PT-13: Coastal and Ocean Engineering ENGI.8751 Undergraduate Student Forum
Faculty of Engineering and Applied Science, Memorial University, St. john’s, NL, Canada MARCH, 13
Paper Code. (PT-13 - Campbell)
Effects of Ice Loads on the Confederation Bridge
Donald Campbell
Memorial University
St. John's, NL A1C 5S7, Canada
[email protected]
ABSTRACT
As the longest bridge in Canada and the longest bridge over ice-covered water in the world, the
design and construction of the Confederation Bridge presented a unique engineering challenge. There
was no precedent for designing for the ice loads experienced by the bridge piers. As such, extensive
studies were conducted by the National Research Council (NRC) and various independent groups to
ensure the Confederation Bridge design was adequate for ice loads in the Northumberland Strait, but
also that the design was not overly robust.
Since the completion of the Confederation Bridge studies have continued on the ice loads
experienced by the bridge piers. These could prove very useful in determining the accuracy of original
calculated expected loads and methods that could be used to calculate ice loads in the future. Also,
engineers will be able to learn more about the effects of ice on fixed structures as the Confederation
Bridge ages.
1
INTRODUCTION
The Confederation Bridge (hereinafter referred to a the Bridge) is a 12.9 km-long bridge
connecting the Canadian Maritime provinces of Prince Edward Island and New Brunswick.
Construction on the Bridge began on October 7, 1993 and it was opened to traffic on May 31 1997.
The 65 piers (44 main piers and 21 approach piers) are designed as double-cantilever post-tensioned
concrete box girders with drop-in sections connecting them. The main piers are spaced 250 meters
apart and reach 40 to 60 meters above average sea level.
Most elements for the Bridge were built in a fabrication yard on the Prince Edward Island side of
the Strait and placed in sections by a large floating crane (one example of a completed pier base before
placement can be seen in Figure 1 below). Due to the shallowness of water near shore, the approach
piers had to be shored and constructed in-place. The construction of the Bridge involved the use of
PT-13 Campbell P.1
478,000 m3 of concrete, 58,500 tonnes of reinforcement and approximately 12,690 km of posttensioning cable.
Figure 1: Typical Frame Dimensions for the Bridge
(Source: Donald McGinn, P.Eng.)
2
ICE CONDITIONS
The Bridge was designed for a 100-year lifetime with a target safety factor of 4.0. As previously
mentioned, there was no precedent for the ice loading conditions on a structure such as the Bridge and
thus extensive studies had to be undertaken to accurately predict the expected loads on the piers. The
Northumberland Strait is ice-covered for four months in an average year, forming what are known as
first-year ice floes. In general, first-year ice is thinner and weaker than old ice that has lasted through at
least one summer thaw.
2.1
Ice Elements
As ice floes collide with each other in the strait, two common elements of ice deformation that
can occur are rafting and ridging. Rafting involves one floe sliding on top of another, doubling the
thickness where the two overlap. It usually occurs early in the season when the floes are relatively thin
and the amount of vertical movement required is minimal.
Ridging occurs when two (or more) ice floes collide and the resulting pressure causes the edges
of the floes to crumble. Ridges usually occur later in the season as the ice becomes too thick for the
floes to raft. They can also occur where an ice floe separates and open water is exposed (e.g. when a
floe fails around a bridge pier). The rubble above-water is called the sail and below-water is called the
keel. Initially this rubble does not have a high level of strength and poses a hazard only due to its
thickness, which is often significantly greater than the floe thickness. Over time a layer of consolidated
ice (up to 3 meters) forms as the water between the ice blocks freezes. Ridges with a thick consolidated
layer often cause the greatest hazard to structures in ice-covered water due to their strength, thickness
and frequency of occurrence.
Figure 2: Typical Ice Ridge Cross-Section (Source: www.ec.gc.ca)
PT-13 Campbell P.2
2.2
Ice Loads on the Bridge
Due to the lack of similar structures, ice loads on the Bridge were calculated using probabilistic
methods. Loads are governed by the lowest of either the driving force of the ice floe or the force
required for the floe to fail. Parameters for both driving force and ice failure load were determined by
researchers and it was determined that the ice failure load was lower. The load required for the failure
of a floe is a function of both the thickness and the ice strength when it comes into contact with the
pier. Due to the thickness and the added strength in the consolidated layer in ice ridges, the extreme
loading case was determined to be most likely to occur in March and April when ridges are often
partially consolidated.
The average number and standard deviation of freezing degree-days were taken (information for
Summerside, approx. 20 kilometres from the Bridge, was used) and used to determine the distribution
of the average thicknesses of ice flows. Field data and thermodynamic analysis were used to model the
thickness of the consolidated ridge core, which can be highly variable over short distances. The
thickness of each type of ice (sheet, consolidated layer, keel, etc.), along with the properties of the ice
can be used to determine the force applied on a pier as an ice floe breaks on it.
Also, researchers observed the concentration (c), mean diameters (!) and standard deviation (σ)
of ice floes for the months of March and April along with the mean velocity (!) of floe movement and
the waterline diameter (d). Assuming the floes to be roughly circular and using the mean diameter and
standard deviation to define the area (!), the density (!) of floes can be found using Eq. (1).
!
!=!
(1)
The density is then used to find the probability of impact in Eq. (2)
Pr ! = !!(! + !)
(2)
By inputting the thicknesses of the ice floes and probability of impact into a simulation system,
researchers were able to come to a conclusion for the design loads one the piers. Due to the conical
shape of the piers at sea level the ice often fails in flexure as is slides up the cone, resulting in
significantly lower loads than if the ice were to fail in shear. The resulting maximum load found for a
60° conical pier over a 100-year period was found to be 18.5 mega-newtons (MN) and 24.3 MN for a
10,000-year period. Analysis was also performed for a 55° conical pier and the 100-year and 10,000year loads were found to be 13.5 MN and 18.0 MN respectively.
2.3
Bridge Design
The Bridge piers were structurally designed and constructed to resist an ice load of 30 MN. To
mitigate shear loads and cause the ice to fail in flexure the piers were designed to be conical in shape
for the area 4 m below and 2.6 m above the mean sea level (M.S.L.). They were designed with a slope
of 52°, significantly lower than both the 60° and 55° piers indicating that the ice loads should be
smaller than those previously mentioned.
PT-13 Campbell P.3
Can. J. Civ. Eng. Vol. 28, 2001
months of March and April, when the
es are prevalent. The ice conditions
e driven by tidal currents, in addition
nd the natural circulation current in
y, the bridge “sees” approximately
nth; of this, approximately 75% will
e bridge axis once, resulting in redge lengths.
on a pier is limited by the force reture, and by the force driving the ice
In the absence of sufficient driving
ary to fail the ice feature cannot be
ce on the pier, X, is the minimum of
ce, V, and the force necessary to fail
Fig. 1. Typical pier arrangement.
he maximum values of W occur durherefore the analysis is focussed on
loads were determined using detercould readily be incorporated in the
of the overall analysis. This then alables in each model to be treated as
the ice failure force, W, the only ice
re partly consolidated ridges. Thus
he ice failure force for the ridge and
onmental driving force.
riving force is determined from the
Figure 3: Typical Main Pier Design Details
e ice floe that contains the ridge unse include wind drag, current
drag,
nature of these
surveys and
the orientation
the survey
(Source:
A probabilistic
approach
to analysis
of ice of
loads
for the Confederation Bridge)
y one of these may, or may not, be
lines provide an excellent record of the scouring resulting
he prevailing conditions, although it
from ridge keels to a water depth of about 20 m, and hence a
One factor
that can
significantly
affect “seen”
the rate
ice is the friction between the
k ice force will be present
in the abrecord
of the keel depths
by of
thatabrasion
portion ofdue
theto
seaicenotand
pier surfaces.
surfaces
canconducted
be quiteinrough
and can
become
her two, and is
present
in open Concrete
bed. The surveys
were
the spring,
as soon
after even rougher as the cement
ee force components
must beaway
addedand ice-out
as possible,
and before
any significant
scour
paste wears
the aggregate
becomes
exposed.
In order
to healing
decrease friction and the slow the
forces depend,
to some
on it was
had occurred.
The keelthat
depth
wasused
derived
from
wear
on theextent,
surface,
recommended
thedistribution
steel forms
when
pouring concrete cones be left
orthumberland Strait, the fetch can be
the number of scours in a given depth range, averaged for
on
as
“ice
shields”.
In
general,
steel
is
a
much
smoother
surface
than
concrete
and is less susceptible to
e direction being considered. For the
the number of scour survey lines, and normalized for the toabrasion,
therefore
the friction
andpassing
wear on
theany
piers.
to the bridge axis,
the available
fetch reducing
tal number
of ridges
along
line. Corrections
neral, however, regardless of condiwere introduced for possible healing of scours during the
driving force to fail the ice features,
course of the winter, variations in the amount of ice on any
3
CURRENT CONDITIONS
ated by the ice feature failure force,
line, and shielding of the seabed. The two winters for which
results were available (1992–1993 and 1993–1994) were relitstocompletion
there
hasicebeen
regular
monitoring
thedistribuBridge and the effects of ice loads
ated ridges, the forceSince
required
fail
atively
severe
seasons,
and hence
the keel of
depth
o components: on
thatthe
resulting
from
the
tion
observed
from
the
scour
record
may
have
been
piers. Most notable are the effects seen on the piers with steel more
ice shields versus those without,
consolidated layer and that from the
severe than the norm. The resulting keel depth distribution,
and
the
magnitude
of
observed
ice
loads
compared
to
the
design
loads.
sually only considering the rubble in
when compared with first-year ridge keel depth distributions
is unlikely that the two maximum
for other regions, showed very similar characteristics, allow3.1 Ice
y are usually assumed
to Shields
do so. This
ing for differences in climatic conditions. This provided
on, as the scale of the larger ridges is
some confidence that the keel depth distribution based on
a number of flexural failure peaks as
observations of seabed scouring was a reliable data set.
When construction
crews began using the steel ice shields simply for formwork instead of
pier, while the maximum keel load
keeping
attached to themethodology
conical surface of the pier shaft it was believed by some to
uch of the same
period.them permanently
3. Probabilistic
be an error.
Steel
ters in the determination
of the
ridgeis significantly less coarse than average concrete, which generally becomes even
layer thickness
and strength,
keelas the
3.1.cement
Procedure,
solution,toand
annual
rougher
over time
pastestructure
has beenof known
erode
and leave a very uneven surface as the
operties, and pile-up geometry and
variations
aggregate is exposed. TheThe
decision
to discontinue the use of steel shields was made base on the finish
l are important, more effort was foice loads were determined using probabilistic methof
the
high-strength
concrete
used
in
construction.
h distribution, as related data became
ods. The simulation
procedure is outlined in Fig. 2. The
r environmental assessment
was years
unquantities that
were considered
being random
shownshields on some of the piers
In recent
significant
damage
has beenas noticed
to thearesteel
ide-scan sonarranging
surveys of
ice scourin Table and
1 (Cammaert
et al.
In thewhere these damages could
from
mild abrasion
corrosion
to1993;
largeBrown
tears.et al.
In 1995).
instances
and Associates 1994, 1995). The
simulation, freezing degree-days and concentration were
possibly prove hazardous to bridge maintenance workers and others who may need to work near the
© 2001 NRC Canada
PT-13 Campbell P.4
piers, it was decided that the ice shields must be removed. Currently ice shields remain on all of the
approach bridge piers and on main piers P1, P2 and P10.
The ultimate failure of the ice shields was caused by two separate conditions. Firstly, as water
that was trapped between the pier and shield froze it would expand; this expansion could have lead to
local deformations in the steel. Secondly, the cathodic protection used to protect the steel from
corrosion was ineffective in some areas leading to corrosion of both the internal and external faces of
shields, significantly weakening the steel. It is likely that the shields on the approach bridge piers have
remained in better condition than those on the main pier because the fast ice located close to shore does
not put as much stress on the shields. Nevertheless, even the ice shields on the approach bridge piers
have corroded significantly.
Figure 4: Corrosion of the Steel Ice Shields on the Approach Bridge Piers
(Source: www.confederationbridge.com)
For those piers that were placed without the ice shields, the cement paste in the concrete has worn
slightly over the years since the Bridge was completed, but studies have shown that the paste is
sufficiently strong for the aggregates to control the abrasion process. This means that the surface
texture has stabilized after slight wear and slowed the rate of abrasion. An abrasion rate of less than
35mm over 100 years was originally predicted and it has been confirmed that the current rate of
abrasion is less than 0.40mm per year. The only areas in which the rate of abrasion has not yet
stabilized to this rate are where the concrete was over-vibrated during placement, decreasing the
quantity of aggregate near the surface. The concrete piers have proven durable thus far and there are
currently no plans to replace the steel ice shields or otherwise mitigate loads on the piers other than
those already in place.
3.2
Observed Ice Loads
The Bridge was designed to resist maximum ice loads on the piers of 30 MN. Two tiltmeters,
located in each pier shaft, track any movement of the pier with an accuracy of 0.1 micro-radians (µrad).
Due to the large side area of the girders at the pier shafts it is also necessary to measure the wind
velocity so that the tilt caused by ice loads is accurate. When the tilt due to wind loads is subtracted
PT-13 Campbell P.5
!
!"
from the measured tilt, the
load on the pier can be calculated using a conversion
"#"$%&$'()!
* factor (units: !"#$) that
is unique for each individual pier. The largest measured ice load on the Bridge in the 16 years since
completion has been 8MN, suggesting that the design load of 30 MN was quite conservative.
!
Figure
5:
Location
of
Tiltmeters
in
the
Pier
Shafts
Figure 3: The bridge piers (L) and P23’s inner cross section (R)
(Source: Response of Confederation Bridge to Ice Forces)
Analyzed4DataCONCLUSIONS
The ice loads on the Bridge were a critical consideration in the design and have potential to cause
serious damage if they are larger than anticipated. Extensive research has occurred both before and
after construction to ensure that the design is sufficient to withstand even extreme ice conditions for the
$"#"!,-.!/001!20-.34516!-17!-1-89.516!7-:-!;4<2!:,0!:58:20:04.!.51=0!>???@!%,0!
area. Given the unique nature of the Bridge for both its size and environmental loads, it was necessary
:04!5.!-/80!:<!20-.340!/5-A5-8!51=851-:5<1!-1680.B!-17!:,0!=<<4751-:0!.9.:02!70;510.!
that many conservative assumptions and estimates had to be made when calculating loads. The design
<3:!:,0!8<165:3751-8!-A5.!<;!:,0!/45760!-.!:,0!A$-A5.!-17!:58:!-/<3:!-!,<45C<1:-8!-A5.!<;!
load of 30 MN is on par with the maximum anticipated load of 24.3 MN, and the maximum observed
-1680!:<!:,0!/45760!-.!:,09$-A5.@!%,0!=<<4751-:0!.9.:02!5.!-8.<!5175=-:07!51!D56340!
load of 8 MN indicates that this should indeed be sufficient for the expected life of the Bridge.
1=0! :,0! G407<251-1:! 5=0! 745;:! 7540=:5<1! 5.! ;4<2! 3G.:40-2! H+IJ! :<K-47.! :,0! /45760B!
The successful design and construction of the Bridge, along with its performance since
2-L<45:9! <;! 5=0! 52G-=:! ;<4=0.! =-3.0.! :,0! G504! :<! :58:! -/<3:!:,0! 9$-A5.@! "<1.0M301:89B!
completion, is a major feat for Canadian Engineering. The methods used in calculating the ice loads
-:-!;4<2!9$-A5.!5.!3.07!:<!=-8=38-:0!:,0!5=0!8<-7@!
and the continuing studies of ice effects on the piers set an important precedent for the design of
structures in ice-infected water worldwide. Engineers will be able to take examples and lessons learned
from the Bridge to improve designs and construction for many years to come.
Tilt Data
PT-13 Campbell P.6
!
REFERENCES
[1] Brown, T.G, Jordaan, I.J., and Croasdale, K.R. (2001). “A probabilistic approach to analysis of ice
loads for the Confederation Bridge”. Canadian Journal of Civil Engineering, Volume 28 (Number
4), pages 562-573
[2] D.J. McGinn, personal communication, March 11, 2013
[3] National Research Council Canada, “Response of Confederation Bridge to ice forces : Winter
2008-2010”, CHC-TR-075, 2010
[4] (2013, March 10). The Confederation Bridge. Retrieved from http://www.confederationbridge.com
[5] Manual of Standard Procedures for Observing and Reporting Ice Conditions (2013, March 10).
Canadian Ice Service. Retrieved from http://www.ec.gc.ca/glaces-ice/
[6] Cammaert, A.B., Jordaan, I.J., Bruneau, S.E, Crocker, G.B., McKenna, R.F., and Williams, S.A.
“Probabilistic Analysis of Ice Loads on Conical Bridge Piers for the Northumberland Strait
Crossing Project”. Contract report for Stanley Atlantic Inc., C-CORE Contract Number 93-C2
PT-13 Campbell P.7