See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/255606763

The undrained inclined load capacity of shallow foundations after con-

solidation under vertical loads

Article · January 2002

DOI: 10.1201/9781439833797-c63

CITATIONS READS

26 303

1 author:

Fraser Bransby
University of Western Australia
169 PUBLICATIONS 4,736 CITATIONS

SEE PROFILE

All content following this page was uploaded by Fraser Bransby on 07 February 2020.

The user has requested enhancement of the downloaded file.

The undrained inclined load capacity of shallow foundations after con-
solidation under vertical loads
M.F. Bransby
Department of Civil Engineering, The University of Dundee, Scotland

ABSTRACT: The effect of pre-loading a surface foundation is examined using finite element analysis. A giv-
en vertical load is applied to the foundation before the resulting undrained failure locus is calculated. It is seen
that the application of vertical dead weight loading while allowing consolidation significantly increases the
capacity of a foundation under combined vertical and horizontal undrained loading. The shape of the resulting
failure loci are examined and it is noted that it is the horizontal capacity that is increased most significantly.
Finally, simple design methods are presented to predict the increase in the size of the failure locus from pre-
loading with differing vertical loads and consolidation times. Allowance of the resulting gradual strength gain
of a foundation with time after installation may allow less conservative foundation design.

1 INTRODUCTION and increases of foundation strength (work harden-

ing) during penetration (e.g. Tan 1990; Martin 1994;
Offshore foundations are frequently subject to com- Schotman 1989).
bined horizontal (H), moment (M) and vertical (V) This paper presents analysis of the undrained ca-
loading due to the self weight of an offshore installa- pacity of surface foundations after they have been
tion and the environmental loads (wind, wave, cur- subject to strength gains due to consolidation under
rent) that are applied. Hence, it is important to be dead weight vertical loading. Finite element analysis
able to characterise the combinations of loads that is used and this allows examination of changing soil
cause foundation failure. In addition to the transient and loading conditions as well as parametric studies.
(often undrained) loads caused by the environmental The behaviour of the footing has been characterised
in terms of a plastic yield envelope. The findings can
loading, there is usually a consistent vertical loading
be used to predict increases of foundation capacity
due to the self-weight of the offshore installation. due to pre-loading events and strength gain during
Over time, this vertical loading will cause the soil the life of a foundation system, and the design use of
beneath a shallow foundation to consolidate. This the findings is discussed.
will increase the capacity of the foundation during
the lifetime of the installation, but this is generally V
not considered in offshore design as it is conserva- H Ho H
tive to ignore this effect. This strengthening can be
done intentionally by the addition of surcharge H h
weights to an installation when it is known as pre- V Yield locus
loading. (M = 0)
Traditionally, predictions of vertical or inclined
load capacity was done using bearing capacity fac-
tors with empirical corrections for load inclination
(e.g. Meyerhof 1963; Hansen 1970). More recently, Vo
2
the plasticity framework has been used to calculate
foundation capacity during combined V-H-M load- V
ings. It has the advantage that full use of the method Figure 1. Definition of footing geometry, loading and the yield
also allows prediction of foundation displacements locus

Bransby: Inclined load capacity of shallow foundations after 1

consolidation under vertical loads
2 PLASTICITY CONCEPTS 3 FINITE ELEMENT METHODOLOGY

Recently, the analysis of offshore foundations such The aim of the finite element analyses was to exam-
as spudcan and shallow foundations have used plas- ine footing-soil interaction behaviour during and af-
ticity concepts to express the bearing capacity of ter pre-loading. Of particular interest was the size
foundations under combined vertical (V), horizontal and shape of the undrained V-H yield loci.
(H) and moment (M) loading (Schotman 1989;
Murff 1994; Tan 1990; Martin 1994). 3.1 Soil conditions and geometry
A plastic yield locus is introduced which express-
es the combination of V-H-M loads that result in A plane strain footing was investigated. This was
foundation failure. This can be given as a mathemat- chosen to simplify analysis, whilst still keeping the
ical expression, f where f(V,H,M) = 0 at yield, or can fundamental aspects of the mechanical behaviour
be shown graphically (Fig. 1). intact. One geometry was investigated: a rough, rig-
The size of the yield locus will vary with soil id surface foundation of width, D. Dimensions are
strength, footing dimensions/shape and foundation defined as shown in Figure 2.
pentration. Bransby & Randolph (1999) suggested The finite element package CRISP, CRItical State
that the yield locus should be calculated after finding Program (Britto & Gunn 1987) was used. In the
the footing capacity under pure vertical (Vo), hori- analyses, 472, 15-noded triangular elements with 27
zontal (Ho) and moment (Mo) load which are easiest degrees of freedom were employed (cubic strain tri-
to calculate. Pure vertical capacity, Vo will be given angles). Normally consolidated original Cam clay
by standard bearing capacity formulations, e.g. Vo = was chosen to describe the soil with  = 0.25,  =
A Nc suo, where A is the base area, suo is the un- 0.05,  = 1.0,  = 3; ' = 0.3, Ko = 0.69 and perme-
drained shear strength of the soil at the base of the ability, k = 1x10-8 m/s. These properties were chosen
foundation and Nc is a bearing capacity factor, which to investigate generic soil behaviour.
will increase with soil non-uniformity or increased A non-uniform and a uniform soil profile were ex-
embedment of the footing. Thus, penetration of a amined. The soil had undrained strength profiles, su
footing into soil is likely to increase Vo because of = suo + kz, where suo is the value of undrained shear
the increasing soil strength with depth and increasing strength at the footing base and k is the rate of in-
Nc (because of the larger embedment). This will crease of su with depth, z. For the non-uniform soil
cause the yield envelope to expand with associated profile, the soil was given ' = 6 kN/m3 and a normal
plastic vertical penetration, vp, and this relationship stress of v = 12 kN/m2 was applied to the soil sur-
is called the plastic hardening law. face. This surcharge represented two metres of over-
The above concept is used when installing a jack- burden and so this may have approximated a skirted
up rig in clays. Pre-loading of the jackup with ballast foundation with a skirt length of 2m (Tani & Craig
tanks causes the spud can foundations to penetrate 1995). This gave an undrained shear strength profile
the clay and thus increases their capacity to resist with a non-uniformity coefficient, kD/suo = 6.
combined loadings when in service (after removal of The uniform soil was weightless but the same
the preload). overburden was applied as before and so the effec-
An alternative form of work-hardening is consid- tive stress in the soil, 'v = 12 kPa at all depths. Con-
ered in this paper. During the lifetime of an installa- sequently, su was the same at all depths (k = 0) and
tion, dead-weight vertical loading will cause the soil so kD/suo = 0.
beneath to work-harden (consolidate) and this will
be reflected in the work-hardening of the system 3.2 Boundary conditions applied
(and the increase in size of the yield locus). The
footing-soil interaction behaviour during this type of The analyses were conducted in four stages. A pre-
event is examined using finite element analysis and load vertical load, Vp was applied with a small time
the results are expressed in terms of changes in plas- step to provoke undrained response (V = Vp, H = 0;
tic yield loci. M = 0; t  0; 'OX' in Fig. 3a). Then, this load was
v = 12 kN/m2 held constant for a time, tp while the soil consolidat-
D = 12 m ed under the vertical load and the foundation settled
(V = Vp, H = 0; M = 0; t = tp; 'X' in Fig. 3a,
2.5 D 'IABCD' in Fig. 3b). The position of the undrained
Cam clay:
' = 6 kN/m3; ' = 0.3; 3D yield locus was then calculated using displacement
 = 0.25,  = 0.05,  = 1.0,  = 3;
controlled analyses with small time steps. The foot-
k = 1x10-8 m/s ing was first displaced vertically until an undrained
plastic failure load was reached (v  0, h, = 0;
t  0; 'XY' in Fig. 3a). The shape of the yield locus
Figure 2. Geometry and soil conditions in the finite element was then followed approximately using a 'sideswipe'
analysis (not to scale). test where the vertical displace-
Bransby: Inclined load capacity of shallow foundations after 2
consolidation under vertical loads
 2.5 250
t=1Ms Ho/Asuo

Capacity/original capacity, %
t=0 200
2
Horizontal load, H/Asuo

t=5Ms
D 150
1.5 t=15Ms
Vo/Asuo
t=30Ms
1 C 100
Z A B
50
(a)
0.5
I
0
X 0
0 5 10 15 20 25 30
0O 5 10 Y 15 20
-0.5 Time, Msecs
300
Vertical load, V/As uo

Capacity/original capacity, %
Predicted Ho
(a) V-H load paths 250

0
I 200 Ho
0 10000000 20000000 30000000 40000000
-0.1
A 150
t=1Ms Vo
t=5Ms 100
-0.2
Settlement, m

t=15Ms
B 50 (b)
t=30Ms
-0.3
0
-0.4 0 20 40 60 80 100
C
Degree of consolidation, %
-0.5
D
Figure 4. Increase of footing strength with time (non-uniform
-0.6
soil, V/Vo = 0.5 during preload).
Time, seconds

(b) Settlement behaviour under vertical preload Figure 4a shows the amount of increase of Vo and Ho
against time for the analyses reported above. Figure
Figure 3. Effect of preload consolidation on the V-H yield lo- 4b shows the same data against degree of consolida-
cus (V/Vo = 0.5 during consolidation, non-uniform soil) tion. The vertical undrained capacity increases by
about 50 % during consolidation, whereas the hori-
ment was fixed and the foundation was displaced zontal capacity increases by about 125 %. Vo in-
laterally (v = 0; h  0; t  0; 'YZ' in Fig. 3a) creases linearly with degree of consolidation, U.
3.5
4 VERTICAL LOADING 3
No preload
Vp/Vo = 0.5
Horizontal load, H/Asuo

2.5 Vp/Vo = 0.22

Increasing
4.1 Non-uniform soil preload
2 Vp/Vo = 0.63

Results from the finite element analyses investigat- 1.5

ing the effect of different amounts of consolidation 1
time under a fixed vertical preload (Vp/Vo = 0.5) are (a)
0.5
given in Figure 3. Figure 3a shows the load paths
calculated in the analyses, and indicates the increas- 0
ing size of the undrained yield loci with increasing -0.5
0 5 10 15 20 25

amounts of consolidation under the preload. The Vertical load, V/Asuo

yield locus for the test without preload gives Vo/Asuo
350
= 11.28 which is 7.5 % greater than the lower bound
Capacity/original capacity, %

value of Vo/Asuo = 10.49 as calculated by Davis & 300 (b) Predicted Ho/As uo
Booker (1973). Horizontal capacity, Ho/Asuo = 0.995 250
Ho/Asuo
which compares very well with Ho/Asuo = 1 from a
200
pure sliding mechanism. The yield locus from each
150 Vo/Asuo
test is of similar shape, but the pure vertical, Vo and
horizontal, Ho capacities have increased significant- 100
ly.
50
Figure 3b shows the settlement under the preload
against time. The average degree of consolidation for 0
the five analyses are: 0 % (labelled 'I' on Fig. 3), 18 0 0.2 0.4 0.6

% ('A'), 55 % ('B'), 89 % ('C') and 100 % ('D') based Preload, V p/Voo

on the settlement data, where degree of consolida- Figure 5. Effect of preload consolidation on the V-H yield lo-
tion, U = v/vfinal. cus (varying V/Vo for full consolidation, non-uniform soil)

Bransby: Inclined load capacity of shallow foundations after 3

consolidation under vertical loads
Figure 5a shows the results of V-H side-swipe anal-
1.8
yses following full consolidation under preloads of t=0
varying magnitude: Vp/Vo = 0, 0.22, 0.5 and 0.63. 1.6
B C
t = 30 Msecs
1.4

Horizontal load, H/Asuo

The yield locus increases in size for larger preloads. 1.2
t = 60 M secs A
This is because more consolidation occurs under the 1
t = 160M secs

footing. Figure 5b shows that the horizontal capacity 0.8 I

is again improved more than the vertical capacity by 0.6
the preloading. 0.4
Figure 6 shows the soil deformation mechanisms 0.2
at failure for undrained vertical loading and horizon- 0

tal loading without (Fig. 6a) and with (Fig. 6b) initial -0.2 0 2 4 6

preloading. Consolidation under the footing changes Vertical load, V/Asuo

the vertical deformation mechanism and forces soil Figure 7. Effect of preload consolidation on the V-H yield lo-
displacements deeper into the soil to avoid the cus (V/Vo = 0.5 during consolidation; uniform soil)
strengthened shallow material. However, the hori-
zontal failure mechanism is unaffected by the pre- Figure 7 shows the improvement of the yield locus
load consolidation as there is no efficient mechanism for degree of consolidation, U = 63 % ('A' in Fig. 7),
to avoid the strengthened material. This suggests that 83 % ('B') and 100 % ('C') under a preload, Vp/Vo =
the horizontal capacity is affected by the preloading 0.5. Note that this corresponds to Vp/Asuo = 2.57
more than the vertical capacity because the soil compared to Vp/Asuo = 5.5 for the non-uniform soil
mechanism at horizontal failure is unable to avoid which has a higher bearing capacity. There is a clear
the particularly improved soil near the surface. improvement in foundation capacity both in terms of
the undrained vertical, Vo and horizontal, Ho capaci-
ty and these increase with time (Fig. 8). However,
there is noticeably less improvement in both Ho and
Vo than in the non-uniform case. This is because
there is less vertical foundation capacity for uniform
Vertical failure, Vo soil and so a preload of Vp/Vo = 0.5 represents a
smaller preload for the uniform soil, and so provokes
less consolidation.

5 DISCUSSION AND ANALYSIS

Horizontal failure, Ho
5.1 Prediction of yield locus changes
(a) No preload
Undrained foundation capacity increases due to
long-term vertical loading because soil consolidates
underneath the footing. The volume and degree of
strengthening will depend on the distribution of the
Vertical failure, Vo stresses within the soil and the changing soil proper-
ties. Both are discussed separately below.
The distribution of contact stresses beneath a rigid
footing (which are transferred to the soil) will be a
Horizontal failure, Ho function of the load level and soil properties. For
180
(b) Following full consolidation under Vp/Vo = 0.63
Capacity/original capacity, %

160 Ho/Asuo
Figure 6. Incremental soil displacement vectors at failure with 140
and without prior vertical load consolidation (kD/suo = 6). 120 Vo/Asuo
100
80
4.2 Uniform soil 60
40
Results from the finite element analyses conducted 20
on uniform strength soil (kD/suo = 0) are shown in 0
Figure 7. The vertical undrained capacity, Vo/Asuo = 0 50 100 150
5.08 from the finite element analysis which com- Time, Msecs
pares well with the exact solution, Vo/Asuo = 5.14 Figure 8. Variation of Ho and Vo with consolidation time for a
suggesting the accuracy of the numerical method. preload with Vp/Vo = 0.5 (uniform soil)

Bransby: Inclined load capacity of shallow foundations after 4

consolidation under vertical loads
 loads on overconsolidated clays.
The soil improvement due to partial consolidation
is likely to be more difficult to predict. Different el-
ements in the soil will have dissipated a different
p = 0.1V/A proportion of their initial excess pore pressure de-
pending on the pore pressures generated and dissi-
0.4 0.2 pated (and stress redistribution). Thus, the changes
0.3 in effective stress (and so changes in strength) will
be difficult to calculate. For example, elements near
the soil surface are likely to dissipate excess pore
Figure 9. Contours of total stress increases during undrained pressure quickly and so the soil will be quickly im-
preloading (V/Vo = 0.63; kD/suo = 6) proved in these areas. This suggests that horizontal
capacity may increase more quickly than vertical ca-
low vertical loads, the soil may remain almost elastic pacity for shallow foundations.
and so there will be stress concentrations on the cor- A simple method to predict the changing horizon-
ners of the footing and a wide distribution of load tal capacity with degree of consolidation utilises the
beneath the footing. For larger pre-loads (such as almost linear relationship between H and degree of
V/Vo = 0.63 for the non-uniform soil shown in Fig. consolidation, U shown in Figure 4b. This suggests
9), the contact stresses will be larger in the middle of that the average soil element beneath the footing ex-
the footing and the total stress increases will be dis- periences a effective stress change of U.Vp/A and so
tributed more locally. During consolidation, as ex- equation 2 is modified to give:
cess pore pressures dissipate, the effective stresses in
the soil will increase. Ho= [U (Vp/A)/('d)] Ho (3)
Footing capacity depends on the undrained shear
strength of the soil. Therefore, it is the changes in Equation 3 has been used to predict the change in
undrained shear strength beneath the footing during horizontal capacity against degree of consolidation
consolidation that needs to be predicted. For a nor- as shown in Figure 4b. Reasonable agreement sug-
mally consolidated soil, su is proportional to effec- gests that the method may be appropriate for approx-
tive stress. Thus, improvement after full consolida- imate design use.
tion To predict the changing failure envelope it ap-
pears necessary to calculate the improved pure load
su = (v/'vi) sui (1) capacities, Vo and Ho (and Mo) and assume that the
shape of the yield locus remains constant. This latter
where su is the initial undrained shear strength; v assumption cannot be true theoretically (the chang-
is the increase in total vertical stress due to the verti- ing undrained shear strength distribution beneath the
cal load; and 'vi and sui are the initial effective stress footing must alter the optimal failure mechanisms
and initial undrained shear strength respectively. somewhat), but may be a reasonable assumption for
Horizontal capacity, Ho involves a sliding failure engineering use. Indeed, the shape of the yield loci
mechanism (e.g. Fig. 6) where Ho = A suo. Hence, calculated in the present finite element analyses sup-
equation 1 can be used to predict the change in hori- port this assumption, although this may be partly due
zontal capacity due to a preloading with full consoli- to the zero rotation conditions investigated.
dation. This gives Single load capacities, Vo, Ho and Mo can be cal-
culated using assumed failure mechanisms (e.g.
Ho= [(Vp/A)/('d)] Ho (2) Bransby & Randolph 1998) after calculating the new
undrained shear strengths of the soil in the deform-
if it is assumed that there is a uniform normal stress ing zones of each mechanism. This will be complex
distribution of Vp/A on the base of the footing and for the vertical and moment capacities, especially if
that the initial vertical effective stress immediately partial consolidation is considered.
beneath the footing, 'vi = 'd, where d is the em- To avoid the complex method, it would be better
bedment depth. This relationship is shown dotted in if the increase in footing capacity could be linked to
Figure 5b and gives a good prediction of horizontal a single variable to form a work-hardening rule.
footing capacity for the range of preloads. Consequently, Figure 10 shows the relationship be-
For overconsolidated soil, the foundation capacity tween the increase of both vertical and horizontal
will be improved less by vertical loading because the footing capacity and consolidation settlement during
effective stress in the soil must exceed its previous the preload consolidation stage. There is a surpris-
maximum in order to significantly change the specif- ingly repeatable relationship between the variables,
ic volume and so increase the undrained shear despite some of the issues discussed above. Indeed,
strength. This suggests that there will be negligible the results suggest that similar work-hardening rela-
improvement of foundation capacity for low vertical tionships may be appropriate for use in design.
Bransby: Inclined load capacity of shallow foundations after 5
consolidation under vertical loads
 350
 Foundations with dead-weight vertical loads are
Capacity/original capacity, %
Varying time Ho/Asuo likely to be increasing strength during their life-
300
(Vp/ Vo = 0.5)
Varying load
time and are therefore are increasingly competent
250
(U = 100 %) to resist environmental loads.
200  Both the horizontal and vertical undrained capac-
150 Vo/Asuo
ity of the foundations may be improved signifi-
cantly by long-term vertical loading, and the hor-
100
izontal capacity will be increased most.
50  The shape of the enhanced V-H yield locus ap-
0 pears similar to the initial locus and so could be
0 0.2 0.4 0.6 0.8 1 expressed normalised by the pure vertical and
Consolidation settlement, m horizontal load capacities, Vo and Ho respective-
Figure 10. Prediction of footing capacity increase from consol- ly.
idation settlements (non-uniform soil).  Prediction of Vo and Ho with time and pre-load
size is complex. However, a simplified method is
5.2 Consideration of consolidation strengthening in presented to predict changes in Ho.
design  An alternative work-hardening law is suggested.
It links the increase in foundation capacity to the
When considering an offshore foundation, worst magnitude of the settlement during consolidation
case (often 100-year storm event) design loads are under vertical load. Preliminary results show
considered and the foundation size is chosen to en- promise for prediction of both vertical and hori-
sure that the loads remain within the yield locus. The zontal capacity.
chosen factor of safety should reflect the probability  It is conservative to ignore consolidation
of occurrence of an ultimate limit state event either strengthening in design, but this may lead to non-
during the design life (economic consideration) or optimal design solutions.
during any period of operation (safety considera-
tions). REFERENCES
From the above analyses, it is clear that in some
circumstances, the capacity of a foundation is in- Bransby, M.F., Randolph, M.F., 1998. Combined loading of
creasing with time after initial installation. Conse- skirted foundations, Geotechnique, Vol. 48, No. 5, pp. 637-
655.
quently, the factor of safety of the foundation against Bransby, M.F., Randolph, M.F., 1999. The effect of embed-
loads caused by a 100-year storm event will increase ment depth on the response of skirted foundations to com-
during the lifetime of the foundation. Thus, the bined loading. Soils and Foundations, Vol. 39 - 4, pp. 19-
probability of encountering an environmental load- 34
ing event causing failure will reduce significantly Britto, A., Gunn, M., 1987. Critical state soil mechanics via
with time. This has two effects: (i) the risk of failure finite element analysis, Chichester, Ellis Horwood Ltd.
Davis, E.H., Booker, J.R., 1973, The effect of increasing
of a foundation during its entire design life will be strength with depth on the bearing capacity of clays. Ge-
significantly less than it would be if calculated using otechnique 23, No. 4, pp. 551-563.
the initial foundation capacity; (ii) the worst case Hansen, J. B., 1970. A revised and extended formula for bear-
probability of failure during any period of operation ing capacity, Danish Geotechnical Institute, Copenhagen,
will only be reduced if consolidation occurs during Denmark, Bulletin 28, pp. 5-11.
installation or before the first storm 'season'. The Martin, C.M., 1994. Physical and numerical modelling of off-
shore foundations under combined loads, PhD Thesis, Uni-
choice of whether to consider foundation strengthen- versity of Oxford.
ing during the design life of an installation thus de- Meyerhof, G.G., 1953. The bearing capacity of footings under
pends on the regulatory framework as much as eco- eccentric and inclined loads, Proc. 3rd ICSMFE, Zurich, pp
nomics or safety considerations. 440-445.
Murff, J.D., 1994. Limit analysis of multi-footing foundation
6 CONCLUSIONS systems, Proc. of the 8th Int. Conf. on Computer Methods
and Advances in Geomechanics, Morgantown, Vol. 1, pp.
223-244.
Finite element analysis has been used to investigate Schotman, G.J.M., 1989, The effects of displacements on the
the amount of improvement of foundation capacity stability of jack-up spudcan foundations. Proc. 21st Offshore
caused by the application of long-term vertical loads Technology Conf., Houston, Paper no. OTC 6026.
such as those experienced by most foundation types Tan, F., 1990. Centrifuge and theoretical modelling of conical
footings on sand, PhD Thesis, The University of Cam-
during their lifetimes. Of particular interest were the bridge.
undrained combined vertical-horizontal load capaci- Tani, K. & Craig, W.F., 1995. Bearing capacity of circular
ties of shallow foundations and how these varied foundaitons on soift clay of strength increasing with depth.
with the time of application and magnitude of the Soils and Foundations. Vol 35, No. 4, pp 21-35.
vertical load.

Bransby: Inclined load capacity of shallow foundations after 6

consolidation under vertical loads
View publication stats