2001Lateral Undrained Resistance of Suction Caisson Anchors

International Journal of Offshore and Polar EngineeringVol. 11, No. 3, September 2001 (ISSN 1053-5381)Copyright © by The International Society of Offshore and Polar Engineers

INTRODUCTION

Suction caissons are large, hollow, cylindrical foundation ele-ments for offshore facilities installed by pressure drawdown with-in the cylinder, referred to as suction, after partial penetration ofthe cylinder due to its dead weight. They can have significantadvantages over conventional piles and have been used as bothfixed foundations and anchors for a variety of offshore structures.Suction caissons avoid a number of drawbacks associated withconventional piles; however, there is a paucity of performancedata on suction caisson behavior and our current level of under-standing is limited. This paper addresses one aspect of the suctioncaisson performance: the capacity of the anchor when subjectedto short-term lateral loads.

This paper presents the development of an algorithm for esti-mating lateral capacity using a simplified lateral resistance factorproposed by Murff and Hamilton (1993). The simplified solutionsare validated through comparisons with finite element solutions.Finally we present a series of sensitivity studies that systematicallydetail the effects of soil strength, caisson aspect (depth-to-diame-ter) ratio, soil-caisson adhesion, soil weight, and suction. Whilethis study was performed primarily with a view toward improvingour understanding of laterally loaded suction caisson anchors, theresults may also be applicable to other rigid, laterally loaded foun-dation elements such as caisson foundations and short piles.

BACKGROUND

The ultimate average unit lateral pressure P on an underwatercaisson can be expressed as (Matlock, 1970; Reese et al., 1975;Murff and Hamilton, 1993):

where Np is a dimensionless bearing factor, su is the local soilundrained shear strength, and s ¢v0 is overburden pressure.

For a deeply embedded, laterally loaded circular caisson in per-fectly plastic material, the problem can be approximated as a 2-dimensional one with soil flowing horizontally around the cais-son. For the cases of perfectly smooth and rough caissons, Ran-dolph and Houlsby (1984) present solutions from plasticity theoryto give Np = 9.14 and Np = 11.94, respectively. These flow-aroundsolutions correspond to conditions of no gap formation behind thecaisson (Eq. 1a).

The influence of a free surface on ultimate lateral pressure istreated by Murff and Hamilton (1993), who propose an upperbound analysis based on a collapse mechanism comprised of a 3-dimensional wedge mechanism near the free surface and a flow-around failure (Randolph and Houlsby, 1984) at depth. By opti-mizing 4 parameters characterizing the kinematics of the failuremechanism, they computed a minimum upper bound load capaci-ty F. By successively increasing the length of the caisson by anincremental length DL and computing the increase in lateralcapacity DF for pure translation of the caisson, they computedbearing capacity factors from the relationship:

where su is the local undrained shear strength, and D is the caissondiameter.

Using this approach, Murff and Hamilton (1993) conductedseveral parametric studies. One study comparing predicted lateralresistance profiles for translating and rotating caissons showedvirtually identical results: Lateral resistance appears to be rela-tively independent of caisson rotation. This is consistent withfindings by Matlock (1970). Murff and Hamilton compared theirNp values to previous semi-empirical models (Matlock, 1970;Reese et al., 1975) and to centrifuge test data (Hamilton et al.,1991) and found favorable agreement.

By conducting a series of analyses for the special case of a lin-

(2)NF

s D Lpu

=D

D

(1b)P N sp u v= ¢+ (gap)s 0

(1a)P N sp u= (no gap or a weightless soil)

Lateral Undrained Resistance of Suction Caisson Anchors

C.P. Aubeny and J.D. MurffTexas A&M University, College Station, Texas, USA

S.K. MoonCalifornia Department of Transportation, Sacramento, California, USA

ABSTRACT

A simplified method of analysis for estimating lateral load capacity of suction caisson anchors based on an upper boundplasticity formulation is presented. The simplification restricts the analysis to caissons in uniform and linearly varyingundrained strength profiles; nevertheless, its computational efficiency permits quick evaluation of a number of parametersaffecting load capacity. The validity and limitations of the simplified formulation are demonstrated through comparisonsto more rigorous finite element solutions. A series of sensitivity studies evaluates the effects of various soil conditions andloading parameters.

Received November 20, 2000; revised manuscript received by the editorsMarch 26, 2001. The original version was submitted directly to theJournal.

KEY WORDS: Suction caissons, anchors, lateral loading, plasticity, finiteelement analysis.

212

ear variation of strength with depth, they obtained a simplifiedempirical (albeit restricted) relation for Np of the form:

where N1 is the limiting value of lateral resistance at depth,(N1 – N2) is the lateral resistance at the free surface, and x is acurve fitting factor depending on the characteristics of the soilstrength profile.

Based on the Randolph-Houlsby analysis, the appropri-ate values for N1 in Eq. 3 can be taken as 9 and 12 for theconditions of smooth and rough caissons, respectively. Atthe free surface, Np is (from Rankine theory) 2 for the caseof a gap forming behind the caisson. N2 is computed sim-ply as the difference between Np at the ground surface andNp at depth; for example, for the case of a smooth caissonwith a gap forming behind the surface failure wedge, N2 =9 – 2 = 7. By least squares curve fitting the results of thenonlinear optimization procedure described above, Murffand Hamilton (1993) define the function x as:

where l = su0 / su1D, su0 is the soil strength at the free surface, andsu1 is the rate of strength increase with depth.

SIMPLIFIED UPPER BOUND ANALYSIS

Using the simplified expression for the Np from Eq. 3, a simpler(while more approximate) upper bound formulation is possible,which reduces the number of optimization variables from 4 to 1(the depth to the center of rotation, L 0) and avoids the complexintegrations required for evaluating the detailed mechanism. Theprocedure is restricted to the special case of linearly increasingshear strength with depth, such as occurs in normally consolidated(e.g., near surface sediments in the Gulf of Mexico) or lightlyoverconsolidated clay deposits; that is:

It is recognized that an analogous lower bound or equilibriumapproach could be used as well.

Side ResistanceIf a caisson undergoes a virtual angular velocity b·

about a cen-ter of rotation O in Fig. 1, the velocity varies linearly with depthv = (1 – z / L 0)v0, where v0 is the mudline velocity. The internalrate of energy dissipation E

·at any point along the caisson is sim-

ply the product of the mobilized pressure (Eq. 1) times the veloc-ity at the point in question times the elemental area. For the caseof a weightless soil (or the case of full adhesion behind the pile)this becomes:

For the case of a weightless soil (or the case of full adhesion)with a linearly varying strength profile, Eqs. 1a, 3 and 5 can besubstituted into Eq. 6, and the resulting expression integratedalong the length of the pile to obtain the total rate of energy dissi-pation due to side resistance:

Eq. 7 can apply to both rotational and translational rigid bodymotions. In the translational case, the center of rotation occurs atinfinite depth or an infinite distance above the top of the caisson;hence, L 0 tends to infinity and the term within the absolute valuesign in Eq. 7 becomes unity.

The effect of soil weight can be accounted for simply by usingEq. 1b rather than Eq. 1a in the above formulation to give the fol-lowing expression for total rate of energy dissipation:

The lateral pressure at depth is defined by the plane straincondition that is unaffected by soil weight; hence, Eq. 8 gov-erns only while:

It should be noted that an alternative equivalent formulation ispossible by considering the soil unit weight effects as externalwork, which they actually are; in this case they are negative due tothe upward velocity of a downward force, resulting in an expres-sion identical to Eq. 8.

End ResistanceThe resistance at the tip of the cylinder may also be significant,

especially for short stubby caissons. A spherical segment on thecaisson tip rotating about the center of rotation (Fig. 1) is anadmissible mechanism. The local dissipation is simply the relativeslip velocity times the local shear strength times the elementalarea. Murff and Hamilton (1993) give the rate of energy dissipa-tion at the end of the cylinder, D

·e , as:

(9)N s z N sp u p u+ £g 'max

(8)˙ 'D v Dz

LN s z dzs o

o

L

p u

f

= - +( )Ú 10

g

(7)˙ expD v D N Nz

Ds s z

z

Ldzs o u u

o

Lf

= - -ÊËÁ

ˆ¯

Ê

ËÁ

ˆ

¯˜ +( ) -

È

ÎÍÍ

˘

˚˙˙Ú 1 2 0 1

0

1x

(6)dE N sz

Lv Ddzp u

oo

˙ = -Ê

ËÁÁ

ˆ

¯˜1

(5)s s s zu u u= +0 1

(4b)x l= <0 55 6.

(4a)x l l= + <0 25 0 05 6. .

(3)N N N z Dp = - -1 2 exp( / )x


Fig. 1 Failure mechanism

213International Journal of Offshore and Polar Engineering

where R = caisson radius, R1 = Lf – Lo = distance from center ofrotation to caisson tip R1 = √——

R2 + R 21

— = radius of spherical surface,

f = angular coordinate about caisson centerline in horizontal plane(0 to 2p), w = angular coordinate from caisson centerline varyingfrom 0 to sin–1(R / R2),

For a translational failure mode, the dissipation rate is simplythe tip area times the local strength times the local velocity:

Total CapacityThe estimate of the collapse load is then found by equating the

external work rate to the internal dissipation rate. Canceling thevirtual velocities we have:

or:

F is then an upper bound to the exact capacity. The best estimateis found by minimizing the right hand side with respect to L 0.

Uniform Weightless SoilIn most cases the load capacity F in Eq. 13 can be evaluated

through numerical integration to evaluate Eqs. 7 (or 8) and 10,respectively. Further, a minimum upper bound of F can beobtained by optimizing with respect to a single parameter, L 0,using numerical procedures. However, a simple analytic solutionof Eq. 13 is possible for the important case of a uniform side resis-tance (constant Np su) and no end resistance (D

·e = 0). The solution

for this special case is presented in the Appendix. For top loading(Li = 0) L 0 reduces to:

The load capacity F for top loading corresponding to this center ofrotation is then:

Noting that the load capacity for a translating caisson is Ft =Npsu DLf , the ultimate load is then 41.4% of the capacity in trans-lation. Further, the critical rotation point is at 0.707Lf below themud line. This solution gives a good baseline from which tounderstand more realistic, hence more complex problems.

FINITE ELEMENT STUDIES

As the plastic limit formulation presented above includes anumber of simplifying assumptions, it is useful to compare it tomore rigorous procedures such as the finite element method(FEM.) To carry out these calculations the program ABAQUS(HKS, 1997) was used. For consistency with the plastic limit for-mulation, a similar constitutive behavior was assumed (von Misesyield criterion, perfect plasticity, normalized flow rule,) exceptthat linear elasticity was assumed for stress states beneath theyield surface to permit implementation of the FEM.

All of the FEM studies presented below consider the case offull adhesion (no gap) behind the caisson and a rough soil-caissoninterface (no relative movement at the boundary) on the sides ofthe caisson. The 3-D analyses were devised to predict load capaci-ty with and without considering tip (end) resistance.

Plane Strain AnalysisThe FEM analyses were first applied to the idealized plane

strain case of an infinitely long translating cylinder. Theexact solution presented by Randolph and Houlsby (1984)for this condition provides a convenient standard for estab-lishing the necessary degree of mesh refinement, the mostsuitable element type, and the overall accuracy of the FEMprocedure. FEM simulations using successively finer meshes

(15)F N s DLp u f= -( )2 1

(14)LL

o

f=2

(13)FD D

L Ls e

i o

=+

-1 /

(12)FL

LD Di

os e1- = +

(11)˙ ( )D v s s LD

et o u u f= +0 1

2

4

p

L

L

R

R

L

Do

f

f= < <0 0 21 to 1,

R

R

L L

D

L

D

L

L

R

R R R

f o f o

f

1

21

222 1

1

1=

-= -

Ê

ËÁÁ

ˆ

¯˜ =

( ) +/,

/

(10)

˙ sin

sin sin cos sin

/ /sin

DR v

Ls L R s

d d

eo

o

R R

u o u= + +( ){ }( ) + ( )

( ) +ÊËÁ

ˆ¯

==

-

ÚÚ23 1 1

0 2 100

2

2 2

121

w

w f w w w f

wf

p

Fig. 2 Comparison between plastic limit analysis and 3D finiteelement analysis for translating caisson

214

showed that the solution converged to Np = 12.56, overpre-dicting the exact Randolph-Houlsby solution by about 5%.Convergence occurred when the element dimensions (b)adjacent to the caisson boundary were smaller than aboutone-tenth the caisson radius, b / R < 0.1. Several element for-mulations were tested, including hybrid and reduced integra-tion schemes, to model the incompressibility conditions.From these studies it was concluded that a linear-displace-ment, full-integration formulation with a Poisson’s ratio m =0.49 provided satisfactory results.

3-D ANALYSES

3-D analyses were performed to evaluate the effects of aspectratio L f / D and load application point L i on lateral capacity F

and the center of rotation L 0 of the caisson. The effect of endresistance was also evaluated. To facilitate comparisons betweencaissons of different aspect ratios, lateral load capacities werenormalized by the soil strength su and the length and diameter ofthe caisson; the result is then equivalent to an average bearingfactor Npavg:

Based on the preliminary 2-D studies described above, a lin-ear-displacement, full integration element was used for thestudy: an 8-noded brick element with 8 integration points. Asthe 3-D analysis involves a substantially larger stiffness matrix,a somewhat coarser mesh, with element dimensions b / R = 0.2at the caisson boundary, was used to economize on storagerequirements. Analysis of the plane strain condition conductedwith this coarser mesh yielded a lateral resistance factor Np

=13.19, or about 10% greater than the exact Randolph-Houlsbysolution of Np =11.94.

Lateral loading of the caisson was modeled using 2 approaches:prescribed displacements and prescribed external forces. Dis-placement control permits evaluation of the hypothetical maxi-mum load capacity Fmax corresponding to conditions of pure trans-lation and is generally more effective in estimating the resistanceF as collapse conditions are approached. Load control was used toeffectively evaluate the optimum center of rotation L 0 under vari-ous loading conditions. In both cases the caisson was modeled asa rigid body with respect to bending, shear and tension.

(16)NF

s DLpavgu f

=


Fig. 3 Comparison between plastic limit analysis and 3D finiteelement analysis for rotating caisson, tip resistance included inanalysis

Table 1 Center of rotation from plastic limit analysis (PLA) andfinite element method (FEM) for rotating caisson considering endresistance for a rough pile with full adhesion on back of caisson

Li / Lf

0

0.25

0.5

0.75

1.0

L0 /Lf – PLA

0.71

0.81

0.99

0.20

0.31

L0 /Lf – FEM

0.75

0.83

1.05

0.20

0.30

L0 /Lf – FEM

0.74

0.84

1.04

0.075

0.19

L0 /Lf – PLA

0.72

0.81

0.98

0.21

0.31

L f / D = 6 L f / D = 10

Fig. 4 Maximum (translational) resistance of caisson


Translational Load CapacityAn initial study compared the average resistance factors Npavg

computed from FEM analyses and plastic limit analyses for thecase of a translating rough pile with full adhesion on the back sideof the pile. The results in Fig. 2 show that the finite element pre-dictions generally exceed plastic limit solutions by about 10%.This result was anticipated from the preliminary studies and thedifference can be attributed in part to the coarseness of the 3-Dfinite element mesh. For small aspect ratios (Lf / D = 1), larger dif-ferences (about 15%) occur between the FEM and plastic limitsolutions. The larger discrepancies at small caisson aspect ratiosare believed to be due to: (1) stress concentrations near the tip thatcan lead to inaccuracies in the FEM analyses, (2) the larger rela-tive influence of tip resistance at small aspect ratios, and (3) theinteraction of tip and side resistances. Overall trends predictedfrom plastic limit analyses with regard to lateral resistance Npavg asa function of caisson aspect ratio Lf / D are supported by the FEManalyses.

Rotational Load CapacityThe conditions studied for the rotating caisson are the same as

for the translating caisson described above. Two load applicationlocations were investigated: at the top (Li / Lf = 0) and at the mid-point (L i / L f = 0.5) of the caisson. Lateral load capacity wastaken as the load corresponding to a lateral deflection at the top ofthe caisson d = D / 10. Due to the small lateral deformations in thevicinity of the center of rotation, this load is somewhat smallerthan a true ultimate load corresponding to a rigid-plastic material.However, finite element simulations carried out to large lateraldeflections suggested that this procedure underestimates the ulti-mate load by no more than about 5%.

Fig. 3 compares Npavg from plastic limit and finite element anal-yses. The FEM solutions do not consistently overpredict the plas-tic limit solutions for the rotational case as they did for the transla-tional case. Rather, the FEM predictions are about 5% less thanplastic limit solutions at the lower caisson aspect ratios Li / D, andexceed the plastic limit solutions by about 5% at higher aspect

Fig. 5 Effect of soil strength profile on lateral load capacity, smooth soil-caisson interface

216

ratios. The improved agreement between solutions is due to someextent to fortuitous compensating errors. On the one hand FEMsolutions tend to overpredict the actual collapse loads. On theother hand, in FEM solutions for an elasto-plastic material, rota-tional resistance cannot be fully mobilized near the center of rota-tion, as strain levels approach zero in this region; this leads to acompensating underprediction of collapse loads.

Optimum Center of RotationAs discussed earlier, the simplified plastic limit formulation

minimizes the lateral load capacity F through a single optimiza-tion variable L 0, the center of rotation. A series of load-con-trolled FEM analyses was performed to compare the optimalcenter of rotation using the 2 methods. The test case was a roughpile in a uniform strength soil profile with full suction behindthe caisson. Tip resistance was considered in both series of anal-yses. Caisson aspect ratios L f / D = 6 and 10 were analyzed.Table 1 shows that for long caissons, L f / D = 10, plastic limitsolutions compare very favorably to FEM solutions for all loadattachment points L i / L f. For the shorter caissons, L f / D = 6, theplastic limit and FEM solutions compare equally favorably forload attachment points in the upper half of the pile, L i / L f < 0.5.However, for deeper load attachment points, in particular whenthe direction of rotation is reversed, larger discrepancies occurbetween the solutions. This behavior may be because: (1) tipresistance is more significant for shorter piles, and (2) at highcenters of rotation (reverse rotation) the FEM analysis most like-ly overestimates the relative contribution of tip resistance, asside resistance is not fully mobilized in the upper portion of thepile. The relationship between load attachment point L i and opti-mum center of rotation L 0 will be discussed subsequently ingreater detail (Fig. 11b).

SENSITIVITY STUDIES

Using the simplified plastic limit formulation presented above,a series of studies was carried out to investigate conditions of gen-eral interest, including the effects of soil strength, soil unit weight,tip resistance, suction and soil-caisson adhesion. In each case, theunknown lateral force, F (obtained for the general rotationalmechanism), was calculated and normalized by the maximumforce, Fmax, corresponding to pure translation:

Fmax for various conditions of interest is plotted in Fig. 4.While these solutions are simplified approximations, they can

provide valuable insights into the effects of various loading condi-tions and site variables, and can be of considerable use in concep-tual and preliminary designs.

Soil StrengthFor the linearly increasing strength case, the soil strength can

be characterized by the dimensionless parameter, su0 / su1D. In thefirst study, the soil was taken as weightless, with no tip resistanceand no suction on the back of the caisson. Results of this study are

(tip resistance) (17b)

F F

D N z s z dz s L D

Trans

p u

L

u f

f

max

( ) ( ) ( ) /

=

= +Ú 2 4p

(no tip resistance) (17a)F F D N z s z dzTrans p u

Lf

max ( ) ( )= = Ú0


Fig. 6 Effect of soil weight, no adhesion at soil-caissoninterface

Fig. 7 Effects of tip resistance, Lf /D = 2 and L f / D = 4 smoothcaissons


shown plotted in Fig. 5 for 6 different strength profile ratios –including the case of a uniform strength profile. Four differentLf / D ratios (1, 2, 4 and 10) and a range of load attachment points(0 < Li / L f < 1) are considered. These plots, particularly the com-bined plot for various L f / D ratios in Fig. 5e, are interesting inthat they show that normalized lateral capacity F / Fmax for a givenstrength ratio is practically independent of Lf / D. Still, it must bepointed out that the magnitude of Fmax (and thus F ) is clearlydependent on Lf / D, as depicted in Fig. 4.

Soil Unit WeightThis study includes the effect of lifting the soil weight in the

failure wedge in front of the caisson. The unit weight is character-ized by the dimensionless parameter, g ¢D / su 0. The soil wasassumed to have no tip resistance and no suction on the back of the

caisson. Two soil strength conditions (uniform strength, su 1D / su 0 =0, and zero strength intercept, su 0 / su iD = 0) were considered. Forthe soil unit weight ratio, typical unit weight was compared withthe weightless case for the study. Results of this series of analysesare presented in Fig. 6. These predictions indicate that soil weightcan have a moderate influence (10% to 15% increases) on F / Fmax,the relative effects being larger for weaker soil.

Tip ResistanceIn this series, the effect of rotational tip resistance was

investigated. The soil was taken as weightless with no suctionon the back of the caisson. As the effect was expected to bemost significant for short caissons, the study focused on lowaspect ratios, L f / D of 2 – 4. Two soil strength profiles(su 1D / su 0 = 0 and su 0 / su iD = 0) were considered. The resultsplotted in Fig. 7 confirmed that tip resistance has little effecton F / Fmax, but again Fmax is strongly dependent on the inclu-sion of tip resistance.

Suction on Back of PileIn short-term, monotonic loading, suction may develop on

the back side of the caisson. When this occurs an active wedgewill develop behind the caisson similar to the passive one infront of the caisson. This essentially doubles the unit resistancealong the caisson due to soil failure within the wedge andnegates any effect of soil weight. That is, the resistance factoris doubled at any point near the soil surface, although totalresistance cannot be larger than the plane strain value. Theextreme soil strength profiles, su 1D / su 0 = 0 and su 0 / su iD = 0,were considered here. Results of this study are shown plottedin Fig. 8. The effect of soil suction on the normalized lateralload capacity F / Fmax is most pronounced for the uniform soilstrength case.

Caisson-Soil AdhesionAnother important consideration is the effect of caisson-soil

adhesion, both in front of and behind the caisson. Here, we con-sider both conditions of no adhesion (smooth caisson) and fulladhesion (rough caisson) with 2 soil strength profiles (su 1D / su 0

= 0 and su 0 / su iD = 0) being considered. Fig. 9 shows a com-parison of normalized lateral force along load attachment pointsfor full adhesion and for no adhesion. Note that the normalizedload F / Fmax capacities are virtually independent of the interfacecondition. However, it must be again emphasized that whileF / Fmax is essentially unaffected by the interface condition, theabsolute magnitudes of Fmax (hence of F ) are impacted asdepicted in Fig. 4.

Combined EffectsThe parameter study here includes considerations of all of

the combined effects mentioned above: tip resistance, suctionon the back of the caisson, caisson-soil adhesion and soil unitweight. Results of this study are shown plotted in Fig. 10 for6 different strength profiles. The normalized force has a simi-lar trend as in the studies where the effects were examinedindependently.

Anchor Attachment PointA key design issue is the attachment depth of the anchor

line. To address this issue, the effect of load attachment pointLi / Lf was investigated for a caisson with aspect ratio L f / D =4, in a weightless soil, assuming no tip resistance and no suc-tion behind the caisson. The solution (Fig. 11a) shows F / Fmax

increasing with increasing Li / Lf . In this case the optimal load

Fig. 8 Effects of suction on back side of caisson, smooth soil-caisson interface

Fig. 9 Effect of soil-caisson adhesion (Lf /D = 4)

218

attachment point, corresponding to a translational failure withF / Fmax = 1, occurs at L i / L f = 0.7. As L i / L f is increasedbeyond the optimum point, a reverse rotation of the caissonoccurs with a concomitant decrease in lateral load capacityF / Fmax . Fig. 11b shows the rotation modes with regard toload attachment points. The depth of the center of rotation,L 0 / L f , increases with increasing attachment point depthLi / Lf until a translational mode is approached. At this criticalpoint, L 0 / Lf jumps to infinite depth. Beyond this point thecenter of rotation L 0 / Lf jumps to the top of the caisson andthen progresses downward.

CONCLUSIONS

A simplified method is presented for upper bound plastic limitanalysis of laterally loaded caissons or short piles that is particu-larly useful for preliminary studies or conceptual design. Themain restrictions on the analysis are: (1) no yielding within thecaisson and (2) a uniform or linearly varying strength profile. Inthe analysis a minimum lateral load capacity F is found by opti-mizing a single parameter L 0. The simplified analyses comparefavorably to more rigorous FEM solutions.

A series of parametric studies using the simplified plasticlimit analysis showed a number of interesting trends withregard to the effects of various caisson geometry and soilconditions on the ratio of rotational to translational lateralcapacity F / Fmax.

• The depth of the load attachment point Li is among the mostsignificant variables—in the example problem analyzed, loadcapacity increased by a factor of nearly 5 when the attachment pointwas moved from the mudline to the point of maximum resistance.

• The normalized capacity F / Fmax is practically independent ofboth caisson aspect ratio L f / D and soil-caisson interface adhe-sion, although both factors will obviously affect the magnitudes ofFmax and F.

• Tip resistance effects on the normalized capacity F / Fmax arerelatively small.

• Soil weight can contribute moderately to normalized lateralresistance F / Fmax in cases of short caissons with no suction devel-oping on the back side of the pile.

• Suction on the back of the caisson can moderately increaseF / Fmax of short caissons (L f / D < about 4).

ACKNOWLEDGEMENTS

The authors would like to acknowledge the support of the Off-shore Technology Research Center and their colleagues at TexasA&M University and the University of Texas. We are particularlyindebted to Seung-Woon Han for his assistance with the finite ele-ment studies.

REFERENCES

Hamilton, JM, Phillips, R, Dunnavant, TW, and Murff, JD (1991).“Centrifuge Study of Laterally Loaded Pile Behavior in Clay,”Proc Int Conf Centrifuge 1991, ISSMFE, pp 285-292.

HKS. (1997) ABAQUS Version 5.7 User’s Manuals.Matlock, H (1970). “Correlations for Design of Laterally Loaded

Piles in Soft Clay,” Proc 2nd Offshore Tech Conf, Houston, pp577-594.

Murff, JD, and Hamilton, JM (1993). “P-Ultimate for UndrainedAnalysis of Laterally Loaded Piles,” ASCE J Geotech Eng, Vol119, No 1, pp 91-107.

Randolph, MF, and Houlsby, GT (1984). “The Limiting Pressureon a Circular Pile Loaded Laterally in Cohesive Soil,” Geotech-nique, London, Vol 34, No 4, pp 613-623.

Reese, LC, Cox, WR, and Koop, RD (1975). “Field Testing andAnalysis of Laterally Loaded Piles in Stiff Clay.” Proc 7th Off-shore Tech Conf, Houston, pp 473-483.

APPENDIX DERIVATIONS:

Consider the mechanism of the caisson failure as shown in Fig.1. The caisson rotates about point O such that it has a virtualvelocity, vo, at the top. The velocity at any point, a distance zbelow the pile top, is then:


Fig. 11 Optimal rotation depth, L0 / Lf = 4 for Lf /D = 4

Fig. 10 Rough caisson with suction, tip resistance and soilweight—Effect of soil strength


The unit resistance at any point is taken as Npsu and thus thetotal rate of energy dissipation is:

Assuming NpsuD = constant by integrating and simplifying, we get:

Equating the work of an external force (F) to internal dissipa-tion and canceling virtual velocities, we get:

Simplifying and normalizing this result by dividing by

NpsuDL f gives:

Minimizing F ¢ with respect to L 0, and solving for L0 gives:

(22)L L L L L L L Lo f i f i f i f/ / / / /= + ( ) - +2

1 2

(21)FF

N s DL

LL

L

L

LL

p u f

o

f

f

o

i

' = =- +

Ê

ËÁÁ

ˆ

¯˜

-

12

10

(20)FL

LN s D L L

L

Li

p u o ff

o

120

2

- = ( ) - +Ê

ËÁÁ

ˆ

¯˜

(19b)E v N s D L LL

Lo p u o ff

o

= ( ) - +Ê

ËÁÁ

ˆ

¯˜

2

2

(19a)Ez

Lv N s D dz

z

Lv N s D dz

oo p u

L

oo p u

L

Lo

o

f

= -Ê

ËÁÁ

ˆ

¯˜ ( )

È

ÎÍÍ

˘

˚˙˙ + -

Ê

ËÁÁ

ˆ

¯˜ ( )

È

ÎÍÍ

˘

˚˙˙Ú Ú1 1

0

(18)vz

Lv

oo= -

Ê

ËÁÁ

ˆ

¯˜1

The Proceedings of The Eleventh (2001) International

OFFSHORE AND POLAR ENGINEERINGCONFERENCE

and The Second (2001) International Deep-Ocean Technology Symposium

ISBN 1-880653-51–6 (Set): ISSN 1098-6189 (Set)Stavanger, Norway, June 17-22, 2001

VOLUME I ISBN 1-880653-52-4

OFFSHORE TECHNOLOGY REVIEWOFFSHORE TECHNOLOGY

SUBSEAOFFSHORE SYSTEMS AND DECOMMISSIONMOB AND VLFSFPSO, TLP AND SPAR

RESOURCES AND ENERGY TECHNOLOGYGAS HYDRATESOCEAN MININGRENEWABLE AND OCEAN ENERGY

POLAR AND ICE ENGINEERINGPOLAR DEVELOPMENTSICINGJOIA PROJECTSICE COVERICE FORCES

VOLUME II ISBN 1-880653-53–2

PLENARY PRESENTATIONSÅSGARD TRANSPORTPIPELINESRISERS, CABLES AND MOORINGAUV AND UNDERWATER CONTROLENVIRONMENT ENGINEERINGGEOTECHNICAL ENGINEERING

SOIL PROPERTIESOFFSHORE INVESTIGATIONSSOIL DYNAMICSSUCTION FOUNDATIONFOUNDATIONSRECLAIMED LANDS

VOLUME III ISBN 1-880653-54-0

PLENARY PRESENTATIONMETOCEANHYDRODYNAMICS

NUMERICAL WAVE TANKFORCESDYNAMIC RESPONSESVORTEX AND VORTEX-INDUCED VIBRATIONSCOUPLED MOORING-STRUCTURE DYNAMICS

COASTAL ENGINEERINGWAVE-STRUCTURE INTERACTIONSPORT AND HARBORWAVESPOROUS BREAKWATERWAVES AND SEABEDTIDE AND CURRENT

VOLUME IV ISBN 1-880653-55–9

PLENARY PRESENTATIONSTUBULAR STRUCTURESFATIGUE: A Joint Industry ProgramCOMPOSITES AND SMART STRUCTURESCOLLISION AND IMPACTMECHANICS AND ANALYSISEARTHQUAKE ENGINEERINGRELIABILITY, RISK AND SAFETYADVANCED SHIP AND OCEAN TECHNOLOGYOFFSHORE ENGINEERING EDUCATIONDEEPWATER RESEARCH STUDIES

THE SECOND (2001) INTERNATIONAL DEEP-OCEAN TECHNOLOGY SYMPOSIUM

Documents

2001Lateral Undrained Resistance of Suction Caisson Anchors