OTC 16294

Design and Installation of the Valhall Piggyback Structures

Torstein Alm, Aker Kværner Jacket Technology; Richard O. Snell, BP Exploration; Kevin M. Hampson, BP Exploration;
and Arne Olaussen, Aker Kværner Jacket Technology

Copyright 2004, Offshore Technology Conference

boreholes to be drilled both inside and adjacent to the piles.
This paper was prepared for presentation at the Offshore Technology Conference held in Rather than piles which had plugged during driving, the
Houston, Texas, U.S.A., 3–6 May 2004.
investigation revealed that the toe of both the refused piles at
This paper was selected for presentation by an OTC Program Committee following review of one corner had collapsed inwards, leaving the toe almost
information contained in an abstract submitted by the author(s). Contents of the paper, as
presented, have not been reviewed by the Offshore Technology Conference and are subject to closed.
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Offshore Technology Conference or officers. Electronic reproduction, As soon as the severity of pile damage had been
distribution, or storage of any part of this paper for commercial purposes without the written
consent of the Offshore Technology Conference is prohibited. Permission to reproduce in print
revealed, the possibility for further driving was soon ruled out
is restricted to an abstract of not more than 300 words; illustrations may not be copied. The as an option for foundation remediation, together with a
abstract must contain conspicuous acknowledgment of where and by whom the paper was
presented. number of other remedial solutions, which would utilise the
jack-up rig. As a consequence, the jack-up rig was then
demobilised. Various possibilities for saving the foundation
Abstract
were investigated. The criticality of jacket location relative to
This paper describes the design development and installation
the existing facilities constrained options to either repair or to
of a novel concept used to repair the piles at Valhall Water
remove and replace. Remove and replace was unattractive due
Injection Platform in the Southern Norwegian Sector of the
to high cost and lengthy delay to the schedule.
North Sea. The paper describes the concept, design principles,
The finally chosen option was based on a BP concept to
the tolerances required to account for known and potential
repair using two remediation structures, each to be connected
dimensional variations, a test programme to verify key
to the two refused piles at two corners. A new pile sleeve was
activities and the installation.
added to this structure, allowing additional piles to be driven.
These structures, named piggyback structures, were connected
Introduction
to the existing piles sleeves and piles already driven, by
The Valhall Water Injection Platform jacket is a 4000 tonne
grouting, see figures 1 and 2.
steel substructure, designed to support a 16000 tonne
integrated deck, situated in the North Sea in 70 meters depth
of water. The platform is located in close proximity to other
existing Valhall structures for drilling and topsides tie in.
During the installation of the jacket in August 2002, five
of the eight skirt piles experienced premature refusal in a very
dense sand layer at a penetration varying between
approximately 45 and 55 meters. The target penetration for
these piles was 67 and 72 meters. The 67 meter target
penetration was reached for the remaining three piles without
abnormal behaviour.
For the refused piles, two located in two corners oppose
to each other, and one in third corner, proper connection to the
jacket could not be made, as there were no pile weld beads on
the piles to match with the pile sleeves. In addition, since the
piles had variable wall thickness make up along depths the
piles bending properties in some depths were now inadequate,
in particular at the mudline elevation. Finally, due to the
shorter pile penetration, the pile capacities were insufficient to
support the design loads, in particular on the tension side of
the jacket.
A jack-up rig was chartered complete with a top-hole
full-bore airlift drilling system and a geotechnical drilling
system. This spread of equipment was mobilised to remove
the soil column within the piles and also enable geotechnical Figure 1 Computer model of the lower part of jacket with
piggybacks
2 OTC 16294

by traditional grouted connection, but with high strength grout.

The two inserts are structurally weld-connected to each other.
Through the grouting between the lower insert and the existing
driven pile, and grouting between the upper insert and the
jacket sleeve, a structural connection between the refused piles
and the jacket is established. In addition, the bending
properties of the refused piles are improved by grouting the
insert to the piles.
At the top of the upper inserts, a new separate pile sleeve
structure is added by connecting shear plates between the two
new pile upper insert tops to the new pile sleeve, enabling a
new pile to be driven outside the existing group. This pile will
then increase the foundation capacity of the group, both in
compression and tension.
Compressive forces from the jacket leg will now be
transferred from the jacket pile sleeves and into the upper
inserts of the piggyback. The majority of this force will then
be transferred directly to the top of the refused piles, through
the grout of a small gap between the upper insert and the
refused pile (that has been cut). The remaining forces will
partly be transferred to the refused piles trough the grouted
plain pipe connection of the lower insert section, and partly to
the new pile through the cluster structure connecting the new
sleeve to the inserts. The overall distribution is of-course
governed by the relative stiffnesses of the different parts of the
system.
Tension forces from the jacket will now primarily be
transferred from the jacket sleeve to the new upper insert and
into the refused piles through the plain pipe grouted
connection. The remaining part of the tensile forces will be
Figure 2 Computer model of one piggyback
transferred to the new pile, again the distribution being
governed by the local stiffnesses.
At the third corner, where only one pile had refused, a single
pile insert was developed, as it could be demonstrated that this
corner had sufficient axial pile capacity without further pile
penetration. The above additions and modifications would
then solve the identified shortcomings in terms of structural
connection to the jacket, increased pile bending properties at
seabed for the refused piles, and increase the overall axial
capacity of the pile group. The solution would however
require the refused piles to be cut inside their respective pile
sleeves. Tools for pile cutting operations were known to exist,
and had already been used to cut similar wall thickness piles.
However it was first time to be used to cut inside sleeves and
expecting that damage to the sleeves would be minimum.
Overall the piggyback concept provided an elegant well-
engineered solution to a difficult foundation problem.

Piggyback concept development

An idealised sketch of the Piggyback concept is shown in
figure no. 3.
The piggyback structure is connected to the refused piles
through a long lower insert that is grouted to the inside of the
piles. Weld beads are present on the outside of these inserts.
This is only to enforce the critical failure surface outwards to
the grout-pile interface, and thus increase the capacity of this
connection by some 15 %.
The upper insert of the piggyback represents the new pile
Figure 3 Idealized sketch of the piggyback bearing principles
top with weld beads, and is connected to the jacket pile sleeve
OTC 16294 3

Critical design considerations shimming, together with the maximum fabrication tolerances
A number of critical design choices had to be taken early in allowed by codes for elements where measurement data were
the project development to ensure a timely delivery of the not availlable. As a result of this study, the upper insert
structures. The project had to implement a novel untried diameter was reduced from the standard 96-inch dimension of
design concept in a very short timescale. In fact, all main steel 2438 mm down to 2400 mm, while the maximum diameter of
was ordered by the end of December 2002, based on the 2 metres was chosen for the lower insert. Also the distance
conceptual design only. A significant part of the design between the two neighbouring inserts within one piggyback
development comprised anticipating and allowing for potential cluster were set out according to the as-built position of the
dimensional variations, and consequently, a large amount of sleeves. Very strict fabrication tolerances was defined for the
robustness was built into the main steel design. new structures.

Pile cutting elevation Details of grouted joint design

The cutting elevation of the refused piles was one such critical The design details of the gap between the top of the refused
design consideration. It had been identified early in the design piles and the new upper insert, was of great importance as this
concept stage that cutting the pile below the lower pile shim detail had impact on the local force distribution in this area.
elevation might result in a sideways shift of that part of the An ultra high capacity grout was chosen, having cube strength
pile remaining in the ground. Such a shift would jeopardise the of 160 MPa. Such a high strength grout was required to safely
re-entry of the piggyback lower insert, and thus raise transfer compressive forces through this gap. But the high
additional uncertainty in the installation. Cutting the pile strength was also beneficial for other grouted connections
inside the sleeve was the only option, even though this would within the piggyback. Minimum gap distance would
introduce an additional risk of damaging the pile sleeve. structurally be beneficial, both as a result of high axial
Also, conflicting factors were present regarding the stiffness for the joint, but also this would minimise the length
height of the cut. A low elevation cut would give little reduction of the strucural elements above for grouting
reduction of the grouted connection length for the new pile capacity. A small nominal gap would however limit the
upper connection, and would be beneficial. However, a pile shimming possibilities in case the piles were cut outside the
wall discontinuity at a low elevation inside the sleeve would tolerance specification. Also, a small gap might affect the
result in transferring high bending stresses of the pile through grout flow during grouting, with increased risk of creating
the grout in the gap above the pile cut. This condition was of voids right at this critical joint.
great concern, and was also the main reason for choice of The final choice was however a rather small gap of 100
using ultra high strength grout. The final choice of a cut mm. There was confidence at time of decision that the pile cut
elevation at 3 meters above mudline was considered to be the could be achieved at the specified elevations, and measures to
best balance between these different factors. increase the accuracy of the cutting were consequently taken.
The 100-mm gap was found sufficient for grouting, taking also
Structural dimensions into account the quite different flow properties for this special
The dimensions for both upper and lower inserts were also of grout. Additionally, a model test of the joint would be done
great importance. Structurally, maximum size diameters are and grouting would be performed to reveal any problem areas.
beneficial for both the inserts. The lower inserts must have
sufficient bending capacity at the cut elevation, i.e. 3 meters Installation aspects
above mudline, and in the upper part of the refused piles close Support of the piggybacks during installation was an important
to mudline. Large plate hicknesses to maintain sufficient issue that had effect on final design. It was soon generally
section module were preferred, but steel plate delivery time agreed that pile driving should be performed prior to any
was long for plates with excessive thickness, thus pushing structural grouting, to eliminate the risk of jeopardising the
towards a maximum diameter. For the upper insert, it was of grouted connections due to excessive vibrations during pile
course beneficial to have a diameter close to the original 96- driving. At the same time, the piggybacks would have to be
inch standard dimension. In particular so, since not only large reasonably fixed during the pile-driving phase, to support
axial compressive forces were transferred between the existing installation and driving forces that inevitably would be
108-inch pile sleve and the new upper insert, but also transmitted into the structure.
maximum grouted connection capacity, as well as a high A connection system utilising pile grippers, connecting
bending capacity are required. the piggyback lower inserts to the inside of the refused piles
A critical factor in the above structural considerations was chosen. The vertical installation forces were designed to
was the need for sufficient clearance to enable a smooth be transmitted directly onto the top of the refused piles, as the
installation. The actual pile inclinations and local positions in piggybacks were designed to rest onto this face through
the sleeves was not known at time of ordering the steel, and carefully sized vertical shims. The gripper system was thus
the design would have to cater for any worst case combination designed primary to hold the structure laterally stable, but
of dimensional tolerance accounting for these factors. would also assist in damping possible vibrations during the
Additionally, there existed uncertainty with regards to the pile driving. To avoid reducing the bending properties of the
actual straightness of the piles. Had these been bent as a result lower inserts, the gripper can was situated 5 meters below the
of hard driving? pile cut elevation. At this elevation, pile bending moments in
A separate tolerance study was undertaken, utilising all inplace condition are significantly less than at mudline.
existing as-built measurement data for the sleeves and local
4 OTC 16294

Design verification The results confirmed in general the range of load

Immediately after the first conceptual layout of the distributions already used in the design.
piggybacks, the structural effects onto the existing jacket were
analysed. During this early engineering, the piggybacks were
only modelled with simple beam members, since the main
interest was to evaluate the effects onto the existing structure,
which could hardly be modified. Both inplace and fatigue
analyses were re-run, and showed only marginal effects. The
bending moments in the jacket legs were slightly increased
due to the overall increase in the eccentricity of the pile cluster
reaction forces. These marginal effects gave the design team
confidence that the piggyback was a promising solution. This
was also confirmed later by more advanced analyses, where
both the original pile cluster, and the new piggybacks were
modelled by finite elements.
Also, new pushover analyses were performed, to
investigate whether the piggybacks would introduce a weak
link in the jacket structure. The analyses results showed
however that the critical failure mechanism would still be
axial pile failure, followed by buckling of the jacket face
braces. A minimum Residual Strength Ratio (RSR) of 2.4 was
documented, again confirming the feasibility of the combined
structures.
The attention of the design team was then to focus on
details, since these were generally of a non-proven nature. In
particular, the grouted joint at the gap needed special attention,
both in terms of force distribution, and in terms of overall
capacity. The following activities were thus implemented:

• A detailed non-linear FEM analysis of the grouted joint,

aiming at disclosing the global force distribution in
compression, as well as the local stress distribution right
at the joint.
Figure 4 Outline of the FEM model of the joint (cut through)
• A model test of the grouted joint aiming to determine the
ultimate compressive capacity of the joint. Model test of grouting and grouted joint
Four identical models of the grouted joint were built at a scale
FEM analysis of grouted joint of 1:4. The connection consisted of scaled copies of the
An advanced FEM analysis of the grouted joint was existing pile sleeve, the upper part of the refused pile, and the
performed. The model included the full length of both lower new lower and upper piggyback inserts that was welded
and upper piggyback inserts, along with all details of the together as in the actual design. Since the main focus of the
enclosing structures, i.e. the grout, the refused pile, the sleeve, testing was to determine only the compressive capacity of the
the upper and lower yoke plates and the shear plates. The non- joint, no weld beads were included. In fact, the potential shear
linear computer program ABAQUS was used, and the capacity was minimised, by wrapping oily plastic around the
elements were either 8 node brick elements or 4 node shell grouted surfaces. The parts were grouted together in the same
elements. The structures were modelled with 2 rows of stages as would occur during installation. The upper part of
elements over the thickness. Figure 4 shows a brief outline of the refused and cut existing pile was first grouted to the sleeve,
the model. before the piggyback inserts were installed, and grouted to the
As noted above, one purpose of the analysis was to other parts. Also, the extreme range of eccentricity between
determine the global load distribution in compression in terms the grouted parts was included to evaluate possible effects of
of the portion taken by shear along the lower insert, and the the installation tolerance. The models were then cured in a
portion going directly through the grout ring on top of the controlled climate, and temperature logging was performed to
refused pile. A series of sensitivity studies were conducted, determine the equivalent degree of maturity during the
where the effect of slip-zones between the steel and the grout subsequent testing. The test rig used along with one test
were included. Such slip zones will inevitably occur at the specimen included for testing is shown in Figure 5.
ends of the grouted connection. The results showed however The results of the tests showed that the grout ring
rather little sensitivity of global force distribution to the retained full integrity up to an axial load of about 6 MN,
lengths of these slip zones, and the force transmitted by shear corresponding to 96 MN in full scale. Furthermore, the grout
through the lower insert was about 30 % of the applied loads. ring could sustain further loading to beyond 13.8 MN (220
MN in full scale) when the tests were stopped due to steel
OTC 16294 5

yielding. A ductile response was observed between these two double acting jacks, was designed to be actuated through two
loads. For comparison to Valhall, the maximum factored load independent hydraulic systems. Each system was connected to
to be taken at the joint was 30 MN for a 100-year extreme four jacks, and the overall design was made so that only one
event. system would be sufficient to resist the installation forces. The
When including the additional effects from bending, the primary system for actuation was through integrated pre-
partial safety factor against first indication of grout ring failure pressurised accumulators, one for each system. As a back up,
is 1.53 minimum for the 100-year event. The partial factor of all systems could also be pressurised by hydraulic hose from
safety against ultimate failure ranged from at least 3.5 to at surface, through a sub-sea hot stab ROV connection.
least 4.9. Grouting of the piggyback to the jacket and piles was
Three specimens were also subjected to cyclic loading performed through 3-inch steel pipes running inside the
after they had been subjected to a load of 3 times the piggyback inserts. 3-inch pipes were chosen as oppose to
equivalent design load. 400 cycles from zero up to the design standard 2-inch pipes, as a result of the flow characteristics for
load was applied, and no progressive growth in strain was this very special grout mix. A primary and a secondary piping
noted. This result demonstrated that even if the grout had system, both with outlets close to the tip were chosen. In
suffered cracking, it was still able to function as intended addition, a third piping system, with outlet immediately below
without a progressive mode of failure developing. the critical joint, was included. In case of damage to the grout
One specimen was machined to simulate potentially pipes, guide tubes were also integrated in the structures. These
damage to the sleeve resulting from the pile cutting operation. allowed for 2-inch stingers to be inserted from the top of the
The simulated damage was severe, being in the form of a piggyback, and down to each of the annuli between the inserts
circumferential groove having a depth of half the sleeve and the pile or sleeve for grouting. Base case for grouting was
thickness. No effect of the damage on test results could be nevertheless that the whole height of the grouting should be
discerned, demonstrating the connection was very tolerant to performed in one go.
sleeve damage.
Preparatory work
A series of preparatory work was required performed before
the piggybacks could be installed, including both onshore and
offshore activities.

Pile cutting trials

A cutting tool that had been previously used to cut similar
piles for a jacket removal project was chosen. This tool hangs
from the top of the pile, and cuts the steel wall from the inside
by the abrasive cutting technique. In detail, this means that a
fluid mixture of water and grit is pointed onto the pile wall
through a very tiny nozzle, at a short distance, and at a
pressure of about 700 bar. (10.000 psi). The continous cutting
speed for an 80 mm wall thickness is 50 mm per minute, and
the full cut of a pile took typically slightly more than 3 hours.
A picture of the cutting tool used is shown in Figure 6.
The success of this project was dependent on a successful
pile cutting operation. The cutting had to be performed at the
exact given depth without seriously damaging the enclosing
pile sleeve. Ensuring that the pile actually is fully cut through,
all along the perimeter, and at the same time that the sleeve
remained intact, required a delicately controlled operation. A
full-scale trial operation was therefore conducted in a
controlled underwater environment, where all variables could
be measured, and the results continuously monitored. The aim
of the trial was to calibrate the prime variables, such as cutting
speed and fluid pressure, and to develop procedures that would
safely ensure a good result.
Figure 5 Picture of test model while being tested It was observed during the trial that if the pressure nozzle
had to pass the same point on the pile more than twice, the risk
Installation system design of cutting holes in the sleeve were significantly increased.
The installation systems, primarily consisting of the grippers Multiple passes were however sometimes required in order to
and the grouting systems, had to be designed with sufficient ensure that full cut-through of the pile wall had been achieved.
redundancy to allow for a number of accidental events to
occur without putting the success of the installation and Offshore preparations
grouting at risk. Each piggyback cluster has two gripper cans, The installation of the piggybacks, including the pile driving
one on each of the inserts. Each gripper can, consisting of 8 and grouting, was to be performed by the S-7000 Heavy Lift
6 OTC 16294

Vessel. However, to minimise installation cost, time for this The results of the measurements confirmed that the
vessel should be minimised. Therefore, all preparatory works various data was within the expected tolerances that had
at the platform site should be conducted using smaller and already been accounted for in the design. However, the local
cheaper vessels. Two separate offshore campaigns were shim dimensions at the inserts could now be adjusted to obtain
carried out in span of several months before the actual additional clearances during installation. Sufficient clearances
installation, to ensure that the jacket structure was ready to during the penetration of the inserts was of major concern for
receive the piggybacks at a minimum of risk. the project, and this adjustments provided additional reduction
in the risk of jamming during the installation.
The second offshore campaign involved performance of
all the remaining preparations including:

• Pile cutting and transport to shore.

• Pile tips bearing plug grouting and soil/gravel back filling
for those piles where soil had been removed during
refusal investigation.
• Soil removal inside piles for those piles not investigated.
• Lower insert tips elevation control grouting.
• Refused pile stick-up to sleeve grouting.
• Internal pile cleaning.
• Soil dredging.

The pile cutting offshore operation was executed according to

plan. The piles were all cut and lifted in a controlled manner,
and the enclosing sleeves were all observed to be without
holes. If such holes had been observed, specially fabricated
magnetic plates were at hand to close these holes, thus
ensuring that the subsequent grouting could be performed
without leakage.
The piles removed were transported to shore for detailed
measurements. These results confirmed that the cutting
elevations were all within 5 centimetres from the planned
positions, and the cutting plane was close to perpendicular to
the pile axis. The results of these measurements were used to
adjust the final vertical shimming of the landing position of
the upper insert. By doing this, it was ensured that the
piggyback would land and rest vertically on all four shims
onto the cut piles.
Soil removal and filling inside the piles was performed
using tailor made equipment. Soil filling had to be performed
into the two piles that had been completely emptied during the
earlier jack-up rig investigation phase. The damaged pile tips
was first secured by filling a 4 meter high concrete plug,
before sand/gravel was added. It was originally planned that
this sand could easily be dredged from the subsurface in the
area, but this proved to be difficult. The final material was
therefore gravel shipped from shore.
Soil removal from the remaining three piles was timely
Figure 6 Picture of cutting tool executed by means of flushing and airlift. At the intended
elevations for the internal soil plug, a concrete plug was set to
The first offshore campaign was purely a survey campaign. assure a fixed elevation for start of final grouting. This plug
The scope was to measure the straightness, the position and elevation was set at 1 meter below the tip elevations for the
inclination angle and direction of all the refused piles. Also, piggyback lower inserts. Knowing the excact position of the
the pile stick-up should properly be paint marked, so that the plug was essential to obtaing control of the grout level at any
cut-offs could later be identified orientation wise, and thus time during the final grouting.
details of the pile cut could be positioned. Additionally, Grouting the annulus between the stick up of the refused
similar measurements of the pile sleeves should also be piles and the lower part of the pile sleeves was also performed
performed, to confirm the as-built data from the fabrication during this offshore campaign. This was done to completely
yard. seal off this part of the annulus, and thus preventing any leaks
and subsequent possible voids during the final and critical
OTC 16294 7

grouting. The same high strength grout was used, even though The actual lift and penetration of the inserts into existing pile
this part was not designed to assist in the structural sleeves and the cut refused piles went extremely smooth, and
connection. all structures penetrated to the planned depths within a few
Finally, the inside of the piles was cleaned by high centimetres of accuracy. The grippers were activated as
pressure flushing, and the area outside the mudmats was planned, and the pressures in all systems were almost spot on
dredged to a pre-scribed elevation. The jacket structure was the expected values. A small leak in one system was detected,
now ready to receive the piggybacks. as fluids were observed coming out from a hot stab receptacle.
Inserting a dummy hot-stab stopped this leak, and the system
Offshore installation was re-pressurised. After this, all systems were observed to
The offshore installation of the Piggybacks commenced in the keep the pressures during the rest phase of the installation.
beginning of August 2003, and lasted for about a week. Pile driving commenced immediately after the grippers
were activated, and this went as planned. The MHU 1700
hammer was used, and the resistance encountered was also
according to the predictions.
The final grouting was the most tedious operation, as a
total amount of 150 m3 of grout was required for the
piggyback inserts, and additionally some 40 m3 for the two
new piles. The grout was delivered in sacks of 1000 kg each,
and was mixed in batches of this size. Exact amount of water
was added to the cement mixture according to the recipe, and
mixed for at least 5 minutes. Typical rate of mixing was 7 to 8
batches per hour, and the pumping rate was thus 3.5 to 4 m3
per hour. Due to the relatively warm weather during the
operation, the complete grout mixing station had to be cleaned
for partly cured grout at the end of each grouting sequence.
Grout strength testing was performed at prescribed time
intervals after mixing, to observe the curing properties. A
strength requirement of 20 MPa had been identified as a
requirement to land the platform topside. Pre testing of the
grout early age strength development had been performed, and
an expected curing time of 34 hours was identified. This was
confirmed by offshore testing, and the topside was lifted and
set onto the jacket almost exactly one year after the initial pile
refusal incident.

Acknowledgements
To Dr. M.S. Lee, Consultant Engineer of BP Norge for
extremely valuable comments and design reviews throughout
the project phase. Acknowledged is also Dr. Adrian Dier of
MSL Engineering for valuable advice and support during
initial concept development, model testing and its result
interpretation.

Figure 7 Piggyback during offshore installation