-
The Institution of Engineers of Ireland
The Design and Construction of the Boyne Bridge Joe ODonovan BE
CEng FIEI MICE MConsEI Managing Director, Roughan & ODonovan
Ltd Keith Wilson MA CEng MICE Chief Bridge Engineer, Roughan &
ODonovan Ltd Pat Maher BE MEngSc MSc Dip Highway Eng CEng MIEI MICE
MIStructE Bridge Manager, National Roads Authority (formerly Chief
Bridge Engineer, Roughan & ODonovan Ltd) Paper presented to the
Institution of Engineers of Ireland Dublin March 4th 2003 Cork
March 11th 2003
-
1
ABSTRACT The Boyne Bridge carries the M1 Northern Motorway over
the River Boyne approximately 3 km west of Drogheda. The paper
briefly describes the environmental issues which influenced the
preliminary design before describing the bridge as finally
designed. The bridge has a number of innovative features, such as
wind shielding and unpainted enclosed steelwork, and these are
described together with the reasons for their inclusion. Other
significant design aspects which arise as a direct result of the
form of structure are discussed. The procurement strategy for the
bridge is described including the development of a supplemental
agreement in which the contractor assumed a much greater degree of
risk in return for a lump sum payment and the opportunity to
develop a deck design which maximised his fabrication and erection
expertise. The main differences between the two designs are
described. Finally, the key aspects of the pylon construction and
the erection procedure for the deck are outlined.
-
2
1. THE NORTHERN MOTORWAY (GORMANSTON TO MONASTERBOICE)
The Boyne Bridge is located on the Gormanston to Monasterboice
section of the M1 Northern Motorway. This 22km long section of
motorway connects the Balbriggan Bypass to the Dunleer Bypass and
is a key element in the major road network of the country. In
February 1998, Northconsult, a consortium of consulting engineers
was appointed by the client to carry out the detailed design and to
supervise the construction stages of the motorway. The consortium
consisted of:
M C OSulllivan Ltd Atkins Ltd Roughan & ODonovan Ltd.
The Boyne Bridge was designed by Roughan & ODonovan Ltd as
part of the above consortium.
2. THE CLIENT The proposed Northern Motorway involved a number
of Local Authorities Meath County Council, Drogheda Corporation and
Louth County Council. The latter two authorities, however, under a
section 59 agreement of the Local Government Act (1995) confirmed
that Meath County Council would exercise all their functions in
relation to the motorway and its associated link roads.
3. SITE OF THE BRIDGE The Boyne Bridge is located about 3km west
of Drogheda, see Figure 1. The bridge crosses the river
approximately at right angles in a north-south direction, see
Figure 2. The River Boyne is approximately 150m wide and consists
of two channels separated by Yellow Island. The river is tidal at
the crossing. At the bridge, the river valley has an asymmetrical
cross-section with a steep escarpment to the south of the river and
a gently rising slope from the river edge going northwards.
-
3
Figure 1 Bridge Location
Figure 2 Bridge Site
4. MOTORWAY AT THE BRIDGE The motorway is about 20m above the
normal water level of the river and the distance between the points
where it departs from ground level on either side of the valley is
approximately 400m. The bridge cross-section will accommodate two
carriageways, each with a road paved width of 11.5m in the short
term and 12.75m in the long term, see Figure 3. The short term
width of 11.5m comprises a 3.0m
-
4
hard shoulder, 2 x 3.75m wide lanes and a 1.0m wide shoulder on
the median side. In the long term, provision is made for an
additional 3.75m wide lane by reducing the width of the median,
reducing the hard shoulder width to 0.5m and adjusting the lane
positions.
Figure 3 Motorway Cross-section at Bridge
5. PRELIMINARY DESIGN OF THE BOYNE BRIDGE
The preliminary design of the Boyne Bridge was completed in June
1994 and in April 1995 the Environmental Impact Statement (EIS) was
published. The EIS for the Boyne Bridge was separate from that done
in respect of the motorway and was the first formal bridge EIS to
be carried out under the Roads Act 1993 and Part 5 of Roads
Regulations 1994. A paper on the preliminary design of the bridge
was presented to the North-East Region IEI in November 19961. That
paper detailed the environmental issues and the constraints which
influenced the bridge design ultimately selected and these may be
summarised as follows: (a) Historical and Heritage The Boyne Valley
is one of the most historically significant sites in Ireland. The
ancient monuments of Newgrange, Knowth and Dowth lie within 3-5km
of the crossing. The three main Williamite crossings of the river
at the Battle of the Boyne in 1690 took place adjacent to the
bridge location. (b) Flora and Fauna The EIS concluded that the
site for the proposed bridge represents the most ecologically
important area in the proposed motorway. Eight
-
5
distinct habitats were identified within the environs of the
river and all show a diversity of flora and fauna species. (c)
Amenity Walkers and anglers make considerable use of the southern
bank of the river along the Oldbridge Road. (d) Physical Within the
river channels, on Yellow Island and in the tidal reed beds to the
north of the river, the presence of soft alluvial conditions,
mudflats and fluctuating water levels would make the construction
of bridge substructures very difficult irrespective of
environmental considerations. (e) Aesthetic Given the outstanding
natural beauty of the Boyne Valley, the scale and height of the
bridge at the crossing, the aesthetic merit of any design was
regarded as probably the single most important factor to be
considered. Accordingly the design team took the view that the
bridge design should as far as possible achieve the following
aims:
be compatible with the river valley, make a positive visual
statement, cause minimal interference with flora and fauna, cause
minimal intrusion onto the Battle of the Boyne Site, cause minimal
( even if temporary) loss of amenity during
construction, and be the design which best meets the design
constraints and not
necessarily the cheapest design.
6. GENERAL DESCRIPTION OF BRIDGE Span Arrangement The elevation
of the bridge is shown in Figure 4. The deck spans, centre to
centre of bearings, are as follows:
Back span 42.5m Main (cable stayed) span 170.0m
-
6
Side Span 1 45.0m Side Span 2 40.0m Side Span 3 30.0m Side Span
4 25.0m Total Length 352.5m
Figure 4 Elevation of Bridge
Pylon The pylon is an inverted Y in elevation, extends some 93m
above the top of the pile cap and is constructed in reinforced
concrete. The legs and the pylon head are hollow being generally
rectangular in section and tapering in both front and side
elevations. The pylon legs vary in section from 5.0m x 4.8m at the
base to 4.2m x 3.865m at the underside of the pylon head. Back
stays and fore stays are stressed and anchored from the inside of
the pylon head. Full access for both personnel and equipment is
provided to the pylon head through the hollows in the pylon legs
and head. There is no access from the pylon into the deck void. The
legs of the pylon are connected with a concrete crossbeam at
underside of deck level. The cross beam is rectangular in section
4.0m wide x 3.9m deep, is hollow inside and supports the deck at
two points transversely through mechanical bearings. The 7m x 7m x
3.5m deep concrete plinths at the bottom of each leg of the pylon
are connected across with a 4m wide x 2.5m deep concrete tie beam.
Each plinth is supported by a 17.5m x 16m x 2.5m deep reinforced
concrete pile cap while each pile cap sits on 16 no. 1.5m diameter
concrete bored piles.
-
7
South Abutment The south abutment provides a gravity anchorage
structure for the back stays of the pylon, see Figure 5. It is a
reinforced concrete multi-cellular structure 32.6m x 41.0m x
approximately 7.5m high. Many of the cells are fully or partially
filled with Class 10/40 concrete. Between the structural concrete
and the infill concrete some 17,000 tonnes of
Figure 5 South Abutment
concrete is provided in the abutment. However, permanent access
is provided to the interior of the abutment so that the anchorages
to the stays can be inspected. The abutment is founded on a spread
footing at an average depth of 5m below existing ground level and
in service conditions exerts a lower ground pressure than the
original insitu condition. Access into the bridge deck is provided
from the abutment. Northern Approach Span Substructures The
intermediate supports for the northern approach spans consist of
pairs of 2m diameter reinforced concrete circular columns. The
columns generally decrease in height as the ground gently rises
going north, column heights varying from 14m to only 5.5m. Columns
are supported on individual reinforced concrete pile caps. The
northern abutment is a simple reinforced concrete cantilevered
abutment and wing walls which is supported on bored piles. Access
into the bridge deck and inspection gallery for bearings and the
movement
-
8
joint are provided at this abutment. All piles for the northern
approach span substructures are 750mm diameter. Bridge Deck The
cross-section of the deck as designed, shown in Figure 6, is
basically the cross-section used for all spans of the bridge. The
deck structure consists of a 230mm thick reinforced concrete slab
acting compositely with a steel space frame giving a structure with
high transverse and torsional stiffness. The space frame is mostly
constructed from circular hollow sections (CHS) and varies in depth
(centre to centre of chords) across the section from 2.7m (on the
longitudinal centre-line) to 2.3m.
Figure 6 Bridge Deck
The space frame module is 5.45m transversely and 5.0m
longitudinally, the latter dimension matching the 10m stay cable
spacing. The deck possesses high torsional stiffness resulting in
good aeroelastic stability. The majority of nodes where members
converge can be fabricated by profiling the ends of the members and
welding them together. At some positions, however, cast steel nodes
are necessary because of the effect of node geometry on the design
stress levels. The manufacture of cast steel nodes is a specialist
activity and the design of the nodes would be carried out by the
specialist supplier of the castings. All of the structural steel is
enveloped by a pultruded glass reinforced plastic (grp) enclosure
so that conditions close to indoor prevail within the enclosure. It
is expected that, in these conditions, the steel will require very
low or possibly no maintenance in the future. The enclosure
provides a ready made access platform for the future inspection of
the steelwork, deck soffit and movement joints.
-
9
Stay Cables The main span of the deck is supported by an
inclined plane of stay cables at each side of the deck. The
fourteen stays on each side fan out in a semi-harp arrangement from
the top of the pylon and pick up the deck at 10m centres with a
22.5m gap between the pylon and the first stay and a further 17.5m
gap between the last stay and the pier support at the end of the
main span. The semi-harp arrangement of stays provides reasonable
space for stressing purposes whilst maintaining the advantage of
steeper inner stays. The number of back span cables are the same as
those of the main span cables and their lines of action intercept
on the centre line of the pylon. The back stay cables are arranged
in two parallel groups in a harp arrangement and are anchored at
each side of the abutment structure. The inclined geometry of the
cables together with the vertical line of anchorages in the pylon
result in a visually interesting warped plane of the cables on each
side of the deck. This feature, in conjunction with the portal
character of the pylon produces a particularly striking appearance
to motorists approaching the bridge from north to south. The design
envisaged the use of either parallel strand or parallel wire stay
cables. Each parallel strand stay was to consist of:
(a) a group of 15.7mm diameter seven-wire galvanised strands,
each strand enclosed within a wax filled high density polyethylene
(hdpe) coating, and
(b) a hdpe outer protective sheath, resistant to weathering and
ultraviolet light radiation, surrounding the grouped strands.
Each parallel wire stay cable was to consist of:
(a) a group of 7mm diameter galvanised prestressing wires, (b) a
petroleum wax corrosion inhibiting medium surrounding
the wires and filling the space inside the outer protective
sheath, and
(c) a hdpe outer protective sheath, resistant to weathering and
ultraviolet light radiation, surrounding the grouped wires.
-
10
7. DESIGN OF THE BRIDGE Wind Considerations The prevailing winds
at the site of the Boyne Bridge blow from a south westerly or
westerly direction, i.e. along the Boyne Valley, which has a
funnelling effect on the wind speeds. The road deck of the bridge
is 20 m above the river level. It is therefore probable that the
wind conditions on the bridge will be worse than elsewhere on the
motorway and could cause problems for drivers of wind susceptible
vehicles. Experience on the West Link Bridge in Dublin and the
Second Severn Crossing in Britain has shown that the provision of
porous windshields along the edges of the bridge deck can
significantly reduce the effects of high winds on traffic using the
bridge without having too much of an adverse effect on the
aerodynamic performance of the bridge. The issues involved are
complex and are succinctly described by Wilson2. A specialist study
was commissioned from FaberMaunsell Ltd which compared the expected
wind regime on the bridge deck with that on the adjacent lengths of
motorway and approach roads, and with other exposed crossings in
Europe. The periods of restriction likely to be experienced by
traffic using the bridge are summarised in Table 1.
Table 1 Likely Restrictions due to Wind (per year)
Limit Ground Level Approach Embankment Bridge Deck
Hrs Days Hrs Days Hrs Days Without Windshields Loss of
directional control 124 56 170 63 600 124 WSVs may overturn 5 4 9 6
66 34
With 2m high Windshields Loss of directional control 197 65 WSVs
may overturn 8 8
Note: WSV = wind susceptible vehicle The report concluded
that:
(a) without windshields or traffic control measures, there is a
higher risk of traffic accidents on the bridge compared with an
equivalent length of ground level road;
(b) this risk can be significantly reduced by the provision of
windshields, to a level approaching that on the approach
embankments; and
(c) 2 m high windshields with a 45% porosity should be provided
along the bridge.
-
11
The Preliminary Report therefore included a recommendation that
wind shielding be provided on the Boyne Bridge because:
(a) the risk of high-sided vehicles overturning would be reduced
to a level which would be comparable to that risk along
considerable lengths of the Northern Motorway, both existing and
proposed. Therefore, the need to divert such vehicles through the
existing local road network would not arise;
(b) there will be a reduction in the periods when difficult
handling would be experienced by drivers of high-sided
vehicles;
(c) there will be no need to rely on observance of warning signs
by drivers of high-sided vehicles for implementation of traffic
control measures; and
(d) there will be no requirement for enforcement by Garda of
traffic control measures.
The layout of the wind shielding is shown in Figure 7. It is 2.1
m high, inclined at an angle of 20 to the vertical to match the
inclined cable stays, and has three transparent polycarbonate
panels 370 mm wide with 380 mm clear gaps between them giving a
porosity of 49%.
Figure 7 Wind Shielding
-
12
The windshields, although porous, present a bluff face to wind
blowing over the bridge and reduce the stability of the deck
against aerodynamic excitation. It is therefore essential that the
aerodynamic response of the deck and windshields is accurately
assessed and this is best achieved by carrying out tests on a
sectional model in a wind tunnel, which also provides the
opportunity to measure the shielding effect of the windshield and
the drag forces on the bridge deck. In early 1996 BMT Fluid
Mechanics Ltd in Britain was commissioned to undertake a series of
tests using a 1:70 scale rigid two-dimensional sectional model in
smooth flow conditions, Figure 8. The full-scale
Figure 8 Sectional Model in Wind Tunnel
properties of the deck are given in Table 2. The scope of the
tests was to:
(a) determine the critical wind speed for the onset of divergent
instability of the deck with windshields at five angles of wind
incidence and three levels of structural damping, for two values of
flexural/torsional frequency ratio;
(b) determine the vortex shedding response of the deck with
windshields;
-
13
(c) determine the static drag force, lift force and moment
coefficients;
(d) determine the static pressure distribution around the deck,
and (e) determine the windshield efficiency.
Table 2 Full Scale Properties of Deck
Mass 40,890 kg/m Polar mass moment of inertia 5,267,000 kg.m2/m
Fundamental bending frequency 0.482 Hz Fundamental torsional
frequency 0.699 Hz Overall width 36.0 m Overall depth 3.725 m
Centre of twist 3.157 m above soffit
The main aerodynamic mechanisms leading to divergent amplitude
oscillations are galloping (or stall flutter) and classical
flutter. The former arises on certain shapes of deck cross section
due to the variation of drag, lift and pitching moment with the
angle of wind incidence or time, while the latter involves coupling
between vertical and bending oscillations and depends primarily on
the separation of the torsional and bending natural frequencies.
This type of motion is potentially catastrophic and must be
avoided. Appropriate factors of safety to be applied to the
predicted extreme mean hourly wind speed are given in BD49/01 as
implemented by the NRA3. For the Boyne Bridge, the critical wind
speed for the onset of divergent amplitude oscillations had to be
greater than 55 m/s. The results of the wind tunnel tests are
summarised in Table 3. The only result which falls below the
critical wind speed is for very low damping with the wind inclined
at 5 below the horizontal. The site is relatively level transverse
to the bridge centre line and there is no reason to expect any
severe inclination of the wind. During the studies for the Severn
Suspension Bridge, measurements of wind inclination were taken at
the site of the Severn Rail Bridge4. These results were reviewed
during the studies carried out for the Second Severn Bridge and the
following conservative generalisation proposed for the variation of
extreme wind speed with angle of incidence:
80% of critical horizontal wind speed (44 m/s for Boyne) at 2.5,
and 50% of critical horizontal wind speed (27.5 m/s for Boyne) at
5.
The results for the inclined winds do not exceed these figures
and hence the Boyne Bridge with windshields will not be susceptible
to divergent amplitude oscillations.
-
14
Table 3 Critical Wind Speeds for Divergent Amplitude
Instability
Frequency Ratio = 1.45 Frequency Ratio = 1.38Wind Incidence
(deg)
Torsion Damping (log dec)
Critical Wind Speed (m/s)
Critical Wind Speed (m/s)
-5 0.01 > 80 > 80 -2.5 0.01 > 80 > 80
0 0.01 > 80 > 80 +2.5 0.01 71 66 +2.5 0.04 > 80 > 80
+5 0.01 56 54 +5 0.04 65 62 +5 0.05 68 67
Limited amplitude oscillations may be excited by the periodic
cross-wind forces arising from the shedding of vortices alternately
from the upper and lower surfaces of the bridge deck. Over one or
more limited ranges of wind speed, the frequency of excitation may
be close enough to a natural frequency of the structure to cause
resonance and consequently cross-wind oscillations at that
frequency. The design of the bridge must consider this type of
response when there are critical wind speeds below the reference
wind speed specified in BD49/01. For the site of the Boyne Bridge,
the reference wind speed is 42.4 m/s. The results from the wind
tunnel tests are summarised in Table 4 and show that vortex-induced
oscillations will only occur when the inclination of the wind is
below the horizontal. For the bending mode there were two distinct
responses, one at approximately twice the wind speed of the other.
The first response is associated with alternate shedding of
vortices in the wake of the deck; the second response is associated
with the shedding of vortices across the deck due to separation at
the leading edge with reattachment further downstream.
Table 4 Vortex Shedding Response (N/Nz =1.45, = 0.04)
Wind Incidence (deg) -3 0 +3 Critical wind speed (m/s) 16.5
Bending
(1st Response) RMS amplitude (mm) stable stable 49
Critical wind speed (m/s) 31 Bending (2nd Response) RMS
amplitude (mm) stable stable 76
Critical wind speed (m/s) 28.1 Torsion RMS amplitude (deg)
stable stable 0.25
-
15
The tests only modelled the fundamental bending and torsional
frequencies, but vortex-induced oscillations can also occur at
higher frequencies at similar values of reduced wind speed (V/NzD).
Table 5 summarises the response of the structure to aerodynamic
excitations using the wind tunnel tests, the natural frequencies
calculated during detailed design and BD49/01. Using the research
described earlier, the reference wind speed for winds inclined at
+3 was estimated to be 0.73 x 42.4 = 31.0 m/s. Hence, the second
bending response and torsion will not be excited for the final
bridge design. The first bending response will be excited at the
first two bending natural frequencies, although the magnitude and
number of oscillations are reasonably low, so that fatigue due to
vortex-induced oscillations will not be critical. It is worth
noting that should the torsional natural frequency fall to 0.7 Hz
(the estimated value at the preliminary design stage) the expected
dynamic loads could be as high as 80% of the HA loading, although
this would occur on less than 100 occasions during the design life
of the bridge.
Table 5 Response to Aerodynamic Excitation (wind incidence =
+3)
1st Bending 2nd Bending Torsion Reduced critical velocity 9.2
17.3 10.8 RMS amplitude 49 mm 76 mm 0.25 1st Bending Mode, Nz = 0.5
Hz Critical velocity (m/s) 17.1 32.2 Dynamic load (kN/m) 19.8 30.7
No. of cycles in 120 yrs 55,000 2nd Bending Mode, Nz = 0.84 Hz
Critical velocity (m/s) 28.8 54.0 Dynamic load (kN/m) 55.8 86.6 No.
of cycles in 120 yrs < 100 1st Torsion Mode, N = 0.70 Hz
(Preliminary Design) Critical velocity (m/s) 28.1 Dynamic load
(kNm/m) 445 No. of cycles in 120 yrs < 100 1st Torsion Mode, N =
0.95 Hz (Final Design) Critical velocity (m/s) 38.2 Dynamic load
(kNm/m) 818 No. of cycles in 120 yrs HA Loading (LL 170 m) 120 kN/m
120 kN/m 550 kNm/m
Exceeds reference wind speed Not applicable
-
16
The measured static force coefficients were lower than those
given in BD37/885 and were adopted for the detailed design.
Measurements were also taken of the static pressure distribution
around the deck section in order to obtain local design wind
pressures to be used in the design of the enclosure system, Figure
9.
Figure 9 Wind Pressure Distribution
The efficiency of the proposed windshields was investigated by
measuring the side force and overturning moment on a model of a
furniture van positioned on the deck, Figure 10. Measurements were
taken with and without the windshields. The results are summarised
in Table 6.
Figure 10 Model Furniture Van
-
17
Table 6 Performance of Windshields
Ratio with wind shields / without wind shields Wind Incidence
(deg) Side Force Overturning Moment
-5 0.74 0.89 -2.5 0.59 0.81
0 0.42 0.69 +2.5 0.25 0.43 +5 0.18 0.32
The results are similar to or better than those assumed in the
wind shielding study for side force at wind incidences greater than
-2.5 and for overturning moment at wind incidences greater than 0.
For winds with more negative angles of incidence (i.e. blowing from
above the horizontal), the barrier is less efficient. However, as
argued above, the ground surface at the site of the bridge is
relatively level and there is no reason to expect any inclination
of the wind other than that due to its turbulence. Further,
reference 4 suggests that although wind inclinations of 5 or more
could occur with wind speeds at which loss of directional control
is possible, 35 mph (15.6 m/s), these are of very short duration
and occur very infrequently. It was therefore concluded that the
proposed windshields were appropriate. Foundations The bridge site
lies on the northern margins of carboniferous limestone known as
the Platin limestone. The Lower Palaeozoic rocks of County Louth
lie less than a kilometre to the north. The Platin limestones form
a belt extending from the southwest near Duleek, to the north-east
through the Platin cement works to Sheephouse and Oldbridge and are
folded into a shallow syncline, which plunges towards the
southwest. The straight contact between the limestone and the
Palaeozoic rocks is indicative of a step and probably faulted
boundary. Earlier borehole investigations indicated that the
limestones in the area are strongly jointed and fissured, with some
of the fissures being filled with a variety of sands, silts and
clays. The presence of fissuring and voiding is indicative of the
karstic nature of the Platin limestone. The presence of karstic
phenomena in the area was well known. Drybridge, to the north east
of the site is indicative of a karstic bedrock. During the
construction of the Platin cement works, a significant number of
fissures and voids were discovered during foundation excavations.
Furthermore, during the construction of the Boyne viaduct,
considerable foundation problems were encountered in the
construction of a river pier because of major variations in rock
head level across the foundation, a phenomenon
-
18
again indicative of karstic limestone. Thus all of the evidence
available to the design team pointed to the potential for difficult
foundation construction arising from the nature of the limestone.
The carboniferous limestone is overlain by glacial drift,
comprising both gravels and boulder clays, within and in the
vicinity of the Boyne Valley. Locally, the river is flanked by
low-lying alluvial deposits on the north bank and at Yellow Island.
Yellow Island itself is an alluvial spit, which is completely
inundated at each high tide, during periods of high flows in the
river. It can be inferred that the valley of the Boyne was at one
stage deeper and wider than currently, but has been infilled with
outwash material from melting glacial ice. Evidence of this is
apparent on both banks of the river where gravel pits are exposed.
In overall terms the geotechnical characteristics of the site can
be summarised as follows:
site underlain by karstic limestone, which was expected to be
fissured and voided;
the northern bank of the river and Yellow Island consisting of
soft alluvial deposits; and
overburden deposits on the northern and southern approaches to
the bridge consisting of a mixture of gravels and boulder
clays.
Preliminary site investigations consisting of geophysical
investigation and nineteen boreholes were undertaken in late 1993.
The cross section of the valley interpreted from the boreholes is
shown in Figure 11. On the south side of the river above the
escarpment, boreholes indicated a
Figure 11 Geological Longitudinal Section
depth of overburden of 10 to 15 metres. On the riverbank at
Oldbridge Road, rock lay close to the surface, at depths of from
1.8m to 4m (0.8 to 2.7m OD). Rock head level fell towards the north
reaching depths of 24m (-18m OD) on the edge of the alluvial
wetlands. Rock coring confirmed karstic characteristics to some
degree in almost all of the boreholes. Empty voids (one up to 5m)
were encountered in a number of
-
19
boreholes. In addition, voids filled with clays and sands were
encountered. Bands of highly weathered rock, with very poor core
recovery, were encountered to considerable depths in the rock. In
the wetlands on the northern bank of the river, approximately 5m of
soft alluvial and organic material overlay from 10 to 15m of
gravels above bedrock. Because of concerns in relation to the
presence of fissures and voids within the rock, a foundation
arrangement consisting of large diameter bored piles, socketed into
rock, was devised. Given the very large loads from the pylon
structure, a pile group of sixteen 1500mm diameter piles, each with
a working load of 1350 tonnes was used to transmit the load from
each pylon leg to the more competent limestone at depth. In the
case of the piers supporting the northern approach spans, an
arrangement of nine 750mm diameter piles, each with a working load
of up to 370 tonnes, was adopted. The northern abutment was also
piled to rock using 750mm diameter bored piles. In the case of the
southern anchorage abutment, the structure was designed as bearing
on the overburden material, on the basis that the bearing pressure
under working conditions was little different from the existing
pressure at that level in the overburden. Notwithstanding the use
of bored piles socketed into rock, and the general improvement of
rock quality with depth, the random and unpredictable occurrence of
voids and fissures was such that there was always a risk that a
void or a zone of very heavily weathered rock could be encountered
below the toe of any given pile. Two expedients were adopted in
order to deal with this design risk. Firstly, the rock sockets were
designed as friction piles only, i.e. no end bearing capacity was
assumed in the design of the rock socket. Secondly, a skin friction
value of 480 kPa was adopted, a value significantly lower than
would normally be used for a rock socket in limestone and
reflecting the possible presence of low RQD rock within the socket.
In order to confirm appropriate pile design depths and prove that
no voids existed in the zone of rock beneath the pile toes, a two
stage probing sequence was specified. The first stage or advance
probing entailed drilling a series of 54mm diameter probe holes
(thirteen in the case of the pylon leg pilecaps and five in the
case of the pier pilecaps) to a depth of six times the pile
diameter below anticipated toe level so as to confirm the absence
of significant voiding below the pile toes. The second stage
probing entailed drilling to a depth confirmed by the advance
probes at each pile location in order to verify the appropriate toe
level.
-
20
Pylon The pylon, see Figure 12, is designed longitudinally as a
slender reinforced concrete stayed strut and transversely as a
triangular frame. The inverted Y-shape enhances the contribution of
the cable stay system to the overall torsional stiffness of the
deck.
Figure 12 General Arrangement of Pylon
Transversely the axial loads in the inclined legs are balanced
through the prestressed buried tie beam, which limits the
horizontal forces transmitted to the piled foundations. The beam is
post-tensioned progressively during construction, as the axial load
in the pylon legs increases. In order to avoid over-designing the
pylon it is imperative that second order (P) effects are properly
evaluated6. For the preliminary design,
the first order moment was magnified by a factor of 1
where is the
buckling factor, i.e. the critical buckling load divided by the
applied load. For the final design a geometric non-linear analysis
was used. In
-
21
both cases it is important to use a flexural stiffness which
takes account of creep and the reduction in stiffness of concrete
as the ultimate limit state is approached. The cable stays are
anchored inside the hollow pylon head. As well as the high local
forces, similar to those occurring at post-tensioning anchorages,
the horizontal component of the stay forces has to be transmitted
across the pylon between the fore and back stays. This produces
very high local bending moments and tensile axial loads which have
to be carried either by reinforcement in the pylon walls, or by a
structural steel assembly, or by a combination of both. The tender
design carried these forces using the reinforcement in the pylon
walls. The contractor re-detailed the pylon head using a structural
steel box assembly which acted compositely with the reinforced
concrete, see Figure 13. This had the advantage of not only
reducing the density of
Figure 13 Pylon Head Detail
reinforcement in the walls, but also allowed the bearing plates
for the stay anchorages to be accurately aligned in the controlled
conditions of the fabrication shop. The steel box, or stem liner,
was divided into a number of sections vertically so that on site it
was only necessary to align each box accurately on top of the
previous box for the stay anchorages to be correctly positioned.
Such an arrangement requires large forces to be transferred between
the concrete walls and the steel box which required careful
detailing with 32 and 40 mm diameter reinforcing bars being welded
to the steel box and fully anchored into the concrete walls.
Personnel access to all levels in the pylon head is provided by
access ladders and rest platforms located in the west leg. A pair
of rails is provided in the east leg so that a trolley can be used
to haul heavy
-
22
equipment up to the base of the pylon head. A lifting beam is
located inside at the top of the pylon head so that, for example,
jacking equipment can be lifted up should the stays require
re-tensioning in the future. The trolley rails are fitted with
rungs which allow them to be used to access the aircraft warning
light in the east leg and to provide an alternative means of escape
to ground level from the base of the pylon head. Ring bolts are
provided on the outside of the pylon to allow for abseiling access.
Deck The bridge deck is built in to the south abutment, which
avoids the need for large thrust bearings to transmit the
horizontal component of the stay forces and reduces the sagging
moment in the back span. The deck is continuous from the south
abutment to the north abutment, where a joint is provided to
accommodate movements due to temperature, creep and shrinkage. It
is supported vertically at the pylon and at the approach span piers
by mechanical pot bearings. One bearing at the pylon, at each
intermediate pier and at the north abutment is guided
longitudinally and resists transverse loads. The Preliminary Report
recommended a tubular steel space frame stiffening girder acting
compositely with a reinforced concrete deck slab. This form of
construction was chosen in preference to an all-concrete deck as it
was some 50% lighter which had a knock on effect of reducing the
sizes of the stays and the loads carried by the pylon, piers and
foundations. Erection of the composite deck could be carried out
using considerably larger segments and in a significantly shorter
time than a concrete deck. A space frame produces a particularly
rigid structure which increases the torsional stiffness of the deck
and therefore improves its stability under aerodynamic excitation.
This was considered to be very important as it was intended to
provide wind shielding along the edges of the bridge. The modular
nature of the space frame lent itself to cantilever erection of the
main span and the design was developed allowing for this form of
erection. The square configuration of the top chords of the space
frame allowed the deck slab to be designed as two-way spanning
thereby allowing a thinner deck slab to be used. The other
innovative feature of the deck is the use of a pultruded grp
enclosure system. This concept had been pioneered by FaberMaunsell
on the refurbishment of the Tees Viaduct in Britain in the nineteen
eighties. Measurements taken by the Transport Research
Laboratory
-
23
(TRL) had shown that the rate of corrosion of untreated steel in
such an enclosure was reduced to under 0.02 mm per annum or about 2
mm over the life of the structure. Pultruded grp has excellent
long-term durability, which combined with its low weight and high
strength makes it very suitable for use as an enclosure envelope
for steelwork. However, it is important that the enclosure is
almost, but not completely, airtight, as this allows the enclosure
to breathe, but not wind and wind-borne pollutants to blow through
the deck void. In order to avoid specifying one particular system,
a performance specification was prepared incorporating the UK
Highways Agency standard BD677. The system was required to carry a
safe working load of 2.5 kN/m2, which permits full access for
operation and plant for inspection and maintenance of deck
steelwork, bearings and deck drainage. The enclosed space is lit
throughout and provided with electrical socket outlets for portable
power tools. A further advantage of the enclosure is that it gives
the deck a smoother, less bluff, cross-section, which partially
compensates for the de-stabilising effect of the windshields under
wind loading. The bridge is generally designed in accordance with
the appropriate parts of BS54008. In addition reference was made to
the draft Eurocode for the design of steel structures9 and the
CIDECT design guide10. A multi-stay cable-stayed bridge is a highly
redundant structure where the permanent vertical load of the deck
is balanced by the vertical component of the cable forces. In
reality the loads in the stays will vary from those assumed by the
designer due to variations in self-weight and adjustments made
during construction in order to obtain an acceptable profile.
Wilson11 discusses the implications of this. As the effect of these
variations depends on the difference between large numbers and
because some of them will be self-compensating, it is important
that their possible magnitude is not over exaggerated. For the
Boyne Bridge the designers adopted the following approach at the
ultimate limit state. The pre-load in the stays was defined as the
difference between the loads in the stays under permanent loading
and those which would be introduced were the permanent loads to be
applied as an external load to the final structure. A partial load
factor fL of 1.05 or 0.95 was applied to the pre-loads and these
were then added to the nominal permanent loads and the partial load
factor applicable to the permanent loads was applied to the sum.
For the preliminary design the forces in the bridge were determined
by linear analysis of a three-dimensional space frame model with
the deck
-
24
modelled as a spine beam. For the final design the model of the
deck was expanded to include the individual members of the space
frame and the cracked section properties of the deck slab used
where appropriate. A linear buckling analysis to determine the
buckling ratio, , and the non-linear analysis to determine second
order effects were carried out using the LUSAS computer program.
The effects of overall temperature changes and differential
temperature through the deck were determined in accordance with
BD37/885. These were combined with the effects of a 10C difference
between the cable stays and the rest of the bridge and between any
two opposite faces of the pylon. The gradients within the pylon
were taken to be those appropriate for concrete box sections as
given in BD37/88. Although the cable stays are located behind P1
bridge parapets there is still a residual risk that one or more
stays could be ruptured should an errant vehicle breach the parapet
railings, so it is important that the bridge is designed to
withstand such an event. Also, there may be a need in the future to
replace one or more cables should they become damaged in some way.
Therefore, the bridge was designed for the following cable-out
scenarios:
(a) Accidental stay rupture The accidental removal of any one
stay cable combined with 10% live load under Load Combination 4 at
the ultimate limit state only, using partial factors of f3 = 1.0
and m = 1.0. The dynamic effects of rupturing a cable are allowed
for by applying forces equal and opposite to twice the load in the
ruptured stay to the structure without that stay. (b) Stay
replacement The planned removal of any one stay with full live load
on the opposite carriageway and a 30 m length of loading of
intensity 1.5 kN/m2 adjacent to the stay being removed, applied
over the width of one notional lane. This scenario was considered
at both the serviceability and ultimate limit states under Load
Combinations 1, 2 and 3.
Cable stay bridges are flexible structures and the method of
construction is such that local deviations of the deck from its
intended vertical line are to be expected. The reinforced concrete
edge beams were precast in order not only to obtain a high quality
finish, but also to allow them to be finally aligned and fixed in
position after the bridge was structurally complete in order to
give a good line to the edge of the bridge.
-
25
Cable Stays The cable stays were designed at the serviceability
and ultimate limit states using partial material factors, m, of
2.25 and 1.75 respectively. During erection the maximum stress in
the strands forming the stays was allowed to increase to 55% of the
characteristic tensile strength of the strands. The fatigue stress
ranges in the stays due to traffic load were determined in
accordance with BS5400:Part 108 and the rain-flow method used to
assess the fatigue life using the design curves from the
Post-tensioning Institutes Recommendations12. The overall diameters
of the stays are relatively small and they would be susceptible to
vortex-induced vibrations should a uniform steady wind blow over a
reasonable length of stay. However, because the stays are inclined
it is unlikely that the wind speed would be sufficiently uniform
over a sufficient length of stay for oscillations to build up.
Further, the dynamic response of the bridge and cables depends on
the interaction between many parameters and on their actual values
in the finished structure. These are impossible to predict
precisely at the design stage and therefore it was specified that
the detail design of the cable stays should allow dampers to be
incorporated at a later stage with little or no modification to the
structure, should this prove necessary. A provisional item for the
supply and installation of six dampers was included in the tender
in order to obtain competitive prices should these subsequently be
required. Each stay consists of between 32 and 68 no. 15.7 mm
diameter galvanised high-tensile 7-wire strands. The hdpe outer
sheathing varies between 180 and 250 mm diameter and was specified
with a 2 mm high double helical rib at approximately 600 mm pitch
to prevent wind-rain induced vibrations. The critical aspect of
cable stay performance is its behaviour under fluctuating loads. In
the Boyne Bridge, the cable stay anchorages are effectively fixed
to the bridge deck and pylon so that the ends of the cables are
subject to flexure under wind, temperature and live loading. Hence,
the stays experience variations not only in axial stress, but also
in flexural stress. Generally only axial loads are applied to cable
stays in standard fatigue tests and therefore any reduction in
fatigue performance due to fluctuating flexural loads would not be
evident. Hence, three fatigue tests on full-size stays were
specified for Boyne Bridge, Table 7, including one in which the
anchorage was rotated through 0.5 for one million cycles. The
specification stated that historical test data on stays and
anchorages of the same construction, assembly and constituent
parts
-
26
and subject to the same stress ranges may be acceptable at the
discretion of the Engineer.
Table 7 Stress Ranges for Fatigue Tests (s = nominal tensile
strength of wire)
Upper Stress Level
Axial Stress Range (MPa)
Anchorage Rotation Range
Cycles
0.45s 162 2 x106
0.45s 197 0.5 1 x 106
0.45s 197 1 x 106
Recent developments in the design of parallel strand cable stay
anchorages has concentrated on attenuating the stresses in each
strand due to rotation of the cable within the transition zone
immediately behind the anchorage block rather than by the overall
behaviour of the deviator, guide pipes and anchorage. This has two
consequences. Firstly, useful and relevant feedback on the fatigue
performance of the complete stay can be obtained by testing a
single strand in a monostrand anchorage, provided the anchorage is
of similar construction and longitudinal dimensions to the
multi-strand anchorages to be used in the actual bridge. Secondly,
such a design allows the deviator to be easily replaced by a
friction damper, should damping of the stay prove necessary in
practice. The Freyssinet HD2000 stay cable system proposed by the
contractor is such a system and an appropriate axial and rotational
fatigue test of a monostrand stay and anchorage, similar to the
test pioneered by the authors on Taney Bridge13, was developed
through discussions between the Engineer and Freyssinet. The test
was carried out at the CEBTP testing laboratories near Paris, see
Figure 14. There were no wire
Figure 14 Test Rig for Single Strand Stay
-
27
breakages during the one million fatigue cycles. In addition,
Freyssinet submitted historical test data on full size stays of
similar construction to those for the Boyne Bridge and, after
examination, the Engineer accepted these as meeting the
requirements of the specification. Architectural Lighting The
asymmetric cable-stayed form of the Boyne Bridge will have a
dramatic visual impact and it is believed that sensitive,
sympathetic lighting will both reinforce the visual design and help
denote this important heritage site within a rural environment.
There is an opportunity to highlight and reveal the bridge from a
number of standpoints:
as a landmark structure indicating an important site, as a piece
of significant architecture, as a balance to the potential
dominance of the proposed roadway
lighting, as an interesting feature in the distant views of the
area, and as a waypoint on the motorway.
A specialist study was therefore commissioned from lightmatters
(formerly Lighting Design Partnership), a British consultancy
specialising in environmental and architectural lighting. The
proposal which evolved was based on the use of two colours:
the white light of the roadway lighting marking the horizontal
deck, and
the lighting of the pylon and cable stays using a deep, subdued
colour to bring out the contrast with the roadway.
Such an approach has the following advantages:
it dramatically reduces the potential for light pollution by
concentrating the effect on a low level of lighting,
it ensures that the level of necessary lighting associated with
the roadway lighting does not dictate the overall level of
lighting,
it brings out the form of the bridge in distinguishing between
the roadway and the pylon/stays, and
it creates a visual effect that is at once dramatic and
interesting. The final choice of colour will be decided following a
site test, but the initial proposal envisages using a deep blue or
purple/mauve colour. A line of side emitting fibre optics in the
architectural colour, located below the edge beam of the deck will
produce a fine sharp line to
-
28
delineate the deck and frame the white roadway lighting. Narrow
beam luminaires mounted on the deck and anchorage abutment and
directed up the cables will pick out the stays in the deep colour.
As the beams converge towards the pylon head they will have the
effect of strongly highlighting the bridge structure. Light will be
projected up the outside of the pylon from locations at the base
and will help to define the pylon shape. The inside faces will be
lit in a similar manner providing a dramatic arch for motorists to
drive through. Finally, the recess at the top between the pylon
legs will be brightly lit to ensure a measure of drama. The
luminaires mounted at deck level will be protected from vandalism
with stainless steel cages. Where the lights face the oncoming
traffic, the cages will be fitted with grp louvres on their ends
and traffic faces to supplement the integral cowls and louvres of
the luminaires in preventing drivers from being distracted by
glare.
8. PILING PROCUREMENT Given the perceived likelihood of
difficulties and delays during the piling operation, it was
considered prudent to carry out the piling works as a separate
operation in advance of the bridge contract. If delays were to
occur in the completion of the piling, it were better that these
would not also cause a delay to the progress of the main
bridgeworks. In June 1998, tenders were sought under the Open
Procedure. However, tender documents required tenderers to submit
specific details in order to demonstrate their technical capacity
to carry out the type and size / scale of piling required. Four
tenders were received at the end of July 1998 and in mid-September
1998, Ascon Ltd. was awarded the contract in the amount of
2,242,000 (1,765,000) excluding VAT.
9. BRIDGE PROCUREMENT
The construction of larger cable-stayed road bridges had not
been carried out previously in Ireland at the time of tender. In
these circumstances it was felt that a pre-qualification process
should be undertaken in order to establish a list of competent
tenderers. The final pre-qualification process, itself would, the
design team believed, allow Irish contractors to form alliances
with appropriate foreign contractors and specialists. The
pre-qualification process was undertaken under the Restricted
Procedure of the EU Works Directive 93 / 37 / EEC. To this end,
each candidate was required to fill in a Pre-qualification
Questionnaire which
-
29
sought information on a wide variety of topics, such as,
company/consortium structure, financial details, technical
expertise, manpower, health and safety, sub-contracting and
programming. Among the stated pre-qualification criteria for
candidates were the following:
(a) minimum annual turnover in civil engineering projects in
each of the last three financial years ,
(b) construction of a cable-stayed road bridge within the
previous five years,
(c) construction of a steel/concrete composite bridge within the
previous five years, and
(d) availability of senior managerial staff with experience in
building structures as in (b) and (c) above.
Expressions of Interest were received from eight candidates in
mid-September 1998 and at the end of October 1998 the design team
recommended that the five best qualified of these be invited to
tender for the bridge. At the beginning of December 1998, the
Employer invited the five recommended candidates to tender. Tender
Documents The tender documents required the tenderer to include the
following with his tender:
(a) name(s) of the structural steelwork fabricator, (b) name(s)
of the supplier of the enclosure systems, (c) name(s) of the
supplier of the steel castings for the bridge deck
steelwork, (d) details of the cable stay system including
protective systems and
end terminations, and (e) details of the enclosure system.
Tenders based only on the tender documents were acceptable, i.e.
alternative tenders were not allowed. The contract documents were
based on the November 1998 Draft of the NRA Manual of Contract
Documents for Road Works.14 The time for completion of the works
under the contract was 24 months. The estimated cost of the works
at the time of the tender was 24.4m (IR19.2m) excluding VAT.
-
30
Tenders were received at the end of March 1999. Further
clarifications were sought from the two lowest tenderers on items
such as:
(a) the cable stays, (b) the enclosures, and (c) steel
castings
in order to ensure that their proposals were in accordance with
the specification. In mid July 1999, Roughan & O'Donovan
recommended the acceptance of the lowest tender in the amount of
25.08m (IR19.75m) excluding VAT from SIAC / Cleveland Bridge
Consortium. The client was considerably delayed in awarding the
tender. The contract award was in fact only made just before
Christmas 1999 and the contract itself was signed at the beginning
of April 2000 with a date for commencement of the works of 3rd May
2000. Supplemental Agreement Post contract, the Contractor put
forward a proposal to alter the layout of the steel deck to suit
his own fabrication strengths at his works in Darlington. Put
simply, a steel deck consisting of open steel sections in a steel
grillage was much more easily fabricated by this contractor than
the steel space frame with its hollow sections and cast or
fabricated steel nodes. The Employer was only prepared to consider
the alternative deck in the context of a Supplemental Agreement in
which the final account for the works would be agreed. In mid May
2000 the basis of a Supplemental Agreement was accepted by both
parties to the contract. A major reallocation of risk from the
Employer to the contractor was an important feature of the
Supplemental Agreement. For example, the agreement does not allow
additional payment or extension of time for:
full or partial suspensions of the works (e.g. arising from
archaeological investigations),
unforeseen physical conditions, instructions which in the
Engineers reasonable opinion are
necessary for the satisfactory completion of the works or the
functioning of the completed works,
legislation enactments and the like, and ambiguities and the
like in the contract documents.
-
31
Basically the only instance where the contractor was entitled to
payment was the in case of an enhancement which was defined as a
variation other than a variation which in the Engineers reasonable
opinion was necessary for the satisfactory completion of the works
or the functioning of the completed works.
10. ALTERNATIVE DECK DESIGN The alternative design prepared by
the contractor as part of the Supplemental Agreement and checked by
the Engineer sought to take advantage of the contractors
fabrication and erection strengths whilst maintaining the same
external appearance. The deck comprises a ladder beam with the
longitudinal girders located on the line of the outer bottom chord
of the space frame and cross girders at 3.333 m centres, see Figure
15. The cable stays are connected between pairs of cross girder
extensions. The longitudinal girders are 1750 mm deep in the main
and back spans, but deepen to 2400 mm at the anchorage abutment and
in the approach spans. The bottom flange is up to 1200 mm wide x
100 mm thick with doubler plates at the anchorage abutment, the
pylon and Piers 1, 2 and 3. The largest doubler plate, located at
Pier 1, is 1100 mm wide by 80 mm thick and extends over a length of
15.4 m.
Figure 15 Alternative Deck
Since there was to be no change to the external appearance of
the bridge, the pier locations were not changed. Transversely these
were located under the next-to-outer longitudinal chords of the
space frame and hence the longitudinal plate girders are out-board
of the piers. This induces very large hogging moments in the
diaphragm beams at the piers which are carried by twin steel cross
girders, 2925 mm deep on the centre-line of the bridge, filled with
reinforced concrete and post-tensioned in stages
-
32
with up to 12 no. tendons, each comprising 27 no. 16 mm diameter
7-wire strands. Each tendon was stressed to a load of 6000 kN. The
230 mm thick reinforced concrete deck slab, which acts compositely
with the steel beams, was constructed using precast Omnia planks as
permanent participating formwork. The weight of steel in the
alternative design is some 30% higher than in the original space
frame design. This together with the extra concrete at the approach
span piers increases the loads on the foundations by between 10 and
15% at Piers 1 to 3 and up to 26% at Pier 4. Following a review of
the trial pile results and an examination of the critical
loadcases, it was concluded that the substructures and foundations
could carry the additional load without any modification. In
addition to the changes to the deck, the contractor also
re-detailed the pylon to suit the erection loads for the
alternative deck and his preferred method of constructing the
pylon. Regions of the pylon where the vertical reinforcement was
not required to carry compression were identified, as in these
locations alternate vertical bars did not need to be restrained and
the transverse link arrangement could be simplified.
11. PILING CONTRACT Works commenced on site in November 1998.
Time for completion of the works was five months. The sequence of
works involved Ascons probing subcontractor drilling advanced
probes. An evaluation of the probes at each pilecap allowed the
Engineer to determine the required toe levels of the pile probes at
each pile location. Pile probes were taken to a point six times the
pile diameter below projected pile toe level. As probe results were
received at each pile location, the situation was reviewed and
amended toe levels advised to site as required. Such an approach
required continuous daily liaison between the Engineers
Representative and head office, but was essential in order to take
due account of the highly variable conditions which were known to
exist on site. Pile toe levels were decided based on an overview of
the probe results at each pilecap location. In overall terms, the
probing confirmed the results of the site investigations data.
Depths to rock head were largely as anticipated, and the rock
consisted of alternate bands of competent and heavily fissured /
weather rock. At pier locations, P1 and P2, the closest and
penultimate piers respectively to the northern bank of the river,
very poor rock conditions were encountered. In addition to voiding
and heavily fractured rock, interconnection between different
probes (as evidenced
-
33
by air flushing blowing up in adjacent probes and by backfill
grout migration into adjacent holes) indicated that voids and
cavities extended over a number of metres. Additionally, during a
period of very wet weather in January 1999, slight artesian effects
were noted with water levels in the boreholes rising to
approximately a metre above ground level. Pile construction was
undertaken by Dutch piling specialist contractor NGT, a subsidiary
of Ascons parent company HBG. A total of 82 no. 750mm diameter rock
socketed bored piles were constructed on the northern side of the
river, and 32 no. 1500mm diameter rock socketed piles for the pylon
foundations on the southern bank of the river. Both the 750mm
diameter and the 1500mm diameter piles were undertaken using a
Sumito LS 60 tonne crawler crane with a Demag D22 piling hammer,
see Figure 16. All piles had a temporary casing extending a minimum
of 2m into rock, with an uncased rock socket below.
Figure 16 Installation of Piles
-
34
Test piling was carried out on two 750mm diameter piles on the
north side of the river and one 750mm diameter pile at the pylon.
All pile tests were satisfactory. Integrity testing was undertaken
on all piles. In the case of the north abutment and piers P3 and
P4, pile toe levels were similar to those indicated in the tender
documents. For pile groups P1E and P2W, the piles were taken down
to levels of -33m and -36m respectively approximately 13m deeper
than anticipated in the tender documents, and reflecting the very
poor rock encountered. The piling contract was completed in July
1999.
12. CONSTRUCTION General The conforming design envisaged the
main span of the bridge being constructed by cantilevering the deck
out from the pylon and installing the permanent stays
progressively. In his tender, the contractor proposed erecting the
main span by launching it from the south abutment and using the
pylon to support the deck with two pairs of temporary stays. The
length of the temporary stays would be adjusted using strand jacks,
while the tension from the hauling cable would match the design
stresses in the steelwork. The contractor adopted the same approach
for the erection of his alternative deck and refined his design to
suit the launching procedure. Pylon The inclined legs of the pylon
were constructed in 6 m high pours using the lightweight Alumna
falsework system similar to that used on the Charles River Bridge
in Boston, USA, see Figure 17. In this system the formwork is
supported from a truss, which is constructed between thelegs and is
initially supported from the ground. As the legs rise upwards,
members of the truss are removed from the bottom and reassembled at
the top of the falsework. In this way the falsework climbs with the
legs and is supported from the legs. The falsework also serves to
keep the legs in their correct lateral positions, obviating the
need for jacking the legs apart prior to constructing the pylon
head, which would have been necessary if the inclined legs had been
constructed in free cantilever.
-
35
For the pylon head, the contractor adopted 3 m high pours with
the steel plates of the stem liners forming the inner faces and the
RMD system of climbing formwork the outer faces.
Figure 17 Construction of the Pylon
Work started on cutting down the piles and constructing the
pilecaps in September 2000 and the pylon was topped out on 28 March
2002. Deck The erection of the deck is shown schematically in
Figure 18. Sections of the longitudinal steel beams of the approach
spans were welded together on the ground and lifted into position
on the approach span piers. The cross girders were then bolted to
the longitudinal girders. The concrete for the composite cross
girders at the piers was cast in two stages and the first stage
post-tensioning installed. The Omnia planks and reinforcement for
the deck slab were placed on the steel framework and the insitu
concrete slab poured, starting from the north abutment. The span
sections of the deck were poured first followed by the sections
over the piers in order to minimise cracking in the slab. As the
concrete slab construction advanced, the second and third stage
post-tensioning was installed in the pier cross beams. The
-
36
final stage post-tensioning in the Pier 1 cross beam will be
installed after the surfacing has been completed.
Figure 18 Construction Sequence
-
37
While the approach spans were being constructed, the steelwork
for the main and back spans was assembled on stillages behind the
anchorage abutment, along the line of the motorway, Figure 19. The
outer
Figure 19 Main Span Steelwork
permanent back stay (B14) was installed and used to stabilise
the pylon during the launch. The steelwork was transferred to two
computer-controlled multi-axle trailers and a skate located at the
rear of the anchorage abutment. A temporary post and stays (T1 and
T2) were installed to support the leading cantilever until it
reached the pylon. A hauling line was anchored behind the north
abutment and connected to the leading end of the steelwork. The
steelwork was launched by driving the trailers forward and pulling
on the hauling line. The trailers also provided braking to control
the launch, Figure 18 (b) and (c). When the steelwork had reached
the pylon, the temporary post was removed, the skate moved to the
back of the abutment and temporary
-
38
stay T1 connected to the top of the pylon, Figure 18 (d). The
launch then continued with the following cycle:
pull, adjust length of temporary stay, survey, and move skate
back.
As temporary stay T2 passed the pylon it was also connected to
the top of the pylon and the launch continued with two temporary
stays, Figure 18 (e). When the steel girders were in the correct
longitudinal position, they were lowered to their final levels at
the anchorage abutment and pylon and the level adjusted at the
closure joint. The steel girders were then welded to anchorage
plates in the south abutment and to the end cantilever of the
approach spans. The concrete deck was constructed starting at the
pylon and working firstly towards the anchorage abutment and then
northwards across the main span. As the slab reached the position
of the first fore stay (M1), the first back stay (B1) was installed
followed by M1, Figure 18 (f). The next 10 m of slab was poured and
stays B2 and M2 installed. Construction continued with a typical
cycle:
pour 20 m of slab, install two pairs of back stays, and install
two pairs of fore stays.
As the construction face reached the temporary stays T2 and T1,
they were removed followed by the hauling cable. Finally, the
precast edge beams and deck furniture were aligned and fixed in
position, the grp enclosure installed and the roadway surfacing
laid, Figure 18 (g). Installation of Stays The stays were installed
using the Freyssinet iso-tension system of strand-by-strand
installation. The hdpe outer sheath is assembled on the deck by
welding together standard lengths of sheathing. The first strand is
installed and lifted into position with the sheathing. It is then
stressed to a predetermined force and attached to the stressing
anchorage with a special device incorporating a load cell. The
second strand is installed, connected to the anchorages and
stressed using a monostrand jack. A second load cell records the
force in the second strand, which is stressed until the forces in
the two strands are equal. The process is repeated until all the
strands are installed in the stay.
-
39
The initial force in the first strand is determined by the
designer taking account of the expected deformation of the
structure. The deck is relatively flexible and therefore a small
change in fore stay force can result in large vertical movement of
the deck. The loads in the stays and the forces and bending moments
in the deck can therefore be more precisely controlled by adjusting
the length of the stay rather than its load. Each fore stay was
initially installed to a load which was consistent with the
unstressed length of the cable being 75 mm longer than its
theoretical value to ensure that the level of the deck at the stay
anchorage would be lower than its required value. The deck was then
surveyed, compared with the expected value and the length of cable
shortening adjusted if necessary. Each strand was then shortened by
this amount. The fore stays were stressed from their upper
anchorages located inside the pylon head. Figure 20 shows the
expected and actual profile of the deck after structural
completion.
The verticality of the pylon was also surveyed after each round
of stay installations. Figure 21 shows the theoretical and actual
profiles at structural completion. The back stays were stressed
from the access galleries within the anchorage abutment. Because
the pylon and anchorage abutment present a relatively stiff
structure to the stay, it was appropriate to install the back stays
to their calculated loads with no subsequent length adjustment.
-200
-100
0
100
200
300
400
500
Rel
ativ
e C
ambe
r (m
m)
Theoretical Actual Corrected for creep + u/s camber
SA Pylon P1
Figure 20 Relative Cambers after Structural Completion
-
40
A final check of the deck profile will be made after the
surfacing has been completed. If required, a final round of stay
length adjustments will be made at that stage, but it is expected
that this will not be necessary as there has been good agreement
between the analysis model and actual behaviour.
13. SUMMARY OF APPROXIMATE QUANTITIES IN BRIDGE
Structural Concrete 19000 cubic meters Reinforcement 2500 tonnes
Structural Steel 2200 tonnes 15.7mm diameter strand in stays 333
tonnes (283,000m total length)
14. ACKNOWLEDGEMENTS The authors gratefully acknowledge the
permission of Mr. Oliver Perkins, County Engineer, Meath County
Council to publish this paper.
Client: Meath County Council Consulting Engineers:
Northconsult
Figure 21 Position of Pylon at Structural Completion
0
20
40
60
80
100
-100 -50 0Displacement (mm)
Hei
ght (
m) Theoretical
Survey
Correctedfor creep
SOUTH
-
41
Consultants for Wind Studies, Wind Shielding and Category 3
check:
FaberMaunsell Ltd
Bridge Contractor: SIAC/Cleveland Bridge J.V Piling Contractor:
Ascon Ltd
The authors wish to acknowledge the support and assistance of
the following:
Mr. Oliver Perkins County Engineer, Meath County Council
Mr. Charles McCarthy Project Engineer, Meath County Council
Mr. Frank Burke Former County Engineer, Louth County Council
Mr Tony Kearon County Engineer, Louth County Council
Mr Donall OCaoimh and Mr John Iliff
Project Resident Engineers
15 REFERENCES
1 ODonovan, P.J, Preliminary design of proposed Boyne Bridge on
Northern Motorway, paper presented to a meeting of North East
Region IEI, Nov. 1996
2 Wilson,K.R, Wind resistance design in large scale bridge
projects, Long-span bridges and aerodynamics (Ed. Miyata T, et al),
Springer, Tokyo,1999.
3 BD49/01 Design rules for aerodynamic effects in bridges, NRA
Design Manual for Roads and Bridges, NRA, Dublin, Dec. 2000.
4 Roberts, Sir G, Severn Bridge: design and contract
arrangements, Proc. ICE, Vol. 41, Sept. 1968, pp.48.
5 BD37/88 Loads for highway bridges, Design Manual for Roads and
Bridges, Highways Agency, London, 1988.
6 ODonovan, P.J, Wilson, K.R and Dempsey, A.T, Second order
effects in the design of concrete pylons for cable-stayed bridges,
to be published in 2003.
7 BD67/96 Enclosure of bridges, Design Manual for Roads and
Bridges, Highways Agency, London, 1996.
8 BS5400:Part 3:1982, Part 4:1990, Part 5:1979 and Part 10:1980,
Steel, concrete and composite bridges, British Standards
Institution, London 1979-1990.
-
42
9 DD ENV 1993-1-1:1992 Eurocode 3: design of steel structures,
Part
1.1 General rules for buildings, British Standards Institution,
London 1992.
10 Wardenier, J et al, Design guide for circular hollow section
joints under predominantly static loading (Ed. CIDECT), Verlag TV
Rheinland GmbH, Kln, 1991.
11 Wilson, K.R, Safety criteria for major bridges, paper
presented to an ICE seminar on Structural Safety, London, July
1999.
12 PTI Committee on Cable-stayed Bridges, Recommendations for
stay cable design, testing and installation, Post-tensioning
Institute, Phoenix, Aug, 1993.
13 ODonovan, P.J, Wilson, K.R, and Dempsey, A.T, The design and
construction of Taney Bridge, Dundrum, IStructE (Republic of
Ireland Branch), Dublin, Jan, 2003.
14 Manual of contract documents for roadworks, National Roads
Authority, Dublin 1998 (draft)