Bangladesh’s biggest and most challenging civil engineering feat is underway – a 6.1km bridge to connect the country’s east to west.
You might know Bangladesh as a land of poverty, famine and flooding but the World Bank expects the country’s growth to be around 7 per cent for 2015-16.
For a growing country situated on a flood delta, there are no shortage of small to medium bridges that need building.
The Padma Multipurpose Bridge Project is seen as particularly significant, as it links the entire south west of the country to the capital Dhaka.
The hope is ithe bridge will boost the country’s second major port, Mongla (in the south west), reducing its current primary dependence on Chittagong (south east), and assisting other major ports of Khulna and Benapole.
The £2.13bn project includes the main bridge (£1.16bn), a huge 13.8km of river training works (£830M), connecting roadworks (£114M), as well as toll plazas and service areas (£20M).
To grasp the relative size of this project, £2.13bn is 1.86% of Bangladesh’s GDP. If the UK were to spend similarly (1.86% of GDP) it translates to about £50bn, which is one estimate of how much the entire HS2 rail project could cost.
Initially, Bangladesh had a £910M loan from the World Bank. But this was withdrawn in 2012 when the bank found evidence of “a high-level corruption conspiracy” among members of government, an engineering company and individuals involved with the bridge. The project is now wholly funded by the Bangladesh Government.
“It’s quite an undertaking for the country,” says Rendel director for assignments Keith Taylor. ”It’s a source of huge pride, that they’re funding this all themselves, and the Prime Minister and ministers are taking a very active interest in its progress.”
International engineering firm Rendel is providing Management Support Consultancy for the Bangladesh Bridge Authority, in in association with BCL Associates Ltd (Bangladesh), Padeco Co. Ltd (Japan) and Katahira & Engineers International (Japan).
The Main Bridge contract was awarded to China Major Bridge Engineering Company which has an extensive record domestically, but Padma marks one of its biggest projects overseas. Another major Chinese contractor SinoHydro Corporation is taking care of the River Training Works.
Rendel has been involved in Bangladesh since the early days of the Eastern Bengal Railway Company (now Bangladesh Railway), formed in 1857. So the company knows better than most the challenges of building in Bangladesh. “It is impossible to be prophetic in dealing with the problems of the river Ganges and of rivers in the alluvial plains of Bengal,” says BL Harvey OBE, MICE in 1933.
Rendel’s Taylor says of the local climate: “Most of the heavy rainfall is July to September, and it’s very hot and humid about 30C. With the monsoon comes the floods, very large flows, which obviously affects the piling that’s being done with floating plant”.
Local scour up to 15m deep on the Padma’s piles has been forecast. Combined with the risk of earthquakes, the bridge requires steel tubular piles down to 115m deep.
Vertical piles are problematic: the larger the number, the higher the blockage, and hence the deeper the local scour will be. This in turn requires more piles, creating a vicious cycle. Computer analyses, together with prior experience, found that, intrinsically, a group of angled piles causes less blockage to the water flow.
Almost 40% of the project’s cost will go into river training works. The vast majority of these works are on the south bank, up to 11.6km, while the revetments on the north bank measure about 1.3km
Works on both banks will see an extended falling apron consisting of 800kg geobags five layers’ thick dumped on to a dredged surface. This gives way to a dredged slope at 1V:6H protected by three layers of 125kg geobags and a 90cm thick layer of hard rock up to a level of -2.4m. Between -2.4m and +2.4m the slope is protected by dumped concrete blocks over geotextile layer, whilst above +2.4m the concrete blocks are placed, again over geotextile material.
“So the first challenge is to get the right profile for the slopes because obviously that’s all done underwater, so it’s quite hard to control the slopes and tolerances for that,” says Taylor.
A large fabrication site sits alongside the bridge: pile cans are being rolled on site, and truss members, after being part-fabricated in China, are being shop welded on site into 150m lengths. These are then floated out and lifted into position.
A pre-feasability study boiled down three contenders for the bridge design: the Warren-type steel truss design, an extradosed bridge, or a twin box concrete girder-design.
In an extradosed bridge, engineers felt it would sit lighter, and with longer spans, but the extra difficulties in construction would have seen a longer build period and higher costs. A twin-box concrete girder model was feasible but the heavy girders led to a need for bigger foundations and shorter spans.
A steel truss bridge with composite concrete top slab – the lightest deck out of the available options – was found to be the optimal solution. It also had a relatively high stiffness-to-mass ratio – useful for freight railway loading and seismic performance.
Earthquake-proofing the structure through a 3D non-linear time history dynamic analysis, designers opted to replace traditional sliding pot bearings with seismic isolation bearings. With conventional pot bearings and shock transmission units, eight piles were needed for each pier; with seismic isolation bearings only six piles were required – a cost saving of more than 20%.
The main bridge works on two levels: provision for a railway at lower deck level and a four lane highway above. A total of 41 steel truss spans, each 150m long, join to make the ‘box’ 12.7m high, with a 22m wide composite upper concrete deck.
The lower level of the truss also carries a 762mm-diameter high pressure gas pipeline and telecommunication facilities.
The main truss is a steel Warren-type construction from welded rectangular hollow sections from steel plates of varying thickness between 20mm to 70mm. The trusses have a typical repeat spacing of 18.75m longitudinally and are 13.5m wide. With the composite concrete deck, structure’s overall depth is 13.6m.
Padmapiles 2 jpg
This bridge deck superstructure is founded on substantial 3m diameter piles driven to depths of between 85m to 115m.
Six piles are driven at each of the 40 piers. An additional two transition piers each have 12 3m diameter bored piles.
On the approaches to the bridge, viaducts consist of 75 piers and abutments, with foundations formed of bored 1.2m diameter piles.
The bridge will be carrying road vehicles only for the forseeable future. The design provides for OLE rail, with potential passenger train speed of 160 km/h and 125 km/h for freight trains.
The project has also required extensive enabling works including temporary access roads and road diversions, temporary working areas including harbours and temporary accommodation for site staff.
With a high rural population, other nearby towns dissapearing due to erosion, and local population densities up to 5000/km2, significant numbers of houses and crops lie in the bridge’s path.
“There’s a whole plan… for the resettlement of people and giving them sustainable livelihoods – for those living on the alignment of roads. And on the main bridge – various people live on the ‘chars’ which are islands in the middle of the river,” Taylor says.
Despite the upheaval, the bridge comes with purported big benefits: reducing travel time from the capital to the south west by 25%, boosting tourism and growing the construction sector by 29%.
With work beginning in November 2014 and expected to finish November, 2018, it can’t come fast enough for the Bangladesh Government. The government calss the current ferry service “unsafe and unreliable”, running with daily traffic of 4,300 people. Within five years this is expected to increase to 12,831 on the bridge, and 45,000 by 2040.
Robin Sham SH, (2015) Design of the Padma road and rail bridge, Bangladesh Proceedings of the Institutions of Civil Engineers, Vol 168, pp. 181-192