Next week's official opening of Bangladesh's Jamuna bridge marks the end of one of the last great civil engineering challenges of the century. The obstacles were formidable. The Jamuna (better known under its Indian name of Brahmaputra) is the world's fourth largest river in terms of flow, reaching 65,000m3/sec or more during the flood season. Sediment loads are the second highest in the world.
Worst of all, the Jamuna is a braided river. With no fixed channels, it runs between shifting islands of unstable micaceous sand, expanding to between 30km and 40km wide when the wet season begins. Even at its narrowest the braid belt is 15km across and during the dry season the river is never less than 5km wide.
River levels vary by 8m over the year. During the annual floods, scour can cut channels up to 45m deep in the soft river bed, where the silt and sands are several kilometres deep.
But there were very strong arguments in favour of a fixed crossing. Bangladesh, one of the world's poorest nations, is effectively cut in half by the Jamuna. Until the late 1970s, only unreliable ferries plied the 500km length of river from Bangladesh's northern border to the Bay of Bengal. Economic development was being crippled.
Rendel Palmer & Tritton's involvement with Jamuna dates back to the 1970s when it was asked to design a 220kV interconnector across the river. It was immediately clear that any crossing would require foundation engineering of heroic proportions. Huge 13m diameter 105m deep caissons were needed to support the 10 interconnector towers which had to withstand earthquakes and the effects of scour.
The interconnector was an economic success and subsequently RPT was asked to design a natural gas connector linking rich farm land in the north-west to gas fields in the east.
It soon became obvious that a multi-purpose bridge carrying road and rail traffic as well as the gas pipeline and a second power line would be a much better bargain than the pipeline alone.
By 1986 the funders - the governments of Bangladesh and Japan, the World Bank and the Asian Development Bank - were convinced, and RPT in joint venture with NEDECO of the Netherlands and in association with Bangladesh Consultants started preparing designs.
Two key problems had to be solved. The first was obvious - foundations had to cope with scour and liquefaction. The second was even more fundamental. Engineers had to persuade the river to flow under the completed crossing, as there was a significant risk that one year when the floodwaters receded the bridge would be left high and dry several kilometres from the wayward Jamuna.
Raking tubular steel piles up to 75m deep took care of the foundation problem. RPT's cure for the fickle nature of the river began with choosing a site 8km downstream of one of the rare fixed node points along the river, the existing flood defences at Sirajgang on one bank and Bhuapur on the other. The idea was to construct guide bunds more than 2km long on each bank upstream of the proposed 4.8km crossing.
This was in 1986, when all available evidence suggested the chosen site was in a relatively stable stretch of the river. Then came the major floods of 1987 and 1988. The equilibrium of the river was disturbed, and by 1992 the banks at the chosen site were 2.5km further apart than before.
This meant that the concept of a near 5km bridge with guide bunds on each bank was gone. But major design changes, extra costs and long delays would have been highly unwelcome at this point. Funding was already in place and tendering in progress. RPT therefore proposed a radical alternative.
The east guide bund would be constructed before the west bund, during the dry season. The west bund - located on either one of the temporary sand islands or in one of the shallower channels - would then be positioned during the following year. This meant the final length and alignment of the bridge could not be fixed until October 1995, when RPT had access to the possible sites after river levels fell - 19 months after the main construction contracts had been signed.
River levels were the key to the construction process. Water levels had to be high enough to allow deep drafted vessels to bring all the heavy equipment and materials up river, but low enough for them to squeeze under the cables of the power interconnector further downstream.
Bunds could only be built during the six or so months of the dry season. Piling was planned to start at the east abutment once the length and alignment of the bridge had been determined. Piling would therefore not reach the western end of the bridge until the west guide bund had been completed, and thus there would be no interference between the separate contracts for the bridge and the guide bunds.
This latter contract went to the HAM-Van Oord ACZ JV for $267M (£167M). A joint venture between Hyundai Engineering and Hyundai Construction was awarded the £154.6M design and construct contract for the bridge and approach viaducts, while local contractors won the much smaller approach road contracts.
Unfortunately, these were not awarded until spring 1994, too late for the JV to mobilise and complete the east guide bund that year.
A year's extension to the project was politically unacceptable. All that could be changed was a partial completion of the east bund in the first dry season, switching to the west bund in the second and back to complete the east bund in the third. This eliminated the potential delay, but at a cost.
Not only was leaving the unfinished east bund to the mercy of two flood seasons a gamble, but the overlap between bridge and bund construction promised endless interface problems.
Hyundai had priced its tender on the concrete deck alternative in the contract documents. It also put forward its own alternative involving the use of a large offshore piling barge to drive fewer but larger diameter piles.
Spans remained at nearly 100m, and balanced cantilever construction was used to build the decks from precast prestressed glued segments 2m long and up to 6.5m deep, weighing a maximum of 180t each.
Hyundai deployed one of the world's largest piling hammers, capable of a mighty 1,700kNm blow to drive piles up to 3.15m diameter and 83m long. Such was the energy of impact that considerable disturbance of the surrounding soft sands was observed. The east bund was still incomplete when piling began, but as construction approached the west bund the trench in front of it was temporarily backfilled to stabilise it against the effects of the final piling operations.
By this time the bund designs had already changed. Basically a combination of trench and bund, the guide bunds' outer faces originally sloped at 1 in 3.5. However, a number of major and minor slips during construction of the west bund prompted the consultants to soften the slope to a more stable 1 in 5 or 1 in 6. On top were sunk geotextile fabric weighed down by rocks - a total area of 880,000m2 on both banks. In all, 26M.m3 of dredging was needed to construct the guide bunds.
Other statistics were equally impressive. Deliveries included 57,000t of cement and 21,000t of pulverised fuel ash, 16,000t of steel reinforcement and 5,200t of prestressing cable, 350,000t of aggregate and 1.5Mt of rock armour. The 121 steel piles alone weighed 34,000t.
Final length is 4.8km, with 49 spans. The deck is 18.5m wide, with four road lanes, a metre gauge railway, and a 600mm diameter gas pipeline. Concrete pylons above the deck carry 232kV power lines and telecommunications links. On each side 17km approach roads on embankments 5m high carry traffic across the flood plains.
Official opening date is 23 June. Projects to link the bridge to new road and rail systems are under way, slightly complicated by the fact that existing rail networks on each side of the river run on different gauge track. But even now the impact of the bridge on the country is becoming evident. In the years to come the importance of the Jamuna bridge can only grow.