FLOAT-IN COFFERDAMS

Robert B. Bittner, P.E., President, Ben C. Gerwick, Inc. San Francisco, CA, USA

Large diameter piles (2 to 3-m diameter, both driven and drilled) are being used
more frequently by designers of bridge foundations. The use of these large
piles allows the pile cap to be positioned off the bottom but usually still below
water. Conventional cofferdams as shown in Figure 1 below are not suited to
this type of pile cap. However, float-in cofferdams offer a viable alternative.
This paper describes the float-in cofferdam system, its initial development for
use on the Bath-Woolwich Bridge in Maine and its most recent application on
the new Carquinez Bridge in California.

Introduction
Cofferdams are temporary structures used in the
construction of bridge piers and other marine
structures. Their primary purpose is to hold out
water and unstable soil from the construction
area, and thereby, allow in-the-dry construction
of the permanent structure below the water line
and quite often below the mud line. Typically,
cofferdams consist of long interlocking steel
sheet piles driven through water into the bottom
of the waterway.
The sheet piles form
temporary exterior walls which are typically
braced internally with wales and struts. The
bottom of the cofferdam is typically sealed with
tremie concrete or an impervious clay layer if
one exists. This type of cofferdam is typically
used for structures positioned at or below the
mud line.

of smaller piles. These larger piles with their

higher stiffness make it possible from a design
standpoint to position the pile cap at any point in
the water column rather than at or below the
midline as required with the use of smaller piles.
In most cases, the pile-top structures are not
positioned completely out of the water because
of aesthetics and concern with splash-zone
corrosion at the pile to pile cap connection.
Therefore, the bottom of the pile cap is typically
located at least a meter below low water line.
For this type of off-the-bottom pile top structure,
conventional bottom founded cofferdams may
not be the best solutions.

The cost of the cofferdam can represent a

significant percentage of the total cost for
constructing the underwater structure.
At
extreme depths, the cost of the cofferdam can
exceed the cost of structure being built in the
cofferdam. In addition, there are practical limits
to the depth to which conventional cofferdams
can be used. As an example, sheet piles over
100-ft in length are extremely difficult to
transport and handle during installation.
However, designers of marine structures are
tending to use larger and larger diameter piles,
both drilled shafts and driven piles. Designers
are finding that piles in the range of 2 to 3-m
diameter are more efficient than a high number

Fig. 1

Conventional braced cofferdam

Float-In Cofferdam Concept

The float-in cofferdam concept has been
developed
to
specifically
address
the
construction requirements for pile-top marine
structures positioned below water but up off the
bottom of the waterway.
This concept is
perhaps best described by its sequence of
construction (See Figs. 2 through 8 for a plan
view of a typical float-in cofferdam and
construction sequence):
1.

2.

3.
4.
5.

6.
7.

8.

Piles or drilled shaft casings are

installed through the water. Following
completion of the pile installation, they
are cut off underwater.
Concurrent with pile installation, a thin
shell of the structure is pre-fabricated at
an off-site location. This thin shell (and
follower cofferdam if needed) is referred
to as the float-in cofferdam.
Following completion of fabrication, the
float-in cofferdam is launched and towed
to the installation site.
Once at site, the float-in cofferdam is
lowered down on to the pre-installed
piles.
A seal is then formed between the piles
and the bottom of the float-in cofferdam,
and grout (or tremie concrete) is placed
to lock the float-in cofferdam to the piles.
The
float-in
cofferdam
is
then
dewatered.
All remaining work is then completed inthe-dry similar to a conventional
cofferdam. This includes completion of
reinforcement and concrete placement
in the piles or drilled shafts, and placing
reinforcement and concrete for the cap
and pier shaft.
After completion of the shaft, if a
follower cofferdam was used, the
cofferdam area is re-flooded and the
follower cofferdam is removed for re-use
on the next pier.

The float-in cofferdam method of underwater

construction offers the following significant
advantages over conventional cofferdams:
It can be used in essentially any water depth
where
the
unsupported
permanent
foundation piles can serve their intended
function.
This is in comparison to
conventional bottom-founded sheet pile
cofferdam systems that are limited by the

efficient lengths that standard sheet piles can

be rolled, shipped and handled at the site.
It offers the potential advantage of a reduced
construction schedule by allowing the pile
installation work to commence immediately
rather than wait until the cofferdam has been
designed, sheet piles ordered, bracing
frames fabricated and cofferdams installed.
This is a significant time advantage in that
pile installation work is typically on the critical
path and is the most common cause of delay
on a marine construction project. Time is of
course required to design, build and launch
the float-in cofferdam. But all of this work
can be performed concurrent with the pile
installation work.
This cofferdam method can offer significant
cost savings by:
1) eliminating or reducing the length of
sheet piles,
2) eliminating the conventional tremie seal
that is used to resist hydrostatic uplift,

Fig. 2

Plan at top of precast pile shell

Fig. 4

Stage 2 Launch and position float-in

cofferdam

Fig. 6

Stage 4 Dewater cofferdam and

construct pile cap

Fig. 3 Stage 1 Pile installation

andrkcut
is off

Fig. 5

Fig. 7

Stage 3 Land cofferdam, seal pile

connection and place tremie concrete

Stage 5 Construct pier shaft

Fig. 8

Stage 6 Strip follower

3)

4)

allowing the pile cap fabrication

operation to be performed more
efficiently at a centrally located
fabrication site on shore or at
dock side rather than at multiple
in-water cofferdam locations, and
allowing for more competitive
pricing for material and labor by
enlarging the geographical area
over which the shells can be
fabricated.

These cost savings are offset to some

degree by the cost of the offsite fabrication
and launch facility, however these are usually
less than the cost associated with the
savings listed above.
It can reduce construction risk by minimizing
the amount of time that the work is exposed
to severe whether and fluctuating water
levels. In addition, the critical operations
such as float-in, positioning and landing can
be timed to fit optimum weather and water
level conditions.
It can provided higher quality for the
underwater concrete by allowing the exterior
of the pile cap to be precast on shore in ideal
conditions under tighter quality control.
From a design point, this system allows
elimination of the conventional tremie
concrete seal. This reduces dead weight of
the structure, and in areas of seismic risk,
significantly reduces seismic loadings as
well.
Float-In Cofferdam Features
The float-in cofferdam concept has the following
typical features:
Off-site prefabrication A thin steel or
concrete shell which forms the exterior of the
pile cap is pre-fabricated in a dry dock or on
the deck of a barge with the barge positioned
at dock side to allow construction support
from a shore based crane. (See Figs. 2 and
4 for details of the float-in pile-cap shell used
on the Bath-Woolwich Bridge.) This offsite
fabrication is typically performed at a central
location that allows an efficient flow of
materials and effective monitoring for quality
control. If the construction site is affected by
severe weather conditions, it may be
preferable to locate the fabrication facility
indoors where higher labor efficiency can be

obtained.
Additionally, selection of the
fabrication site should take into consideration
the method of launching.
(See launch
method below for specific details.)
Temporary
bulkheads

Temporary
bulkheads are typically steel ribbed elements
used to close off openings in the shell
structure, making the shell water tight and
allowing it to float after launch.
These
bulkheads are typically salvaged and re-used
on following caps.
The bulkheads are
typically bolted in place and use
compressible neoprene or rubber seals at the
mating surfaces to maintain water tightness.
(See Figs. 23 and 24 for a view of the
temporary bulkheads used on the Carquinez
Bridge float-in cofferdams.)
Pile-top bulkheads Pile-top bulkheads can
be classified as a special type of temporary
bulkhead. They are inverted open-bottomed
cans that are centered directly over the top
of the circular openings in the bottom of the
pre-fabricated shells.
They allow the
foundation piles to penetrate up into the
interior of the shell while maintaining the
watertight integrity of the float-in cofferdam.
(See Figs. 10 and 22 for examples of pile-top
bulkheads used on the Bath-Woolwich and
Carquinez bridges respectively.)
These
blockouts are of course watertight except for
the open bottom and can act as individual
ballast compartments if compressed air
hoses are attached to each compartment.
By regulating the air pressure in each pile-top
blockout, it is possible to level or trim the
float-in cofferdam during launch and
transport. (See Fig. 12 for a view of the air
piping used on the Bath-Woolwich Bridge.)
Follower steel cofferdams Follower
cofferdams are required for pile-top
structures that are positioned completely
under water. Follower cofferdams extend the
height of the prefabricated shell walls
sufficiently to allow the pile cap to be lowered
to final grade without allowing water to
overtop the cofferdam. These are typically
stripped at completion of the pier and re-used
on following piers. (See Figs. 14, 15 and 18
for pictures of the follower cofferdam used on
the Bath-Woolwich Bridge.) These follower
cofferdams are very similar to conventional
sheet pile cofferdams. They have the same

type of internal bracing consisting of steel

wale frames and struts. The walls of the
follower cofferdams can be steel ribbed
panels,
corrugated
panels
or
even
conventional steel sheet piles. All three
types typically terminate in a flat bottom plate
that bolts directly to the top of the pile-cap
shell structure.
Concurrent foundation and pile-cap shell
construction The float-in cofferdam
method of construction allows installation of
foundation piles to run concurrently with
fabrication of the pile-cap shell. This feature
allows for a shorter construction schedule
compared to conventional pier construction.
Temporary
pile-top
cofferdams
for
underwater cut off of piles In order to
float-in over the pre-installed piles, it is
necessary to cut the permanent piles off to
grade under water. This can be performed
with divers, however it is expensive and risky
to do so. It is preferable to cut the piles off
in-the-dry using a temporary cofferdam that
slips down around the permanent pile and
seals to the side of the pile, thus allowing the
annulus between the outer cofferdam and the
exterior wall of the pile to be dewatered.
Once the interior of the pile and the annulus
have been dewatered above the cut off point,
a worker is lowered into the interior of the
pile, and the pile is cut off to precise grade.
An alternate method which can be used
when the drill casings are torqued into
position is to use drilled shaft casings that
have a reverse torque disconnect built into
the wall of the casing. The disconnect point
is pre-determined at each pile location, and
each casings is fabricated with a disconnect
at a pre-planned location. This method was
used on the drilled shafts for the BathWoolwich Bridge.
Cofferdam launch method There are
many different potential ways to launch the
float-in cofferdams. The launch method is
usually tied to the method and location of the
shell fabrication or casting. If the shell is
fabricated in a dry dock, launching is
performed by flooding the dry dock and
exiting through a removable gate at the
entrance to the dry dock. If the shell is cast
on the deck of a flat-deck barge, it is typically
launched by ballasting the barge out from
under the shell and allowing the shell to float

free. If this method is used, special

consideration needs to be given to
maintaining stability of the barge during all
ballasting operation. The stability analysis
for the launch operations should be
performed and monitored by an engineer
experienced
at
performing
stability
calculations or by a naval architect. The
stability can be controlled by setting one end
of the barge down on a pre-leveled bottom,
or by attaching wing-tanks to the barge in
order to provide sufficient water plane during
all stages of launch. If the shell is light
enough and
heavy lift equipment is
available, the shell can be lifted off and set in
the water. (For an example of this method,
see Fig. 11 for a photo of the float-in pile-cap
shell being lifted off the deck of the casting
barge at the Bath Ship Yard. If launch ways
are available close to the installation site,
they cap provide an alternate launch method.
Method of transport Tugs are typically
used to assist the floating cofferdams during
their transit to the installation site.
Positioning and guidance system
Positioning and guidance systems are
needed during the mating of the float-in
cofferdam to the pre-installed foundation
piles. These are particularly important when
the tops of the foundation piles are several
meters below water and out of site. These
systems include mooring lines, anchors, and
tapered guide structures to direct the
cofferdam onto the permanent support piles.
Temporary ballast system In order to sink
the prefabricated pile-cap shell down to the
cut-off piles, it is necessary to partially flood
the cofferdam. During the sinking operation
the cofferdam should be maintained relatively
level. This requires an ability to add water to
different areas of the pile-cap shell. The
ballast system performs this function.
Temporary support system In order to
land the float-in cofferdam on the cut- off
piles, a means of load transfer must be
developed to carry the partial dead weight of
the cofferdam. The required capacity for the
system is relatively low and is only temporary
until it is replaced by the tremie or grout
closure pour. (See Figs. 19 and 20 for
details of the temporary support beam,

shown in yellow, used as the temporary

landing support on the Carquinez Bridge.)
Bottom sealing system After landing of
the float-in cofferdam, the foundation piles
typically penetrate up into the bottom of the
pre-fabricated pile- cap shell, and a grout or
concrete placement is made in the bottom of
the shell to lock the shell to the foundation
piles. In order to place the grout or tremie
concrete, it is first necessary to seal the
bottom of the annulus between the outside
wall of the pile and the hole in the bottom of
the pre-fabricated shell. This seal does not
have to resist the total hydrostatic head at
that point but only the buoyant head of a few
feet of the grout or tremie concrete. (See
Fig. 20 for details of the temporary sealing
system used on the Carquinez Bridge.)
Grout or tremie closure The grout or
tremie concrete placement is the primary
sealing system to allow dewatering of the
float-in cofferdam. It is also the primary load
transfer element to resist the uplift buoyancy
of the cofferdam after the cofferdam is
dewatered and to support any net downward
load immediately after placement of the first
lift of concrete in the pile-cap shell. (See Fig.
5 for a schematic of the tremie concrete
closure that was used on the Bath-Woolwich
Bridge float-in cofferdams.)
Specific Considerations In The Design Of
Float-In Cofferdams
Tolerances The float-in cofferdam method
requires the mating of the float-in cofferdam to
the top of the pre-installed piles. In order for this
mating to succeed without punching a hole in
the bottom of the float-in cofferdam, it is
essential that the pile be properly located within
a specified tolerance.
Piles installed from floating equipment on anchor
lines or spuds can typically meet installation
tolerances of +/- 15 cm horizontally at the
surface of the water and 1% in verticality. Piles
installed through a bottom founded template can
be installed to a tolerance of typically 6 cm at the
surface of the water and 0.5% verticality. These
tolerances need to be taken into account when
designing the net clearance of the openings in
the bottom of the float-in cofferdam relative to
outside diameter of the pile. These clearances

need to be large enough to take into account

possible misalignment of the piles while
maintaining the structural integrity of the bottom
slab and providing a void space that can be
effectively sealed off for grouting or tremie
placement.
The cut-off tolerance of the piles under water
using a pile top cofferdam for access is typically
about +/- 3mm.
This tolerance may be
necessary if the support bearing on landing is
directly onto the pile cut-off surface. However
an alternate way of leveling the float-in
cofferdams after landing is to use flat-jacks
attached at the bearing points in the float-in
cofferdam. These jacks should be plumbed
together to allow a three point landing system.
Water Tightness During launch, transport and
positioning of the pile-top cofferdams it is
essential that the inflow of water into the
cofferdam be controlled at all times. The joints
between the float-in pile top structure and water
tight bulkheads, follower cofferdam and pile top
blockouts are all typically fitted with strips of
compressible neoprene or rubber.
These
continuous strips are typically 5 to 8 cm wide
and 1 cm thick. The durometer of the seal
should be approximately 40 to 60. Compression
of the sealing strip is provided by bolts on a
typically about 30 cm.
The strips should be
positioned to compress with application of
increased bolt tension and increased hydrostatic
load on the outside of the cofferdam.
If conventional steel sheet piles are used for the
follower cofferdam, the interlocks should be
treated with hydrophilic strips prior to attachment
to the float-in cofferdam.
If ballasting operations during sit-down
operations require internal bulkheads to
maintain stability, the water tightness of these
bulkheads is essential and should be treated
with the same level of importance as exterior
bulkheads.
Leaking interior bulkheads can
cause loss of stability during ballasting
operations and result in loss of the floating pilecap shell.
Draft Limitations There may be several
factors that will limit the allowable draft of the
float-in cofferdam at launch. These may include
the depth of dry dock where the float-in
cofferdam is being fabricated, the depth of the
navigation channel leading to the installation site

and finally, the cut-off elevation of the foundation

piles at the installation site. If excessive draft is
a problem, there are several potential ways of
minimizing the draft:
Decrease the thickness of the concrete walls
and bottom slab.
Use lightweight or semi-lightweight concrete
for the shell.
Blow all water out of the open-bottom pile top
bulkheads.
Attach supplemental buoyancy tanks.
Floating Stability A floating stability check
should be performed for all stages of launch,
transport, positioning and landing of the float-in
cofferdam. This is performed using standard
naval architecture principles for calculation of
meta-centric height and reserve stability. These
calculations may show that a float-in cofferdam
is unstable during the set-down or landing
operation when water is pumped into the
cofferdam in order to increase draft. This
typically occurs when water is allowed to freely
flow from one side of the cofferdam to another.
This sloshing can create an unstable situation
and is referred to as excessive free-surface
effect. The standard way of reducing the freesurface effect is to partition the interior of the
cofferdam with watertight bulkheads. If interior
bulkheads are necessary their water tightness is
critical to the stability of the cofferdam and tests
should be conducted on their watertightness
prior to launch of the float-in cofferdam.
Installation Stresses The float-in cofferdam
should be designed for all stages of launch,
transport, positioning and landing as well as
buoyancy forces and full hydrostatic forces that
develop when the cofferdam is sealed off and
dewatered. In a typical installation sequence for
a float-in cofferdam, the vertical loading on the
temporary support system will reverse during the
course of construction. Upon landing on the
foundation piles, the support system must
support the partial dead weight of the float-in
cofferdam. Once the grout or tremie concrete
seal has been placed, cured and the interior of
the cofferdam dewatered, the support system
must then be able to resist the full upward
buoyancy force of the float-in cofferdam. During
placement of the first lift of footing concrete, the
support system must resist the net difference
between the buoyancy and the dead weight of
the first lift.

Examples
Bath-Woolwich Bridge
Ben C. Gerwick, Inc. (BCG) of San Francisco
working for Flatiron Structures Company of
Longmont, Colorado designed the floating
cofferdam system used on the Bath-Woolwich
Bridge across the Kennebec River at Bath,
Maine.
The
Maine
Department
of
Transportation awarded the design-construct
contract for the 1,000m long four lane bridge to
Flatiron Structures Company (FSC) and Figg
Engineers in August of 1997, and shortly
thereafter, FSC contracted with BCG to design a
cofferdam system for the six main river piers.
Figgs design for the six main river piers called
for two size footings: 10.9m by 9.9m by 3.6m
deep supported by four 2.44m diameter steel
cylinder piles socketed into rock and 9.6m by
9.6m by 3.6m deep supported by three piles of
the same size. The typical footings had a 2.7m
deep tremie seal and was located in 13m of
water with the top of the footing about 1.4m
below mean tide. This cap design significantly
reduced the water depth at which the footing
were to be constructed, however, it created the
problem of how to construct an underwater
footing suspended 6m above the river bottom.
In order to address this problem and reduce the
amount of work in the river, BCG proposed the
use of a floating cofferdam system (See plan
view in Fig. 2 and sequence of construction
shown in Figs. 3 through 8):
Pre-install the drilled shafts using a two stage
template.
Construct a precast pile-cap shell on shore and
attach a reusable temporary steel follower
cofferdam.
Launch the cofferdam and tow it to the bridge
site.
Position the cofferdam over the drilled shafts
and fix it in position with four spud piles.
Lower the cofferdam down over the preinstalled drilled shafts with jacks located on top
of the spud piles.
Lock the pile-cap shell to the drilled shafts by
placing a 1.3m deep tremie seal.

 Dewater the cofferdam and complete the pile

cap and pier shaft in the dry.
Flood and remove the follower cofferdam for
reuse on the next pier.
The floating cofferdam system and the above
sequence substantially reduced in-water work
and improved quality control by transferring the
work to an offsite dock location. The system
also saved 4 months in the construction of the
foundations by allowing the drilled shaft
installation to start immediately after contract
award and to proceed concurrently with
fabrication of the floating cofferdams.
The final design for the precast pile-cap shell
was a 5.182m deep box with 288mm thick walls
and bottom slab.
The typical precast box weighed 340 tons and
was cast on the deck of a barge at dock side.
(See Fig. 9.) Two 4.267m high steel followers
were constructed of interlocking steel sheet piles
and assembled on the deck of another barge.
Two local shipyard cranes were used to lift the
precast pile-cap shells off the deck of the casting
barge and into the water. (See Fig. 11.) The
precast shell was then leveled by using
compressed air to vary the water level in the
open bottomed 3m diameter pile top block outs.
(See Fig. 12.) Once the segment was afloat and
trimmed, one of the steel follower was lifted into

Fig. 9

Casting pile-cap shell on deck of barge,

Bath-Woolwich Bridge

position on top of the precast pile-cap shell and

bolted to the top lip of the shell. (See Fig. 14.)
The floating cofferdam was then towed into
position over the top of the preinstalled piles
(See Fig. 13.) and locked in horizontal position
with four 1m diameter spud piles driven through
the pockets, attached to the corners of the
cofferdam. (See Fig. 15.)
One of the keys to the successful use of this
technique was the installation tolerances of the
2.44m diameter drilled shafts. The drilled shafts
had to be installed to a relatively tight tolerance
or the holes in the bottom of the pre-cast shell
would not have matched the layout of the preinstalled piles. For the Bath-Woolwich Bridge,
the inside diameter of the holes in the bottom of
the pile-cap shell was 3.15m. This provided a
pile clearance of 355mm. The drilled shafts
where successfully installed to a horizontal
tolerance of +/- 150mm at the point of cut off, 5
m below mean water elevation. This tolerance
was obtained by using a two piece template.
The bottom steel frame was positioned with
anchor winches to a tolerance of +/- 300mm and
fixed in vertical and horizontal position with four
1m diameter spud piles. The upper guide slides
on top of the bottom frame and were positioned
to a tolerance of +/- 50mm before it was fixed to
the lower frame. The drill casing was then
stabbed through a guide sleeve attached to the
top guide.

Fig. 10 Pile-top bulkhead inside precast shell,

Bath-Woolwich Bridge

Fig. 11 Lifting pile-cap shell off deck barge,

Bath-Woolwich Bridge
A second key to the successful installation was
the initial temporary underwater connection
between the float-in cofferdam and the preinstalled drilled shafts. This connection was
made with a 1.22m deep tremie seal, and the
vertical load was transferred through bond of the
tremie seal to the sides of the drilled shaft
casing. This bond was sufficient to resist the
buoyancy of the dewatered float-in cofferdam at
high tide and allowed placement of the pile cap
reinforcing steel and first concrete lift in the dry.
(See Figs. 16 and 17.) When the footing
concrete was being placed in the pile-cap shell,
and the tide was out, the vertical loading was
reversed, and the bond supported the dead load
of the fresh concrete. The allowable bond stress
between the tremie concrete and the drilled
shaft casing was calculated using the American
Petroleum Institute (API) Code RP 2A,
Recommended Practice for Planning, Designing
and Constructing Fixed Offshore Platforms.

Fig. 12 Hooking up air lines to pile-top

bulkheads, Bath-Woolwich Bridge

Fig. 13 Pre-installation of drilled shafts using two-stage

template to set casings,
Bath Woolwich
Bridge

Fig. 14 Mating follower cofferdam to pile-cap shell,

Bath-Woolwich Bridge

Fig. 15 Positioning and lowering float-in cofferdam,

Bath-Woolwich Bridge

With this type of support system, if additional

capacity is required, the bond area can be
increased by deepening the tremie placement or
the allowable bond stress be increased by
adding weld beads to the outside of the casing
in the zone of tremie concrete.

installed erection frames, or falsework. The

footing shells were to be used as temporary
cofferdams and as templates for the installation
of drill shaft casings.

Following completion of the pile cap, the pier

shaft was completed to a point above the high
water line and the follower cofferdam was
stripped. (See Fig. 18.)
New Carquinez Bridge
The contract for the new bridge was awarded in
early 2000 to a joint venture between FCI
Constructors and Cleveland Bridge California,
(FCI/CBC). Ben C. Gerwick, Inc. was selected
by FCI/CBC to modify Caltrans float-in
cofferdam concept and further reduce the
amount of onsite work.

Fig. 17 Placing pile-cap reinforcement, BathWoolwich Bridge

Fig. 16 Dewatering float-in cofferdam,

Bath-Woolwich Bridge
The foundation design for the new bridge
required six 3m diameter drilled shafts under
each of the four main tower legs. Because of the
very high lateral stiffness of these piles, it was
possible to elevate the four tower footing blocks
- 5.4m deep and 18m by 20m in plan - up off the
bottom of the strait. At the north tower, the
footing block is positioned 30m off the bottom
with the lower edge of the footing at 2.34m
below mean sea level. The original design by
Caltrans was to cast the footing shells offsite,
float them to the bridge and land them on pre-

10

Fig. 18 Stripping follower cofferdam from

completed pier shaft, Bath-Woolwich
Bridge
BCG modified this concept by designing a
temporary support system that allowed the floatin cofferdams to be landed directly on the preinstalled drill shaft casings. (See Figures 19 and
20 for details of the temporary support system,

shown in yellow.) This modified design offered

significant advantages. First, it allowed drilled
shaft installation, an activity which was on the
critical path, to begin sooner, by allowing casing
installation to start before the float-in cofferdams
had been completed.
Secondly, the
modifications meant that the construction
schedule was shortened by allowing casing
installation to run independently of and
concurrently with cofferdam fabrication.

Fig. 21 Casting float-in cofferdam on deck of

launch barge, Carquinez Bridge

Fig. 19 Float-in cofferdam concept for Carquinez

Bridge

Fig. 20 Pile-top details for Carquinez Bridge

Construction cost was reduced by completely
eliminating the need for erection frame or
temporary falsework support for the float-in
cofferdam and infill concrete, and risk was also
reduced as a result of decreasing exposure
time of a partially completed cofferdam and
footing to storm conditions .

Fig. 22 Completed float-in cofferdam with pile-top

bulkheads on deck of launch barge,
Carquinez Bridge

11

Cofferdams and connecting ties were cast on a

flat-deck barge, 22m wide by 76m long and
4.6m deep, which was positioned at the
dockside within reach of a shore-based crane.
(See Figs. 21, 22 and 23.) The surface of the
barge was prepared by casting a 15cm concrete
slab to provide a level and smooth casting
surface. In order to ensure that the cofferdams
would lift off the barge at launch, the surface of

the casting slab was sprayed with a bond

breaker that created a polished surface. As an
added precaution, water pressure was applied to
the contact surface between the top of the slab
and the underside of the cofferdams. This was
performed about a week after casting and before
the six 3.8m diameter pile top blockouts were
flooded with a head of 3m for launching.
Complete release from the casting bed was
confirmed when water seeped out at the
perimeter of the cofferdam.

The modified design incorporated a number of

features essential to the success of the project:
First, the cut-off elevation of the drilled shaft

Fig. 24 Float-in cofferdams floating in storage at

dock side after launch, Carquinez Bridge
casings was set at 1.55m below mean sea level,
it was therefore necessary to limit the draught of
the float-in cofferdams to 2m. This was
accomplished by following modifications:
Fig. 23 Completed precast shell of link-beam on
deck of barge, Carquinez Bridge

For the launch operation, a site was selected in

the Carquinez Strait approximately 2-km
upstream from the bridge. The primary criteria
were that the site should have a level bottom
and a water depth at low tide of about 6.7m. The
barge was towed to the site and positioned
alongside a derrick barge on anchors. In order
to maintain sufficient stability of the casting
barge and cofferdams during the launch, the
barge was ballasted down one end at a time.
The stern of the barge was first set down on to
the bottom during low tide, and then, as the tide
turned, the bow was gradually ballasted down
allowing the cofferdams to float free one at a
time.
Following launch, each cofferdam was towed to
the dockside and stored afloat until the drilled
shafts at the first tower were completed. (See
Fig. 24.)

12

the wall thickness of the float-in cofferdam

was decreased,
semi-lightweight concrete with a specific
gravity of 2 was used,
Styrofoam blocks were used to form the
bottom ribs of the cofferdam, and finally,
each pile-top block-out in the bottom slab
(located over the top of each pile position)
was capped with airtight 3.8m diameter steel
cylinder and the water was forced out of the
pile compartments with compressed air.
Secondly, Caltrans set the horizontal positioning
tolerance for the 3m diameter drill shaft casings
at +/- 150mm, but site conditions including water
depths of up to 32m and tidal currents up to
3m/s made meeting these criteria difficult. In
order to meet the specified tolerances under
these conditions, a steel guide template with
four spud piles was installed at the south tower
and a template attached to the pile driving barge
was used on the north tower. As an added
precaution against fit-up problems during
landing of the cofferdams, the size of the
blockouts in the bottom of the float-in cofferdams
was increased to 3.66m. This provided a

theoretical clearance around the pre-installed

drilled shafts of 330mm.
Third, the system for cutting off the drilled shafts
under water included a 4.5m diameter cylindrical
cofferdam that fitted over the top of the drilled
shaft. The cofferdam sealed to the casing at
about 1m below cut-off elevation by means of an
inflatable rubber O-ring that fit inside a circular
recess at the bottom of the cofferdam. Following
inflation of the seal, the cofferdam was
dewatered and the drilled shaft casing was cut
off in the dry to a precise elevation. This cut-off
cofferdam also allowed access for welding a
steel ring to the exterior of the casing; the ring
was used to provide the watertight seal between
the float-in cofferdam and the drilled shaft
casings. It also provided the necessary shear
resistance for the support of the cofferdam
during the concrete infill operation. (See Figs.
19 and 20 for details of the bottom seal.)
Fourth, high tidal currents in the Carquinez Strait
meant there was only about an hour of slack tide
during which the cofferdams could be landed.
Hence it was necessary to install a guide system
to allow rapid positioning and landing. Because
the water was more than 32m deep, the guide
system was attached directly to the 3m diameter
drilled shaft casings. The guides were L-shaped
wide flange brackets attached to the casing just
below the landing support bracket and extended
above high tide so they could be used for visual
positioning of the cofferdams. (See Fig. 25.)

elevation. This provided an excellent bearing

surface for the landing of the cofferdams.
Caltrans initial design showed the top of the
casing protruding through the 3.2m diameter
holes in the bottom of the cofferdam and cut off
flush with the bottom floor of the precast
cofferdam. Landing support was provided by
steel beams spanning across four of the bottom
slab openings and bolted flush to the floor of the
precast cofferdam. The actual bearing surface
was the top edge of the drilled shaft casing and
the underside of the steel beam. (See Figs. 19
and 20 for details of the bearing surface.)

Fig. 26 Landed float-in cofferdam and

accessing seals through through piletop bulkheads, Carquinez Bridge

In order to dewater the cofferdam it was

necessary to create a watertight seal between
the six drilled shaft casings and oversized holes
in the bottom of the cofferdam. This was
accomplished by pre-welding doughnut-shaped
steel rings - 12mm thick by 700mm wide - to the
casings just below the cut-off elevation, at the
finished bottom elevation of the cofferdam.
These plates were attached in-the-dry by using
the same 4.5m diameter cofferdam used for
casing cut-off.

Fig. 25 Positioning and landing float-in

cofferdam at N. Tower, Carquinez Bridge
The 4.5m diameter cylindrical cofferdam allowed
the tops of the casings to be cut off to a precise

13

At the outer edge of the steel doughnut, an

inflatable O-ring was bonded to the plate. As the
void space around each casing was dewatered
the hydrostatic pressure pushed up on the
bottom of the steel plates and the rubber O-rings
provided the necessary seal. In areas where the

3.6m diameter holes in the bottom of the float-in

cofferdam. (See Fig. 20 for a detail of the void
space.) The 3.8m diameter cans positioned
over the top of each hole provided access to this
area and the seal system described above
allowed the cans to be dewatered and the infill
concrete to be placed in-the-dry. (See Fig. 27,
showing removal of the lids of the pile top
bulkheads.) Hanger rods 25mm in diameter and
spaced at 600mm along the center of the void
space provided shear reinforcement and the
fillet weld connecting the 12mm plate to the
drilled shaft casing provided the necessary
shear resistance.
Fig. 27 Float-in cofferdams and link-beam shell
in place, ready for placement of
reinforcing steel, Carquinez Bridge

After the concrete in the void space had reached

the specified strength, the entire cofferdam was
dewatered, the 3.8m diameter cans were
removed from each drilled shaft location, and
the initial support beams were removed. At this
point the area over the top of each drilled shaft
casing was free and open for extending the
drilled shaft vertical reinforcement up into the
footing, and installation of the footing bottom
reinforcement mat. (See Fig. 28 for a view of
the cleared area inside the float-in cofferdam.)
Conclusion
The successful use of the float-in cofferdam
system on two major bridges has demonstrated
the effectiveness of this method for constructing
pile-caps in deep water where the caps are
located below water but elevated off the bottom.

Fig. 28 Placing reinforcing cages in drilled

shafts at Carquinez Bridge
seal did not seal properly, vertical hanger bolts
were available to pull the 12mm plate tight to the
bottom of the cofferdam and compress the seal.
(See Fig. 20 for details of the seal.)
The initial support system for landing the
cofferdams was simple and allowed a quick
landing, but it was directly in the path of both the
vertical extension of the drilled shaft
reinforcement and the horizontal reinforcement
in the bottom footing mat, hence it had to be
replaced before the reinforcement could be
installed.
Final support was provided by placing infill
concrete in the void spaces between the 3m
diameter drilled shaft casings and the oversized

14

Bibliography
1. Anderson, H. V.: Underwater
Construction Using Cofferdams,
2. Bittner, R. B.: Tapping Technology,
BRIDGE design & engineering, Issue
No. 14, First Quarter 1999, pp. 40-41.
3. Bittner, R. B. : Fast-track float-in,
BRIDGE design & engineering, Issue No.
24, Third Quarter 2001, pp. 38-40.
4. Ratay, R. T.: Handbook of Temporary
Structures in Construction , Chapter 7, by
Ben C. Gerwick, Jr.
5. Stubb, F. W., Jr.: Handbook of Heavy
Construction, 3rd Edition, Chapter 23,
Cofferdams and Caissons, by Ben C.
Gerwick, Jr.