th

9 Austroads Bridge Conference, Sydney, New South Wales 2014

PEDESTRIAN BRIDGES ON THE REGIONAL RAIL LINK

PROJECT, VICTORIA
Joakim Dupleix, VSL Technical Centre Asia (Formerly SKM
Australia), Singapore
Bruce Gibbens, Jacobs Engineering Group Inc., Australia

ABSTRACT
The Regional Rail Link (RRL) Project involves the construction of new regional tracks from
Southern Cross Station in Melbourne, through to Sunshine then west to Werribee. Within the
Footscray to Deer Park section of the project, major pedestrian bridge works were undertaken
at Nicholson Street and HV McKay Reserve to replace existing bridges by new structure
spanning across the enlarged rail corridor. This section of the works was delivered by an
alliance consisting of Balfour Beatty, Metro Trains Melbourne, Parsons Brinckerhoff, the
Regional Rail Link Authority, Sinclair Knight Merz, Thiess and V/Line. The architectural input for
the design was provided by Hassell Studio.

This paper focuses on the special design features of these two pedestrian bridges in the
brownfield environment, including the impact of installation in an operating rail corridor,
limitations in space and operating conditions imposed due to the location of the bridges,
aesthetically pleasing design and the dynamic analysis involved in the design.

INTRODUCTION
For the RRL project, two pedestrian links over the existing metro and regional tracks were
required to be demolished and rebuilt to allow for the widening of the corridor. The bridges are
pinpointed on the map in Figure 1 below, at number 2 and 7.

The HV McKay Reserve Bridge is a historically-listed link between the HV McKay Memorial
gardens and Harvester Road in Sunshine town centre. It was built by Hugh Victor McKay circa
1920 between the Harvester Housing Estate and the Harvester Factory. HV McKay wanted to
facilitate the access to his factory from the Sunshine Town Centre. The old bridge consists of
multiple spans built from a mixture of timber and steel material. A new bridge was required to
span over the enlarged corridor and to comply with Australian Standard AS5100 Bridge Design.

The new bridge is a 66 m single span 3.8 m deep Warren Truss structure which minimises the
structural depth and avoid locating piers immediately adjacent to the rail tracks. The bridge
provides full access to people with disability with lift access from ground level. The bridge
screens incorporate a representation of the HV McKay gate symbol to remind users of the
bridge history.

The Nicholson Street Bridge in Footscray links the town centre to Victoria University main
campus and to the main road access to Melbourne-Footscray Road. It is also an important bus
route. The old bridge was a two span concrete bridge carrying road and pedestrian traffic
reinforced with rail sleepers. In order to allow for services to be relocated ahead of time a
separate pedestrian bridge was built first, fitted with 34 conduits within its structural depth. The
tight geometric constraints to maintain both the current road and rail levels only allowed for a
structural depth of 1.1 m to span the 33.5 m wide rail corridor. A composite steel and concrete
structure was designed to provide the required strength and allow for the conduits to be fitted
between the steel girders. Anti-throw screens were provided to protect the public from the rail
corridor below.

© ARRB Group Ltd and Authors 2014 1

 th
9 Austroads Bridge Conference, Sydney, New South Wales 2014

Figure 1: Overview map of the Footscray - Deer Park section of the project

HV MCKAY RESERVE BRIDGE

To minimise the number of vertical members and maximise the visibility from the inside of the
bridge, the truss design is a square Warren Truss type which uses square hollow sections for
both the chords and the braces. High strength steel with a yield strength greater than 450 MPa
was specified to minimise the structural depth for this large span. The concrete deck does not
contribute to the strength of the truss but provides a continuous walking surface. To minimise
the height of the abutment piers the truss is designed with a pronounced camber.

Figure 2: HV McKay bridge elevation

As shown in the general arrangement in Figure 2 above, the diagonal brace members are out-
of-phase, one side to the other. This was advantageous to disguise the fact that the abutments
were not parallel, and thus the north and south truss were different length.

The construction of the truss was as follows:

• Fabrication, assembly and painting of the truss by Haywards in Tasmania of three 22 m long
segments.
• Delivery of the segments on site, and assembly of the segments with bolted splices.

© ARRB Group Ltd and Authors 2014 2

 th
9 Austroads Bridge Conference, Sydney, New South Wales 2014

• Erection of the truss in a single lift over the existing rail corridor, using three cranes and
midair transfer.
• Casting of the bridge deck over Bondek sheeting.

Figure 3: 3D modelling of the steel truss

The requirement to eliminate site welding introduced bolted splices for both the chords and
diagonal braces, and special attention was given to keep them as aesthetically pleasing as
possible.

Furthermore the truss design followed the design guidelines of the Comité International pour le
Développement et l’Etude de la Construction Tubulaire (CIDECT) manual [1], considered
internationally to reflect the ‘state of the art’ in steel hollow section joint design practices.

Figure 4: Erection of the truss

Truss member joint design

The joint behaviour of hollow-section gantry trusses has to be considered at the early stages of
the member design. In comparison to the rigid moment connections used in portal frame
warehouses – which typically consist of I-beams and stiffened endplates – the connections
between hollow sections are generally flexible. This flexibility gives rise to the following:

 Large stress concentrations – which can magnify the stresses calculated on the basis of

by up to five times. Stress concentrations are affected by truss and hollow-section

© ARRB Group Ltd and Authors 2014 3

 th
9 Austroads Bridge Conference, Sydney, New South Wales 2014

geometries, the state of loading, metallurgical effects, residual stresses, the welding
process and geometry of welds, and any post-weld improvements.
 Modes of failure which are not otherwise a problem with heavily-stiffened moment
connections. For example hollow-section trusses may be subject to – “chord plastification”,
“chord punching”, or “chord buckling”.

Hollow-section connections are generally identified as “T”, “L”, “X”, “N”, and “Y” connections,
with some further variation or combination. Connections may be “gap”, or “overlap”
connections. The former may be easier to weld, however are most often weaker, and more
prone to fatigue, than overlap connections; overlap connections are generally stiffer.

It is incorrect to design the connections in hollow-section trusses on the basis of standard beam

theory, i.e. . The designer requires a method which recognises the concentrations of
stress at various points around the perimeter of each connection, as well as the possible modes
of failure.

Design principles

One of the key principles to optimise the design of square hollow section joints in trusses is to
use thick-wall chords and thin-wall braces. The stiffer walls in the chord resist loads from the
brace members more effectively, and the joint resistance thereby increases as the width-to
thickness ratio decreases. The thin walls in the brace require smaller welds as a result. Due to
the high strength steel used in the design, a reduction factor of 0.9 to the joint capacity is
required by the CIDECT manual.

For aesthetics, 45 degrees braces were preferred. To simplify fabrication all joints were
designed as K gap joints, as per Figure 5 below, since they are easier to prepare, fit and weld.
For this type of joint, a deflection allowance of 15 % of the total calculated truss deflection
should be used to account for the flexibility of the joint.

The fabricator was given the liberty to choose between fillet and complete penetration butt weld
to achieve the most economical outcome. Only the minimum required weld throat thickness was
stipulated on the drawings.

Figure 5: Typical K gap joint

Bolted splices design

The design of the splices needed to be as visually unobtrusive as possible and work in both
tension and compression. Typical bolted chord splices include flange-plate joint with bolts on all
sides of the hollow section. This splice type was however rejected due to its poor visual aspect.

© ARRB Group Ltd and Authors 2014 4

 th
9 Austroads Bridge Conference, Sydney, New South Wales 2014

Figure 6: Flange-plate joint connection

Alternative options were considered, such as bolts going through the section and resisting the
force in shear. The bolts however cannot be fully tensioned without the addition of compression
tubes inside the spliced section, as they would crush the walls. Refer to Figure 7 below. Due to
the large number of bolts required this option was not practical.

Figure 7: Spliced connection example using compression tubes

Chord splice

Figure 8: Chord splice detail

© ARRB Group Ltd and Authors 2014 5

 th
9 Austroads Bridge Conference, Sydney, New South Wales 2014

An innovative solution was developed to “hide” the bolts and maintain the profile of the chord. A
splice plate was introduced inside the hollow section on all sides with the adequate number of
bolts on each face of the section to resist the design compression or tension force. Because the
bolt nuts will not be physically accessible for tensioning of the bolt when the splice is closed, the
holes in the splice plate were tapped to replace the nut.

Figure 9: Chord splice before assembly

Interior splice plates on all four faces of the splice, used as a ‘blind nut’, were made of high
strength Bisalloy steel. BISPLATE® 80 typically achieves tensile strength of 800 MPa which is
equivalent to the tensile strength of a high strength nut, and is of similar thickness to have an
equivalent thread length. The high strength steel also allows the tension force to be transmitted
through the plates even after taking into account the bolt holes reduction.

The bolts are fully tensioned grade 8.8/TF. The faying surfaces are painted with inorganic zinc to
provide a reliable friction coefficient and prevent slip in the serviceability limit state.

Figure 10: Chord splice after assembly

Brace splice (webs)

Similarly, the brace splices design was driven by the desire to achieve minimal visual
disturbance. The design consists of a splice plate welded to two walls of the Square Hollow
Section (SHS) brace member, bolted to a mirror plate on the other end of the brace. To prevent
shear lag failure, the welded length to the SHS member must be maximised. This can be
achieved by providing slots through the walls to fit the splice plate. Once the connection has
been bolted up, the joint is finished by additional non-structural cover plates in the shape of the
brace member.

© ARRB Group Ltd and Authors 2014 6

 th
9 Austroads Bridge Conference, Sydney, New South Wales 2014

Figure 11: Brace splice detail

Figure 12: Chord and brace splices before cover plates are installed

© ARRB Group Ltd and Authors 2014 7

 th
9 Austroads Bridge Conference, Sydney, New South Wales 2014

Figure 13: Completed HV McKay bridge

NICHOLSON STREET PEDESTRIAN AND SERVICES BRIDGE

The structure is made of 925 mm deep steel I girders composite with a 200 mm thick concrete
deck, made of the precast layer (Transfloor) and an in situ topping. In plan, one edge of the
bridge is straight and this other follows a 39 m radius, with a minimum trafficable width at mid-
span of 4 m. During construction and before the composite action, the steel girders are
stabilised and prevented from lateral-torsional buckling with the U-frame action of Parallel
Flange Channel (PFC) member welded to the web stiffeners. Composite action is achieved
using shear studs designed in accordance with AS5100.6 Section 6.

Figure 14: Typical bridge cross section

Lateral bracing to the bottom flanges was provided to resist rail impact loading to the
superstructure as per AS5100.2.

© ARRB Group Ltd and Authors 2014 8

 th
9 Austroads Bridge Conference, Sydney, New South Wales 2014

Figure 15: Underside view of the bridge

As mentioned previously the brownfield environment constraint imposed a limited structural

depth to maintain the existing rail clearance and road level. A composite structure was an
economical design which allows for maximum strength with minimum depth required.
Nevertheless the critical aspects of this bridge design include:

• Stability of the girders during construction and before composite action. The effective length
for bending was calculated as per AS5100.6 Clause 5.6.3. To maximise the girders depth the
concrete deck thickness was minimised to 200 mm, the minimum required to achieve the
reinforcement cover.

Figure 16: Typical concrete deck section

• By nature and due to the large span-to-depth ratio (approx. 30) the bridge is relatively
flexible. The dynamic behaviour analysis revealed the first natural vertical frequency of the
bridge to be 2.2 Hz which is in the critical range of resonant frequencies of 1.5 Hz to 3.5 Hz
provided in AS5100.2 to be investigated as likely to be excited by pedestrian traffic. A
transient dynamic analysis was therefore developed to calculate the theoretical dynamic
deflections from a pedestrian movement across the bridge.

© ARRB Group Ltd and Authors 2014 9

 th
9 Austroads Bridge Conference, Sydney, New South Wales 2014

Figure 17: Side view of the Nicholson Road and Pedestrian bridge with anti-throw screen

Dynamic behaviour and site testing

AS5100 describes the pedestrian load as a 700 N load traversing the bridge at a given range of
walking speeds. This behaviour can be modelled using the inbuilt transient dynamic capabilities
of Strand7 software.

Modelling and findings

The procedure adopted here follows recommendations by Wheeler [2]. Different walking speeds
can typically be linked to a corresponding stride length, and therefore an applied vertical
frequency in footfalls per second as per Table 1 below from Bachmann and al [3].

Table 1: Pedestrian walking pattern

Walking speed Forward speed Stride length Equivalent vertical

(m/sec) frequency (Hz)
(m)
Slow walk 1.1 0.60 1.82
Normal walk 1.5 0.75 2.00
Fast walk 2.2 1.00 2.22
Slow run (jog) 3.3 1.30 2.50

The calculated bridge natural frequency corresponds to a “fast walk”. As suggested by Wheeler,
this dynamic loading can be modelled by time-dependant half sinusoidal loading applied at
successive locations along the bridge axis, distant from each other by the stride length.

Using Strand7 transient analysis (time-history), the dynamic response can be calculated by
assuming a critical damping ratio. Literature review suggested very low value for composite
steel-concrete structure. Sétra Technical guide [4] quotes the Comité européen du béton (CEB)
information bulletin No. 209 which provides average damping value of 0.6 %. Using this data
during modelling it was observed that the maximum dynamic deflection was marginally greater
than 1.8 mm, which is marginally greater than the maximum dynamic deflection allowed by
AS5100.2.

However, these results do comply with the British and the Eurocode standards, which present
acceptability criteria based on acceleration instead of displacement. Besides, it has been shown
that the installation of screen mesh fencing (such as the anti-throw screen fitted on the bridge

© ARRB Group Ltd and Authors 2014 10

 th
9 Austroads Bridge Conference, Sydney, New South Wales 2014

edge) generally contribute to a significant increase of the damping ratio, because of the friction
generated between wires and bolted connections during vibrations and the additional stiffness
provided reduces the amplitude of the vibrations. Likewise, the conduits were also likely to
modify the response of the bridge.

Figure 18: Pedestrian load model from Wheel [2]

It was thus proposed to build the bridge as designed, and undertake further testing upon
completion of the bridge to determine the “actual” natural vertical frequency and damping
coefficient. Provision was made to fix tuned mass dampers to the web of the main girders in the
event the measured vibration was deemed not satisfactory.

Testing and outcome

Vibration testing was undertaken on the complete but yet to be opened bridge by SKM Sydney
Advanced Analysis and Test group. Three accelerometers were positioned across the bridge
deck at mid-span of the bridge so as to measure vertical vibrations due to flexure as well as
torsional vibration responses of the bridge at the edges of the deck. The dynamic properties of
the bridge were measured using model testing and the maximum vibration response using
staged pedestrian walk-by tests.

Figure 19: Measured Bridge Vertical Acceleration Levels for Walk-By Test

The first significant finding was that the lowest measured natural frequency of 2.3 Hz closely
matched the calculated one. However, the measured critical damping of 1.7 % was notably
higher than the damping used in modelling.

© ARRB Group Ltd and Authors 2014 11

 th
9 Austroads Bridge Conference, Sydney, New South Wales 2014

After the pedestrian walk by test, the maximum measured vibration level at mid-span of the
bridge due to a single 700N pedestrian walk by was approximately 0.8 mm which is less than
50% of the maximum allowable vibration level according to AS5100.2. Even allowing for some
test variability, this vibration level is well within the acceptable range. An output of the testing
data is presented in Figure 19 below.

Although the vibrations are noticeable to a stationary person on the bridge, the testing
concluded the dynamic response of the bridge complied with AS5100.2 and that no tuned mass
dampers were required.

CONCLUSION
The new pedestrian bridges built as part of the RRL project have overcome difficult design and
construction constraints in a brownfield environment.

For the Nicholson Services and Pedestrian Bridge, the existing conditions required a small
structural depth which made the structure flexible and therefore sensitive to vibration. In-depth
dynamic analysis and site testing post construction allowed to justify the new bridge compliance
with AS5100.

The HV McKay Reserve Bridge link has high heritage significance to the local community. The
new bridge structure was carefully detailed to be aesthetically pleasing and satisfies the
strength and serviceability requirements of AS5100. The structural design with steel hollow
section follows the latest recommendations from international research, by considering the joint
design from the beginning, and using proven methods to check the adequacy of the chosen
arrangements.

REFERENCES
1. PACKER J.A. and al, “Design Guide for Rectangular Hollow Section (RHS) Joints under
Predominantly Static Loading”, CIDECT, Second Edition, 2009
2. WHEELER J E, “Pedestrian-Induced Vibrations in Footbridges”, ARRB Proceedings,
Volume 10, Part 3, 1980
3. BLANCHARD J, DAVIES B. L. and SMITH J.W., “Design Criteria and Analysis for Dynamic
Loading of Footbridges”, Symposium on Dynamic Behaviour of Bridges, Transport Road
Research Laboratory U.K. TRRL Supplementary Report SR275, 1979
4. Service d'Études Techniques des Routes et Autoroutes (Sétra), "Footbridges: Assessment
of vibrational behaviour of footbridges under pedestrian loading", Technical Guide, 2006

AUTHOR BIOGRAPHIES
Joakim Dupleix had completed a Master Degree of Engineering in both France and Australia
by 2008. He has worked for 5 years with SKM in Melbourne in structural design specialised in
bridge design and its associated structures. He has experienced different aspects of the bridge
life-cycle: tender to detailed designs and large scale strengthening of existing bridges. Joakim
was the Design Package Lead on the HV McKay Reserve Bridge and the Nicholson Street
Services Bridge. Early 2014 he moved on to take a new role with VSL Technical Centre Asia in
Singapore specialising in the design of construction related equipment for large scale
infrastructure projects worldwide.
Bruce Gibbens has almost 20 years’ experience in all aspects of the bridge life-cycle: from
concept studies and independent verification, to the load-rating, strengthening, and material-
condition assessments of existing bridges, as well as providing technical support during the
construction phase. He specialises in segmental box girders, including balanced cantilever
design. He has also supervised the design of steel box girders, arches, and integral bridges, as
well as complex bridge widenings and railway structures. Bruce once lived and worked in
California, enduring the rigors of high-level seismic design. Bruce was Technical Lead for
Bridges on Regional Rail Project.

© ARRB Group Ltd and Authors 2014 12

 th
9 Austroads Bridge Conference, Sydney, New South Wales 2014

Copyright Licence Agreement

The Author allows ARRB Group Ltd to publish the work/s submitted for the 9th Austroads Bridge
Conference, granting ARRB the non-exclusive right to:

• publish the work in printed format

• publish the work in electronic format
• publish the work online.

The Author retains the right to use their work, illustrations (line art, photographs, figures, plates) and
research data in their own future works

The Author warrants that they are entitled to deal with the Intellectual Property Rights in the works
submitted, including clearing all third party intellectual property rights and obtaining formal permission from
their respective institutions or employers before submission, where necessary.

© ARRB Group Ltd and Authors 2014 13