200

250

300

350

400

Pres

sure

(kPa

)

Axial LoadLid PPT

Figure 13 Time history for Test 21

-20

-10

0

10

20

30

40

50

60

70

80

180 185 190 195 200 205 210 215 220 225 230

Displacement (mm)

Load

(kN

)

-100

-50

0

50

100

150

200

250

300

350

400

Pres

sure

(kPa

)

Axial LoadLid PPT

Figure 14 Axial load and pressure beneath the caisson lid vs displacement

16

-40

-20

0

20

40

60

80

1000 3000 5000 7000 9000 11000 13000

Time (s)

Load

(kN

)

-200

-100

0

100

200

300

400

Pres

sure

(kPa

)

Axial LoadLid PPT

Figure 15 Time history for Test 22

-40

-20

0

20

40

60

80

180 190 200 210 220 230 240 250

Displacement (mm)

Load

(kN

)

-200

-100

0

100

200

300

400

Pres

sure

(kPa

)

Axial LoadLid PPT

Figure 16 Axial load and pressure beneath the caisson lid vs displacement

17

-40

-20

0

20

40

60

80

1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Time (s)

Load

(kN

)

-200

-100

0

100

200

300

400

Pres

sure

(kPa

)

Axial LoadLid PPT

Figure 17 Time history for Test 23

-40

-30

-20

-10

0

10

20

30

40

180 185 190 195 200 205 210 215 220

Displacement (mm)

Load

(kN

)

-200

-150

-100

-50

0

50

100

150

200

Pres

sure

(kPa

)

Axial LoadLid PPT

Figure 18 Axial load and pressure beneath the caisson lid vs displacement

18

-20

0

20

40

60

80

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Time (s)

Load

(kN

)

-100

-50

0

50

100

150

200

250

300

350

400

Pres

sure

(kPa

)

Axial LoadLid PPT

Figure 19 Time history for Test 24

-20

-10

0

10

20

30

40

50

60

70

80

180 190 200 210 220 230 240 250

Displacement (mm)

Load

(kN

)

-100

-50

0

50

100

150

200

250

300

350

400

Pres

sure

(kPa

)

Axial LoadLid PPT

Figure 20 Axial load and pressure beneath the caisson lid vs displacement

19

-20

0

20

40

60

80

100

0 2000 4000 6000 8000 10000 12000

Time (s)

Load

(kN

)

-100

0

100

200

300

400

500

Pres

sure

(kPa

)

Axial LoadLid PPT

Figure 21 Time history for Test 25

-20

0

20

40

60

80

100

180 190 200 210 220 230 240 250

Displacement (mm)

Load

(kN

)

-100

0

100

200

300

400

500

Pres

sure

(kPa

)

Axial LoadLid PPT

Figure 22 Axial load and pressure beneath the caisson lid vs displacement

20