timeline item
Here is the information we have
on the item you selected
More like this
| |
sign up for our newsletter
© 2017 Engineering Timelines
engineering timelines
explore ... how   explore ... why   explore ... where   explore ... who  
home  •  NEWS  •  search  •  FAQs  •  references  •  about  •  sponsors + links
Gladesville Bridge
Parramatta River, Gladesville, Sydney, New South Wales, Australia
Gladesville Bridge
associated engineer
Anthony Francis (Tony) Gee
Guy Anson Maunsell
G Maunsell & Partners
date  1959 - 1964, opened 2nd October 1964
era  Modern  |  category  Bridge  |  reference  Te791998
photo  Julian Evans
On its completion, Gladesville Bridge in Sydney was the longest span concrete bridge in the world, and is believed to be the first designed using a computer. Its construction embodied many techniques not seen before in Australia. The bridge is a vital link in the city’s trans-harbour traffic route. Perhaps uniquely, one of its designers, Tony Gee, took part in its 50th anniversary celebrations.
The New South Wales coastline, from Byron Bay in the north to Nadgee in the south, is characterised by numerous rivers, inlets and lakes. As a result, the state led the way in Australian bridge building.
Before Sydney Harbour Bridge was completed in 1932, the original Gladesville Bridge (constructed 1878-81) was the only road crossing east of Parramatta. A 273m long iron lattice truss bridge with an opening swing span, it carried two road lanes and later a tramline, until 1949. It was the first in the Five Bridges route across some of Sydney Harbour's inlets, along with Pyrmont Bridge, Glebe Island Bridge, Iron Cove Bridge and Fig Tree Bridge.
By the 1950s, the route was heavily congested and the bridge could not accommodate the volume of traffic trying to cross, particularly at peak times. Often the swing span was difficult to close in hot weather, as the iron truss expanded while it was open.
For greater traffic capacity, the New South Wales Department of Main Roads (DMR) planned a complex of three new bridges — Gladesville, Tarban Creek and Fig Tree — to form part of a north western expressway serving the northern suburbs of Sydney. Although the expressway project was abandoned in 1977, the bridges, approaches and traffic exchanges were constructed. The three bridges carry the A40 trunk road. The complex is about 2.1km in length.
The present Gladesville Bridge is sited some 350m downstream (east) of the original, 6km north west of Sydney’s central business district. It links the suburbs of Huntleys Point in the north and Drummoyne in the south. The Parramatta River is about 335m wide here, with a spring tidal range of 1.37m. Its banks are of Sydney sandstone, overlain by up to 27m of mud and clay in the river bed.
The DMR had been working on a balanced cantilever design in steel for Gladesville since the early 1950s. In 1957, international firms were invited to submit tenders based on the DMR design, with the option to offer an alternative. British engineer Guy Anson Maunsell (1884-1961) was in Australia at the time and thought a concrete arch would work.
Back in London, Maunsell’s preliminary sketches and design concepts landed on the desk of Anthony Francis Gee (b.1934), then a graduate engineer with G. Maunsell & Partners. Transforming Maunsell’s ideas into a viable design might have been viewed as a speculative exercise and Gee chosen because his time was less expensive to the company. But he rose to the challenge and later described it as "an incredible opportunity".
In October 1957, four steel bridge tenders were received by the DMR, ranging from £2.51m to £3.87m. However, the lowest (£2.395m) was for the Maunsell concrete bridge, accompanied by drawings, calculations and a scale model. The submission was a British-Australian joint venture with contractors Reed & Mallik supplying project manager, agent and engineering staff, and Stuart Brothers providing the site foreman and labour. Ernest Alexander (Sandy) McKenzie (b.1923) would be the DMR’s resident engineer.
The bridge's waffle slab concrete deck is supported on a soaring arch composed of four prestressed concrete box ribs, and intermediate slender piers. In Gee's original design (arch span: 277.4m), the central 152.4m of each rib was to be prefabricated on floating centring and lowered (by flooding the supporting pontoon) onto the outer sections of the corresponding rib, already cast in situ on fixed falsework. This method was chosen to fit the constraints of the DMR’s limits for crown roadway level and navigation clearance during construction.
In 1958, the design was checked by Prof. Jack William Roderick (1913-90), Head of the School of Civil Engineering at the University of Sydney 1951-78, Kevin Forrester of the DMR and prestressed concrete pioneer Eugène Freyssinet (1879-1962) of the Société Technique pour l’Utilisation de la Précontrainte. Their analysis confirmed the design's viability but they recommended deepening the arch ribs at the crown and installing longitudinal prestressing cables between the ribs to resist buckling, such as could occur during an earthquake.
The stipulated navigational headroom was later reduced, so fixed falsework could be used over the whole span. The method finally adopted used continuous centring to support precast voussoir units, echoing the Roman method of using temporary formwork to support segmental components until an arch is complete.
Site investigation revealed the bedrock on the Gladesville side to be unsatisfactory, and the length of the arch was increased from 277.4m to 304.8m. This simplified the cofferdam design for the northern thrust block foundation, from a circular double-walled structure to a three-sided one. It also gave the bridge the world’s longest concrete arch. At the same time, the approach spans were extended from 27.7m to 30.5m.
The tender sum increased to £2.56m, and Gee had to submit detailed construction drawings. To analyse the arch, and enable an accurate setting out, he wrote three computer programs to handle the vast number of calculations required. At the time, design codes and specifications were somewhat rudimentary, and desktop PCs with proprietary software a distant dream.
Gee says, "We had to design almost everything from first principles and since several aspects of the design were not addressed in any of the specifications, this allowed us a latitude not available today. Although a limited number of main frame computers existed, there were no commercially available engineering programmes so we had to write our own. The potential benefits of computerisation outweighed the effort of learning basic programming and ultimately all aspects of the analysis and detailing of the arch were executed using application programmes specially written for the purpose”.
From late 1956 onwards, he used a room-sized Ferranti Pegasus computer at the London Computer Centre in Portland Place. Its computations used plug-in modules of valves (vacuum tubes) instead of transistors or microprocessors, and data input and output was via punched paper tape. It had no keyboard or screen. According to Vince Taranto, leader of Road Network Analysis at Roads & Maritime Services, Ove Arup & Partners used the same computer two years later for the structural analysis of the Sydney Opera House shells.
The design was subject to certain constraints. Once the centring was removed, the arch profile had to be "perfectly funicular under its own weight". The pattern of visible joints between units had to be aesthetically pleasing and all units close to equal weight, with neighbouring units not differing noticeably in width.
The first program calculated the geometric curves of the intrados and extrados, and the centroid profile, then deriving the area, moment of inertia, section modulus and other properties for numerous sections through the arch. The data produced enabled the second program to calculate the funicular (curved like a rope would curve only inverted) profile, as well as influence lines for unit loads, and moments due to unit thrust and strain. The third program calculated the dimensions of the precast voussoir units and the co-ordinates of their positions in the arch.
As a result, the geometric curves used in setting out the arch did not deviate more than 9.5mm from the funicular profile over its entire length. For navigational clearance, the underside of the arch is a maximum of 40.8m above high water level, and more than 36.6m above water for the central 61m of the span.
As built, the bridge measures 578.5m between rock-filled reinforced concrete box abutments. It crosses the Parramatta River on an arch spanning 304.8m, founded on thrust blocks cut into the bedrock, overflying the river banks. The central 91.4m section of the deck is supported directly on the fixed crown of the arch, with eight spans of 30.5m to either side.
In three months from December 1959, some 994 cu m of soil and 6,881 cu m of sandstone were excavated for the thrust block foundations, and 2,600 cu m for the abutments and approach piers. The thrust blocks were completed in 1961, with deeper footings than anticipated, increasing the cost to around £3.6m.
The thrust blocks are up to 18.8m long, founded on 2m by 1.7m steps cut some 12m into the sandstone. A corridor runs through each for access to the rib interiors. The blocks contain 11,087 cu m of mass concrete, poured in layers one step deep. Three concrete mixes were used in the blocks, two of low strength (Class A 17.2 MN per sq m and Class AA 20.7 MN per sq m) and one of high strength (Class PS 41.4 MN per sq m, cement-water ratio 0.35). The high strength mix includes sand and crushed gravel from the Nepean River, and is used also in the piers, deck beams and arch rib units.
Concrete and rock testing was carried out throughout the project in a purpose-built on-site laboratory. Tests on the steel reinforcement, high tensile steel bars and steel wire strands were undertaken at the DMR’s Central Testing Laboratory.
Meanwhile, piling for the centring and the casting of the arch ribs began. The prefabrication yard, downstream at Woolwich, was large enough to allow all the elements of a rib to be laid out and cast. Each rib is a voussoir arch 6.1m wide, and forms a tube of 108 box units stiffened by 19 diaphragms. The box units are hollow, rectangular with chamfered corners, and weigh about 51t apiece. They increase in height from 4.3m at the crown to 6.7m at the springings, with flanges 381mm deep and webs 305mm thick. The diaphragms, placed at 15.2m intervals, are 610mm thick and solid except for oval access holes.
By November 1961, the fixed steel falsework was ready to support erection of the first rib. The falsework was founded on 298 driven piles and 140 piles recessed into the bedrock, and carried on tubular columns and girders mostly of 18.3m spans, with a 67.1m truss span for navigational access. The steelwork was tied together and anchored by Macalloy bars set into the arch thrust blocks.
Though the pile bents (bearing piles plus pile caps) ran the full width of the bridge, the column and girder superstructure was only one rib wide. The ribs were constructed sequentially. As each was completed, the centring slid on rails on the pile bents to the next position. A full-width braced tower at centre span prevented sideways movement of the ribs before post-tensioning and contained lifting gear for the arch units. The floating steam crane Titan from Cockatoo Island Dockyard lifted the falsework columns and trusses.
Between 23rd February and 31st July 1962, the units for the first rib were craned onto the crown of the falsework and winched into position. Mass concrete was poured into the 76mm gaps between units to make the rib monolithic. The public watched from a viewing platform on the Drummoyne side.
A technique pioneered by Freyssinet was used to separate rib from formwork. Flat jacking systems are set into the ribs at the arch quarter points. Each contains 224 mild steel plates surrounded by rubber gaskets, forming a rectangular array of jacks in four layers of 56. Once all the rib units were placed, the gaskets were inflated with synthetic hydraulic fluid (oil), expanding to put the rib in compression and lift it away from the centring. After adjustments to correct the alignment, the fluid was driven out by filling the gaskets with cement grout.
The ribs were jacked to (calculated) levels above their true funicular profiles to allow for movement caused by concrete creep and shrinkage. Jacking operations were carried out between midnight and dawn to ensure uniform temperatures and avoid transverse thermal movement. The first rib became self-supporting in September 1962, and the remaining three in January, March and June 1963, respectively.
Transverse stressing cables are cast into the rib diaphragms. Once all four ribs were in place, the diaphragm cables were post-tensioned to tie them together into a single arch.
The bridge deck is supported on precast piers positioned at 30.5m centres, rising from alternate arch diaphragms. Each is a portal frame of two columns joined by a crosshead. The 3.1m wide columns are rectangular in section, with a gap of 9.75m between each pair. The crossheads are 24.4m wide and 3.1m deep, with chamfers to the undersides of the outer edges.
The piers above the thrust blocks are 762mm thick and the others are 610mm thick, with rockers at the quarter points (hinges at the bases of the columns). Concrete hinges are located at the tops of the two columns above the thrust blocks on the Gladesville side.
Beyond each end of the arch are four girder spans, supported on three piers also 610mm thick. The pier columns on the Drummoyne side are in cantilever, with concrete hinges built into the tops of the second pair. Columns on the Gladesville side are fixed. The column foundations measure 5.3m by 2.6m.
The stresses in the columns induced by longitudinal movement and horizontal forces are much greater than the axial ones. Because of their slenderness ratio, the columns are prestressed to counteract tensile stresses, using cast-in Macalloy bars wrapped in Denso tape. Taping the bars means ducting and grouting are not needed, and prevents a bond developing between the bars and the concrete — not a disadvantage here as high ultimate bending strength is not necessary. To avoid excessive bending moments, the crossheads are not prestressed but constructed in heavily reinforced concrete.
The deck rises to the centre of the arch at a gradient of 6 percent, and comprises 143 prestressed precast concrete T beams weighing 60-65t each. Its general arrangement is eight T beams side by side, spanning 30.5m longitudinally between piers. Each beam is 1.8m high, with a top flange 2.8m wide and web 152mm thick ending in a chamfered base 381mm thick. They are post-tensioned by four steel cables of 12 x 13mm diameter strands, located in a vertical row in the base of the beams.
The webs are stiffened using four vertical stub diaphragms running transversely under the deck at 6.1m centres. In situ concrete between the flanges unites the deck into a grillage, though no concrete was poured over the precast beams. The flanges are tapered at the ends of the beams to make room for sufficient reinforcement in the concrete to counteract the live loading, ensuring longitudinal continuity over the piers.
To accommodate longitudinal movement, small precast concrete hinges at the ends of the beams are recessed into the crossheads. Expansion joints with roller bearings cross the deck above the ends of the arch.
The original design was for a six lane roadway 21.9m wide between kerbs with 1.8m footways behind protective barriers on either side. The plan was modified during construction by the inclusion of a traffic interchange between the Gladesville and Tarban Creek Bridges, necessitating a wider carriageway over the four northernmost spans.
The deck width was increased from 21.9m to 36.6m by using T beams with narrower top flanges and only three prestressing cables, and tapering the in situ concrete slab infill between beams. South to north, the four spans contain 10, 11, 12 and 14 of the smaller beams, supported on proportionately wider piers. The concrete hinges adopted elsewhere in the deck beams are replaced here by heavy prestressing, allowing longitudinal movement to be absorbed in flexure.
The operation to lift the deck beams, using a special launching truss, and position them was completed in February 1964. The in situ concrete between the beams was cast as erection of the piers and deck beams progressed, after which the falsework was dismantled. The deck was finished by casting the cantilevered footways, installing railings and lighting, and laying an asphaltic concrete wearing course on the roadway.
On 2nd October 1964, Gladesville Bridge was opened by HRH Princess Marina Duchess of Kent with New South Wales Premier John Brophy (Jack) Renshaw (1909-87) and Deputy Premier and Minister for Highways Patrick Darcy (Pat) Hills (1917-92) in attendance. The final cost of the bridge and approaches was about £4.5m.
The construction was recorded by two official bridge artists, both previously official war artists during World War II (1939-45). Rhys Williams (1894-1976) produced three oil paintings of the bridge and Robert Emerson Curtis (1898-1996) made 16 pencil sketches.
In 1965, the bridge received a Civic Design Award from the Royal Australian Institute of Architects for a work of outstanding environmental design. By this time, 49,400 vehicles were using it daily.
In the 1960s and 70s, cracks appeared as a result of concrete creep. These were sealed with epoxy to protect the reinforcement. Further cracking occurred where the deck sits over the arch crown, possibly caused by the hinges in the Gladesville pier columns not flexing enough. Spalled concrete was replaced and epoxy grouting used for the cracks.
In the 1970s, the carriageway was reconfigured to accommodate an extra lane by reducing the width of the upstream footway and reconstructing the median barrier. There are now three northbound and four southbound lanes.
In July 1980, Gladesville Bridge ceased to be the longest span concrete bridge in the world with the opening of Krk Bridge in Croatia — span 416m.
In 1974, Gee founded the practice Tony Gee & Partners in the UK. He retired in 1988 to live and work in America. In 1992, he was awarded the Telford Gold Medal by the Institution of Civil Engineers. In 1995, he was likely the only person to be a Fellow of the Institution of Civil Engineers, the Institution of Structural Engineers and the Institution of Mechanical Engineers at the same time.
In 2000, an oral history of the bridge expressed its technical, historic and social significance through interviews with those involved in all stages of the project.
On 21st March 2014, Engineering Heritage Australia approved it as an Engineering Heritage National Marker. On 15th December 2015, the American Society of Civil Engineers recognised Gladesville Bridge as an International Historic Civil Engineering Landmark. By that time, over 81,000 vehicles a day were using the bridge.
Resident engineer: Ernest Alexander (Sandy) McKenzie
Contractor agent: Howard James
Contractor: Reed & Mallik, Salisbury, UK
Contractor: Stuart Brothers, Sydney, Australia
Research: ECPK
"Nomination of Gladesville Bridge NSW as an Engineering Heritage International Marker" by Michael Clarke, Sydney Engineering Heritage Committee, New South Wales, April 2014
"Construction of the Gladesville Bridge, Summary Report" by Martha Ansara, Roads and Traffic Authority Oral History Program, New South Wales, August 2001
"Gladesville Bridge" by John Walter Baxter, Anthony Francis Gee, and Howard Baikie James, Proceedings of the Institution of Civil Engineers, Vol.30, Issue 3, pp.489-530, London, March 1965

Gladesville Bridge